TOWARDS THE USE OF PROXY REPORTED OUTCOMES IN MS
JUDITH SONDER
The studies described in this thesis were performed at the department of Neurology of the VU University Medical Center Amsterdam (VUmc), the Netherlands. MS research at the VUmc is organized within the VUmc MS Center Amsterdam and supported by a grant (06-594) of the Stichting MS Research (Dutch MS Research Foundation, Voorschoten, the Netherlands). About the cover
The team of hand in hand paper dolls represents the team of MS patient, partner (proxy), family and MS specialists in research and medical care. The patient and proxy are highlighted as subjects of this thesis.
Printed by
Ridderprint, Ridderkerk
ISBN
978-90-5335-828-3
© 2014 Judith Sonder, the Netherlands All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, photocopying, or otherwise, without the permission of the author or, when appropriate, of the publishers of the publications.
VRIJE UNIVERSITEIT
Towards the use of Proxy Reported Outcomes in MS
ACADEMISCH PROEFSCHRIFT
ter verkrijging van de graad Doctor aan de Vrije Universiteit Amsterdam, op gezag van de rector magnificus prof.dr. F.A. van der Duyn Schouten, in het openbaar te verdedigen ten overstaan van de promotiecommissie van de Faculteit der Geneeskunde op woensdag 21 mei 2014 om 11.45 uur in de aula van de universiteit, De Boelelaan 1105
Door
Judith Maria Sonder
geboren te Leidschendam
promotoren:
prof.dr. B.M.J. Uitdehaag prof.dr. C.H. Polman
Voor mijn ouders en grootouders
CONTENTS Chapter 1
General introduction
Chapter 2
Validation of a patient reported outcome scale for patient and proxy use
23
2.1
25
Chapter 3
Chapter 4
Chapter 5
Validation and interpretation of the Dutch version of the Multiple Sclerosis Neuropsychological Screening Questionnaire
9
Agreement and disagreement on five patient reported outcome scales
43
3.1
Proxy measurements in multiple sclerosis: agreement on different patient-reported outcome scales
45
3.2
Analyzing differences between patient and proxy on patient reported outcomes in multiple sclerosis
57
Longitudinal analyses of patient and proxy responses
69
4.1
Do patient and proxy agree? Multiple Sclerosis physical impact and walking ability change on patient reported outcome scales
71
4.2
Towards the use of proxy reports for estimating long term patient reported outcomes in multiple sclerosis
85
Summary and general discussion
97
Nederlandse samenvatting
109
Appendix
115 List of abbreviations Overview of patient reported outcome scales
About the author List of publications Biografie Dankwoord
116 118 129 130 131 133
CHAPTER 1
General Introduction
Chapter 1
Multiple Sclerosis (MS) is a progressive disease of the central nervous system that is characterized by lesions in the brain and spinal cord. These lesions affect the myelin sheath which causes inhibition of axonal transmission and axonal loss. MS is the most common chronic disabling neurological disease in young adults and results in a broad range of symptoms which including motor, sensory, visual and cognitive problems. Although the exact etiology of MS remains to be elucidated, it involves environmental exposure and genetic susceptibility. Approximately 2.3 million individuals worldwide are affected with MS.
1-3
Globally, the prevalence of MS is 33 per 100.000, whereas the highest prevalence was found in North America and Europe, being more than 100 per 100.000.
4,5
MS is more common
among women than men and symptoms occur at around 30 years of age, when people are economically most active and when they would be most likely to start or support a family.
DIAGNOSIS AND DISEASE COURSE The disease typically presents with an acute episode of neurological dysfunction (relapse) with symptoms as loss of vision, sensory and motor disturbances, cognitive problems, mood disorders, loss of bladder and bowel control, sexual dysfunction and others like pain and fatigue. The diagnosis of MS is based on clinical findings and magnetic resonance imaging (MRI) of the central nervous system. The diagnostic criteria for MS have been revised over the 6-9
years.
The disease may start with a single clinically evident episode caused by
demyelination in the brain or spinal cord without any preceding episodes. This is called a clinically isolated syndrome (CIS). In these cases dissemination in time (and often also in 1
space) is lacking and therefore CIS patients are not yet diagnosed with MS. The majority of these patients will eventually develop MS and show a relapsing remitting (RR) disease course characterized by relapses (i.e. episodes of neurological dysfunction) from which they usually recover. With time, recovery from relapses can be incomplete, and persistent symptoms may accumulate. In due time the majority of RRMS patients enters the secondary progressive 3
phase (SP), characterized by steady progression of disability, with or without superimposed relapses. In around 20% of the MS patients the disease course is characterized by slow progression from the onset of the disease, without relapses, but with continuous
10
General Introduction
accumulation of neurological disability, called primary progressive (PP). The clinical courses of the disease types are shown in Figure 1.1.
Figure 1.1 Schematic overview of the disease courses of MS subtypes The X-axis represents time, the Y-axis represents neurological disability
CLINICAL ASSESSMENT IN MS For the evaluation of disease activity and disease progression, various methods can be applied to assess disability, impairment and impact of MS. These assessments can be physician based, for instance neurological examination, neuropsychological evaluation and functional tests, or patient derived, by using patient reported outcome scales or interviews. Quantification of MS symptoms is difficult due to the variety of neurological symptoms and the lack of one outcome parameter that covers the full range of MS symptoms. Therefore a combination of physician based and patient based measures can be useful. The physician based outcome measures that are often used in MS research are described below. EDSS 10
The Expanded Disability Status Scale (EDSS) uses information from neurological examination to evaluate several functional systems: visual, brainstem, pyramidal, sensory, cerebellar, bladder/bowel and cerebral function. The ordinal scale ranges from 0 to 10, in which a high score refers to more severe disability. The EDSS is internationally the most common used outcome measure of disability in MS, despite its limited reliability, the limited ability to reflect change and the psychometric disadvantage of an ordinal scale with non-linear steps.
11-13
11
1
Chapter 1
MSFC The Multiple Sclerosis Functional Composite (MSFC)
14
is a functional measure combining
three quantitative tests assessing dimensions that are possibly affected by MS. The Timed 25Foot Walk (T25FW) evaluates ambulation, the 9-Hole Peg Test (9HPT) evaluates arm dexterity and cognition is assessed by the Paced Auditory Serial Addition Test (PASAT). Although the tests are often used separately, the designed composite, a combination of three tests, has not gained wide acceptance. BRB-N The Brief Repeatable Battery for Neuropsychological Tests (BRB-N)
15-17
is widely used to
evaluate cognitive impairment. A recent study comparing neuropsychological test batteries concluded that the BRB-N is reliable and sensitive to use in MS patients.
18
The BRB-N was
developed as a short observational instrument to identify disturbances of cognitive domains 19
in MS patients and consists of five subtests: The Selective Reminding Test (SRT); The Spatial 20
Recall Test (SPART); The Symbol Digit Modalities Test (SDMT); 16,22
Addition Test (PASAT);
17,21
The Paced Auditory Serial
and the Word List Generation (WLG) test.
16
The indication of
possible cognitive impairment of patients in this thesis was based on normative values of a Dutch sample of healthy controls.
23
PATIENT REPORTED OUTCOMES Patient reported outcome scales (PROs) are increasingly used as measures of disease impact, physical functioning, psychological functioning, cognition and quality of life in MS.
24-28
The
advantages of PROs, compared to more former clinical tests, are that they are time efficient and easy to complete. Using PROs the patients’ disease can be monitored during routine clinical care or responses can be evaluated at group level in cohort studies, evaluating the short and long term effects of MS. The quality of PROs can be determined by the evaluation of psychometric criteria: The reliability (consistent performance on test items), validity (measure of the intended underlying construct), interpretability (sensitivity and specificity of the scale) and responsiveness (ability to detect change).
13
The PROs used in this thesis are
described in the following sections and presented in the Appendix. 12
General Introduction
MSIS-29 29
The Multiple Sclerosis Impact Scale (MSIS-29) evaluates disease impact of MS on daily life and can be divided into two subscales; a physical scale which consists of 20 items and a psychological scale with 9 items. The range of item scores is 1 (no impact on daily life) to 5 (extremely influencing daily life). Two separate scores for the subscales can be calculated. The MSIS-29 satisfied the psychometric criteria for patient and informant use.
30
MSWS-12 The Multiple Sclerosis Walking Scale (MSWS-12)
31
is a patient-based measure of walking
ability in MS with a 12-item scale, ranging from 1 (no problems with walking at all) to 5 (extremely difficult). The scale is restricted to patients who can walk and can also be completed by an informant. The MSWS-12 satisfies the standard criteria as a reliable and valid patient-based measure of the impact of MS on walking ability. MSNQ 32
The Multiple Sclerosis Neuropsychological Screening Questionnaire (MSNQ)
is a brief self-
administered questionnaire with 15 items that reflects neuropsychological competence in patients with MS during activities of daily living. The questionnaire can be completed by the patient or an informant. Every item has 5 response options, 0 (does not occur) to 4 (very often, very disruptive), a higher score indicates more cognitive problems. The psychometric properties of the MSNQ are evaluated in one of the studies in this thesis.
33
GNDS Guy’s Neurological Disability Scale (GNDS)
34
is an interview-based questionnaire measuring
neurological disability. The GNDS contains 12 items in which an interviewer comes to a score with one or more questions to the patient or informant. The items include cognition, mood, vision, speech, swallowing, upper limb function, lower limb function, bladder function, bowel function, sexual function, fatigue and 'others'. All items are scored on a 0 (no problems) to 5 (disabled) scale. The GNDS was found to be a reliable, responsive and valid measure of neurological disability.
13
1
Chapter 1
PROXY MEASUREMENTS Research projects are often driven by information obtained directly from patients. Unfortunately, when the patients’ response is less reliable or missing due to disease status, age, cognitive functioning or mood disorders, it can be useful to obtain data from another perspective, for instance a proxy respondent. A proxy respondent can be a partner, close family member or healthcare provider of the MS patient. Proxy measurements have recently been evaluated in several patient populations such as eldery, cancer
39
and patients with Alzheimers disease.
40
35,36
children,
37
stroke,
37,38
These studies report different results
according to the agreement between patient and proxy. In most studies proxies rated patients as more severely affected than patients themselves and agreement was better on physical scales compared to evaluation of (neuro)psychological and cognitive status. According to these studies, proxy responses cannot be used as a direct substitute for missing patient data without considering other patient and proxy related variables. PATIENT-PROXY AGREEMENT IN MS In order to explore the use of proxy measurements in MS, the accuracy of proxies (i.e. partners of MS patients) rating the impact of MS and treatment induced change was evaluated in groups with small sample sizes.
30,41-44
The accuracy was determined by the extent 41
to which proxy ratings agree with responses provided by the patients on the MSIS-29. On the MSIS physical scale good to perfect agreement was found between patients and proxies, but larger differences were seen on the MSIS psychological scale. There was a tendency for proxies to overestimate, to report more disease impact of MS than patients themselves did.
43
The advise out of these studies was to extend the number of patients and proxies and focus on factors that may cause discrepancies in proxy versus self-report measurements, for example cognition and mood disorders. THE ACCURACY OF PATIENT REPORTS Cognitive deficits are present in 40-70% of the MS patients.
45,46
Disturbances in the domains
of attention, information processing speed, memory and executive skills are major features of 14
General Introduction
the cognitive profile and can often be detected early in the course of the disease.
47-49
These
deficits affect many aspects of daily life, such as social and emotional functioning, ability to run a household, employment status and the overall quality of life. Patients fatigue, anxiety and depression have been identified as important contributors to cognitive performance and 50-52
the accuracy of patient reports.
As cognitive dysfunction, fatigue and mood may cause
methodological problems, such as loss to follow-up or reduction of the reliability of patient scores, the use of proxy reports could be beneficial. THE ACCURACY OF PROXY REPORTS Researchers aiming to explain the differences between patients' and proxies' assessments have focused on patient related factors, such as cognition and mood, which can reduce the reliability of MS patients' self-assessment.
53,54
However, coping with MS symptoms is not only
challenging for patients. A range of psychological, physical and financial changes may also affect the MS patient's partner (proxy respondent) and family. Thereby the disease can change relationships and caregiving roles.
55
These factors can cause long-term stress in
proxies and influence their physical health, mental health and quality of life.
55-58
Hence, it is
important for researchers to consider both the patients' disease status and their proxies' physical health, mental health, mood and caregiver strain when examining the agreement between the patient and proxy rating on MS PROs.
AIMS AND OUTLINE OF THIS THESIS Researchers in MS often have to deal with loss of information and loss to follow up as a result of disease burden. The time points of this methodological problem are not random, but are often critical stages in the disease of the patient in which important information about the health status and disease progression should be collected. Moreover, a full understanding of self-report forms and the ability of the patient to respond adequately can be influenced by disease worsening, fatigue, mood disturbances and cognitive problems due to MS.
59
Therefore including proxy respondents (i.e. partners of patients) to evaluate the disease state 60
of the patient might be useful to continue reliable follow-up research. In earlier studies the 15
1
Chapter 1
use of proxy responses on the MSIS-29 was evaluated.
30,41-44
The goal in this thesis was to
explore whether other types of MS PROs are also suitable to be used by proxy respondents of MS patients. The aim of this thesis was to estimate patients’ long term outcomes using proxy responses. Therefore we evaluated patient and proxy responses on several patient reported outcome scales that are often used in MS research. We analyzed different patient and proxy related variables that could be of influence on (dis)agreement between patient and proxy scores to establish the most accurate prediction of long term patient outcomes. When using proxy responses, one should be able to rely on the measurement scales to yield the same results for patients and proxies. Since earlier studies comparing MS patients’ and informants’ responses on the MSNQ showed opposing results
32,61-63
the validity and
interpretability of the MSNQ patient and informant form were evaluated for the Dutch population of MS patients and proxies in chapter 2. In chapter 3 the exploration of patient-proxy agreement is described on five PROs that are frequently used in MS research, the MSIS-29 physical and psychological scale, the MSWS-12, the GNDS and the MSNQ. In chapter 3.1 the mean differences between patients and proxies on these five scales were analyzed. The agreement between patients and proxies on scale level and item level was evaluated for all PROs. Following these results the patient and proxy related factors that could be of influence on the (dis)agreement were analyzed and described in chapter 3.2. The difference between patient and proxy was used as outcome variable and several patient and proxy related variables such as age, gender, education, health status, type of MS, cognitive status, mood and caregiver burden were used as independent variables in linear regression analyses. This resulted in a model explaining a part of the differences between patients and proxies for each PRO separately. The study so far did not include patient-proxy measurements over time. In chapter 4 the responses of patients and proxies in a longitudinal setting are described. Data of both patient and proxy were collected at a 6 months and a 2 year follow up. Because results in the previous sections has shown that physical MS scales acquire more agreement between 16
General Introduction
patient and proxy, the focus in chapter 4.1 is on longitudinal changes in patient and proxy responses and explanation of agreement and disagreement over 2 years on the MSIS physical scale and the MSWS. Demographic, disease related and proxy related variables that could potentially explain (dis)agreement between patients and proxies were investigated using linear regression analyses. Finally, in chapter 4.2 the search for the most accurate prediction of the patient follow up score on the MSIS physical scale is described using proxy follow up scores and potential other contributing patient and proxy related scores or variables.
17
1
Chapter 1
REFERENCES 1
Coles A. Multiple sclerosis. Pract. Neurol. 2009; 9: 118-26.
psychometric methods to clinical intuition. Brain 2000; 123 ( Pt 5): 1027-40.
2
Compston A, Coles A. Multiple sclerosis. Lancet 2002; 359: 1221-31.
3
Compston A, Coles A. Multiple sclerosis. Lancet 2008; 372: 1502-17.
12 Sharrack B, Hughes RA, Soudain S, Dunn G. The psychometric properties of clinical rating scales used in multiple sclerosis. Brain 1999; 122 ( Pt 1): 141-59.
4
Koutsouraki E, Costa V, Baloyannis S. Epidemiology of multiple sclerosis in Europe: a review. Int. Rev. Psychiatry 2010; 22: 2-13.
5
Multiple Sclerosis International Federation. Atlas of MS database 2013. http://www.atlasofms.org.
6
McDonald WI, Compston A, Edan G et al. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann. Neurol. 2001; 50: 121-7.
7
Polman CH, Reingold SC, Edan G et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the "McDonald Criteria". Ann. Neurol. 2005; 58: 840-6.
8
Polman CH, Reingold SC, Banwell B et al. Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria. Ann. Neurol. 2011; 69: 292-302.
9
Poser CM, Paty DW, Scheinberg L et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann. Neurol. 1983; 13: 227-31.
10 Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983; 33: 1444-52. 11 Hobart J, Freeman J, Thompson A. Kurtzke scales revisited: the application of
18
13 Thompson AJ, Hobart JC. Multiple sclerosis: assessment of disability and disability scales. J. Neurol. 1998; 245: 189-96. 14 Cutter GR, Baier ML, Rudick RA et al. Development of a multiple sclerosis functional composite as a clinical trial outcome measure. Brain 1999; 122 ( Pt 5): 871-82. 15 Bever CT, Jr., Grattan L, Panitch HS, Johnson KP. The Brief Repeatable Battery of Neuropsychological Tests for Multiple Sclerosis: a preliminary serial study. Mult. Scler. 1995; 1: 165-9. 16 Rao SM in collaboration with the Cognitive Function Study Group of the Nations Multiple Sclerosis Society. A manual for the Brief Repeatable Battery of Neuropsychological Tests in multiple sclerosis. 1990. 17 Rao SM. A manual for the brief, repeatable battery of neuropsychological tests in multiple sclerosis. 1991. 18 Strober L, Englert J, Munschauer F et al. Sensitivity of conventional memory tests in multiple sclerosis: comparing the Rao Brief Repeatable Neuropsychological Battery and the Minimal Assessment of Cognitive Function in MS. Mult. Scler. 2009; 15: 1077-84. 19 Buschke H, Fuld PA. Evaluating storage, retention, and retrieval in disordered memory and learning. Neurology 1974; 24: 1019-25.
General Introduction
20 Barbizet J, Cany E. Clinical and psychometrical study of a patient with memory disturbances. Int. J. Neurol. 1968; 7: 44-54.
29 Hobart J, Lamping D, Fitzpatrick R et al. The Multiple Sclerosis Impact Scale (MSIS-29): a new patient-based outcome measure. Brain 2001; 124: 962-73.
21 Smith A. Symbol digit modalities test. Los Angeles: Western Psychological Services., 1982.
30 van der Linden FA, Kragt JJ, Klein M et al. Psychometric evaluation of the multiple sclerosis impact scale (MSIS-29) for proxy use. J. Neurol. Neurosurg. Psychiatry 2005; 76: 1677-81.
22 Gronwall DM. Paced auditory serialaddition task: a measure of recovery from concussion. Percept. Mot. Skills 1977; 44: 367-73. 23 Boringa JB, Lazeron RH, Reuling IE et al. The brief repeatable battery of neuropsychological tests: normative values allow application in multiple sclerosis clinical practice. Mult. Scler. 2001; 7: 263-7. 24 Bosma L, Kragt JJ, Polman CH, Uitdehaag BM. Walking speed, rather than Expanded Disability Status Scale, relates to long-term patient-reported impact in progressive MS. Mult. Scler. 2013; 19: 326-33. 25 Giordano A, Pucci E, Naldi P et al. Responsiveness of patient reported outcome measures in multiple sclerosis relapses: the REMS study. J. Neurol. Neurosurg. Psychiatry 2009; 80: 1023-8. 26 McKenna SP, Doward LC, Twiss J et al. International development of the patientreported outcome indices for multiple sclerosis (PRIMUS). Value. Health 2010; 13: 946-51. 27 Schaffler N, Schonberg P, Stephan J et al. Comparison of patient-reported outcome measures in multiple sclerosis. Acta Neurol. Scand. 2013. 28 Twiss J, Doward LC, McKenna SP, Eckert B. Interpreting scores on multiple sclerosisspecific patient reported outcome measures (the PRIMUS and U-FIS). Health Qual. Life Outcomes. 2010; 8: 117.
31 Hobart JC, Riazi A, Lamping DL et al. Measuring the impact of MS on walking ability: the 12-Item MS Walking Scale (MSWS-12). Neurology 2003; 60: 31-6. 32 Benedict RH, Munschauer F, Linn R et al. Screening for multiple sclerosis cognitive impairment using a self-administered 15item questionnaire. Mult. Scler. 2003; 9: 95-101. 33 Sonder JM, Mokkink LB, van der Linden FA et al. Validation and interpretation of the Dutch version of the Multiple Sclerosis Neuropsychological Screening Questionnaire. J. Neurol. Sci. 2012; 320: 91-6. 34 Sharrack B, Hughes RA. The Guy's Neurological Disability Scale (GNDS): a new disability measure for multiple sclerosis. Mult. Scler. 1999; 5: 223-33. 35 Middleton LE, Kirkland SA, Mitnitski A, Rockwood K. Proxy reports of physical activity were valid in older people with and without cognitive impairment. J. Clin. Epidemiol. 2010; 63: 435-40. 36 Yasuda N, Zimmerman S, Hawkes WG, et al. Concordance of proxy-perceived change and measured change in multiple domains of function in older persons. J. Am. Geriatr. Soc. 2004; 52: 1157-62. 37 Irwin DE, Gross HE, Stucky BD et al. Development of six PROMIS pediatrics proxy-report item banks. Health Qual. Life Outcomes. 2012; 10: 22.
19
1
Chapter 1
38 Hilari K, Owen S, Farrelly SJ. Proxy and selfreport agreement on the Stroke and Aphasia Quality of Life Scale-39. J. Neurol. Neurosurg. Psychiatry 2007; 78: 1072-5. 39 Giesinger JM, Golser M, Erharter A et al. Do neurooncological patients and their significant others agree on quality of life ratings? Health Qual. Life Outcomes. 2009; 7: 87. 40 Jensen-Dahm C, Vogel A, Waldorff FB, Waldemar G. Discrepancy between selfand proxy-rated pain in Alzheimer's disease: results from the Danish Alzheimer Intervention Study. J. Am. Geriatr. Soc. 2012; 60: 1274-8. 41 van der Linden FA, Kragt JJ, Hobart JC et al. Proxy measurements in multiple sclerosis: agreement between patients and their partners on the impact of multiple sclerosis in daily life. J. Neurol. Neurosurg. Psychiatry 2006; 77: 1157-62. 42 van der Linden FA, D'hooghe MB, Nagels G et al. Proxy ratings from multiple sources: disagreement on the impact of multiple sclerosis on daily life. Eur. J. Neurol. 2008; 15: 933-9. 43 van der Linden FA, Kragt JJ, van Bon M et al. Longitudinal proxy measurements in multiple sclerosis: patient-proxy agreement on the impact of MS on daily life over a period of two years. BMC. Neurol. 2008; 8: 2. 44 van der Linden FA, Kragt JJ, Hobart JC et al. The size of the treatment effect: do patients and proxies agree? BMC. Neurol. 2009; 9: 12. 45 Bobholz JA, Rao SM. Cognitive dysfunction in multiple sclerosis: a review of recent developments. Curr. Opin. Neurol. 2003; 16: 283-8. 46 Rao SM, Leo GJ, Bernardin L, Unverzagt F. Cognitive dysfunction in multiple sclerosis.
20
I. Frequency, patterns, and prediction. Neurology 1991; 41: 685-91. 47 Amato MP, Portaccio E, Goretti B et al. Cognitive impairment in early stages of multiple sclerosis. Neurol. Sci. 2010; 31: S211-S214. 48 Rao SM, Leo GJ, Ellington L et al. Cognitive dysfunction in multiple sclerosis. II. Impact on employment and social functioning. Neurology 1991; 41: 692-6. 49 Rogers JM, Panegyres PK. Cognitive impairment in multiple sclerosis: evidencebased analysis and recommendations. J. Clin. Neurosci. 2007; 14: 919-27. 50 Chiaravalloti ND, DeLuca J. Cognitive impairment in multiple sclerosis. Lancet Neurol. 2008; 7: 1139-51. 51 Julian L, Merluzzi NM, Mohr DC. The relationship among depression, subjective cognitive impairment, and neuropsychological performance in multiple sclerosis. Mult. Scler. 2007; 13: 816. 52 van der HK, Spliethoff-Kamminga NG, Ruimschotel RP, Middelkoop HA, Visser LH. The relationship between self-reported executive performance and psychological characteristics in multiple sclerosis. Eur. J. Neurol. 2012; 19: 562-9. 53 Bruce JM, Arnett PA. Self-reported everyday memory and depression in patients with multiple sclerosis. J. Clin. Exp. Neuropsychol. 2004; 26: 200-14. 54 Carone DA, Benedict RH, Munschauer FE, III, Fishman I, Weinstock-Guttman B. Interpreting patient/informant discrepancies of reported cognitive symptoms in MS. J. Int. Neuropsychol. Soc. 2005; 11: 574-83.
General Introduction
55 Courts NF, Newton AN, McNeal LJ. Husbands and wives living with multiple sclerosis. J. Neurosci. Nurs. 2005; 37: 20-7. 56 Buchanan RJ, Radin D, Huang C. Burden among male caregivers assisting people with multiple sclerosis. Gend. Med. 2010; 7: 637-46. 57 Dunn J. Impact of mobility impairment on the burden of caregiving in individuals with multiple sclerosis. Expert. Rev. Pharmacoecon. Outcomes. Res. 2010; 10: 433-40. 58 Figved N, Myhr KM, Larsen JP, Aarsland D. Caregiver burden in multiple sclerosis: the impact of neuropsychiatric symptoms. J. Neurol. Neurosurg. Psychiatry 2007; 78: 1097-102. 59 Tripoliti E, Campbell C, Pring T, Taylor-Goh S. Quality of life in multiple sclerosis: should clinicians trust proxy ratings? Mult. Scler. 2007; 13: 1190-4.
60 Sprangers MA, Aaronson NK. The role of health care providers and significant others in evaluating the quality of life of patients with chronic disease: a review. J. Clin. Epidemiol. 1992; 45: 743-60. 61 Benedict RH, Cox D, Thompson LL, Foley F, Weinstock-Guttman B, Munschauer F. Reliable screening for neuropsychological impairment in multiple sclerosis. Mult. Scler. 2004; 10: 675-8. 62 O'Brien A, Gaudino-Goering E, Shawaryn M, Komaroff E, Moore NB, DeLuca J. Relationship of the Multiple Sclerosis Neuropsychological Questionnaire (MSNQ) to functional, emotional, and neuropsychological outcomes. Arch. Clin. Neuropsychol. 2007; 22: 933-48. 63 Vanotti S, Benedict RH, Acion L, Caceres F. Validation of the Multiple Sclerosis Neuropsychological Screening Questionnaire in Argentina. Mult. Scler. 2009; 15: 244-50.
21
1
22
CHAPTER 2
Validation of a patient reported outcome scale for patient and proxy use
24
CHAPTER 2.1
Validation and interpretation of the Dutch version of the Multiple Sclerosis Neuropsychological Screening Questionnaire
Judith M Sonder, Lidwine B Mokkink, Femke AH van der Linden, Chris H Polman, Bernard MJ Uitdehaag
Journal of the Neurological Sciences 2012;320:91–96
Chapter 2.1
ABSTRACT Background: The Multiple Sclerosis Neuropsychological Screening Questionnaire (MSNQ) was developed for screening of MS patients at risk for cognitive impairment with a patient self-report (MSNQ-P) and an informant version (MSNQ-I). The objective of this study was to validate the Dutch versions and determine their interpretability. Methods: The MSNQ was completed by 121 MS patients and their partners (informants). We investigated the factor structure, internal consistency and construct validity. Interrater reliability between MNSQ-P and MSNQ-I was investigated with the intraclass correlation coefficient (ICC) and Cohen’s kappa. For interpretability of both MSNQ versions we calculated sensitivity, specificity and cut-off scores. Receiver operating characteristic (ROC) curves with related areas under the curve (AUC) were used to evaluate the added value of combining both versions. Results: We found an unidimensional factor structure. Cronbach’s alphas were 0.93 and 0.94 for MSNQ-P and MSNQ-I, respectively. The ICC was 0.59 and kappas were ≤0.50. No cut-off score could be defined for the MSNQ-P because of low sensitivity. For the MSNQI, sensitivity was 0.75 and specificity 0.71 (AUC 0.80). The cut-off score was 21. ROC curve analyses showed no added value of the MSNQ-P when used in combination with the MSNQ-I. Conclusions: The MSNQ-I is preferred over the MSNQ-P to screen MS patients for cognitive impairment.
26
Validation and interpretation of the MSNQ
INTRODUCTION Cognitive impairment in multiple sclerosis (MS) is common, occurring in 40-70% of MS 1
patients. It can occur in any stage of the disease and it is associated with problems in
2.1
2
activities of daily living, employment, social functioning and a reduced quality of life.
Decline in cognitive functioning can be tested using neuropsychological tests. However, there are several disadvantages of neuropsychological testing. First, it is time-consuming and may be burden for some patients. Second, it is expensive since there is a need for qualified staff. Therefore a short screening instrument is needed to identify patients who might be cognitively impaired and in whom additional formal neuropsychological testing is useful.
3
Such a short screening instrument may even be used to indicate possible cognitive decline in MS patients in (long term) follow-up studies. 3
The Multiple Sclerosis Neuropsychological Screening Questionnaire (MSNQ) is a self-report questionnaire that was developed to screen MS patients at risk for cognitive impairment. The MSNQ consists of 15 items and reflects neuropsychological competence during activities in daily living. In addition to the MSNQ for patient self-report (MSNQ-P), an informant form was developed (MSNQ-I) to compare the opinion of the patient about cognitive problems in daily life with the opinion of an informant. 3-5
Studies in the United States
6
and Argentina tested the psychometric properties of both
versions of the MSNQ. In these studies moderate to high Cronbach’s alphas were found.
3-6
Test-retest correlations were only calculated in one study and they were also found to be 4
high. Construct validity was investigated by calculating the relationship of the MSNQ with neuropsychological testing and depression. All previous studies found high correlations between the MSNQ-P score and depression and low correlations between MSNQ-P score and cognitive impairment. In contrast, the MSNQ-I was positively correlated with cognitive impairment and not with depression of the patient. Comparison of MSNQ scores with an instrument that measures cognitive impairment provides information about the validity of the (change) scores of the MSNQ. Interpretability is 7
the degree to which one can assign qualitative meaning to quantitative scores. The Brief 27
Chapter 2.1
Repeatable Battery for Neuropsychological Tests (BRB-N) is widely used to measure cognitive impairment.
8-10
A recent study comparing neuropsychological test batteries concluded that
the BRB-N is reliable and sensitive to use in MS patients.
11
The BRB-N measures different
aspects of cognition in five subtests. In MSNQ validation studies cut-off scores were defined, based on measures of cognitive impairment, for the interpretability of MSNQ scores.
3,4,6
These cut-off scores were different in all studies. In most of the validation studies no discriminative value was found for the MSNQ-P because of low sensitivity.
3,5,6
The first objective of this study was to validate the Dutch version of the MSNQ-P and MSNQ-I in a large sample of Dutch MS patients and informants. The psychometric properties were investigated by determining the structure of the MSNQ, internal consistency and construct validity. In addition, interrater reliability and agreement between patient and informant scores were investigated. The second objective was to determine the interpretability of the MSNQ. Using sensitivity and specificity we tried to calculate cut-off scores for both versions to identify patients at risk for cognitive impairment. For that purpose, the BRB-N subtest scores were used as an external criterion. Receiver operating characteristic (ROC) curves with related area under the curve (AUC) were used to evaluate the added value of combining both versions. Table 2.1.1 Characteristics of the study sample
Malea Femalea Age in yearsb Relapsing Remitting (RR)a Secondary Progressive (SP)a Primary Progressive (PP)a Clinically Isolated Syndrome (CIS)a Disease duration in yearsb EDSSb High education (college/university)a Moderate education (secondary school)a Low education (primary school)a Completion of MSNQ in hospital settinga N (%) bMedian (IQR)
a
28
Patient
Proxy
46 (38) 75 (62) 53 (45-63) 43 (36) 40 (33) 34 (28) 4 (3) 15 (9.0-22.0) 4.5 (3.0-6.5) 37 (30) 40 (33) 42 (35) 121 (100)
73 (60) 48 (40) 55 (45-63) na na na na na na 38 (31) 42 (35) 40 (33) 87 (72)
Validation and interpretation of the MSNQ
METHODS Study sample The subjects in this study were 121 MS patients and 121 informants, partners of MS patients. Patients were participating in ongoing research projects at the MS Center of the VU University Medical Center Amsterdam and their visit was already scheduled for these projects. The medical ethical committee of the VU University Medical Center approved the study protocol. The characteristics of the study sample are shown in table 2.1.1. Translation of MSNQ A forward and backward translation procedure of the original English version was followed to 3
develop the Dutch version of the MSNQ patient and informant form. Both versions were translated into Dutch by two independent researchers followed by an independent backward translation into English. The backward translation was compared with the original version. In only three items minor differences were observed which were dealt with in consensus by adapting the wording of the Dutch version. Tests and procedures Patients were asked to complete the MSNQ-P and other questionnaires (described below) and were invited to perform a number of neuropsychological tests. The informant (partner of patient) was asked to complete the MSNQ-I in the hospital waiting room or it was completed at home and returned using a prepaid envelope. The MSNQ versions for patient self report and informant contain the same items but from another perspective. The 15 MSNQ items have 5 response options, 0 (does not occur) to 4 (very often, very disruptive). A total score is computed with a range from 0 to 60, a higher score indicates more cognitive problems. 12
The Hospital Anxiety and Depression Scale (HADS) is a 14-item self-report screening scale to indicate the possible presence of anxiety (7 items) and depressive states (7 items). For both continuous scales the range of scores is 0 to 21. The HADS is validated in different groups of
29
2.1
Chapter 2.1
Dutch subjects (three groups of healthy controls >18 years and general, somatic and psychiatric patient groups).
13
The BRB-N was developed as a short observational instrument to identify disturbances of cognitive domains in MS patients. The BRB-N (version A) consists of five subtests. Immediate 14
and delayed recall memory is assessed by the Selective Reminding Test (SRT); Visuospatial immediate and delayed recall memory is assessed by the Spatial Recall Test (SPART); Rao version of the Symbol Digit Modalities Test (SDMT)
10,16
15
The
examines processing speed and
concentration by primarily assessing complex and visual scanning and tracking; The Rao version of the Paced Auditory Serial Addition Test (PASAT)
9,17
measures information
processing speed and interference suppression in the 2 and 3 second tests; and the Word List 9
Generation (WLG) test assesses semantic verbal fluency. For each of the five subtests a score is obtained. An abnormal score for each subtest was defined with the most stringent criterion, 18
i.e. two standard deviations below the mean reported for Dutch healthy subjects. Patients were classified in three levels of cognitive functioning, no impaired subtest score (no cognitive impairment), 1 or 2 impaired subtest score(s) (group at risk) and 3 or more impaired subtest scores (cognitively impaired).
19
The Expanded Disability Status Scale (EDSS)
20
was used as a standardized neurological
examination. The EDSS assesses neurological impairment and disability and was performed by a trained doctor. The EDSS score varies between 0 and 10, a high score indicates more severe disability.
DATA ANALYSES Validity Factor analyses (principal component analyses) were done to investigate the factor structure of each MSNQ version separately. We hypothesized that the MSNQ items load on one general factor. In addition to the principal component analysis we used parallel analysis to compare the unrotated eigenvalues from a random sample, with the same number of cases and variables, with our data. If the eigenvalues from the principal component analysis were found
30
Validation and interpretation of the MSNQ
to be higher than the eigenvalues from the parallel analysis, it was significant and we could retain the factor. We calculated Cronbach’s alpha for both scales to determine their internal consistency. Cronbach’s alpha above 0.70 was considered to be appropriate.
21
2.1
Construct validity was tested by defining hypotheses. It examines whether the MSNQ score represents the theoretical concept of interest, i.e. neuropsychological competence. tested the following hypotheses based on results from earlier studies.
3-6
22
We
Since the data were
not normally distributed, Spearman correlations were used. 1.
There is a small positive correlation (i.e. correlation between 0 and 0.40) between MSNQP score and amount of impaired subtests on the BRB-N.
2.
There is a moderate and positive correlation (i.e. correlation between 0.40 and 0.70) between MSNQ-I score and amount of impaired subtests on the patient BRB-N.
3.
There is a moderate and positive correlation (i.e. correlation between 0.40 and 0.70) between MSNQ-P and patient HADS score.
4.
There is a small positive correlation (i.e. correlation between 0 and 0.40) between MSNQI and patient HADS score.
5.
The correlation between MSNQ-I and amount of impaired cognitive tests is higher than the correlation between MSNQ-P and impaired cognitive tests (difference ≥ 0.10).
6.
The correlation between MSNQ-P and patient depression is higher than the correlation between MSNQ-I and patient depression (difference ≥ 0.10).
Interrater reliability between patient and informant on scale level was evaluated using the intraclass correlation coefficient two-way mixed model for absolute agreement (ICCagreement).
23
Cohen’s weighted kappa (κw) was calculated to evaluate interrater reliability between patient 24
and informant on item level, using quadratic weights. The weighted kappas were calculated for all items. ICC and kappa above 0.70 were considered to be appropriate.
21
Agreement was measured with the Limits of Agreement. In a Bland and Altman plot the differences between patient and proxy scores were plotted against the mean score of each 25
pair on the MSNQ. The limits of agreement (LoA) were calculated using the mean difference per scale (đ) and the standard deviation of the mean difference (SDdifference): 31
Chapter 2.1
LoA = đ ± 1.96 SDdifference . Interpretability One of the objectives of this study was to calculate cut-off scores for the MSNQ that best classifies patients in the study sample as cognitively impaired. The scores on the BRB-N subtests were used as a criterion to classify patients in three levels of cognitive functioning. We calculated sensitivity and specificity for each MSNQ version separately and for that we used 3 or more impaired subtest scores as cutoff for cognitive impairment. In addition we repeated the analyses using 1 or more impaired subtest score(s) as cut off for cognitive impairment. We considered the optimal cut-off score for both MSNQ versions to be the value for which [1-sensitivity] + [1-specificity] is the smallest. Because the goal of this instrument is to screen for cognitive impairment, we considered sensitivity to be more important than specificity. Receiver operating characteristic (ROC) curves were constructed using logistic regression models in which the presence of cognitive impairment was the dependent variable and MSNQ-P and MSNQ-I were entered either alone or both together as independent variables. We compared the area under the curve (AUC) of MSNQ-P, MSNQ-I and the combined line. All statistical analyses were performed using the Statistical Package for Social Sciences (SPSS) version 14.0 (SPSS, Inc., Chicago, IL., USA) and Stata 11.1 (StataCorp, College Station, TX 2009).
RESULTS Missing items Three patients and 4 informants had one missing MSNQ item. In these cases, the mean item score of the scale was imputed. In one case more than a single item was missing and this case was excluded from analyses.
32
Validation and interpretation of the MSNQ
Psychometric properties The results of the factor analysis (principal component analysis) are displayed in table 2.1.2. The items of both MSNQ versions loaded high on their first factor. The eigenvalue of this factor was 7.5 for MSNQ-P and explained 50% of the variance. The highest loading was 0.83 for item 11 (i.e. failing to track two things at once) and the lowest loading was 0.50 for item 15 (i.e. excessive egocentric speech). The first factor of the MSNQ-I explained 56% of the total variance and the eigenvalue was 8.4. The highest loading for the informant version was 0.86 for the items 7, 8 and 11 (i.e forgetting instructions, needing frequent reminders and failing to track two things at once) and the lowest was 0.52 for item 13 (i.e. problems controlling impulses). In the parallel analyses was seen that the eigenvalues of this sample were much higher than the eigenvalue of a random sample. Some items also loaded on a second factor in the principal component analysis (eigenvalue MSNQ-P 1.35 and MSNQ-I 1.17), but when looking at the parallel analysis we could not retain this factor. Therefore, we can confirm the unidimensional structure of the MSNQ. Cronbach’s alpha analyses showed good internal consistency; 0.93 for MSNQ-P and 0.94 for MSNQ-I. Spearman correlations were calculated between MSNQ-P and MSNQ-I, and all other variables. The results of these correlations are presented in table 2.1.3 with variable means and standard deviations (SD). All hypotheses were confirmed concerning the construct validity: 1.
The correlation between MSNQ-P score and amount of impaired subtests on the BRB-N was 0.26, small positive.
2.
The correlation between MSNQ-I score and amount of impaired subtests on the BRB-N was 0.39, almost moderate and positive, although just below the hypothesized value of 0.40.
3.
Both patient depression (0.49) and anxiety (0.47) scales of the HADS showed a moderate and positive correlation with the MSNQ-P score.
4.
The correlations between MSNQ-I and patient anxiety and depression score were 0.36 and 0.33 respectively, both small positive.
33
2.1
Chapter 2.1 Table 2.1.2 Factor loadings from principal component analysis
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
Distractibility Problems with listening to others Slowed processing Forgetting appointments Forgetting what is read Forgetting shows or programs Forgetting instructions Needing frequent reminders Failing to follow through on planned activities Failing to answer questions coherently Failing to track two things at once Failing to follow conversations Problems controlling impulses Laughing or crying without cause Excessive egocentric speech
MSNQ-P
MSNQ-I
0.71 0.70 0.74 0.64 0.78 0.78 0.79 0.76 0.59 0.78 0.83 0.78 0.59 0.57 0.50
0.69 0.78 0.77 0.72 0.83 0.80 0.86 0.86 0.69 0.80 0.86 0.81 0.52 0.55 0.57
Table 2.1.3 Spearman correlations Correlations MSNQ-P MSNQ-I Age Gender patient Education (level) Disease duration (years) EDSSa HADS Anxiety HADS Depression BRB-N Global (amount impaired out of 5) SRT – immediate recall, long term storage SRT – immediate recall, cons. long term storage SRT – delayed recall SPART – immediate recall (3) SPART – delayed recall SDMT Pasat 3” – correct Pasat 2” – correct WLG a
Median, IQR
34
Mean 18.4 17.1 52.8 na na 16.9 4.5 4.9 4.1 0.6 46.0 35.6 46.1 29.0 8.6 19.8 6.8 47.1 24.4
SD 10.3 11.2 11.7 na na 8.8 3.0 - 6.5 3.2 3.5 1.2 14.8 15.9 16.0 15.5 2.9 5.1 2.4 15.1 7.4
MSNQ-P na 0.52 0.20 -0.03 -0.13 0.26 0.33 0.47 0.49 0.26 -0.28 -0.22 -0.17 -0.21 -0.13 -0.06 -0.12 -0.27 -0.17
MSNQ-I 0.52 na 0.19 -0.24 -0.13 0.17 0.38 0.33 0.36 0.39 -0.29 -0.34 -0.36 -0.37 -0.32 -0.02 -0.14 -0.34 -0.30
Validation and interpretation of the MSNQ
5.
The correlation between MSNQ-P and impaired BRB-N subtests was 0.26, for MSNQ-I this correlation was 0.39. The difference in correlation between the MSNQ versions and amount of impaired BRB-N subtests was 0.13, more than 0.10.
6.
The correlations between MSNQ-P and patient anxiety and depression were respectively 0.47 and 0.49, for the MSNQ-I these correlations were 0.33 and 0.36 respectively. The differences in correlations were both more than 0.10.
Interrater reliability between the patient and informant version on scale level was moderate. The ICC of MSNQ-P total score and MSNQ-I total score was 0.59 (95% CI: 0.46-0.69), which indicated that the MSNQ-P and MSNQ-I were not able to discriminate patients similarly. Interrater reliability on item level was also moderate. None of the kappa values reached the limit of 0.70. Weighted kappa values (κw) for item scores are shown in table 2.1.4 with their confidence intervals (CI). The Bland and Altman plot is displayed in figure 2.1.1. The mean difference was +1.3, meaning that the mean score of patients was 1.3 points higher than the mean score of informants (i.e. systematic error). The limits of agreement (i.e. random error) were between -17.7 and +20.4 (range -60 to +60).
Table 2.1.4 Weighted kappa: Interrater reliability of MSNQ items
MSNQ 1 MSNQ 2 MSNQ 3 MSNQ 4 MSNQ 5 MSNQ 6 MSNQ 7 MSNQ 8 MSNQ 9 MSNQ 10 MSNQ 11 MSNQ 12 MSNQ 13 MSNQ 14 MSNQ 15
κw
95% CI
0.32 0.36 0.25 0.39 0.50 0.50 0.43 0.43 0.28 0.47 0.49 0.46 0.42 0.42 0.35
0.11–0.49 0.17–0.52 0.07–0.44 0.22–0.58 0.33–0.62 0.35–0.65 0.27–0.58 0.24–0.58 0.11–0.44 0.30–0.62 0.32–0.64 0.28–0.61 0.26–0.57 0.20–0.60 0.17–0.55
35
2.1
Chapter 2.1
Interpretability Patients were classified in three levels of cognitive functioning, no impaired subtest score (no cognitive impairment), 1 or 2 impaired subtest scores (group at risk) and 3 or more impaired subtest scores (cognitively impaired). Of all MS patients in this sample, 39 patients (32%) had one or more abnormal scored subtests and 12 patients (10%) were classified as cognitively impaired because of 3 or more abnormal subtests (table 2.1.5). We could not define a cut-off score for the MSNQ-P because sensitivity (0.42) was found to be too low and not discriminative for cognitive impairment (3 or more impaired subtest scores). The optimal cut-off score in our data for the MSNQ-I was 21, with a sensitivity of 0.75 and a specificity of 0.71. The ROC curves are shown in figure 2.1.2. The AUC of MSNQ-I was 0.80 (95% CI 0.65-0.95), equal to the combined line of MSNQ-P and MSNQ-I. Also when including the group at risk (1 or 2 abnormal subtest scores) in the cognitively impaired group, the MSNQ-I was more discriminative compared to the MSNQ-P and equal to the combination of MSNQ-I and MSNQ-P. Figure 2.1.1 Limits of Agreement MSNQ
36
Validation and interpretation of the MSNQ Table 2.1.5 Classification of BRB-N subtest results BRB-N
N
no impaired subtest score 1 or 2 impaired subtest scores 3 or more impaired subtest scores
82 27 12
68% 22% 10%
2.1
Figure 2.1.2 ROC curves
DISCUSSION In this study about the validation and interpretation of the Dutch version of the MSNQ we found a clear unidimensional factor structure for both versions. Internal consistency was good for both scales. Assessment of construct validity showed that all hypotheses based on previous studies were confirmed.
3-6
As expected, we found a higher correlation between the
MSNQ-P and patient anxiety and depression (measured with the HADS) than between MSNQP and impaired cognitive subtests (measured with the BRB-N). In contrast, MSNQ-I correlates 37
Chapter 2.1
higher with the amount of impaired subtests compared to the correlation with anxiety or depression. Therefore, we recommend to use the MNSQ-I version. The interrater reliability of the total score and the item scores between the patient and informant versions was moderate. The MSNQ-P is not able to identify cognitive disabled patients as good as the MSNQ-I does. Interrater agreement was poor. Although the systematic error between patient and informant version was low, the random error (i.e. Limits of Agreement) was large. Ninety-five percent of the differences between patient and informant were between -17.7 and +20.4. Since the range of the scale was 0 to 60, and therefore the difference between patient and informant could have a maximum variation between -60 and 60, the random error is huge. No significant differences were found between the group of informants participating in the hospital or at home (data not shown). However, it cannot be excluded that better results might have been obtained under more strictly controlled conditions. In some earlier published studies about the MSNQ researchers found low sensitivity for the MSNQ-P and no cut-off score could be defined.
3,5,6
Also in this study we found a sensitivity
that was too low to define a clear cut-off score for the patient version. The sensitivity and specificity of the informant version were better than the patient version. We could define a cut-off score of 21 for the Dutch version of the MSNQ-I, which was almost equal to results of 4
earlier studies in North America. Analyses showed that when combining MSNQ-I and MSNQP there was no added value of the patient version. Some challenges were encountered when comparing our data with existing data because of the differences in the design and execution of previous studies. There were differences in measurement instruments and patient groups. Cognitive tests, used as the external standard, were not the same in all published studies about the sensitivity and determining cut-off 6
scores for the MSNQ. Only the Argentinean study used the BRB-N, but the classification of impaired patients was different compared to our study. In our study only a small number of patients were classified as cognitively impaired according to neuropsychological testing. Patients were participating in ongoing research projects and it cannot be excluded that this may have caused some selection bias towards less cognitively impaired patients. Furthermore
38
Validation and interpretation of the MSNQ
it should be noted that the BRB-N as used in this study is missing a measure of higher executive functioning, which could have influenced the results. There were also differences in measuring depression in the published studies. Other researchers used Beck’s Depression Inventory – Fast screen (BDI-FS),
26
some in combination with the Center for Epidemiologic
27
Studies Depression scale (CES-D) while we used the HADS for measuring depression as well 3-6
as anxiety.
These issues may explain some of the observed differences in the results.
Future studies would benefit from selecting more MS patients with cognitive impairment, so that the relationship between cognitive testing and the MSNQ can be tested more specifically. In addition we would recommend to collect test-retest data. These studies should also address the origin of disagreement between patients and informants on the MSNQ. With the growing interest in self assessment scales, understanding the background of disagreement is of increasing importance. It has been subject to research before and possible explaining factors are depression, anxiety and caregiver burden.
28-30
In summary, in this study we investigated the validity and interpretability of the MSNQ. The main outcome is that the MSNQ-I is more promising to screen for cognitive impairment in MS patients. The patient version has no added value, so when screening for cognitive impairment in MS the MSNQ informant version is preferred.
39
2.1
Chapter 2.1
REFERENCES 1
Rao SM, Leo GJ, Bernardin L, Unverzagt F. Cognitive dysfunction in multiple sclerosis. I. Frequency, patterns, and prediction. Neurology 1991; 41: 685-91.
2
Phillips LH, Saldias A, McCarrey A et al. Attentional lapses, emotional regulation and quality of life in multiple sclerosis. Br. J. Clin. Psychol. 2009; 48: 101-6.
3
Benedict RH, Munschauer F, Linn R et al. Screening for multiple sclerosis cognitive impairment using a self-administered 15item questionnaire. Mult. Scler. 2003; 9: 95-101.
4
Benedict RH, Cox D, Thompson LL et al. Reliable screening for neuropsychological impairment in multiple sclerosis. Mult. Scler. 2004; 10: 675-8.
5
O'Brien A, Gaudino-Goering E, Shawaryn M et al. Relationship of the Multiple Sclerosis Neuropsychological Questionnaire (MSNQ) to functional, emotional, and neuropsychological outcomes. Arch. Clin. Neuropsychol. 2007; 22: 933-48.
6
Vanotti S, Benedict RH, Acion L, Caceres F. Validation of the Multiple Sclerosis Neuropsychological Screening Questionnaire in Argentina. Mult. Scler. 2009; 15: 244-50.
7
Lohr KN, Aaronson NK, Alonso J et al. Evaluating quality-of-life and health status instruments: development of scientific review criteria. Clin. Ther. 1996; 18: 97992.
8
Bever CT, Jr., Grattan L, Panitch HS, Johnson KP. The Brief Repeatable Battery of Neuropsychological Tests for Multiple Sclerosis: a preliminary serial study. Mult. Scler. 1995; 1: 165-9.
40
9
Rao SM in collaboration with the Cognitive Function Study Group of the Nations Multiple Sclerosis Society. A manual for the Brief Repeatable Battery of Neuropsychological Tests in multiple sclerosis. 1990.
10 Rao SM. A manual for the brief, repeatable battery of neuropsychological tests in multiple sclerosis. 1991. 11 Strober L, Englert J, Munschauer F et al. Sensitivity of conventional memory tests in multiple sclerosis: comparing the Rao Brief Repeatable Neuropsychological Battery and the Minimal Assessment of Cognitive Function in MS. Mult. Scler. 2009; 15: 1077-84. 12 Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr. Scand. 1983; 67: 361-70. 13 Spinhoven P, Ormel J, Sloekers PP et al. A validation study of the Hospital Anxiety and Depression Scale (HADS) in different groups of Dutch subjects. Psychol. Med. 1997; 27: 363-70. 14 Buschke H, Fuld PA. Evaluating storage, retention, and retrieval in disordered memory and learning. Neurology 1974; 24: 1019-25. 15 Barbizet J, Cany E. Clinical and psychometrical study of a patient with memory disturbances. Int. J. Neurol. 1968; 7: 44-54. 16 Smith A. Symbol digit modalities test: manual. Los Angeles: Western Psychological Services; 1982. 17 Gronwall DM. Paced auditory serialaddition task: a measure of recovery from concussion. Percept. Mot. Skills 1977; 44: 367-73.
Validation and interpretation of the MSNQ
18 Boringa JB, Lazeron RH, Reuling IE et al. The brief repeatable battery of neuropsychological tests: normative values allow application in multiple sclerosis clinical practice. Mult. Scler. 2001; 7: 263-7. 19 Sepulcre J, Vanotti S, Hernandez R et al. Cognitive impairment in patients with multiple sclerosis using the Brief Repeatable Battery-Neuropsychology test. Mult. Scler. 2006; 12: 187-95. 20 Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983; 33: 1444-52. 21 Nunally J, Bernstein I. Psychometric theory. New York: McGraw-Hill, 1994. 22 Guyatt GH, Feeny DH, Patrick DL. Measuring health-related quality of life. Ann. Intern. Med. 1993; 118: 622-9. 23 McGraw K, Wong S. Forming inferences about some intraclass correlation coefficients. Psychol Methods 1996; 1: 3046. 24 Cohen J. Weighted kappa: nominal scale agreement with provision for scaled disagreement or partial credit. Psychol. Bull. 1968; 70: 213-20.
25 Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: 307-10. 26 Beck A, Steer R, Brown G. BDI-Fast Screen for medical patients: manual. San Antonio, TX: Psycholocial corporation, 2000. 27 Radloff L. The CES-D scale: a self report depression scale for research in the general population. Appl Psychol Meas 1977; 1: 385-401. 28 Carone DA, Benedict RH, Munschauer FE, III et al. Interpreting patient/informant discrepancies of reported cognitive symptoms in MS. J. Int. Neuropsychol. Soc. 2005; 11: 574-83. 29 Sonder J, Bosma L, van der Linden F et al. Proxy measurements in multiple sclerosis: agreement on different patient-reported outcome scales. Mult. Scler. 2012; 18: 196201. 30 van der Hiele K, Spliethoff-Kamminga NG, Ruimschotel RP et al. The relationship between self-reported executive performance and psychological characteristics in multiple sclerosis. Eur. J. Neurol. 2012; 19: 562-9.
41
2.1
42
CHAPTER 3
Agreement and disagreement on five patient reported outcome scales
44
CHAPTER 3.1
Proxy measurements in multiple sclerosis: Agreement on different patient reported outcome scales
Judith M Sonder, Libertje VAE Bosma, Femke AH van der Linden, Dirk L Knol, Chris H Polman, Bernard MJ Uitdehaag
Multiple Sclerosis 2012;18:196-201
Chapter 3.1
ABSTRACT Background: Patient-reported outcome (PRO) scales are often used in multiple sclerosis (MS) research. Full understanding of items can be influenced by disease worsening, mood disturbances and cognitive problems of the MS patient. Earlier research with the Multiple Sclerosis Impact Scale (MSIS-29) showed that proxy respondents (i.e. partners of patients) can provide useful information. Objective: To determine agreement between patients and proxy respondents on different MS PRO scales. Methods: 139 Patients and partners completed the MSIS-29 (Physical and Psychological scale),
Multiple
Sclerosis
Walking
Scale
(MSWS-12),
Multiple
Sclerosis
Neuropsychological Screening Questionnaire (MSNQ) and Guy’s Neurological Disability Scale (GNDS). We calculated the mean difference and intra-class correlation coefficients (ICC) on scale level and weighted kappas (κw) on item level. Results: On all scales, except MSNQ, the partner score was higher. ICCs were good for MSWS, GNDS and MSIS Physical, and moderate for MSNQ and MSIS Psychological. κw was excellent for MSWS items, fair to good for GNDS, MSIS Physical and MSIS Psychological items, and poor for MSNQ items. Conclusion: Partners of patients with MS can be a useful source of information for several PRO scales, especially when the focus is on physical functioning. For psychological functioning this seems to be less reliable.
46
Agreement on different PRO scales
INTRODUCTION Multiple Sclerosis (MS) is a chronic disease with a variable pattern. Many instruments and scales have been developed to look at effects of MS on physical state, psychological state, cognition, mood and quality of life. The measurement instruments can be ratings of clinicians or patient reported outcome (PRO) scales. For some of these latter instruments there are also informant versions available, developed to be completed by healthcare providers or proxy respondents. Researchers in MS often have to deal with loss of information and loss to follow-up as a result of disease burden. The time points of this methodological problem are not random, but are often critical points in the disease of the patient in which important information about the health status should be collected. Moreover, a full understanding of self-report forms and the ability of the patient to respond adequately can be influenced by disease worsening, fatigue, mood disturbances and cognitive problems due to MS.
1
Therefore including proxy
respondents (i.e. partners of patients) to evaluate the disease state of the patient might be useful to continue reliable follow-up research.
2
3-7
We described in previous studies with relative small groups of patients
that there is
8
agreement on the Multiple Sclerosis Impact Scale (MSIS-29) between patients and proxy respondents on group level. The objective of this study was to extend this research to other PRO scales concerning MS. To investigate this objective we used a range of MS scales from what we hypothesized was more easy to observe (walking ability, physical functioning), intermediate difficult to observe (overall disability) and difficult to observe (psychological functioning). We hypothesized that agreement was better on well-observable scales and more differences would be seen on psychological scales. We were also interested in the size and direction of possible differences and the influence of the level of disability.
47
3.1
Chapter 3.1
MATERIALS AND METHODS Patients and test procedures The subjects (18 years or older) in this study were MS patients and their partners, functioning as proxy respondents. When the partner could not come to the hospital, the PRO scales were completed at home and the partner was interviewed by phone. The medical ethical committee of the VU University Medical Center approved the study protocol. Tests The MSIS-29 measures disease impact of MS on daily life and can be divided into two subscales; a physical scale (MSIS Physical) which consists of 20 items and a psychological scale 9
(MSIS Psychological) with 9 items. The Multiple Sclerosis Walking Scale (MSWS-12) is a patient-based measure of walking ability in MS with a 12-item scale, restricted to patients who can walk. Guy’s Neurological Disability Scale (GNDS)
10
is an interview-based
questionnaire measuring disability in MS. The items include cognition, mood, vision, speech, swallowing, upper limb function, lower limb function, bladder function, bowel function, sexual function, fatigue and ‘others’. The Multiple Sclerosis Neuropsychological Screening 11
Questionnaire (MSNQ) is a brief self-administered questionnaire with 15 items that reflects neuropsychological competence in patients with MS during activities of daily living. The Expanded Disability Status Scale (EDSS)
12
assesses neurological impairment and disability.
EDSS score ranges between 0 and 10, a higher score indicates more severe disability. Data analyses All statistical analyses were performed using the Statistical Package for Social Sciences (SPSS) version 14.0 (SPSS, Inc., Chicago, Ill., USA) and Stata 11.1 (StataCorp, College Station, TX 2009). Agreement at scale level As an indicator of agreement at scale level the mean difference (đ) between patient and proxy was calculated for all scale scores. A positive mean indicated a higher scale score for the
48
Agreement on different PRO scales
patient on a scale, a negative mean was a higher score reported by the proxy respondent. To be able to compare the different scales, all scores were converted to a 0-100 scale using the formula:
13
100 x (observed score – minimum score) maximum score – minimum score For all scales paired sample t-tests were used to see whether the patient and proxy scores were significantly different. With analyses of variance (ANOVA) we tested whether there were differences between the proxy respondents participating in the hospital and at home and also between more (EDSS≥6) and less (EDSS<6) disabled patients. In Bland & Altman plots the differences between patient-proxy paired scale scores were plotted against the mean score of each pair. Furthermore limits of agreement (LoA) were calculated using the mean difference per scale (đ) and the standard deviation of the 14
difference (SDdifference):
đ ± 1.96 SDdifference .
Internal consistency of each scale was assessed using Cronbach’s alpha. Pearson correlations were computed as a measure of the statistical dependence between the different patient scales and separate for the proxy scales. Interobserver reliability between ratings of patients and proxy respondents was calculated using the intraclass correlation coefficients (ICC) twoway mixed model for absolute agreement. acceptable.
15
An ICC higher than 0.70 was considered to be
16
Agreement at item level Cohen’s weighted kappa was used to evaluate inter-rater agreement on item level.
17
For
interpretation, the mean weighted kappa’s of the five scales were classified according to Fleiss as excellent (κw>0.75), fair to good (0.40≤κw≤0.75) and poor (κw<0.40).
18
49
3.1
Chapter 3.1
RESULTS Study sample The characteristics of the study sample are shown in table 3.1.1. Of the proxy respondents, 66% was attending the patient in the hospital visit and 34% participated at home. Data analyses Agreement at scale level The results of the agreement between patient and proxy respondent at scale level are shown in table 3.1.2. A significant difference between patient and proxy score (p<0.05) was found for the MSIS-29 Psychological scale and the GNDS. There were no significant differences in agreement between couples in which the proxy completed the questionnaires at home compared to those who came to the hospital and also no differences were found between groups with more or less disabled patients.
Table 3.1.1 Characteristics of the study sample
Malea Female a Age in years b Relapsing Remitting (RR)a Secondary Progressive (SP)a Primary Progressive (PP)a Clinically Isolated Syndrome (CIS)a Disease duration in yearsb EDSSb EDSS <6a EDSS ≥6a MSIS-29 Physical (20 items)b MSWS (12 items)b GNDS (12 items)b MSNQ (15 items)b MSIS-29 Psychological (9 items)b a
N (%) bMedian (IQR) cProxy report of patient data
50
Patient
Proxy
53 (38.1) 86 (61.9) 53.0 (43.0-62.0) 58 (41.7) 41 (29.5) 36 (25.9) 4 (2.9) 15.0 (9.0-22.0) 4.5 (2.5-6.5) 84.0 (60.0) 54.0 (39.0) 33.8 (12.5-53.8) 39.6 (12.5-69.8) 21.7 (12.9-33.3) 28.3 (20.0-40.0) 19.4 (8.3-33.3)
84 (60.4) 55 (39.6) 54.0 (44.0-62.0) na na na na na na na na 36.3 (16.3-54.1)c 44.8 (10.4-74.5)c 25.0 (12.9-33.3)c 23.3 (15.0-38.3)c 25.0 (13.2-44.4)c
Agreement on different PRO scales Table 3.1.2 Agreement between patient and proxy respondent at scale score level
MSIS Physical MSWS† GNDS MSNQ MSIS Psychological †
N (pairs)
đ
SDdifference
Range
138 115 137 134 138
-1.1 -1.3 -1.6 1.8 -6.6
14.7 13.9 7.6 15.9 18.8
-67.5 ; 38.8 -39.6 ; 31.3 -35.0 ; 21.7 -58.3 ; 45.0 -66.7 ; 44.4
P 0.37 0.32 0.02 0.19 <0.001
Scale is restricted to patients who can walk
3.1
Figures 3.1.1 and 3.1.2 show examples of patterns in Bland and Altman plots. On the MSIS Physical scale the lower points of agreement were found around the median, indicating that agreement was better for higher and lower scores. The mean difference was found to be small. The MSWS (figure 3.1.1) was scored on the total scale range, the plot showed that there was better agreement on the beginning (no problems with walking) and at the end of the scale (walking is extremely difficult). The GNDS was scored in small limits of agreement. The differences between patient and proxy scores on the MSNQ (figure 3.1.2) were variable and there were fewer scores in the higher part of the scale. The MSIS Psychological scale showed good agreement on low scale scores, but agreement decreased with an increasing scale score. Proxy respondents scored higher than patients, with a large mean difference compared to the other scales (-6.6).
Table 3.1.3 Pearson correlations of patient and proxy scales and Cronbach’s alpha on the diagonal MSIS Physical Patient scales MSIS Physical MSWS GNDS MSNQ MSIS Psychological Proxy scales MSIS Physical MSWS GNDS MSNQ MSIS Psychological
MSIS Psychological
MSWS
GNDS
MSNQ
0.97 0.85 0.81 0.45 0.73
0.97 0.73 0.38 0.52
0.82 0.49 0.66
0.92 0.66
0.88
0.96 0.89 0.81 0.52 0.66
0.98 0.75 0.40 0.48
0.80 0.65 0.64
0.94 0.61
0.90
51
Chapter 3.1 Figure 3.1.1 Bland and Altman MSWS
Figure 3.1.2 Bland and Altman MSNQ
52
Agreement on different PRO scales
Table 3.1.3 presents the correlations of the patient and proxy scales. All correlations were positive and showed that there were strong similarities between the observed correlations for patient and informant scales. The tables also contain Cronbach’s alpha, all alpha’s were ≥0.80. The ICC’s are shown in table 3.1.4 with their 95% confidence interval (CI). The coefficients were good for MSWS, MSIS-Physical and GNDS. For MSIS-Psychological and MSNQ the
3.1
correlations between patient and partner ratings were small.
Table 3.1.4 Intraclass correlation coefficients (ICC) and mean weighted kappas (κw)
MSIS-29 Physical (20) MSWS (12) GNDS (12) MSNQ (15) MSIS-29 Psychological (9)
N (pairs)
ICC
95% CI
κw mean
Range
138 115 137 134 138
0.82 0.91 0.86 0.59 0.55
0.76 – 0.87 0.87 – 0.94 0.80 – 0.90 0.47 – 0.69 0.40 – 0.67
0.61 0.78 0.64 0.39 0.44
0.47 – 0.77 0.69 – 0.89 0.39 – 0.92 0.28 – 0.50 0.18 – 0.68
Agreement at item level Mean Cohen’s weighted kappa varied per scale (table 3.1.4). Excellent agreement was found for MSWS items, the lowest kappa for an item on this scale was 0.69. Fair to good agreement was found for GNDS, MSIS Physical and MSIS Psychological. The MSNQ showed poor agreement, no kappa higher then 0.50 was found and 8 of 15 items showed a kappa smaller than 0.40.
DISCUSSION The aim of this study was to explore the agreement between patients and proxy respondents on different PRO scales for MS. As we hypothesized agreement was seen on scales measuring physical impact and walking ability of the MS patient, with high ICC values and good to excellent weighted kappa’s. Our data supports previous findings that proxy respondents can
53
Chapter 3.1 3
rate the disease status of patients in an accurate way, when the focus is on specific and 2
observable disease aspects or (physical) limitations.
For psychological impact and functioning (MSIS Psychological and MSNQ) and overall disability (GNDS) the differences between patients and proxy respondents were large and dependent of scale score. Factors influencing the differences on psychological scales should be further investigated using longitudinal data of PRO scales, cognitive testing and evaluation of mood and anxiety. We found no differences between groups with more or less disabled patients, indicating that proxy respondents can provide useful information in both groups. However, it is important to stress that results of this study may not be applicable to individual patient-proxy couples. Further research to identify individual variables that can predict agreement or disagreement more accurate is ongoing. Concluding, we confirmed that partners of MS patients can be a useful source of information for several PRO scales, especially when the focus is on physical functioning. Further research on factors that can influence (dis)agreement in comparison with additional variables and follow-up data is required.
54
Agreement on different PRO scales
REFERENCES 1. Tripoliti E, Campbell C, Pring T, Taylor-Goh S. Quality of life in multiple sclerosis: should clinicians trust proxy ratings? Mult Scler 2007; 13:1190-1194. 2. Sprangers MA, Aaronson NK. The role of health care providers and significant others in evaluating the quality of life of patients with chronic disease: a review. J Clin Epidemiol 1992; 45:743-760. 3. van der Linden FA, Kragt JJ, Klein M et al. Psychometric evaluation of the multiple sclerosis impact scale (MSIS-29) for proxy use. J Neurol Neurosurg Psychiatry 2005; 76:1677-1681. 4. van der Linden FA, Kragt JJ, Hobart JC et al. Proxy measurements in multiple sclerosis: agreement between patients and their partners on the impact of multiple sclerosis in daily life. J Neurol Neurosurg Psychiatry 2006; 77:1157-1162. 5. van der Linden FA, D'hooghe MB, Nagels G et al. Proxy ratings from multiple sources: disagreement on the impact of multiple sclerosis on daily life. Eur J Neurol 2008; 15:933-939. 6. van der Linden FA, Kragt JJ, van Bon M et al. Longitudinal proxy measurements in multiple sclerosis: patient-proxy agreement on the impact of MS on daily life over a period of two years. BMC Neurol 2008; 8:2. 7. van der Linden FA, Kragt JJ, Hobart JC et al. The size of the treatment effect: do patients and proxies agree? BMC Neurol 2009; 9:12. 8. Hobart J, Lamping D, Fitzpatrick R et al. The Multiple Sclerosis Impact Scale (MSIS-29): a new patient-based outcome measure. Brain 2001; 124:962-973.
9. Hobart JC, Riazi A, Lamping DL et al. Measuring the impact of MS on walking ability: the 12-Item MS Walking Scale (MSWS-12). Neurology 2003; 60:31-36. 10. Sharrack B, Hughes RA. The Guy's Neurological Disability Scale (GNDS): a new disability measure for multiple sclerosis. Mult Scler 1999; 5:223-233. 11. Benedict RH, Munschauer F, Linn R et al. Screening for multiple sclerosis cognitive impairment using a self-administered 15item questionnaire. Mult Scler 2003; 9:95101. 12. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: An expanded disability status scale (EDSS). Neurology 1983, 33; 1444-1452 13. Hobart JC, Riazi A, Lamping DL et al. Improving the evaluation of therapeutic interventions in multiple sclerosis: development of a patient-based measure of outcome. Health Technol Assess 2004; 8:iii, 1-iii,48. 14. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1:307-310. 15. McGraw KO, Wong SP. Forming inferences about some intraclass correlation coefficients. Ps.meth 1996; 1:30-46. 16. Nunally JC. Psychometric theory. New York: McGraw-Hill, 1978. 17. Cohen J. Weighted kappa: nominal scale agreement with provision for scaled disagreement or partial credit. Psychol Bull 1968; 70:213-220. 18. Fleiss JL.. Statistical methods for rates and proportions, second edition. New York: John Wiley & Sons, 1981.
55
3.1
56
CHAPTER 3.2
Analyzing differences between patient and proxy on patient reported outcomes in multiple sclerosis
Judith M Sonder, Rebecca Holman, Dirk L Knol, Libertje VAE Bosma, Chris H Polman, Bernard MJ Uitdehaag
Journal of the Neurological Sciences 2013;334;143-147
Chapter 3.2
ABSTRACT Background: Proxy respondents, partners of multiple sclerosis (MS) patients, can provide valuable information on the MS patients’ disease. In an earlier publication we found relatively good agreement on patient reported outcomes (PROs) measuring physical impact and functioning, but we found large differences on (neuro)psychological scales. Objective: To identify patient and proxy related variables explaining differences between patients’ and proxies’ ratings on five PROs. Methods: We report on data from 175 MS patients and proxy respondents. Regression analyses were performed, using as dependent variable the mean differences on five scales: Physical and Psychological scale of the Multiple Sclerosis Impact Scale (MSIS-29), the Multiple Sclerosis Walking Scale (MSWS), Guy’s Neurological Disability Scale (GNDS) and the Multiple Sclerosis Neuropsychological Screening Questionnaire (MSNQ). The independent variables were patient, proxy and disease related variables. Results: Caregiver strain was significantly related to differences between patient and proxy scores for all five PROs. A higher level of patient anxiety on the HADS was linked to larger differences on all PROs except the GNDS. In addition, cognitive functioning, proxy depression, walking ability, proxy gender and MS related disability were contributing to the discrepancies. Conclusion: We found several patient and proxy factors that may contribute to discrepancies between patient and proxy scores on MS PROs. The most important factor is caregiver burden.
58
Analyzing differences between patient and proxy
INTRODUCTION Researchers considering multiple sclerosis (MS) are becoming increasingly interested in patient reported outcomes (PRO) as measures of disease impact, physical functioning, psychological functioning and quality of life.
1-5
Many PROs can be completed by the MS
patient or a proxy respondent. A proxy respondent is the partner or a close family member of 6
the MS patient. Proxy respondents can provide valuable information on MS patients’ disease when patients cannot participate in research projects due to their disease status, cognitive functioning or mood disorders.
7,8
In a previous study, we investigated the agreement between patients and proxy respondents 7
on MS PROs measuring physical functioning, psychological functioning and impact of MS. We found relatively good agreement on scales measuring physical impact and functioning, with only small and statistically non-significant differences between patients and proxies. However, we found large and statistically significant differences on PROs measuring psychological and neuropsychological functioning and impact of MS. Researchers aiming to explain the differences between patients’ and proxies’ assessments have focused on patient related factors, such as cognition and mood, which can reduce the reliability of MS patients’ self-assessment.
9,10
However, coping with MS symptoms is not only
challenging for patients. A range of psychological, physical and financial changes can also affect an MS patient’s partner and family and change relationships and caregiving roles.
11
These factors can cause MS patients’ partners long-term stress and influence their physical health, mental health and quality of life.
11-14
Hence, it is important for researchers to consider
both the patient’s disease status and their proxy’s physical health, mental health and caregiver strain when examining discrepancies between patient and proxy ratings on PROs. The objective of this study was to examine whether MS patients’ cognitive status, physical status and levels of depression and anxiety and their proxies’ health, levels of stress and depression explained the differences between patient and proxy ratings of MS patients’ health and the impact of MS on five MS PROs.
7
59
3.2
Chapter 3.2
METHODS Subjects and MS patient reported outcomes As previously described, the subjects were MS patients and their partners. The partners 7
functioned as proxy respondents. All subjects were aged 18 years or older. The patients had been diagnosed as having MS or clinically isolated syndrome ascertained by either the Poser or the revised McDonald criteria.
15,16
They visited the outpatients’ clinic of the VU University
Medical Center Amsterdam between November 2008 and December 2010. We contacted the patients and proxies to ask them to participate in this study before the patient visited the hospital. If they agreed, they completed the questionnaires during the patient’s visit to the hospital. If a proxy could not come to the hospital, he or she completed the questionnaires at home and a member of the research team interviewed him or her by telephone. The medical ethical committee of the VU University Medical Center approved the study protocol and we obtained informed consent from all participants. Patients and proxies completed five MS PROs with respect to the patients’ health status and the impact of MS on their daily lives. These were: (1) the physical scale of the Multiple 17
18
Sclerosis Impact Scale (MSIS-29); (2) the Multiple Sclerosis Walking Scale (MSWS-12); (3) 19
the Guy’s Neurological Disability Scale (GNDS); (4) the Multiple Sclerosis Neuropsychological 20
Screening Questionnaire (MSNQ);
and (5) the psychological scale of the MSIS-29.
17
On all
PROs, a higher score indicates a worse clinical situation. Patients’ cognitive status, physical status and levels of depression and anxiety We assessed patients’ cognitive status using the Brief Repeatable Battery for Neuropsychological Tests (BRB-N),
21,22
their physical status using the Expanded Disability
23
Status Scale (EDSS) and their levels of depression and anxiety using the Hospital Anxiety and Depression Scale (HADS).
24
Proxies’ general health status, levels of depression and anxiety and caregiver strain 25-
We assessed the proxies’ own general health status using Short-Form Health Survey (SF-36) 27
and their own levels of depression and anxiety using the HADS. We measured caregiver 28
strain using the Caregiver Strain Index (CSI) and the Self-Rated Burden scale (SRB).
60
29
Analyzing differences between patient and proxy
Socio-demographic variables Patients’ and proxies’ education was divided in three categories, low education (primary school), average education (secondary school) and high education (college or university). Statistical analyses As previously described, we converted all scores on the MS PROs to a 0 to 100 scale to enable us to compare the results on the five PROs.
7,30
We calculated the mean differences between
patient and proxy scores and limits of agreement for the five MS PROs. We defined the limits of agreement as the mean difference between patients’ and proxies’ ratings plus or minus 1.96 times the standard deviation of the difference. We also used paired sample t-tests to test whether the means of the patients’ and the proxies’ scores were significantly different. We used unpaired t-tests to examine whether there were significant differences between couples where the proxy filled in the questionnaires at home or at the hospital and between patients with an EDSS score less than six or greater or equal to six. We used a two stage procedure and a general linear model to examine the relationship between the differences between the patient and proxy ratings on each of the five PROs and the independent variables. We used three groups of independent variables in the analysis: (1) the patients’ raw scores on the EDSS and the HADS anxiety and depression sub-scales and 31
their standardized z-scores on the BRB-N global cognition, verbal memory, visual memory, attention or executive functioning and fluency sub-scales; (2) the proxies’ raw scores on the SF-36, HADS, CSI and SRB; and (3) socio-demographic variables. The socio-demographic variables were: patient age, gender, time since MS diagnosis and the type of MS; patient and proxy’s levels of education; the number of years they had been in a relationship; and whether any children lived with them. In the first stage, we examined the univariate regression coefficient between the differences on each PRO and each of the independent variables. If the regression coefficient was significantly different from zero (p < 0.05), we retained the independent variable for that PRO. In the second stage, we constructed a multivariable model for the differences on each PRO using the independent variables retained in the first stage. Then, using a backward selection procedure, we deleted the least significant variables one at a time until all regression 61
3.2
Chapter 3.2
coefficients were significantly different from zero. We performed all statistical analyses using the Statistical Package for Social Sciences (SPSS) version 14.0 (SPSS, Inc., Chicago, Ill., USA).
Table 3.2.1 Study population
Age of patient in years Age of proxy in years Disease duration in years Duration of patient-proxy relationship in years
N
Mean
SD
Range
175 175 175 175
49.4 50.3 14.2 23.8
11.9 12.1 8.7 13.1
24-80 25-80 5-42 0-55
N (Patient)
N (Proxy)
Male Female
62 113
35% 65%
110 65
63% 37%
Relapsing Remitting (RR) Secondary Progressive (SP) Primary progressive (PP) Clinical Isolated Syndrome (CIS)
90 42 38 5
51% 24% 22% 3%
na na na na
Low education (primary school) Average education (secondary school) High education (college/university) Missing education
52 63 54 6
30% 36% 31% 3%
54 62 58 1
31% 35% 33% 1%
Completion of PROs in hospital setting
175
100%
109
62%
RESULTS Study sample We report on data from 175 MS patients and their proxy respondents. We have previously 7
reported on 139 of these couples. We present patient and proxy characteristics in Table 3.2.1. A total of 174 couples consisted of one man and one women and one couple consisted of two women. Forty-two patient-proxy couples (24%) had no children, 72 (41%) had children living at home and 61 couples (35%) had children not living with them. We present patient and proxy scale scores in Table 3.2.2. Systematic discrepancies between patient and proxy response We present the discrepancies between patient and proxy responses on the MS PROs in Table 3.2.3. We found a significant difference between mean patient and proxy scores (p<0.05) on
62
Analyzing differences between patient and proxy
the MSIS-29 psychological scale (p<0.001) and the GNDS (p=0.02). There were no significant differences in discrepancies between couples for whom the proxy completed the scales at home or at the hospital or between couples for whom the patient had an EDSS score less than six or greater or equal to six.
Table 3.2.2 Patient and proxy scale scores Patient
MSIS Physical MSWSb GNDS MSNQ MSIS Psychological HADS Anxiety HADS Depression EDSSc CSI SRB a
N 175 154 175 172 175 170 169 175 na na
Proxy report of patient data
b
Mean 45.7 29.7 12.8 17.2 16.6 4.5 3.6 4.0
Proxy
SD 19.5 15.4 8.4 10.1 6.7 3.0 3.5 2.5-6.0
Range 20-98 12-60 0-35 0-57 9-42 0-16 0-18 0-8
N 172a 149a 174a 173a 173a 171 170 na 175 175
This scale is restricted to patients who can walk
c
Mean 46.6 30.0 13.7 15.8 18.7 4.1 2.4
SD 19.6 15.5 8.9 10.7 7.5 3.1 2.8
Range 20-100 12-59 0-38 0-54 9-41 0-16 0-16
3.9 2.1
3.1 2.2
0-14 0-10
Median (Inter Quartile Range)
Table 3.2.3 Discrepancies between patient and proxy responses on the standardized scale scores Difference in standardized scale score (patient – proxy)
MSIS Physical MSWSa GNDS MSNQ MSIS Psychological a
N (pairs) 172 147 174 171 173
Mean difference -0.9 -0.6 -1.3 2.2 -5.9
SD (difference) 14.4 13.8 7.5 15.2 18.0
Limits of agreement -29.1 ; 27.3 -27.7 ; 26.5 -16.0 ; 13.4 -27.6 ; 32.0 -41.2 ; 29.4
p 0.42 0.58 0.02 0.06 <0.001
This scale is restricted to patients who can walk.
Regression analysis We present the independent variables that fulfilled the preset criteria (not adjusted for multiple testing) and remained in the regression model as explanatory variables for the difference between patient and proxy on each MS PRO in Table 3.2.4. Caregiver strain was
63
3.2
Chapter 3.2
significantly related to differences between patient and proxy scores for all five PROs. A higher level of caregiver strain was linked to the proxies rating the patients’ status as worse than the patients. A higher level of patient anxiety on the HADS was linked to larger differences between patient and proxy ratings on all PROs except the GNDS. Patient neuropsychological functioning as measured by the BRB-N was linked to larger differences between patient and proxy ratings on the MSIS physical scale (BRB-N visual memory), the GNDS (BRB-N visual memory), the MSIS psychological scale (BRB-N attention/executive functioning) and the MSNQ (BRBN global score). In addition, proxy depression on the HADS was linked to larger differences between patient and proxy ratings of the physical and psychological impact of MS as measured by the MSIS. There were larger differences between patient and proxy reports on the patients’ walking ability as measured by the MSWS for couples with children living at home. In couples where the proxy was female, the differences between patient and proxy reports on general MSrelated disability as measured by the GNDS were larger. A higher level of patient disability as expressed by the EDSS resulted in larger differences between patient and proxy reports of neuropsychological functioning as measured by the MSNQ.
DISCUSSION This study shows that both patient and proxy related factors have an impact on the agreement between MS patients and their proxies when assessing their disease using PROs. On four out of five instruments, proxies indicated that the MS patients had more difficulties or that the MS had a greater impact than the patients indicated themselves. We found several variables that explained part of the differences between patient and proxy ratings. Caregiver strain explained a significant amount of the differences between patients’ and proxies’ ratings on all PROs. Proxies who experienced a higher level of caregiver strain reported that the patients had more problems and the impact of MS was greater. Previous research demonstrated that there is a high level of correlation between caregiver distress, 14
disability, cognition and neuropsychiatric symptoms. It seems logical that that caregiver 64
Analyzing differences between patient and proxy Table 3.2.4 Results of the regression analyses Dependent
MSIS Physical
MSWS
GNDS
MSNQ
MSIS Psychological
Independent variable HADS anxiety (patient) BRBN – visual memory (patient) Caregiver strain index (proxy) HADS depression (proxy) HADS anxiety (patient) Caregiver strain index (proxy) Children living at home EDSS (patient) Caregiver strain index (proxy) Proxy is female BRBN – visual memory (patient) Years of relationship HADS anxiety (patient) EDSS (patient) BRBN – global (patient) Caregiver strain index (proxy) Self-rated burden scale (proxy) HADS anxiety (patient) Caregiver strain index (proxy) BRBN – attention (patient) HADS depression (proxy)
β
SE (β)
p
1.97 -2.54 -1.53 -1.17 1.01 -1.18 -5.02 0.95 -0.85 -3.40 -1.49 -0.09 1.37 2.24 5.46 -1.97 -1.76 1.77 -2.65 -3.85 -1.59
0.35 1.18 0.38 0.40 0.41 0.40 2.54 0.20 0.20 1.10 0.61 0.05 0.39 0.72 2.39 0.51 0.75 0.43 0.49 1.58
<0.001 0.033 <0.001 0.004 0.015 0.003 0.050 <0.001 <0.001 0.002 0.016 0.042 0.001 0.002 0.024 <0.001 0.021 <0.001 <0.001 0.016 0.002
0.50
Model R2 0.29
0.13
0.22
3.2 0.26
0.33
strain would increase as the impact of MS, in terms of limitations or handicap, increases. However, our results indicate that patients’ and proxies ratings on MS PROs diverge as caregiver strain increases. This might indicate that proxies experience an increase in their partners need for care as a worsening of the patients’ health. In contrast to all other scales, the patients’ mean score on the MSNQ was significantly higher than the proxies’. The MSNQ is designed to measure neuropsychological competencies. In our model, the global BRB-N score explained a part of this difference. A possible explanation for this is that although there were only a few patients with serious cognitive problems in this sample, these problems may affect patients more than their partners. In the literature, patient depression is strongly associated with subjective cognitive impairment independently of objective neuropsychological functioning.
10,32
In our study, anxiety among MS patients
explained part of the differences between patient and proxy ratings on the MSNQ, but patients’ levels of depression did not. Discrepancies in MSNQ scores need to be interpreted
65
Chapter 3.2
with caution since previous studies have shown that patient and proxy ratings on the MSNQ do not reflect the same construct.
10,33
Our study has some limitations. We did not use objective measurements of proxy health. This means that it is impossible to determine whether proxies actually had physical or mental problems or illnesses. In addition, we have no data on the use of medication that could have influenced patient and/or proxy ratings. Finally, although we did not find any significant differences between proxies, who filled in the questionnaires at home and in the hospital, we cannot exclude that this difference has influenced the results. In conclusion, we found several factors that may contribute to discrepancies between patient and proxy scores on MS PROs. The most important proxy factor is caregiver burden. Since the differences between patient and proxies were not constant, it is not currently possible to use proxy ratings with a simple correction as a substitute for patient ratings. Proxy scores can provide additional information, but more research is needed on the dynamics of factors influencing agreement between patients and proxies on MS PROs.
66
Analyzing differences between patient and proxy
REFERENCES 1 Bosma L, Kragt JJ, Polman CH, Uitdehaag BM. Walking speed, rather than Expanded Disability Status Scale, relates to long-term patient-reported impact in progressive MS. Mult. Scler. 2013; 19: 326-33. 2 Giordano A, Pucci E, Naldi P et al. Responsiveness of patient reported outcome measures in multiple sclerosis relapses: the REMS study. J. Neurol. Neurosurg. Psychiatry 2009; 80: 1023-8. 3 McKenna SP, Doward LC, Twiss J et al. International development of the patientreported outcome indices for multiple sclerosis (PRIMUS). Value. Health 2010; 13: 946-51. 4 Schaffler N, Schonberg P, Stephan J et al. Comparison of patient-reported outcome measures in multiple sclerosis. Acta Neurol. Scand. 2013. 5 Twiss J, Doward LC, McKenna SP, Eckert B. Interpreting scores on multiple sclerosisspecific patient reported outcome measures (the PRIMUS and U-FIS). Health Qual. Life Outcomes. 2010; 8: 117.
patients with multiple sclerosis. J. Clin. Exp. Neuropsychol. 2004; 26: 200-14. 10 Carone DA, Benedict RH, Munschauer FE III et al. Interpreting patient/informant discrepancies of reported cognitive symptoms in MS. J. Int. Neuropsychol. Soc. 2005; 11: 574-83. 11 Courts NF, Newton AN, McNeal LJ. Husbands and wives living with multiple sclerosis. J. Neurosci. Nurs. 2005; 37: 20-7. 12 Buchanan RJ, Radin D, Huang C. Burden among male caregivers assisting people with multiple sclerosis. Gend. Med. 2010; 7: 637-46. 13 Dunn J. Impact of mobility impairment on the burden of caregiving in individuals with multiple sclerosis. Expert. Rev. Pharmacoecon. Outcomes. Res. 2010; 10: 433-40. 14 Figved N, Myhr KM, Larsen JP, Aarsland D. Caregiver burden in multiple sclerosis: the impact of neuropsychiatric symptoms. J. Neurol. Neurosurg. Psychiatry 2007; 78: 1097-102.
6 van der Linden FA, Kragt JJ, Klein M et al. Psychometric evaluation of the multiple sclerosis impact scale (MSIS-29) for proxy use. J. Neurol. Neurosurg. Psychiatry 2005; 76: 1677-81.
15 Poser CM, Paty DW, Scheinberg L et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann. Neurol. 1983; 13: 227-31.
7 Sonder J, Bosma L, van der Linden F et al. Proxy measurements in multiple sclerosis: agreement on different patient-reported outcome scales. Mult. Scler. 2011.
16 Polman CH, Reingold SC, Edan G et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the "McDonald Criteria". Ann. Neurol. 2005; 58: 840-6.
8 Tripoliti E, Campbell C, Pring T, Taylor-Goh S. Quality of life in multiple sclerosis: should clinicians trust proxy ratings? Mult. Scler. 2007; 13: 1190-4.
17 Hobart J, Lamping D, Fitzpatrick R et al. The Multiple Sclerosis Impact Scale (MSIS-29): a new patient-based outcome measure. Brain 2001; 124: 962-73.
9 Bruce JM, Arnett PA. Self-reported everyday memory and depression in
18 Hobart JC, Riazi A, Lamping DL et al. Measuring the impact of MS on walking
67
3.2
Chapter 3.2
ability: the 12-Item MS Walking Scale (MSWS-12). Neurology 2003; 60: 31-6.
reliability across diverse patient groups. Med. Care 1994; 32: 40-66.
19 Sharrack B, Hughes RA. The Guy's Neurological Disability Scale (GNDS): a new disability measure for multiple sclerosis. Mult. Scler. 1999; 5: 223-33.
27 Ware JE, Jr., Sherbourne CD. The MOS 36item short-form health survey (SF-36). I. Conceptual framework and item selection. Med. Care 1992; 30: 473-83.
20 Benedict RH, Munschauer F, Linn R et al. Screening for multiple sclerosis cognitive impairment using a self-administered 15item questionnaire. Mult. Scler. 2003; 9: 95-101.
28 Robinson BC. Validation of a Caregiver Strain Index. J. Gerontol. 1983; 38: 344-8.
21 Rao SM in collaboration with the Cognitive Function Study Group of the Nations Multiple Sclerosis Society. A manual for the Brief Repeatable Battery of Neuropsychological Tests in multiple sclerosis. 1990. 22 Rao SM. A manual for the brief, repeatable battery of neuropsychological tests in multiple sclerosis. 1991. 23 Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983; 33: 1444-52. 24 Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr. Scand. 1983; 67: 361-70. 25 McHorney CA, Ware JE, Jr., Raczek AE. The MOS 36-Item Short-Form Health Survey (SF-36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs. Med. Care 1993; 31: 247-63. 26 McHorney CA, Ware JE, Jr., Lu JF, Sherbourne CD. The MOS 36-item ShortForm Health Survey (SF-36): III. Tests of data quality, scaling assumptions, and
68
29 van Exel NJ, Scholte op Reimer WJ, Brouwer WB et al. Instruments for assessing the burden of informal caregiving for stroke patients in clinical practice: a comparison of CSI, CRA, SCQ and self-rated burden. Clin. Rehabil. 2004; 18: 203-14. 30 Hobart JC, Riazi A, Lamping DL et al. Improving the evaluation of therapeutic interventions in multiple sclerosis: development of a patient-based measure of outcome. Health Technol. Assess. 2004; 8: iii, 1-iii,48. 31 Sepulcre J, Vanotti S, Hernandez R et al. Cognitive impairment in patients with multiple sclerosis using the Brief Repeatable Battery-Neuropsychology test. Mult. Scler. 2006; 12: 187-95. 32 Julian L, Merluzzi NM, Mohr DC. The relationship among depression, subjective cognitive impairment, and neuropsychological performance in multiple sclerosis. Mult. Scler. 2007; 13: 816. 33 Sonder JM, Mokkink LB, van der Linden FA et al. Validation and interpretation of the Dutch version of the Multiple Sclerosis Neuropsychological Screening Questionnaire. J. Neurol. Sci. 2012; 320: 91-6.
CHAPTER 4
Longitudinal analyses of patient and proxy responses
70
CHAPTER 4.1
Do patient and proxy agree? Long term changes in multiple sclerosis physical impact and walking ability on patient reported outcome scales
Judith M Sonder, Lisanne J Balk, Libertje VAE Bosma, Chris H Polman, Bernard MJ Uitdehaag
In press: Multiple Sclerosis 2014
Chapter 4.1
ABSTRACT Background: Patient reported outcome scales (PROs) are useful in monitoring changes in multiple sclerosis (MS) over time. Although these scales are reliable and valid measures in longitudinal studies with MS patients, it is unknown what the impact is when obtaining longitudinal data from proxies. Objective: To compare longitudinal changes in patient and proxy responses on PROs assessing physical impact of MS and walking ability. Methods: In a prospective observational study, data on the Multiple Sclerosis Impact Scale (MSIS-29 physical) and Multiple Sclerosis Walking Scale (MSWS-12) was obtained from 137 patient-proxy couples at baseline and 2 year follow up. Demographic and disease related variables explaining agreement or disagreement between patients and proxies were investigated using linear regression analyses. Results: Full agreement was found in 56% (MSIS) and 62% (MSWS) of the patient-proxy couples. Complete disagreement was very rare for both scales (2% MSIS, 5% MSWS). When patients were more positive than proxies, a higher age, longer disease duration, longer patient-proxy relationship and increased levels of depression, anxiety and caregiver burden in proxies were observed. Conclusion: In the majority of the patient-proxy couples there was agreement. Proxies can serve as a valuable source of information for assessing changes in impact of physical disability and walking ability in MS patients, but caution remains essential when using scores from proxies.
72
Do patient and proxy agree?
INTRODUCTION Patient reported outcomes (PROs) gain more and more importance in the assessment of 1
different disease aspects and phases of multiple sclerosis (MS). Different PROs have been developed and evaluated to assess physical functioning, psychological functioning, disease impact and quality of life in MS.
1-6
patients, but also for proxy use.
7-9
Some of these scales are not only validated to be used in A proxy is a partner or close family member of the MS
patient and can give important additional information about the patients disease state or impact of MS on their daily life.
9
The Multiple Sclerosis Impact Scale (MSIS-29)
10
was considered as a reliable tool to use in 9
proxies of MS patients for assessing disease impact. However, in studies with patient and proxy scores on cognitive functioning, using the Multiple Sclerosis Neuropsychological 7
Screening Questionnaire (MSNQ) it was found that patient and proxy ratings do not reflect the same construct.
8,11
A recent evaluation of proxy scores on five MS PROs, including the 12
Multiple Sclerosis Walking Scale (MSWS-12), showed that patients and proxies agree more on physical MS PROs in comparison to (neuro)psychological scales and that discrepancies between patient and proxy can be explained by patient-related, proxy-related and sociodemographical factors.
13,14
When monitoring disease progression in MS, one must rely on longitudinal repeated assessments and PROs are suitable tools to do so. The MSWS-12 is the best validated PRO for walking ability in MS patients, and can monitor changes over time in longitudinal studies.
15-17
The MSIS-29, measuring the physical and psychological effect of MS on activities of daily living is also suitable for use in long term studies with MS patients.
3,9,10,18-21
Although these scales
are reliable and valid measures in longitudinal studies with MS patients, it is unknown to what extent they provide reliable longitudinal data from proxies. Therefore, we performed a prospective study over two years collecting PRO data from patients and proxies. Given the results on the cross sectional studies we focused on the physical aspects of the disease, physical impact (MSIS physical) and walking ability (MSWS). We compared longitudinal changes as assessed by patients and proxies on these scales and
73
4.1
Chapter 4.1
investigated which factors could be of influence when patients and proxies disagreed about change over time.
METHODS Subjects As previously described, the subjects in this study were MS patients and their proxies.
13,14
All
subjects were aged 18 years or older. The patients were eligible for inclusion if they had been diagnosed as having MS or clinically isolated syndrome, according to the Poser or the revised McDonald criteria.
22,23
All included patients visited the outpatients’ clinic of the VU University
Medical Center Amsterdam for their baseline visit between November 2008 and December 2010. PROs were completed by patient and proxy during the patient’s visit to the hospital. If a proxy could not come to the hospital, he or she completed the questionnaires at home. For the follow-up after 2 years, PROs were completed at home and additionally the patient was interviewed by phone. The medical ethical committee of the VU University Medical Center approved the study protocol and informed consent was obtained from all participants. Physical impact and walking ability The MSIS-29 gives information about disease impact of MS on daily life and can be divided into a physical and psychological subscale. For the present study we only used the physical subscale which contains 20 items. The range of each item score is 1 (no impact on daily life) to 5 (extremely influencing daily life). The MSWS-12 is a patient-based measure of walking ability in MS. The scale comprises of 12 items, all ranging from 1 (no problems at all) to 5 (extremely difficult). The use of this scale is restricted to patients who can walk, no scores can be obtained from non-walking patients. Both scales were completed by patients and proxies rating the physical impact for patients and walking ability of patients. All scale scores were converted to a 0-100 scale.
24
For MSIS
physical and MSWS scores we used the clinical relevant change score of 10% to define higher, lower or equal (within 10% higher or lower) scores at 2 years after baseline. 74
3,25
Do patient and proxy agree?
Patients’ physical status and levels of depression and anxiety Neurological impairment and disability were assessed using the Expanded Disability Status 26
Scale (EDSS). The EDSS score ranges between 0 and 10, where a higher score indicates more severe disability. At baseline the EDSS was performed by a trained physician, at follow up a trained member of the research team administrated the EDSS by phone.
27
Patients’ levels of depression and anxiety were assessed using the Hospital Anxiety and Depression Scale (HADS).
28,29
The HADS is a 14-item self-reported screening scale to indicate
the possible presence of anxiety (7 items) and depressive states (7 items). For both continuous scales the range of scores is 0 to 21. Proxies’ levels of depression, anxiety and caregiver strain We assessed the proxies’ own levels of depression and anxiety also using the HADS. Caregiver strain was assessed using the Caregiver Strain Index (CSI),
30
a list of 13 stressors about the 31
daily care for the patient. In addition the Self-Rated Burden scale (SRB) was used to measure the overall burden of informal care for the patient as perceived by the proxy, with a visual analogue scale (VAS) ranging from (0) not straining at all to (10) much too straining. Statistical analyses Changes in scores on the MSIS physical and MSWS between baseline and the 2 year follow up (2Y) were calculated for patients and proxies. With paired sample t-tests the mean baseline and 2Y scores of patients and proxies were compared. For both patient and proxy MSIS physical and MSWS scores at 2Y compared to baseline, there were three possibilities: deterioration (≥10% increase), unchanged (<10% change in score), and improvement (≥10% decrease). This resulted in nine combinations of patient-proxy changes at 2Y including 3 with full agreement, both patient and proxy reporting deterioration, equal scores or improvement. The association between changes indicated by patients and changes indicated by proxies was analysed using Fishers exact test. The couples in which there was disagreement were divided into 2 groups: those in which the patient was more positive about the (lack of) change regarding walking ability or the physical impact of MS in
75
4.1
Chapter 4.1
two years than the proxy and those in which the patient was more negative than the proxy. The differences between these two groups were evaluated with respect to demographic variables (age, gender, length of patient-proxy relationship), disease related variables (MS subtype, disease duration, EDSS), anxiety and depression of both patient and proxy and caregiver burden using linear regression analyses. When necessary we adjusted for the age of the patient respectively proxy. All statistical analyses were performed using the Statistical Package for Social Sciences (SPSS) version 20.0 (SPSS, Inc., Chicago, Ill., USA). Given the exploratory nature of the study we did not correct for multiple testing.
RESULTS Out of a total of 137 patient-proxy couples, 128 patients and 126 proxies completed the MSIS physical at baseline and follow up and 102 patients and 112 proxies completed the MSWS at both time points. This resulted in a complete dataset (full assessment for patient and related proxy at both time points) for 119 couples for the MSIS physical and 97 for the MSWS. Median EDSS at baseline was 3.5 (interquartile range 2.0-5.5), and this value remained the same after 2 years (median 3.5, interquartile range 3.0-6.0). Additional subject characteristics are shown in table 4.1.1.
Table 4.1.1 Subject characteristics at baseline
Age patient, years Age proxy, years Disease duration, years Duration of patient-proxy relationship, years
Mean
(SD)
50.1 50.6 14.6 24.9
(11.7) (11.9) (8.8) (12.8)
% (N) Gender patient Male Female Disease type Relapsing Remitting Secondary Progressive Primary progressive Clinical Isolated Syndrome
76
36% (49) 64% (88) 51% (70) 28% (38) 18% (25) 3% (4)
Range 24 - 80 25 - 79 5 - 42 0 - 55
Do patient and proxy agree?
Change scores The change scores were defined as the difference in score of patient or proxy between baseline and 2Y. Table 4.1.2 shows the mean scores for patients and proxies at both time points and the change scores. The mean scores at 2Y were significantly higher compared to the baseline scores (p<0.05), except for the MSWS proxy scores (p=0.10). Although on average changes in scores obtained from patients were larger than those obtained from proxies, there was no significant difference at a group level. Table 4.1.2 Outcome of PROs at baseline and 2 year follow-up N
Baseline Mean (SD)
2 year Mean (SD)
Change score
119 119
32.0 (24.2) 32.9 (24.4)
37.2 (26.0) 37.1 (26.4)
+ 5.2* + 4.2*
97 97
34.5 (29.5) 35.5 (29.7)
39.3 (32.2) 37.8 (32.3)
+ 4.8* + 2.3
MSIS physical Patient Proxy MSWS§ Patient Proxy
4.1
§ MSWS is restricted to patients who can walk *Significant difference between baseline and 2 year mean scores (p<0.05)
Agree or disagree? On the MSIS physical scale there were complete scores of 119 patient-proxy couples of which 67 couples (56%) fully agreed and 52 (44%) disagreed about changes over 2 years (table 4.1.3). Complete disagreement (‘higher’ versus ‘lower’ scores) was exceptional (5%). The MSWS was completed by 97 couples of which 60 couples fully agreed (62%) and 37 (38%) disagreed (table 4.1.4). Complete disagreement on change on the MSWS was rare (2%). The results of Fishers exact test confirmed that there was a significant association (p<0.01) between changes indicated by patients and proxies, for both MSIS physical scale and MSWS. What causes disagreement? We further explored the patient-proxy couples that disagreed on the change of physical impact of MS (table 4.1.5) and walking ability (table 4.1.6) over two years. We compared the
77
Chapter 4.1
group in which the patients were more positive than the proxies to the group in which the patients were more negative than the proxies for both scales separately. Patients who were more positive than the proxy were significantly older. This was true for both scales. The group in which the patients were more positive, i.e. the proxies more negative about the (lack of) change regarding physical impact or walking ability, had a significantly longer disease duration (MSIS physical p=0.05) and a significantly longer duration of their relationship (MSWS p=0.02). In addition, the proxies’ levels of anxiety, depression and caregiver burden (CSI and SRB, both age corrected) at 2 year were higher (although not reaching statistical significance) in the group in which the proxies were more negative regarding the change over two years on physical disease impact. The two groups did not differ on gender, MS subtype or EDSS score (not shown).
Table 4.1.3 Agreement between patient and proxy for the MSIS physical Change score proxy
Change score patient
Lower Equal Higher
Total
Lower 5a 7c 1c 13
Equal 8b 43a 17c 68
Higher 5b 14b 19a 38
Total 18 64 37 119
Higher 1b 7b 13a 21
Total 11 56 30 97
a
Agreement Disagreement: higher proxy score at 2Y c Disagreement: higher patient score at 2Y b
Table 4.1.4 Agreement between patient and proxy for the MSWS Change score proxy
Change score patient
Lower Equal Higher
Total a
Agreement Disagreement: higher proxy score at 2Y c Disagreement: higher patient score at 2Y b
78
Lower 5a 7c 1c 13
Equal 5b 42a 16c 63
Do patient and proxy agree? Table 4.1.5 Comparison of MSIS physical disagreement over 2 years
Age (patient) Age (proxy) Gender patient Gender proxy Disease duration (years) Duration of patient-proxy relationship ( years) HADS (patient) anxiety baseline anxiety 2Y depression baseline depression 2Y HADS (proxy) anxiety baseline anxiety 2Y depression baseline depression 2Y CSI baseline (proxy) CSI 2Y (proxy) SRB baseline (proxy) SRB 2Y (proxy)
Proxy more positive, higher patient score (N=25) Mean (SD) 47.0 (11.6) 46.7 (12.1) Na Na 13.0 (7.2)
Patient more positive, higher proxy score (N=27) Mean (SD) 54.2 (9.7) 54.9 (10.1) Na Na 18.0 (10.6)
p-value 0.02 0.01 0.44 0.82 0.05
23.4 (11.8)
27.2 (14.7)
0.30
4.6 (3.1) 5.9 (3.5) 3.9 (3.8) 5.0 (4.1)
4.8 (3.0) 5.2 (3.8) 4.2 (3.3) 4.4 (3.9)
0.88a 0.62a 0.89a 0.58a
2.8 (2.2) 3.7 (2.9) 1.0 (1.4) 1.8 (2.4) 3.8 (3.2) 3.8 (3.3) 1.9 (2.0) 2.3 (1.9)
4.3 (2.8) 4.5 (3.4) 2.9 (2.5) 3.2 (3.4) 4.5 (3.1) 5.6 (3.0) 2.5 (2.3) 3.3 (2.2)
0.13 0.08 0.24 0.09 0.24b 0.08b 0.87b 0.10b
4.1
Table 4.1.6 Comparison of MSWS disagreement over 2 years
Age (patient) Age (proxy) Gender patient Gender proxy Disease duration (years) Duration of patient-proxy relationship (years) HADS (patient) anxiety baseline anxiety 2Y depression baseline depression 2Y HADS (proxy) anxiety baseline anxiety 2Y depression baseline depression 2Y CSI baseline (proxy) CSI 2Y (proxy) SRB baseline (proxy) SRB 2Y (proxy) a
adjusted for age (patient)
b
Proxy more positive, higher patient score (N=24) Mean (SD) 51.0 (11.2) 50.7 (10.6) na na 16.6 (10.5)
Patient more positive, higher proxy score (N=13) Mean (SD) 58.4 (11.7) 58.5 (11.6) na na 13.7 (10.1)
p-value <0.01 0.05 0.52 0.52 0.43
25.8 (12.7)
35.6 (10.6)
0.02
3.9 (2.7) 4.6 (3.0) 3.4 (2.7) 4.6 (3.4)
3.9 (1.4) 4.1 (2.0) 3.5 (2.0) 3.5 (2.4)
0.97a 0.64a 0.66a 0.36a
3.7 (2.6) 4.6 (3.5) 1.9 (1.9) 2.0 (2.6) 3.2 (2.9) 3.6 (2.6) 1.8 (1.9) 1.9 (1.5)
3.4 (2.2) 4.0 (2.5) 2.3 (3.1) 3.2 (2.9) 4.0 (3.0) 5.0 (2.9) 2.1 (1.4) 3.0 (2.1)
0.99b 0.39b 0.54b 0.62b 0.46 0.76b 0.96b 0.76b
adjusted for age (proxy)
79
Chapter 4.1
DISCUSSION Our prospective study showed that, in the majority of cases, patients and proxies agreed on the change of the physical impact of MS and the walking ability of MS patients over 2 years as assessed by PROs. Although there were few strong opposing differences, there were also couples that disagreed. In our study 44% of the patient-proxy couples disagreed with respect to the change of experienced physical impact of MS, while 38% of couples disagreed about changes in walking ability of the patient over a 2-year period. However, complete disagreement (‘higher’ versus ‘lower’ scores) was rare for both scales (2% for MSIS, 5% for MSWS). In the group in which the patients were more positive on the change than their proxies, the age of both patients and proxies was higher. We also observed a longer duration of the patient-proxy relationship in the group in which the patients were more positive than their proxies on the change of the patients’ walking ability in 2 years and a longer disease duration in the same group when rating physical impact. An earlier study which explored proxy responses showed that proxies reported greater MS 32
related physical disability compared to the patients’ responses. The more negative ratings of proxies were associated with poorer mental health of the proxy whereas more positive ratings of proxies were associated with lower emotional distress.
32
Changes in disability,
social, emotional and mental health of the patient and changes in perceived burden of the caregiver were associated with changes in the degree of depression in caregivers in earlier studies.
33,34
Our study confirms these findings with increased levels of anxiety, depression and
caregiver burden in the group in which proxies were more negative than patients on the change of physical impact of MS in 2 years.
A phenomenon that could limit the reliability and validity of these measures over time is 25
response shift. Response shift refers to the variability in people and their values that may reflect informative shifts in an individual’s standards, priorities and conceptualisation of 35
constructs, in addition to changes in actual health status. These factors may lead to patients reassessing their perceived limitations of everyday life and resetting goals so that they may
80
Do patient and proxy agree?
consider that the impact of MS is less marked than they thought formerly.
25
Patient and
proxy, as different individuals with different perspectives, behave and adapt differently to the disease over time and response shift may play a role in assessments from either individual. Although impossible to proof using the present data, it is tempting to speculate that the reported improvement over time and the opposed change scores of patients and proxies on the physical impact of MS and walking ability in this study, could in fact be partially explained by response shift. A finding that could have limited the results of this study is that the MSIS appeared to be less sensitive to change in less disabled patients (EDSS range 0-5), as in this study, compared to 25
more disabled patients. Furthermore, because the MSWS is restricted to patients who can walk, individuals were instructed to only complete the scale when the patient was able to walk. Though, instructions were not always well understood, some patients or proxies completed the scale with an (almost) maximum score while their partners did not complete the MSWS because the patient could not walk. This resulted in a discrepancy between complete patient (102) and complete proxy reports (112) . We only analysed the 97 couples with full assessment for patient and related proxy at both time points. Finally we would like to mention that using this method and splitting the group based on agreement and further splitting into different patterns in patient and proxy responses, the final sample sizes of the groups to perform statistical analyses (table 5 and 6) were small thus limiting the power of the study. In conclusion, proxies can serve as a valuable source of information for assessing changes in impact of physical disability and walking ability in MS patients. In the majority of the patientproxy couples there was agreement on both of these scales and complete disagreement was rare. However, increased depression, anxiety and caregiver burden in proxies may affect the scores and lead to an overestimation of disease impact scores. Therefore caution remains essential when using scores from proxies.
81
4.1
Chapter 4.1
REFERENCES 1
Schaffler N, Schonberg P, Stephan J et al. Comparison of patient-reported outcome measures in multiple sclerosis. Acta Neurol. Scand. 2013.
2
Bosma L, Kragt JJ, Polman CH et al. Walking speed, rather than Expanded Disability Status Scale, relates to long-term patientreported impact in progressive MS. Mult. Scler. 2013; 19: 326-33.
3
4
5
6
7
8
9
82
Giordano A, Pucci E, Naldi P et al. Responsiveness of patient reported outcome measures in multiple sclerosis relapses: the REMS study. J. Neurol. Neurosurg. Psychiatry 2009; 80: 1023-8. McKenna SP, Doward LC, Twiss J et al. International development of the patientreported outcome indices for multiple sclerosis (PRIMUS). Value. Health 2010; 13: 946-51. Twiss J, Doward LC, McKenna SP et al. Interpreting scores on multiple sclerosisspecific patient reported outcome measures (the PRIMUS and U-FIS). Health Qual. Life Outcomes. 2010; 8: 117. Riazi A. Patient-reported Outcome Measures in Multiple Sclerosis. Int. MS J. 2006; 13: 92-9. Benedict RH, Munschauer F, Linn R et al. Screening for multiple sclerosis cognitive impairment using a self-administered 15item questionnaire. Mult. Scler. 2003; 9: 95-101. Sonder JM, Mokkink LB, van der Linden FA et al. Validation and interpretation of the Dutch version of the Multiple Sclerosis Neuropsychological Screening Questionnaire. J. Neurol. Sci. 2012; 320: 91-6. van der Linden FA, Kragt JJ, Klein M et al. Psychometric evaluation of the multiple
sclerosis impact scale (MSIS-29) for proxy use. J. Neurol. Neurosurg. Psychiatry 2005; 76: 1677-81. 10 Hobart J, Lamping D, Fitzpatrick R et al. The Multiple Sclerosis Impact Scale (MSIS-29): a new patient-based outcome measure. Brain 2001; 124: 962-73. 11 Carone DA, Benedict RH, Munschauer FE et al. Interpreting patient/informant discrepancies of reported cognitive symptoms in MS. J. Int. Neuropsychol. Soc. 2005; 11: 574-83. 12 Hobart JC, Riazi A, Lamping DL et al. Measuring the impact of MS on walking ability: the 12-Item MS Walking Scale (MSWS-12). Neurology 2003; 60: 31-6. 13 Sonder J, Bosma L, van der Linden F et al. Proxy measurements in multiple sclerosis: agreement on different patient-reported outcome scales. Mult. Scler. 2011. 14 Sonder JM, Holman R, Knol DL et al. Analyzing differences between patient and proxy on Patient Reported Outcomes in multiple sclerosis. J. Neurol. Sci. 2013; 334: 143-7. 15 McGuigan C, Hutchinson M. Confirming the validity and responsiveness of the Multiple Sclerosis Walking Scale-12 (MSWS-12). Neurology 2004; 62: 2103-5. 16 Motl RW, Snook EM. Confirmation and extension of the validity of the Multiple Sclerosis Walking Scale-12 (MSWS-12). J. Neurol. Sci. 2008; 268: 69-73. 17 Motl RW, McAuley E, Mullen S. Longitudinal measurement invariance of the Multiple Sclerosis Walking Scale-12. J. Neurol. Sci. 2011; 305: 75-9.
Do patient and proxy agree?
18 Hobart JC, Riazi A, Lamping DL et al. How responsive is the Multiple Sclerosis Impact Scale (MSIS-29)? A comparison with some other self report scales. J. Neurol. Neurosurg. Psychiatry 2005; 76: 1539-43. 19 Hoogervorst EL, Zwemmer JN, Jelles B et al. Multiple Sclerosis Impact Scale (MSIS-29): relation to established measures of impairment and disability. Mult. Scler. 2004; 10: 569-74. 20 McGuigan C, Hutchinson M. The multiple sclerosis impact scale (MSIS-29) is a reliable and sensitive measure. J. Neurol. Neurosurg. Psychiatry 2004; 75: 266-9. 21 Riazi A, Hobart JC, Lamping DL et al. Multiple Sclerosis Impact Scale (MSIS-29): reliability and validity in hospital based samples. J. Neurol. Neurosurg. Psychiatry 2002; 73: 701-4.
status scale (EDSS). Neurology 1983; 33: 1444-52. 27 Lechner-Scott J, Kappos L, Hofman M et al. Can the Expanded Disability Status Scale be assessed by telephone? Mult. Scler. 2003; 9: 154-9. 28 Spinhoven P, Ormel J, Sloekers PP et al. A validation study of the Hospital Anxiety and Depression Scale (HADS) in different groups of Dutch subjects. Psychol. Med. 1997; 27: 363-70. 29 Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr. Scand. 1983; 67: 361-70. 30 Robinson BC. Validation of a Caregiver Strain Index. J. Gerontol. 1983; 38: 344-8.
22 Polman CH, Reingold SC, Edan G et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the "McDonald Criteria". Ann. Neurol. 2005; 58: 840-6.
31 van Exel NJ, Scholte op Reimer WJ, Brouwer WB et al. Instruments for assessing the burden of informal caregiving for stroke patients in clinical practice: a comparison of CSI, CRA, SCQ and self-rated burden. Clin. Rehabil. 2004; 18: 203-14.
23 Poser CM, Paty DW, Scheinberg L et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann. Neurol. 1983; 13: 227-31.
32 Schwartz L, Kraft GH. The role of spouse responses to disability and family environment in multiple sclerosis. Am. J. Phys. Med. Rehabil. 1999; 78: 525-32.
24 Hobart JC, Riazi A, Lamping DL et al. Improving the evaluation of therapeutic interventions in multiple sclerosis: development of a patient-based measure of outcome. Health Technol. Assess. 2004; 8: iii, 1-iii,48.
33 Pozzilli C, Brunetti M, Amicosante AM et al. Home based management in multiple sclerosis: results of a randomised controlled trial. J. Neurol. Neurosurg. Psychiatry 2002; 73: 250-5.
25 Costelloe L, O'Rourke K, Kearney H et al. The patient knows best: significant change in the physical component of the Multiple Sclerosis Impact Scale (MSIS-29 physical). J. Neurol. Neurosurg. Psychiatry 2007; 78: 841-4. 26 Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability
34 Aronson KJ. Quality of life among persons with multiple sclerosis and their caregivers. Neurology 1997; 48: 74-80. 35 Schwartz CE, Sprangers MA. Methodological approaches for assessing response shift in longitudinal healthrelated quality-of-life research. Soc. Sci. Med. 1999; 48: 1531-48.
83
4.1
84
CHAPTER 4.2
Towards the use of proxy reports for estimating long term patient reported outcomes in multiple sclerosis
Judith M Sonder, Lisanne J Balk, Femke AH van der Linden, Libertje VAE Bosma, Chris H Polman, Bernard MJ Uitdehaag
Under revision: Multiple Sclerosis
Chapter 4.2
ABSTRACT Background: Assessment of disease impact in multiple sclerosis (MS) is usually driven by information obtained directly from patients using patient reported outcomes. However, when the patient’s response in longitudinal studies is less reliable or missing, data from other sources, for instance proxy respondents, may be used. Objective: To evaluate whether long term patient scores can be reliably estimated using scores obtained from proxies. Methods: Baseline, 6 months and 2 year data was collected from 155 patients and proxies on the physical scale of the Multiple Sclerosis Impact Scale (MSIS-29). Linear regression analyses were used with the patient 2 year scores as outcome and proxy 2 year scores as predictor. Thereby other variables were added to the analyses to investigate whether they contributed to a better prediction of the patient follow up score. 2
Results: The patient follow up score could be predicted rather accurately (R = 0.74) using the patient baseline score and the proxy follow up score. The correlation between observed and predicted scores was 0.86. The model did not improve by adding more variables and performed good in different follow up durations. The model was validated in an external cohort in which it performed even better, with a high correlation between observed and predicted scores (0.92). Conclusion: A simple model of a constant value (intercept), the patient baseline score and the proxy follow up score can predict patients’ follow up scores on the physical impact of MS.
86
Towards the use of proxy reports in MS
INTRODUCTION Research on disease impact in multiple sclerosis (MS) is usually based on information obtained directly from patients. Yet, when the patients’ response is less reliable or missing due to disease status, age, cognitive functioning or mood disorders, it might be necessary to obtain data from other sources, for instance proxy respondents.
1-3
A proxy respondent can be
a partner, close family member or healthcare provider of the MS patient. 4-10
Proxy measurements have recently been evaluated in MS populations such as eldery,
11,12
children,
13
stroke,
13,14
4
and in several other patient
15
cancer and patients with Alzheimers
16
disease. These studies reported different results according to agreement between patients and proxies. In most of the studies the differences between patients and proxies were small, but proxies rated patients as more severely affected than patients themselves and agreement was better on well observable or objective scales (activities of daily living, physical functioning) compared to evaluation of more subjective scales (quality of life, psychological impact and cognitive status). Longitudinal studies in MS may benefit from the use of proxies when patients’ responses are not available or less reliable due to disease circumstances. In a prospective study we described changes in scores of patients and proxies rating the impact of physical disability and 17
walking ability of patients over 2 years. We found agreement on the change over 2 years in the majority of the patient-proxy couples. In the group in which the proxies were more negative about the change, we found a significantly higher age of patients and proxies and increased depression, anxiety and caregiver burden in 2 years in proxies. In a previous study we also found that caregiver strain, proxy depression, patient anxiety and some neuropsychological components were causes of differences between patient and proxy.
6
The objective of this study was to evaluate whether patient scores in longitudinal MS studies can be reliably estimated using proxy outcomes. Based on the results we found in earlier studies we first focused on the physical impact of MS (Multiple Sclerosis Impact Scale – physical).
18
We were interested in the variables that, in addition to the longitudinal proxy
scores, contributed to the most accurate prediction of longitudinal patient scores.
87
4.2
Chapter 4.2
METHODS Subjects As previously described, the subjects in this study were MS patients and their proxies.
5,6
All
subjects were aged 18 years or older. The patients were eligible for inclusion if they had been diagnosed as having MS or clinically isolated syndrome, according to the Poser or the revised McDonald criteria.
19,20
All included patients visited the outpatients’ clinic of the VU University
Medical Center Amsterdam for their baseline visit between November 2008 and December 2010. The PROs, as described below, were completed by patient and proxy during the patient’s visit to the hospital. If a proxy could not come to the hospital, he or she completed the questionnaires at home. For the follow-up after 6 months and 2 years, PROs were completed at home and additionally the patient was interviewed by phone. For external validation of the analyses, data from an independent longitudinal cohort of 56 8
MS patients and their partners was used. Data was collected at baseline and after two years between December 2003 and April 2006. The medical ethical committee of the VU University Medical Center approved both study protocols and informed consent was obtained from all participants. Physical impact of MS The MSIS-29 gives information about disease impact of MS on daily life and can be divided into a physical and psychological subscale. For the present study we only used the physical subscale which contains 20 items. The range of each item scores is 1 (no impact on daily life) to 5 (extremely influencing daily life). The scale was completed by patients and proxies at baseline and after 6 months and 2 years. Proxies rated the actual physical impact for the patients. Patients’ physical status and levels of depression and anxiety Neurological impairment and disability was assessed using the Expanded Disability Status 21
Scale (EDSS). The EDSS score varies between 0 and 10, a higher score indicates more severe
88
Towards the use of proxy reports in MS
disability. At baseline the EDSS was performed by a trained physician, for follow up after 2 years a trained member of the research team administrated the EDSS by phone.
22
Patients’ levels of depression and anxiety were assessed using the Hospital Anxiety and Depression Scale (HADS).
23,24
The HADS is a 14-item self-reported screening scale to indicate
the possible presence of anxiety (7 items) and depressive states (7 items). For both continuous scales the range of scores is 0 to 21. Proxies’ levels of depression, anxiety and caregiver strain We assessed the proxies’ own levels of depression and anxiety also using the HADS. Caregiver strain was assessed using the Caregiver Strain Index (CSI),
25
a list of 13 stressors about the 26
daily care for the patient. In addition the Self-Rated Burden scale (SRB) was used to measure the overall burden of informal care for the patient as perceived by the proxy, with a visual analogue scale (VAS) ranging from (0) not straining at all to (10) much too straining. Proxies’ anxiety, depression as well as caregiver strain and burden were only evaluated in the present cohort (not in the external validation cohort).
4.2
Statistical analyses To predict the patients’ follow up score, we used linear regression analyses with the patient 2 year MSIS physical score as outcome and proxy 2 year MSIS physical score as predictor. Additional analyses were performed on the predictive value of: MSIS physical patient baseline score; MSIS physical difference between patient and proxy at baseline; physical disability (EDSS); depression and anxiety scores of patient and proxy at baseline and follow up (HADS); and caregiver strain or burden at baseline and follow up (CSI, SRB). First, a regression model was created based on the proxies’ follow up 2 year scores of the present cohort to predict the patients’ 2 year MSIS physical scores. Second, the other potentially predictive variables were added to this model to evaluate whether they contributed significantly. Based on the intercept and the regression coefficient(s) (β) of the variables that retained significant in the model, a function could be created to calculate the patients’ 2 year MSIS physical scores. As a measure to express how good the observed outcomes were predicted by the model we used R squared (R2), a proportion of the total 89
Chapter 4.2
variation of outcomes explained by the model. Additionally, the correlation coefficient between the observed and predicted 2 year MSIS physical patient scores was calculated. Following this we used the final and best predictive regression model, as described above, to calculate predicted patient follow up scores in the same cohort, but for different follow-up durations: Six months (baseline to six months) and 18 months (6 months to 2 year). The predicted scores were evaluated by correlations with the observed patient follow up scores. The last step was to validate the regression model, based on the data of the present patientproxy cohort, in an independent sample. In the external validation cohort with the same follow up duration (baseline to 2 year), the predicted patient 2 year MSIS physical scores were calculated and again evaluated by correlations between the observed and predicted scores.
RESULTS In the present cohort a total of 155 patients and proxies participated at baseline and follow up 6 months and/or 2 years. This resulted in a complete dataset of MSIS physical scores (full assessment for patient and related proxy at both time points) for 140 couples at follow up 6 months and 120 couples at follow up 2 years. In the external validation cohort, 56 patients and proxies completed the MSIS physical scale at both time points. Subject characteristics are shown in table 4.2.1.
Table 4.2.1 Subject characteristics at baseline Present cohort Age patient, years Age proxy, years Disease duration, years Gender patient Male Female Disease type Relapsing Remitting Secondary Progressive Primary Progressive Clinically Isolated Syndrome Unknown
90
N 155 155 155
External validation cohort Mean (SD) 49.8 (12.0) 50.7 (12.3) 14.6 (8.9)
N 56 50 54
54 (35%) 101 (65%)
16 (29%) 40 (71%)
78 (50%) 40 (26%) 32 (21%) 5 (3%) -
26 (46%) 18 (32%) 10 (18%) 1 (2%) 1 (2%)
Mean (SD) 48.1 (9.1) 49.5 (9.1) 12.7 (7.4)
Towards the use of proxy reports in MS
Patient reported outcomes In table 4.2.2 the mean scores on the different PROs and the EDSS are shown. Data in the external validation cohort had already been collected in the past, thus not all PROs could be evaluated in this group. On all PROs that could be evaluated in the external validation cohort we observed higher mean scores.
Table 4.2.2 Scores on patient reported outcome scales and EDSS
MSIS physical patient baseline MSIS physical proxy baseline HADS anxiety patient baseline HADS anxiety patient follow up 2Y HADS anxiety proxy baseline HADS anxiety proxy follow up 2Y HADS depression patient baseline HADS depression patient follow up 2Y HADS depression proxy baseline HADS depression proxy follow up 2Y CSI proxy baseline CSI proxy follow up 2Y SRB proxy baseline SRB proxy follow up 2Y EDSS baseline (median, IQR)
Present cohort
External validation cohort
N
Mean (SD)
N
Mean (SD)
155 153 152 108 152 109 152 108 152 110 146 103 155 109 155
32.1 (24.1) 33.6 (24.6) 4.5 (3.0) 4.7 (3.8) 4.3 (3.2) 4.3 (3.5) 4.5 (3.0) 4.7 (3.8) 2.5 (2.9) 2.7 (3.3) 3.8 (3.0) 4.1 (3.1) 2.2 (2.3) 2.9 (2.2) 4.0 (2.5-6.0)
56 56 56 56
42.2 (24.2) 42.9 (25.3) 5.8 (3.9) 5.2 (3.7) na na 6.0 (4.1) 5.2 (3.9) na na na na na na 4.6 (3.5-6.0)
56 56
56
4.2
Regression The results obtained from the regression analyses based on the baseline and 2 year data from the present cohort are shown in table 4.2.3. Proxies’ follow up score (β=0.47, p<0.001) and patients’ baseline score (β=0.47, p<0.001) were significantly related to patients’ follow up score on the MSIS physical scale. Further analyses showed that none of the other suggested variables (disability, patient-proxy difference, mood, caregiver burden) contributed to a better prediction of the patients’ follow up score (p>0.05 for all variables). Further analyses to evaluate the model for different follow up durations and in an independent cohort were performed by calculating the predicted patient follow up scores 91
Chapter 4.2
using the function displayed in table 4.2.3. Correlations between the observed and predicted patient follow up scores were again calculated, the results are presented in table 4.2.4. High correlations ranging from 0.86 to 0.92 were found between the observed and predicted scores, independent of cohort and follow up length. Interestingly, the highest correlation (0.92) was found between the observed and predicted MSIS physical patient follow up 2Y scores in the external validation cohort.
Table 4.2.3 Results of the regression analyses Dependent
Independent/constant
MSIS physical Patient follow up 2Y
Constant (intercept) MSIS physical proxy follow up 2Y MSIS physical patient baseline
β
SE(β)
p-value
Model R2
4.43 0.47 0.47
2.14 0.07 0.08
0.041 <0.001 <0.001
0.74
MSIS Physical patient follow up score = 4.43 + (0.47 * MSIS physical proxy follow up) + (0.47 * MSIS physical patient baseline)
Table 4.2.4 Correlation between observed and predicted MSIS physical patient follow up scores N
Observed score Mean (SD)
Predicted score Mean (SD)
Correlation
Present cohort MSIS physical patient 6M (0-6M) MSIS physical patient 18M (6M-2Y) MSIS physical patient 2Y (0-2Y)
140 109 120
33.4 (24.8) 36.6 (26.5) 39.6 (25.0)
36.5 (22.1) 37.6 (22.3) 38.9 (21.2)
0.88 0.90 0.86
External validation cohort MSIS physical patient 2Y (0-2Y)
56
45.1 (25.3)
47.7 (22.1)
0.92
Predicted follow up score
DISCUSSION The aim of this study was taking a first step towards the use of proxy reports in longitudinal MS studies. For calculation of the most accurate patient follow up scores on the physical impact of MS, the use of proxy follow up scores and other potentially contributing patient or proxy related variables were evaluated. The results demonstrated that the most accurate prediction of the patient follow up score on the MSIS physical scale was based on a constant value (4.43) and a combination of the proxy follow up score (β=0.47) and the patient baseline
92
Towards the use of proxy reports in MS
score (β=0.47). Correlations between the observed and predicted patient follow up scores were high, even when follow up lengths varied. Remarkably, the highest correlation between the observed and predicted patient follow up score of physical MS impact was found by the external validation of the model in an independent cohort of MS patients and proxies. This indicates that the accuracy of model we provided was not the result of overfitting and might be considered promising for its use in other cohorts. Although these findings are important regarding the use of proxy data, caution must be applied as these findings might not be perfectly convertible to other MS populations. The first evaluation in an independent cohort of MS patients and proxies was positive, but apart from the fact that this cohort was independent and included different patients and proxies, it should be mentioned that these patients came from the same MS center and setup of the studies was almost identical. In earlier publications we found that proxies with increased levels of anxiety, depression and caregiver burden were more negative than patients on the physical impact or change of physical impact of MS in 2 years.
6,17
It could be expected that in a regression model to predict
patients’ follow up score out of proxies’ follow up score, a correction should be made for proxy mood or caregiver strain. However, the results of the current study showed a clear and simple model to predict a patient follow up score in which these variables showed no significant added value. Concluding, the purpose of the current study was to take a first step towards the use of proxy respondents in longitudinal MS research. We found that a model of proxy follow up score and patient baseline score can predict a patient follow up score on physical impact of MS, which correlates high with the observed patient score. This calculated score gives a reliable estimation of the real patient score and can therefore be of importance when patient scores are either not available or unreliable due to different frequently occurring disease related circumstances in longitudinal MS studies.
93
4.2
Chapter 4.2
REFERENCES 1
Bruce JM, Arnett PA. Self-reported everyday memory and depression in patients with multiple sclerosis. J. Clin. Exp. Neuropsychol. 2004; 26: 200-14.
2
Carone DA, Benedict RH, Munschauer FE III et al. Interpreting patient/informant discrepancies of reported cognitive symptoms in MS. J. Int. Neuropsychol. Soc. 2005; 11: 574-83.
3
Tripoliti E, Campbell C, Pring T, Taylor-Goh S. Quality of life in multiple sclerosis: should clinicians trust proxy ratings? Mult. Scler. 2007; 13: 1190-4.
4
van der Linden FA, Kragt JJ, Klein M et al. Psychometric evaluation of the multiple sclerosis impact scale (MSIS-29) for proxy use. J. Neurol. Neurosurg. Psychiatry 2005; 76: 1677-81.
5
Sonder J, Bosma L, van der Linden F et al. Proxy measurements in multiple sclerosis: agreement on different patient-reported outcome scales. Mult. Scler. 2012;18:196201.
6
Sonder JM, Holman R, Knol DL et al. Analyzing differences between patient and proxy on Patient Reported Outcomes in multiple sclerosis. J. Neurol. Sci. 2013; 334: 143-7.
7
van der Linden FA, Kragt JJ, Hobart JC et al. Proxy measurements in multiple sclerosis: agreement between patients and their partners on the impact of multiple sclerosis in daily life. J. Neurol. Neurosurg. Psychiatry 2006; 77: 1157-62.
8
94
van der Linden FA, Kragt JJ, van Bon M et al. Longitudinal proxy measurements in multiple sclerosis: patient-proxy agreement on the impact of MS on daily life over a period of two years. BMC. Neurol. 2008; 8: 2.
9
van der Linden FA, D'hooghe MB, Nagels G et al. Proxy ratings from multiple sources: disagreement on the impact of multiple sclerosis on daily life. Eur. J. Neurol. 2008; 15: 933-9.
10 van der Linden FA, Kragt JJ, Hobart JC et al. The size of the treatment effect: do patients and proxies agree? BMC. Neurol. 2009; 9: 12. 11 Middleton LE, Kirkland SA, Mitnitski A, Rockwood K. Proxy reports of physical activity were valid in older people with and without cognitive impairment. J. Clin. Epidemiol. 2010; 63: 435-40. 12 Yasuda N, Zimmerman S, Hawkes WG et al. Concordance of proxy-perceived change and measured change in multiple domains of function in older persons. J. Am. Geriatr. Soc. 2004; 52: 1157-62. 13 Irwin DE, Gross HE, Stucky BD et al. Development of six PROMIS pediatrics proxy-report item banks. Health Qual. Life Outcomes. 2012; 10: 22. 14 Hilari K, Owen S, Farrelly SJ. Proxy and selfreport agreement on the Stroke and Aphasia Quality of Life Scale-39. J. Neurol. Neurosurg. Psychiatry 2007; 78: 1072-5. 15 Giesinger JM, Golser M, Erharter A et al. Do neurooncological patients and their significant others agree on quality of life ratings? Health Qual. Life Outcomes. 2009; 7: 87. 16 Jensen-Dahm C, Vogel A, Waldorff FB, Waldemar G. Discrepancy between selfand proxy-rated pain in Alzheimer's disease: results from the Danish Alzheimer Intervention Study. J. Am. Geriatr. Soc. 2012; 60: 1274-8.
Towards the use of proxy reports in MS
17 Sonder J, Balk L, Bosma L et al. Do patient and proxy agree? Long term changes in multiple sclerosis physical impact and walking ability on patient reported outcome scales. Submitted, 2014. 18 Hobart J, Lamping D, Fitzpatrick R et al. The Multiple Sclerosis Impact Scale (MSIS-29): a new patient-based outcome measure. Brain 2001; 124: 962-73. 19 Poser CM, Paty DW, Scheinberg L et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann. Neurol. 1983; 13: 227-31. 20 Polman CH, Reingold SC, Edan G et al. Diagnostic criteria for multiple sclerosis: 2005 revisions to the "McDonald Criteria". Ann. Neurol. 2005; 58: 840-6. 21 Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983; 33: 1444-52.
22 Lechner-Scott J, Kappos L, Hofman M et al. Can the Expanded Disability Status Scale be assessed by telephone? Mult. Scler. 2003; 9: 154-9. 23 Spinhoven P, Ormel J, Sloekers PP et al. A validation study of the Hospital Anxiety and Depression Scale (HADS) in different groups of Dutch subjects. Psychol. Med. 1997; 27: 363-70. 24 Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr. Scand. 1983; 67: 361-70. 25 Robinson BC. Validation of a Caregiver Strain Index. J. Gerontol. 1983; 38: 344-8. 26 van Exel NJ, Scholte op Reimer WJ, Brouwer WB et al. Instruments for assessing the burden of informal caregiving for stroke patients in clinical practice: a comparison of CSI, CRA, SCQ and self-rated burden. Clin. Rehabil. 2004; 18: 203-14.
4.2
95
96
CHAPTER 5
Summary and General Discussion
Chapter 5
98
Summary and General Discussion
Methodological issues like the loss of information and the potentially reduced reliability of longitudinal patient reports in MS research can partially be solved by using proxy respondents. In this thesis we outlined agreement and disagreement between patients and proxy respondents and gave a comprehensive insight in the underlying factors of disagreement. Finally a model was developed to estimate patients’ long term outcomes on the physical impact of MS, using proxy scores. Results described in this thesis can be considered as a step towards the use of proxy reported outcomes in MS research and clinical trials.
DISCUSSION OF OUR FINDINGS VALIDATION OF PATIENT REPORTED OUTCOME SCALES FOR PROXY USE When using proxy responses, one should be able to rely on the measurement scales to yield the same results for patients and proxies. In this thesis we assessed patient and proxy responses on five PROs: The MSIS-29 physical and psychological scale, the MSWS-12, the GNDS and the MSNQ. 1
The MSIS-29 was evaluated for proxy use in earlier studies with good results. The MSWS is an objective scale with a construct (walking) that is relatively easy to observe, in which we did not expect problems when asking questions to a proxy respondent in the third phrase (“was he/she limited in his/her ability to walk?”). The GNDS is an interview in which a member of the research team asks questions to a proxy and in which understanding of the questions can be directly checked. With respect to the MSNQ, we found opposing results in the existing literature and therefore we aimed to test the validity and interpretability of the Dutch version of the MSNQ patient and informant form.
2-5
The results in chapter 2 demonstrated that we could confirm the unidimensionality of the MSNQ. It assesses one underlying construct, which is believed to be neuropsychological competence for both the patient and the informant version. The internal consistency was high. Concerning the construct validity, the results confirmed earlier findings that the patient
99
5
Chapter 5
version correlates higher with anxiety and depression of the patient than with neuropsychological testing (BRB-N). The informant form on the other hand showed the desired result, higher correlations were found with impaired subtest scores of the patient on the BRB-N compared to anxiety and depression of the patient. Inter rater agreement on scale and item level were moderate, indicating that the patient and informant form were not able to discriminate cognitive impaired patients similarly. Regarding the interpretability, the sensitivity of the patient version was too low to discriminate for cognitive impairment. We could define a cut-off point for the informant version and further analyses showed that when combining both versions, there was no added value of the patient version upon the use of the informant version. Although these results on the MSNQ were in favor of the use of proxy ratings, they were not very promising for the analyses on agreement between patient and proxy in the next chapter. The population of MS patients in this thesis had a good cognitive status. Only about as few as 10% of our patients could be classified as cognitively impaired, whereas cognitive deficits are described to be present in 40-70% of the MS patients and can often be detected early in the disease.
6-10
This finding indicates that our patient population may not reflect the general
population of MS patients or that perhaps others were more liberal in defining cognitive impairment. For most of the work in this thesis it was an advantage to work with cognitively preserved patients, since the assessments that were obtained from them could be considered as reliable. However, with respect to assessment of cognitive dysfunction and the possible influence of cognition on (dis)agreement between patients and proxies, an underrepresentation of cognitive impaired patients is clearly a disadvantage, making any estimation of the value of scales for cognition less reliable.
AGREEMENT BETWEEN PATIENT AND PROXY In chapter 3 we described agreement between patients and proxies in a cross sectional study. From the perspective of an earlier study with a relatively small group of patients and proxies with the MSIS-29, we aimed to extend this to other PROs that are frequently used in MS research with larger groups of MS patients and proxies. We used a range of MS scales from 100
Summary and General Discussion
what we hypothesized was ‘more easy to observe’ (walking ability, impact of physical functioning), ‘intermediate difficult to observe’ (overall disability) and ‘difficult to observe’ (psychological impact and functioning). In chapter 3.1 we could confirm our hypothesis that there was agreement between patients and proxies on PROs measuring physical impact of MS and walking ability of the MS patient. Our data supports previous findings that proxies can rate the disease status of patients in an 1
accurate way, when the focus is on specific and observable disease aspects or (physical) limitations.
11
However, for psychological scales (MSNQ and MSIS Psychological) and overall
disability (GNDS) the differences between patients and proxy respondents were large and dependent of scale score. Although these results were promising for the use of physical MS scales rated by proxy respondents, we should stress that this was an overall impression of mean differences between patients and proxies for five scales on group level, which could not automatically be generalized to individual patient-proxy couples. To achieve that goal, we aimed to gain more insight in the underlying factors of patient-proxy agreement.
5
HOW TO EXPLAIN DIFFERENCES BETWEEN PATIENT AND PROXY? We suggested that factors influencing the differences between patients and proxies on MS PROs should be further investigated. The objective in chapter 3.2 was to examine whether MS patients’ cognitive status, physical status and levels of depression and anxiety and their proxies’ health, levels of caregiver strain and depression explained the differences we found between patient and proxy ratings. The results demonstrated that on all PROs except MSNQ, proxies reported more difficulties or impact of MS than the patients themselves. Caregiver strain explained a significant part of the differences, proxies who had a higher level of caregiver strain were more negative than the patients on the disease related complaints and disease impact. It is an important observation that patients’ and proxies ratings on MS PROs diverge as caregiver strain increases. This might indicate that proxies experience stress related to their partners need for care as worsening of the patients’ health or as increasing 101
Chapter 5
impact on their daily lives. Besides caregiver strain, we also observed that patients’ anxiety and proxies’ depression were explanatory for a substantial amount of the rating differences. The results according to the differences on the MSNQ were opposite to the other PROs under investigation. In chapter 3.2 different explanations are mentioned, it could be related to starting cognitive problems which cannot be observed by the neuropsychologist or the proxy, but in the meantime really bother patients. Thereby part of the differences between patients and proxies on the MSNQ can be explained by anxiety of patients, which could be a reasonable reaction on incipient cognitive problems or fear for future problems. However, according to the results reported in chapter 2, we interpret MSNQ patient-proxy results with caution.
DO PATIENT AND PROXY AGREE ABOUT CHANGE OVER TIME? Although we know from previous studies that we can use PROs to monitor changes in MS over time, only a little is known about how proxies behave in rating the patients’ MS status over time.
12-16
Keeping in mind the positive results with patient-proxy measurements on
physical scales, as seen in the previous chapters, chapter 4.1 focused on the physical impact of MS and walking ability of the MS patient rated by patients and proxies over a 2-year period. We compared longitudinal changes as assessed by patients and proxies on these scales and investigated which factors could be of influence when finding disagreement about change over time. The majority of the couples agreed about the change over time, indicating that patient and proxy both scored higher, lower or equal after 2 years in comparison to their baseline scores. However, 44% of the patient-proxy couples disagreed with respect to the change of experienced physical impact of MS and 38% of the couples disagreed about changes in walking ability of the patient over a 2-year period. Despite these contradictions, complete disagreement (‘higher’ versus ‘lower’ scores) was rare for both scales (2% for MSIS, 5% for MSWS). In the group in which the patients were more positive on the change than their proxies, a higher age of patients and proxies was found on both scales and we observed a
102
Summary and General Discussion
longer duration of the patient-proxy relationship (MSWS) and a longer disease duration (MSIS physical) for this group. And although these results did only reach statistical significance at 0.10 level, we found increased levels of anxiety, depression and caregiver burden in the group in which proxies were more negative than patients on the change of physical impact of MS in 2 years. These results are promising for the use of proxy reports in (longitudinal) MS research. However, it must be emphasized that these results are found on scales that assess ‘easy to observe’ aspects of the disease. Reaching agreement over time for scales that assess ‘more difficult to observe’ aspects of MS is probably more challenging.
A FIRST STEP TOWARDS THE USE OF PROXY MEASUREMENTS IN MS STUDIES In all previous studies and chapters we concluded that proxy reports should be interpreted with caution and that proxy responses cannot serve as a direct substitute for patients’ responses in (longitudinal) MS studies. However, in daily practice researchers have to deal with the loss of information in longitudinal studies, as we described earlier. Hence, the question keeps coming up: To what extent are proxies able to rate patients’ follow up scores when patients are not able to participate in follow up measurements? If we would be able to answer this question, we could also be able to write a clear conclusion of this thesis: Is it worth continuing research on proxy measurements in MS? In chapter 4.2 we focused on the MSIS physical scale, aiming to evaluate whether (missing) patient scores in longitudinal MS studies on this scale can be estimated using proxy outcomes. We were aware of the fact that a proxy follow up score on itself would not be sufficient, therefore we were interested in the variables that, in addition to the longitudinal proxy scores, contributed to the most accurate estimation of longitudinal patient scores on the physical impact of MS. The results demonstrated that the most accurate prediction of the patient follow up score on the MSIS physical scale was based on a constant value (4.43) and a combination of the proxy follow up score (β=0.47) and patient baseline score (β=0.47). Correlations between the real and predicted patient follow up scores were high (≥0.86), even
103
5
Chapter 5
when follow up lengths varied. The model was validated in an external cohort of MS patients and proxies with good result. Although these findings can be considered as a first step towards the use of proxy reported outcomes in future studies, caution must be applied as these findings might not be perfectly convertible to other MS populations and further validation is needed. Moreover it is unknown whether this will be feasible for other PROs.
PRESENT POSITION The results of this thesis illustrate the feasibility of using assessments obtained from proxies in MS research. It shows the limitations which were in part predictable (easy to observe ‘objective’ aspects are more easy to assess from a different perspective than hard to observe ‘subjective’ aspects) and in part unpredictable (caregiver burden causes discrepancies in the assessments between patients and proxies). However, both limitations are meaningful and face valid. In cross sectional settings it is hard to deduce patients scores from proxy assessments. Even when taking into account aspects that may account for differences, agreement on a group level is moderate and on an individual level poor. However, for longitudinal measurements our first results are promising. Factors that may induce differences between patients and proxies might filter out when using changes over time. This was confirmed in an independent sample of MS patients and proxies. Nevertheless, although progress is made in the research reported in this thesis, it is too early to use this approach in intervention studies or clinical practice.
FUTURE PERSPECTIVES CONDITIONS OF PATIENTS AND PROXIES We are aware of the limitation in our project that, at baseline, proxies were able to participate in the hospital, during the patient visit, or from home. The conditions at home
104
Summary and General Discussion
cannot be controlled, so there is a possibility that the proxy consulted the patient when completing the questionnaires or that the proxy would have had questions for a member of the research team while completing the PROs. Although we didn’t find any significant differences between the baseline groups completing PROs in the hospital or at home, we would advice to keep the conditions equal for the whole study population in future studies.
SCREENING OF PROXIES In this thesis we assessed not only patients’, but also proxies’ anxiety and depression. Thereby we used two scales to indicate the strain and burden of partners in taking care of the MS patient, which seemed to be important variables in explaining the differences between patient and proxy reports. Future research might need to go beyond that and include more information on proxy health. Despite the use of the Short Form Health Survey (SF-36), we didn’t get an objective impression of proxies’ health and cognitive state and we didn’t collect information on medication use of proxies. It might be important to take these aspects into account in future studies.
5
CARE FOR PROXIES? We found high levels of caregiver strain/burden and depression in MS patients’ partners, which had impact on their assessment of the disease of their partner. High level of burden and depression could raise the question whether it shouldn’t be more common to also provide care to proxies? If so, the question may be: Who should be responsible for that care? Should that be their general practitioner or the neurologist who takes care of the MS patient? Apart from being an interesting question for future researchers (what is the need for care in partners of MS patients?) if provided this may affect the caregiver burden and the depression. In turn this may affect the (dis)agreement between patients and proxies and thus potentially impact the results of this study.
105
Chapter 5
FUTURE USE OF PROXY REPORTS The model applied in chapter 4.2 for use of proxy reports regarding the physical impact of MS should be tested in other MS patient populations. In addition, in future research should be evaluated whether the same approach can be successful for other PROs in MS. Based on our own results we are convinced that proxy responses, especially for physical aspects of the disease, do have value. The question remains how to handle the situation where we can use proxy reports on physical (impact) scales but not on psychological or cognitive scales in the future? Will proxy measures than still retain value? Longitudinal studies in MS do not only concern the physical status of MS patients, many researchers are currently focused on the psychological and cognitive aspects of the disease. If they are not able to use proxy responses on these scales, their data collection would still be incomplete. What is the consequence of using proxy ratings for only part of the spectrum? An additional question is whether we can use patient and proxy ratings interchangeably? To answer these questions we advise to continue the collection of both proxy and patient data in larger cohorts with diverse profiles and different follow up periods.
CONCLUSION This thesis was focused on the evaluation of patient and proxy responses on patient reported outcome scales that are often used in MS research, to optimize the quality of long term data collection. We can conclude that proxies, rating the patients’ disease state or disease impact, can be a valuable source of information, especially on PROs measuring physical aspects of MS. For the physical impact of MS we developed a model to estimate missing patient scores in longitudinal studies with good accuracy. However, we also observed that there are some factors causing discrepancies between patients and proxies, especially on (neuro)psychological scales. We found patients’ anxiety, some neuropsychological test components and proxies’ depression and caregiver burden to be explanatory for differences between patients and proxies. However, we are hopeful that in the future, with good models to estimate long term patient scores, proxy reported outcomes can be a valuable source of information.
106
Summary and General Discussion
REFERENCES 1
van der Linden FA, Kragt JJ, Klein M et al. Psychometric evaluation of the multiple sclerosis impact scale (MSIS-29) for proxy use. J. Neurol. Neurosurg. Psychiatry 2005; 76: 1677-81.
2
Benedict RH, Munschauer F, Linn R et al. Screening for multiple sclerosis cognitive impairment using a self-administered 15item questionnaire. Mult. Scler. 2003; 9: 95-101.
3
Benedict RH, Cox D, Thompson LL et al. Reliable screening for neuropsychological impairment in multiple sclerosis. Mult. Scler. 2004; 10: 675-8.
4
O'Brien A, Gaudino-Goering E, Shawaryn M et al. Relationship of the Multiple Sclerosis Neuropsychological Questionnaire (MSNQ) to functional, emotional, and neuropsychological outcomes. Arch. Clin. Neuropsychol. 2007; 22: 933-48.
5
6
7
8
Vanotti S, Benedict RH, Acion L, Caceres F. Validation of the Multiple Sclerosis Neuropsychological Screening Questionnaire in Argentina. Mult. Scler. 2009; 15: 244-50. Amato MP, Portaccio E, Goretti B et al. Cognitive impairment in early stages of multiple sclerosis. Neurol. Sci. 2010; 31: S211-S214. Bobholz JA, Rao SM. Cognitive dysfunction in multiple sclerosis: a review of recent developments. Curr. Opin. Neurol. 2003; 16: 283-8. Rao SM, Leo GJ, Bernardin L, Unverzagt F. Cognitive dysfunction in multiple sclerosis. I. Frequency, patterns, and prediction. Neurology 1991; 41: 685-91.
9
Rao SM, Leo GJ, Ellington L et al. Cognitive dysfunction in multiple sclerosis. II. Impact on employment and social functioning. Neurology 1991; 41: 692-6.
10 Rogers JM, Panegyres PK. Cognitive impairment in multiple sclerosis: evidencebased analysis and recommendations. J. Clin. Neurosci. 2007; 14: 919-27. 11 Sprangers MA, Aaronson NK. The role of health care providers and significant others in evaluating the quality of life of patients with chronic disease: a review. J. Clin. Epidemiol. 1992; 45: 743-60. 12 McGuigan C, Hutchinson M. Confirming the validity and responsiveness of the Multiple Sclerosis Walking Scale-12 (MSWS-12). Neurology 2004; 62: 2103-5. 13 Giordano A, Pucci E, Naldi P et al. Responsiveness of patient reported outcome measures in multiple sclerosis relapses: the REMS study. J. Neurol. Neurosurg. Psychiatry 2009; 80: 1023-8. 14 McGuigan C, Hutchinson M. The multiple sclerosis impact scale (MSIS-29) is a reliable and sensitive measure. J. Neurol. Neurosurg. Psychiatry 2004; 75: 266-9. 15 Motl RW, McAuley E, Mullen S. Longitudinal measurement invariance of the Multiple Sclerosis Walking Scale-12. J. Neurol. Sci. 2011; 305: 75-9. 16 van der Linden FA, Kragt JJ, van Bon M et al. Longitudinal proxy measurements in multiple sclerosis: patient-proxy agreement on the impact of MS on daily life over a period of two years. BMC. Neurol. 2008; 8: 2.
107
5
108
SAMENVATTING
Het gebruik van proxy gerapporteerde uitkomstmaten in MS onderzoek
Samenvatting
ACHTERGROND Multipele sclerose (MS) is een chronische ziekte van het centrale zenuwstelsel (hersenen en ruggenmerg) waarbij beschadigingen zorgen voor aantasting van de myelineschede, een isolerende beschermlaag rondom de zenuwen. Dit veroorzaakt vertraagde geleiding van de informatie in de zenuwen en verlies van zenuwen en axonen, de uitlopers van zenuwcellen die zorgen voor informatieoverdracht. Daardoor kunnen op verschillende plaatsen in het lichaam neurologische klachten optreden. MS is de meest voorkomende chronische ziekte e
onder jong volwassenen. De eerste symptomen treden vaak op rond het 30 levensjaar. MS komt meer voor bij vrouwen dan bij mannen. Naar schatting zijn er wereldwijd 2,3 miljoen mensen met MS, waarvan de meesten leven in Noord Amerika of Europa. Helaas is de oorzaak van de ziekte nog altijd niet duidelijk. De diagnose MS is gebaseerd op klinische bevindingen en MRI-scans van het centrale zenuwstelsel. Het beloop van MS verschilt per patiënt. Er kunnen episodes van klachten optreden (zogenaamde schubs) met bijvoorbeeld uitval van het gezichtsveld, spierzwakte, gevoelsstoornissen, coördinatie- en balansproblemen, blaasproblemen, geheugenstoornissen en vermoeidheid. Bij een eenmalige episode van klachten spreken we van CIS (clinically isolated syndrome) en heeft men nog niet de diagnose MS. De meeste CIS-patiënten ervaren uiteindelijk meerdere episodes van klachten en dan is er sprake van MS, waarbij periodes van klachten gevolgd worden door, al dan niet volledig, herstel (relapsing-remitting MS). Een gedeelte van deze mensen gaat over naar de secundair progressieve vorm. Hierbij vindt een geleidelijke achteruitgang plaats, onafhankelijk van schubs (die verderop in het beloop ook weg kunnen blijven). Ongeveer 20% van de patiënten ervaart vanaf het begin van de ziekte een geleidelijke achteruitgang zonder schubs en zonder herstel, dit subtype wordt primair progressieve MS genoemd. Om ziekteactiviteit en ziekteprogressie te kunnen evalueren zijn er verschillende methoden ontwikkeld die beperkingen, handicaps en de impact van MS op het dagelijks leven meten. Sommige metingen worden gedaan door een arts of psycholoog, bijvoorbeeld neurologisch en neuropsychologisch onderzoek, andere metingen zijn rechtstreeks afkomstig van de patiënt zelf, de zogenaamde zelf-rapportages of vragenlijsten (PRO: Patient Reported 110
Samenvatting
Outcomes). Om de hele range van MS symptomen goed in beeld te kunnen brengen is de combinatie van bovengenoemde maten het beste. Patiënt gerapporteerde uitkomstmaten worden steeds meer gebruikt om de ziekte impact, het fysiek en psychologisch functioneren, cognitie en kwaliteit van leven bij MS in kaart te brengen. Het voordeel van deze klinische schalen is dat ze gemakkelijk in te vullen zijn en dat het meestal niet veel tijd kost. Met deze schalen kan de ziekte per patiënt en op groepsniveau geëvalueerd worden, om zo de korte en lange termijn effecten van MS te meten. Het is daarbij wel belangrijk dat de klinische schalen aan bepaalde kwaliteitseisen voldoen. In dit proefschrift wordt met name gebruik gemaakt van de volgende patiënt gerapporteerde uitkomstmaten: De MSIS-29, waarbij de fysieke en psychologische impact van MS op het dagelijks leven gemeten wordt, de MSWS-12, die de loopproblemen van de patiënt in kaart brengt, de MSNQ, een vragenlijst die screent op neuropsychologische klachten en de GNDS, een interview waarbij algemene neurologische klachten worden uitgevraagd. Bovengenoemde vragenlijsten en het interview zijn terug te vinden in de Appendix. Bij een ziekte als MS kan men er helaas niet altijd van uit gaan dat de respons van patiënten volledig betrouwbaar is, vooral vanwege problemen met het cognitief functioneren of door stemmingsstoornissen. Er kunnen in onderzoeken over langere tijd ook gegevens missen, door het verergeren van de ziekte of omdat patiënten niet meer deel willen nemen aan omgeving van de patiënt, de zogenaamde proxy respondenten. In dit proefschrift is de proxy altijd de partner van de MS patiënt. Uit eerder onderzoek is gebleken dat de mening van patiënten en proxies niet altijd overeenkomt en dat proxies soms meer klachten of een hogere impact van de ziekte rapporteren dan de patiënten zelf. Een individuele proxy score kan dus niet direct als een vervangende patiëntscore gebruikt worden. Daarnaast moeten we er ook rekening mee houden dat MS niet alleen invloed heeft op de patiënt, ook de proxy (zeker als het de partner is) moet omgaan met de fysieke, psychologische en bijvoorbeeld ook financiële gevolgen die de ziekte met zich meebrengt. Hierdoor kunnen rollen binnen de relatie of het gezin veranderen. Daarom is het belangrijk
111
NL SV
onderzoek. In die gevallen kan het nuttig zijn om gebruik te maken van personen in de directe
Samenvatting
om ook de fysieke en mentale gezondheid, stemming en zorglast van de proxy mee te nemen in het onderzoek.
DOEL VAN DE STUDIE In dit proefschrift maken we gebruik van de respons van patiënten en proxies (die rapporteren over de MS-klachten van de patiënt) op vragenlijsten die veelvuldig gebruikt worden in wereldwijd MS onderzoek. We onderzoeken de overeenkomst van de antwoorden van patiënten en proxies en de factoren die daarop van invloed zijn. Het uiteindelijke doel is om lange termijn uitkomsten van patiënten te kunnen berekenen door middel van proxy uitkomsten. Dit is met name belangrijk als we uitkomsten van patiënten missen of wanneer deze minder betrouwbaar zijn.
RESULTATEN VAN HET ONDERZOEK In hoofdstuk 2 is een studie beschreven waarin werd onderzocht of de MSNQ gelijke resultaten geeft voor patiënten en proxies met betrekking tot het neuropsychologisch functioneren van de patiënt. Uit de resultaten blijkt dat de antwoorden van de patiënten meer samenhangen met de angst en depressie van de patiënt, dan met het neuropsychologisch functioneren zoals gemeten door de psycholoog. Bij de proxy versie zien we wel een hogere samenhang met neuropsychologische testen. De conclusie is dat de patiënt en proxy versie niet hetzelfde meten en dat de proxy versie van de MSNQ een beter beeld geeft van het neuropsychologisch functioneren van de patiënt. In hoofdstuk 3 wordt de overeenstemming tussen patiënten en proxies op één tijdstip op vijf verschillende uitkomstmaten (vragenlijsten) beschreven. In de studie beschreven in hoofdstuk 3.1 worden de verwachtingen bevestigd: Proxies kunnen de ziekte van de patiënt goed en in overeenstemming met de patiënt beoordelen wanneer het gaat om specifieke en goed te observeren ziekteaspecten of fysieke beperkingen. Voor psychologische schalen en het algemeen functioneren lijkt dit lastiger te zijn. In hoofdstuk 3.2 wordt beschreven hoe we de verschillende antwoorden van patiënten en proxies kunnen verklaren. Het belangrijkste resultaat is dat wanneer de proxy een hogere zorglast ervaart, deze de situatie van de patiënt 112
Samenvatting
negatiever beoordeelt dan de patiënt zelf. Dit geldt voor alle onderzochte schalen. Ook angst bij patiënten en depressie bij proxies kunnen uiteenlopende reacties gedeeltelijk verklaren. Met het doel om lange termijn uitkomsten van patiënten te kunnen berekenen, waren we ook geïnteresseerd in hoe patiënt- en proxy uitkomsten zich over de tijd verhouden. In het onderzoek beschreven in hoofdstuk 4.1 ligt de focus op de fysieke impact van MS en de loopfunctie van MS-patiënten. We bekeken de verschillen over de tijd tussen de antwoorden van patiënten en proxies en onderzochten welke factoren daarop van invloed zouden kunnen zijn. De meerderheid van de patiënt-proxy koppels was het eens over verbetering, stabilisatie of verslechtering over 2 jaar tijd. Toch verschilt ook een aanzienlijk deel van mening over de veranderingen. Patiënten die positiever zijn dan hun partner zijn vaak ouder en langer samen met hun partner. Wanneer proxies negatiever zijn dan de patiënten blijken opnieuw angst, depressie en zorglast een rol te spelen. Vervolgens is bekeken (hoofdstuk 4.2) of voor de veranderingen van de fysieke impact van MS over tijd een model opgesteld kon worden om de patiënt score te berekenen aan de hand van de proxy score en eventuele andere variabelen. Op die manier kunnen de proxy scores in de toekomst ook echt gebruikt worden. Uit de resultaten blijkt dat de fysieke impactscore van de patiënt op een later tijdstip (bijvoorbeeld 2 jaar) vrij nauwkeurig bepaald kan worden aan de hand van de proxy score op dat tijdstip en de patiënt score waarmee is gestart (baseline).
In dit proefschrift wordt het gebruik van proxy metingen binnen MS onderzoek beschreven. Enerzijds waren de beperkingen van het gebruik van proxy metingen
te voorspellen,
bijvoorbeeld dat de verschillen tussen patiënt en proxy groter zijn wanneer de schaal minder goed observeerbare factoren (zoals psychologisch functioneren) meet. Anderzijds kan een aanzienlijk deel van de verschillen tussen patiënt en proxy verklaard worden door middel van patiënt- en proxy gerelateerde variabelen zoals zorglast, angst en depressie. Op één tijdstip blijft het lastig om patiëntmetingen te vervangen door proxymetingen, maar bij het in kaart brengen van veranderingen over de tijd lijken proxymetingen een goede benadering van patiëntmetingen weer te kunnen geven.
113
NL SV
CONCLUSIE
114
APPENDIX
List of abbreviations Overview of patient reported outcome scales
Appendix
2Y
2 years
6M
6 months
ANOVA
analysis of variance
AUC
area under the curve
BRBN
brief repeatable battery of neuropsychological tests
CI
confidence interval
CIS
clinically isolated syndrome
CSI
caregiver strain index
d
mean difference
EDSS
expanded disability status scale
GNDS
guy’s neurological disability scale
HADS
hospital anxiety and depression scale
ICC
intraclass correlation coefficient
IQR
interquartile range
κw
weighted kappa
LOA
limits of agreement
MRI
magnetic resonance imaging
MS
multiple sclerosis
MSIS-29
multiple sclerosis impact scale
MSNQ
multiple sclerosis neuropsychological screening questionnaire
MSNQ-I
multiple sclerosis neuropsychological screening questionnaire – informant version
MSNQ-P
multiple sclerosis neuropsychological screening questionnaire – patient version
MSWS-12
multiple sclerosis walking scale
116
N
number
na
not available
p
probability
PASAT
paced auditory serial addition test
PP
primary progressive
PROs
patient reported outcome scale(s)
ROC
receiver operating characteristic
RR
relapsing remitting
SD
standard deviation
SDMT
symbol digit modalities test
SF-36
short-form health survey
SP
secondary progressive
SPART
spatial recall test
SPSS
statistical package for social sciences
SRB
self-rated burden scale
SRT
selective reminding test
WLG
word list generation test APPENDIX
List of abbreviations
117
Appendix
Multiple Sclerosis Impact Scale (MSIS-29) • • •
De volgende vragen hebben betrekking op uw mening over de invloed van MS op uw dagelijks leven gedurende de afgelopen 2 weken. Omcirkel bij elke vraag het nummer dat uw situatie het beste beschrijft. S.V.P. alle vragen beantwoorden.
Gedurende de afgelopen twee weken, in welke mate heeft MS u beperkt in….
Helemaal niet
Een beetje
Matig
Tamelijk veel
Heel erg
1.
Het uitvoeren van lichamelijke inspanning?
1
2
3
4
5
2.
Het stevig vastpakken van dingen? (bijvoorbeeld opendraaien van kranen)
1
2
3
4
5
3.
Het dragen van dingen? (bijvoorbeeld boodschappentas)
1
2
3
4
5
Helemaal niet
Een beetje
Matig
Tamelijk veel
Heel erg
Gedurende de afgelopen twee weken, in welke mate werd u gehinderd door…. 4.
Evenwichtsstoornissen?
1
2
3
4
5
5.
Moeilijkheden om uzelf te verplaatsen in huis?
1
2
3
4
5
6.
Onhandigheid?
1
2
3
4
5
7.
Stijfheid?
1
2
3
4
5
8.
Zwaar gevoel in armen en/of benen?
1
2
3
4
5
9.
Trillen van armen en/of benen?
1
2
3
4
5
10.
Spasmen in armen en/of benen?
1
2
3
4
5
11.
Dat uw lichaam niet doet wat u zou willen dat het doet?
1
2
3
4
5
12.
Afhankelijkheid van anderen om dingen voor u te doen?
1
2
3
4
5
13.
Beperkingen in uw sociale activiteiten en vrije tijd bestedingen thuis?
1
2
3
4
5
14.
Het meer aan huis gebonden zijn, dan u zou wensen?
1
2
3
4
5
Controleert u alstublieft of u alle vragen heeft beantwoord voordat u naar de volgende pagina gaat.
118
Patient Reported Outcome scales
15. 16. 17.
Problemen met het gebruik van uw handen tijdens het uitvoeren van dagelijkse bezigheden? Noodgedwongen minder tijd moeten gaan besteden aan werk of andere dagelijkse bezigheden? Problemen met het gebruik van vervoermiddelen? (bijvoorbeeld auto, bus, trein, taxi, etc.)
Helemaal niet
Een beetje
Matig
Tamelijk veel
Heel erg
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
18.
Meer tijd nodig hebben om dingen te doen?
1
2
3
4
5
19.
Problemen met spontaan dingen te doen? (bijvoorbeeld: opeens ergens naar toe gaan)
1
2
3
4
5
20.
Het moeten haasten naar het toilet?
1
2
3
4
5
21.
U onwel voelen?
1
2
3
4
5
22.
Slaapstoornissen?
1
2
3
4
5
23.
Geestelijk moe voelen?
1
2
3
4
5
24.
U zorgen maken in verband met uw ziekte MS?
1
2
3
4
5
25.
Angstig of gespannen voelen?
1
2
3
4
5
26.
Snel geïrriteerd, ongeduldig of driftig zijn?
1
2
3
4
5
27.
Concentratieproblemen?
1
2
3
4
5
28.
Gebrek aan zelfvertrouwen?
1
2
3
4
5
29.
Depressiviteit?
1
2
3
4
5
APPENDIX
Gedurende de afgelopen twee weken, in welke mate werd u gehinderd door….
Controleert u alstublieft of bij ALLE vragen ÉÉN cijfer heeft omcirkeld.
119
Appendix
Multiple Sclerosis Walking Scale (MSWS-12) • • • •
De volgende vragen hebben betrekking op uw beperkingen in het lopen als gevolg van MS gedurende de afgelopen 2 weken. Omcirkel bij elke vraag het ene nummer dat uw mate van beperking het beste beschrijft. S.V.P. alle vragen beantwoorden, ook al lijken sommige vragen op elkaar of lijken ze niet op u van toepassing. Als u helemaal niet kunt lopen, kruis dan dit vakje aan: □
Gedurende de afgelopen twee weken, in welke mate heeft de MS …..
Helemaal niet
Een beetje
Matig
Tamelijk veel
Heel erg
1.
U beperkt in uw mogelijkheid te lopen?
1
2
3
4
5
2.
U beperkt in uw mogelijkheden te rennen?
1
2
3
4
5
3.
U beperkt in uw mogelijkheden de trap op en af te gaan?
1
2
3
4
5
4.
Het u moeilijker gemaakt om te staan terwijl u dingen deed?
1
2
3
4
5
5.
U beperkt in uw balans als u stond of liep?
1
2
3
4
5
6.
U beperkt in hoe ver u kon lopen?
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
7.
8.
9.
Ervoor gezorgd dat lopen u meer moeite kostte? Het noodzakelijk gemaakt dat u steun gebruikte bij het lopen in huis (b.v. vastpakken van meubels of gebruik van een stok, etc)? Het noodzakelijk gemaakt dat u steun gebruikte bij het lopen buitenshuis (b.v. gebruik van een stok of looprekje, etc.)?
10.
Ervoor gezorgd dat u langzamer ging lopen?
1
2
3
4
5
11.
Invloed gehad op hoe soepel u liep?
1
2
3
4
5
12.
Ervoor gezorgd dat u zich moest concentreren op het lopen?
1
2
3
4
5
Controleert u alstublieft of bij ALLE vragen ÉÉN cijfer heeft omcirkeld.
120
Patient Reported Outcome scales
Multiple Sclerosis Neuropsychological Screening Questionnaire (MSNQ Patient) Nogal vaak stoort het dagelijks leven
Soms zelden een probleem
Zeer zelden geen probleem
Nooit komt niet voor
Kruis het best passende antwoord aan
Heel vaak zeer storend
INSTRUCTIES: De volgende vragen peilen de problemen die u mogelijk ondervindt. Geef aan hoe vaak deze problemen voorkomen EN hoe ernstig deze zijn. Baseer uw antwoorden op hoe het met u ging in de afgelopen 3 maanden.
4
3
2
1
0
1. Bent u gemakkelijk afgeleid? 2. Dwalen uw gedachten af terwijl u naar iemand luistert? 3. Bent u traag in het oplossen van problemen? 4. Vergeet u gemakkelijk afspraken of verplichtingen? 5. Vergeet u gemakkelijk wat u leest? 6. Heeft u moeite om programma’s te beschrijven die u onlangs gezien heeft? 7. Is het nodig dat instructies voor u worden herhaald? 8. Moet u eraan herinnerd worden om taken uit te voeren? 9. Vergeet u boodschappen en klusjes die gepland waren? 10. Heeft u moeite om vragen te beantwoorden? APPENDIX
11. Heeft u moeite om twee dingen tegelijk te volgen? 12. Mist u soms de kern van wat iemand probeert te zeggen? 13. Heeft u soms moeite om u te beheersen? 14. Lacht of huilt u nagenoeg zonder aanleiding? 15. Praat u te veel of bent u te veel gericht op uw eigen zaken? Copyright © 2004 Ralph H.B. Benedict
121
Appendix
GUY'S NEUROLOGICAL DISABILITY SCALE (GNDS) 1 Cognitie Heeft u problemen met het geheugen en/of concentreren? ja Denken uw familie of vrienden dat u zo'n probleem heeft? ja Indien het antwoord op een van de vragen ja is: Heeft u hulp van anderen nodig om uw dagelijkse activiteiten te plannen, omgaan met geld of besluiten te nemen? ja Indien ja: (voor de onderzoeker) is de patiënt georiënteerd in plaats, tijd en persoon? ja, helemaal ja, gedeeltelijk* nee, totaal gedesoriënteerd*
nee nee nee
*Als de patiënt niet volledig georiënteerd is moeten alle antwoorden gecontroleerd worden met de (belangrijkste) verzorger en moeten zijn/haar antwoorden gebruikt worden. 2 Stemming Heeft u zich angstig, geïrriteerd, depressief gevoeld de laatste maand, of ja nee andere stemmingswisselingen gehad? Gebruikt u hier medicatie voor? ja nee Indien het antwoord op de eerste vraag ja is: Beïnvloed het probleem u in het doen van een van uw dagelijkse activiteiten zoals werk, huishouden, normale sociale activiteiten met familie/vrienden? ja nee Indien ja: Is het probleem zo erg dat het u weerhoudt al uw gebruikelijke activiteiten te doen? ja nee Bent u de laatste maand opgenomen geweest in een ziekenhuis voor behandeling van dit probleem? ja nee 3 Visus Heeft u een probleem met uw visus dat niet gecorrigeerd kan worden met een bril? ja nee Indien ja: kunt u normale (kranten)letters lezen (zo nodig met een bril, maar geen vergrootglas oid.)? ja nee Indien nee: kunt u grote (kranten)letters/krantenkoppen lezen? ja nee Indien nee: kunt u uw vingers tellen als u uw hand voor u uit steekt? ja nee Indien nee: kunt u uw hand zien als u uw hand beweegt voor u? ja nee 4 Spraak en communicatie Heeft u problemen met spreken? Indien ja: moet u woorden herhalen als u spreekt met uw familie of goede vrienden? Indien ja: moet u dingen opschrijven, of gebruik maken van een verzorger om te zorgen dat anderen u begrijpen? Indien ja: (voor de onderzoeker) is de patiënt in staat om adequaat te communiceren met behulp van deze methode(s)?
122
ja ja
nee nee
ja
nee
ja
nee
Patient Reported Outcome scales
5 Slikken Moet u voorzichtig zijn met het doorslikken van vast of vloeibaar voedsel? Indien ja: moet u voorzichtig zijn bij de meeste maaltijden? Indien ja: heeft u een speciaal dieet nodig? Indien ja: verslikt u zich bij de meeste maaltijden? Indien ja: heeft u een voedingssonde? 6 Armen/handen Heeft u problemen met uw armen en/of handen Indien ja: heeft u problemen met het dichtdoen van een rits of knoopjes ? Indien ja: kunt u al uw ritsen en knoopjes dichtdoen zonder hulp? Heeft u een probleem met het strikken van veters of touw? Indien ja: kunt u veters strikken of touw knopen zonder hulp? Heeft u een probleem met het wassen en kammen van uw haar? Indien ja: kunt u uw haar wassen en kammen zonder hulp? Heeft u problemen met zelfstandig eten? Indien ja: kunt u zelfstandig eten? Indien de patiënt geen van de bovengenoemde handelingen kan verrichten Kunt u uw handen of armen gebruiken voor andere handelingen?
ja ja ja ja ja
nee nee nee nee nee
ja ja ja ja ja ja ja ja ja
nee nee nee nee nee nee nee nee nee
ja
nee
8 Blaas Heeft u problemen met uw blaas? Gebruikt u medicijnen voor dit probleem Indien het antwoord op de eerste vraag ja is: Moet u zich haasten naar het toilet, frequent gaan, of heeft u problemen met starten van plassen? Bent u incontinent geweest afgelopen maand? Indien ja: ben u incontinent geweest de afgelopen week? Indien ja: was u incontinent elke dag? Gebruikt u een katheter om uw blaas te legen? Heeft u een verblijfskatheter of gebruikt u een condoom-katheter om de urine op te vangen?
ja ja
nee nee
ja
nee
ja ja ja ja ja
nee nee nee nee nee
123
APPENDIX
7 Benen Heeft u problemen bij het lopen? ja nee Indien ja: gebruikt u een loophulpmiddel? ja nee Indien ja: zonder hulpmiddel A. Hoe verplaatst u zich gewoonlijk met een stok of kruk of aan de arm van iemand anders buitenshuis? of of met twee stokken of krukken of een stok of kruk en aan de arm van iemand anders of met een rolstoel B. Hoe verplaatst u zich gewoonlijk zonder hulpmiddel binnenshuis? of met een stok of kruk of aan de arm van iemand anders of met twee stokken of krukken of een stok of kruk en aan de arm van iemand anders of met een rolstoel Indien u een rolstoel gebruikt: Kunt u staan en een paar passen lopen met hulp? ja nee
Appendix
9 Darm/ontlasting Heeft u problemen met uw ontlasting? Gebruikt u medicijnen voor dit probleem? Indien het antwoord op de eerste vraag ja is: Heeft u last van obstipatie? Indien ja: gebruikt u laxantia of zetpillen hiervoor? Gebruikt u klysma's? Moet u de ontlasting manueel verwijderen? Moet u zich haasten naar het toilet voor de ontlasting? Bent u incontinent (geweest) voor ontlasting de afgelopen week? Indien ja: was u incontinent elke week? 10 Sexueel De volgende vragen hebben betrekking op seksuele functies. Heeft u er bezwaar tegen dat ze worden gesteld? Indien de patiënt akkoord gaat: Heeft u problemen met betrekking tot uw seksuele functies? Indien ja: is uw seksuele verlangen verminderd? Heeft u problemen met het bevredigen van uzelf of uw partner? Worden uw seksuele functies beïnvloed door lichamelijke problemen, bijvoorbeeld veranderd gevoel, spasme, pijn etc.? Heeft u problemen met: Mannen: erectie/ejaculatie; Vrouwen: vochtig worden /orgasme Indien er lichamelijke of seksuele problemen zijn: Weerhoudt een van deze problemen u totaal van seksuele activiteit? 11 Vermoeidheid Heeft u zich moe gevoeld, of was u snel moe de afgelopen maand Indien ja: Heeft u zich de meeste dagen moe gevoeld? Heeft de vermoeidheid invloed gehad op uw dagelijkse activiteiten zoals werk, huishouden of normale sociale activiteiten met familie/vrienden? Indien ja: is de vermoeidheid zo erg dat het u weerhoudt van al uw gebruikelijke dagelijkse activiteiten? Indien ja: is de vermoeidheid zo ernstig dat het u weerhoudt van alle lichamelijke activiteiten? 12 Anders Heeft u nog andere problemen veroorzaakt door MS die we nog niet genoemd hebben (bijvoorbeeld pijn, spasme, duizeligheid, dubbelzien)? Gebruikt u hier medicijnen voor? Indien het antwoord op een van de vragen ja is: Wat is uw ergste probleem: …………………………………… Beïnvloed dit probleem u in uw dagelijkse activiteiten? Is dit probleem zo erg dat het u weerhoudt van al uw dagelijkse activiteiten? Bent u opgenomen geweest in een ziekenhuis voor de behandeling van dit probleem?
124
ja ja
ja ja ja ja ja ja ja
nee nee
nee nee nee nee nee nee nee
ja nee niet van toepassing (celibaat)
ja ja ja ja
nee nee nee nee
ja
nee
ja
nee
ja
nee
ja ja
nee nee
ja
nee
ja
nee
ja
nee
ja
nee
ja ja ja
nee nee nee
Patient Reported Outcome scales
SCORE 1 Cognitie 0-Geen cognitieve problemen. 1-Cognitieve problemen, niet opgemerkt door familie of vrienden. 2-Cognitieve problemen, opgemerkt door familie of vrienden, geen hulp van anderen nodig. 3-Cognitieve problemen, hulp van anderen nodig om dagelijkse activiteiten te plannen; volledig georiënteerd in plaats, tijd en persoon. 4-Cognitieve problemen, hulp van anderen nodig om dagelijkse activiteiten te plannen, niet volledig georiënteerd. 5-Volledig gedesoriënteerd in plaats, tijd en persoon. 2 Psychisch 0-Geen stemmingsproblemen. 1-Stemmingsproblemen asymptomatisch met huidige medicatie. 2-Stemmingsproblemen aanwezig, geen invloed op het uitvoeren van dagelijkse activiteiten. 3-Stemmingsproblemen die het uitvoeren van een aantal van de dagelijkse activiteiten beïnvloeden. 4-Stemmingsproblemen die de patiënt weerhouden van het uitvoeren van alle dagelijkse activiteiten. 5-Stemmingsproblemen die opname in een ziekenhuis voor behandeling vereisen. X-Onbekend (als score het gemiddelde (afgerond) van cognitie en vermoeidheid nemen). 3 Visus 0-Geen visus problemen. 1-Visusproblemen (dubbelzien, scotoom, wazig zien), nog wel in staat normale krantenletters te lezen. 2-Niet in staat normale letters te lezen. 3-Niet in staat grote letters te lezen. 4-Niet in staat vingers te tellen. 5-Niet in staat handbewegingen te zien. 4 Spraak 0-Geen spraakproblemen. 1-Spraakproblemen, de patiënt hoeft woorden niet te herhalen als hij met vreemden praat. 2-Spraakproblemen, de patiënt moet woorden herhalen als hij met vreemden praat. 3-Spraakproblemen, de patiënt moet woorden herhalen wanneer hij met familie/vrienden praat. 4-Spraakproblemen, spraak moeilijk te verstaan, patiënt kan goed communiceren d.m.v. schrift of met hulp van verzorger. 5-Spraakproblemen, spraak moeilijk te verstaan, patiënt is niet goed in staat adequaat te communiceren d.m.v. schrift of hulp van verzorger. APPENDIX
5 Slikken 0-Geen slikproblemen. 1-Moet voorzichtig zijn met doorslikken van vast of vloeibaar voedsel, maar niet bij de meeste maaltijden. 2-Moet voorzichtig zijn met doorslikken van vast of vloeibaar voedsel bij de meeste maaltijden. 3-Moet het eten speciaal bereiden om de consistentie te veranderen. 4-Neiging zich te verslikken bij de meeste maaltijden. 5-Dysphagie, voedingssonde nodig.
125
Appendix
6 Armen/handen 0-Geen problemen armen en handen. 1-Problemen in 1 of beide armen, geen invloed op het uitvoeren van de genoemde activiteiten. 2-Problemen in 1 of beide armen, die invloed hebben op een aantal van de genoemde activiteiten, maar ze niet onmogelijk maken. 3-Problemen in 1 of beide armen, beïnvloeden alle genoemde activiteiten of maken 1 of 2 van de genoemde activiteiten onmogelijk. 4-Problemen in 1 of beide armen, maken 3 of alle genoemde activiteiten onmogelijk. 5-Niet in staat beide armen te gebruiken voor een doelgerichte handeling. 7 Benen 0-Lopen wordt niet beïnvloed. 1-Lopen wordt beïnvloed, maar patiënt kan onafhankelijk lopen. 2-Gebruikt gewoonlijk 1 hulpmiddel (unilateraal) buitenshuis, maar loopt onafhankelijk binnenshuis. 3-Gebruikt gewoonlijk 2 hulpmiddelen (bilateraal) buitenshuis of 1 hulpmiddel (unilateraal) binnenshuis. 4-Gebruikt gewoonlijk een rolstoel buitenshuis of 2 hulpmiddelen (bilateraal) binnenshuis. 5-Gebruikt gewoonlijk een rolstoel binnenshuis. 8 Blaas 0-Geen blaas problemen. 1-Asymptomatisch met huidige medicatie. 2-Frequente mictie, urgentie, hesitatie zonder incontinentie. 3-Soms incontinent voor urine (1 keer of vaker de afgelopen maand, maar niet elke week) of intermitterend katheteriseren zonder incontinentie. 4-Frequent urine incontinentie (1 keer of vaker per week de afgelopen maand, maar niet dagelijks) of soms incontinent voor urine ondanks intermitterend katheteriseren. 5-Dagelijks incontinent voor urine of verblijfskatheter of condoomkatheter. 9 Darm/ontlasting 0-Geen ontlastingsproblemen. 1-Asymptomatisch met huidige medicatie of obstipatie die geen behandeling nodig heeft. 2-Obstipatie, wel laxantia of zetpillen nodig, of faecale urgentie. 3-Obstipatie, vereist gebruik van klysma's. 4-Obstipatie, manuele verwijdering van faeces nodig of soms incontinent voor faeces (1 keer of vaker de afgelopen maand, maar niet wekelijks). 5-Wekelijks incontinent voor faeces. 10 Sexueel 0-Normale seksuele functies of patiënten die vrijwillig celibaat zijn. 1-Verminderd seksueel verlangen. 2-Problemen met zelfbevrediging of bevrediging van de partner. 3-Lichamelijke problemen die het seksuele functioneren beïnvloeden, maar niet tegen houden. 4-Autonome problemen die het seksuele functioneren beïnvloeden, maar niet tegen houden. 5-Lichamelijke of autonome problemen die patiënt totaal weerhouden van seksuele activiteiten. X-Onbekend.(score als het afgeronde gemiddelde van benen-, blaas- en darm score)
126
Patient Reported Outcome scales
11 Vermoeidheid 0-Afwezig. 1-Soms vermoeid (een aantal dagen aanwezig). 2-Vaak vermoeid (de meeste dagen aanwezig). 3-Moedheid beïnvloed de patiënt in het uitvoeren van een aantal van zijn dagelijkse activiteiten. 4-Moeheid die ervoor zorgt dat de patiënt al zijn dagelijkse activiteiten moet laten. 5-Moeheid die de patiënt weerhoudt van alle lichamelijke activiteit. X-Onbekend. (score als het afgeronde gemiddelde van cognitie en stemming)
APPENDIX
12 Anders 0-Afwezig. 1-Asymptomatisch met huidige medicatie. 2-Problemen aanwezig, maar hebben geen invloed op het uitvoeren van dagelijkse activiteiten. 3-Problemen die invloed hebben op het uitoefenen van enkele van de dagelijkse activiteiten. 4-Problemen die de patiënt weerhouden van het uitoefenen van al zijn dagelijkse activiteiten. 5-Problemen die ziekenhuisopname vereisen voor behandeling.
127
128
ABOUT THE AUTHOR List of publications Biografie
List of publications
LIST OF PUBLICATIONS Sonder JM, Balk LJ, van der Linden FA, Bosma LV, Polman CH, Uitdehaag BM. Towards the use of proxy reports for estimating long term patient reported outcomes in multiple sclerosis. (under revision: Mult Scler. 2014) Sonder JM, Balk LJ, Bosma LV, Polman CH, Uitdehaag BM. Do patient and proxy agree? Long term changes in Multiple Sclerosis physical impact and walking ability on patient reported outcome scales. In press: Mult Scler. 2014. Bosma LV, Sonder JM, Kragt JJ, Polman CH, Uitdehaag BM. Detecting clinically relevant changes in progressive multiple sclerosis. (submitted) Mokkink LB, Knol DL, van der Linden FA, Sonder JM, MSc; D'hooghe M, Uitdehaag BM. The Arm Function in Multiple Sclerosis Questionnaire (AMSQ): Development and validation of a new tool. (submitted) Sonder JM, Burggraaff J, Knol DL, Polman CH, Uitdehaag BM. Comparing long-term results of PASAT and SDMT scores in relation to neuropsychological testing in multiple sclerosis. Mult Scler. 2014;20(4):481-8 Sonder JM, Holman R, Knol DL, Bosma LV, Polman CH, Uitdehaag BM. Analyzing differences between patient and proxy on Patient Reported Outcomes in multiple sclerosis. J Neurol Sci. 2013;15:143-147. Sonder JM, Mokkink LB, van der Linden FA, Polman CH, Uitdehaag BM. Validation and interpretation of the Dutch version of the Multiple Sclerosis Neuropsychological Screening Questionnaire. J Neurol Sci. 2012;15:91-96. Balk LJ, Sonder JM, Strijbis EM, Twisk JW, Killestein J, Uitdehaag BM, Polman CH, Petzold A. The physiological variation of the retinal nerve fiber layer thickness and macular volume in humans as assessed by spectral domain-optical coherence tomography. Invest Ophthalmol Vis Sci. 2012:1251-1257. Sonder JM, Bosma LV, van der Linden FA, Knol DL, Polman CH, Uitdehaag BM. Proxy measurements in multiple sclerosis: agreement on different patient-reported outcome scales. Mult Scler. 2012;18:196-201.
130
Biografie
BIOGRAFIE Judith Maria Sonder werd op 12 december 1983 geboren in Leidschendam. Ze groeide tot haar zesde jaar op in Zoetermeer waarna het gezin verhuisde naar Badhoevedorp. In 2002 behaalde ze haar Gymnasium diploma aan het Keizer Karel College in Amstelveen en startte haar studie psychologie aan de Vrije Universiteit in Amsterdam. In 2004 werd ze als student assistent aangenomen bij het onderzoek naar Multiple Sclerose op de polikliniek neurologie van het VU medisch centrum, onder leiding van prof. dr. C.H. Polman. In 2006 behaalde ze cum laude haar master ontwikkelingspsychologie. Na haar studie werkte ze nog enige tijd binnen het MS onderzoek, waarna ze in 2007 naar Suriname vertrok om daar te werken als psycholoog, verbonden aan het Streekziekenhuis Nickerie, en wetenschappelijk onderzoek te doen in Paramaribo, onder leiding van prof. dr. T. Graafsma. In 2009 keerde ze terug naar het MS centrum van het VU medisch centrum Amsterdam en startte haar promotieonderzoek dat leidde tot dit proefschrift, onder begeleiding van prof. dr. B.M.J. Uitdehaag en prof. dr. C.H. Polman. Naast het promotieonderzoek heeft Judith in 2008 een eigen bedrijf opgezet waarin ze als weddingplanner verschillende huwelijken en evenementen heeft georganiseerd. Ook begeleidt ze sinds 2007 als freelancer vanuit het persoonsgebonden budget kinderen met ontwikkelingsproblemen. Judith is in 2011 getrouwd met Raymond Swager. Ze hebben een dochter Sofia (2012) en in
AUTHOR
juli 2014 verwachten zij hun tweede kind.
131
132
DANKWOORD
Dankwoord
Uiteraard was dit proefschrift niet tot stand gekomen zonder de hulp van heel veel plezierige, lieve en dierbare collega’s, vrienden en familieleden die ik (op de meest gelezen pagina’s van dit proefschrift) kort probeer te bedanken. Maar laten we beginnen bij de meest belangrijke personen: De MS patiënten en hun partners. De meeste patiënten kwamen al voor andere onderzoeksprojecten naar de VU en vonden het leuk dat hun partner ook bij het onderzoek betrokken werd. Veel partners waardeerden het feit dat nu ook naar hun mening werd gevraagd. Gelukkige, verdrietige en heftige verhalen passeerden in de jaren daaropvolgend. Hartelijk dank voor jullie inzet en openhartigheid! Daarnaast zou ik graag de Stichting MS Research, Novartis Pharma, Bayer HealthCare, Teva Nederland en BiogenIdec International willen bedanken voor de financiële ondersteuning van de kosten die dit proefschrift met zich mee heeft gebracht. Grote dank aan mijn promotoren, prof. dr. Bernard Uitdehaag en prof. dr. Chris Polman. Beste Bernard, dank dat je dit project aan mij hebt willen toevertrouwen. Door jou begon ik op de afdeling Epidemiologie & Biostatistiek waar ik leuke herinneringen aan heb. Naast het proxy onderzoek heb je me ook organisatorische taken toevertrouwd, zoals afdelingsuitjes en afscheidsborrels (met self made kalenders en dergelijke…), een welkome afwisseling wat mij betreft! Bedankt voor je betrokkenheid bij het project, je bijdrage aan mijn professionele ontwikkeling, je nauwkeurige correcties op mijn manuscripten en de prettige samenwerking door de jaren heen! En hartelijk dank voor de flexibiliteit en vrijheid waarmee ik heb mogen werken, waardoor ik in de afgelopen jaren meer dan een proefschrift heb kunnen realiseren. Beste Chris, het bijzondere aan jou is dat je iets in mij zag wat ik zelf nog niet had gezien. Niet meteen hoor, want we ontmoetten elkaar pas nadat ik een jaar lang, zwoegend in een hokje op de poli, achterstallige data had ingevoerd. Maar vanaf dat moment heb ik altijd een bijzonder soort vertrouwen gevoeld. Jij hebt mijn organisatietalent als een van de eersten benoemd, waardoor ik in de afgelopen jaren niet alleen een proefschrift heb geschreven, maar ook als weddingplanner successen heb behaald. Daarnaast heb je me vaker in vertrouwen genomen over uiteenlopende onderwerpen, waardoor ik me erg gewaardeerd voelde. Bedankt voor het vertrouwen en de leuke samenwerking! 134
Dankwoord
Alle leden van de promotiecommissie: Prof. dr. Neil Aaronson, prof. dr. Riekie de Vet (tevens inspirator in de klinimetrie), prof. dr. Frank Snoek, dr. Patricia van Oppen, dr. Bob van Oosten en dr. Femke van der Linden, hartelijk dank voor jullie beoordeling van mijn proefschrift en jullie bereidheid om zitting te nemen in de promotiecommissie. Beste Femke, zonder jou had ik hier nooit gestaan! Doordat jij en Jolijn op zoek waren naar een student, ben ik op de poli neurologie terecht gekomen en na jouw promotie mocht ik het proxy onderzoek voortzetten. Bedankt voor de inspiratie en de vele gezellige koffie momenten! Mijn lieve collega’s, zonder jullie had ik deze 5 jaar echt nooit vol kunnen houden! Alle vrouwelijke collega’s met wie ik de laatste jaren samen heb mogen werken: Alexandra, Anke, Eva, Jessica, Laura, Libertje, Lisanne, Madeleine en Marieke, bedankt voor jullie inspiratie, hulp en gezelligheid, ook naast het werk. Een avondje kaasfonduen, een ballet of concert, babycadeautjes, hilarische optredens op huwelijken, een extra vrijgezellenfeestje, shopsessies tijdens MS congressen…. jullie zijn overal voor in! Lieve Anke, dank voor je gezelligheid als ‘wervelwind’ op onze kamer. Het is met jou nooit saai en nooit rustig! Voor mij een welkome afwisseling, met name de laatste tijd. Ik wens je alle geluk en hoop dat jij ook snel zult promoveren! Lieve Libertje, door onze tijd samen (zwanger) op de kamer boven hebben we veel met elkaar gedeeld en beleefd. Dat zetten we nu buiten werk voort in sport- en koffiedates in Haarlem, ook met onze kindjes. Dank voor je enorme interesse en betrokkenheid. Ook dank voor je input en precisie als mede auteur van een heel aantal artikelen. Lieve Lisanne, dank dat ik altijd op je terug kon vallen, als mijn persoonlijke statistiek helpdesk, maar met name als heel fijn persoon om een eerlijke en oprechte mening van te kunnen vragen, om samen koffie mee te drinken en ervaringen als niet-artsen te delen. Heel jammer dat je er vandaag niet kunt zijn, maar op naar jouw promotie! Lieve Marieke, ik ben heel blij en trots dat je vandaag naast mij wilt staan als mijn paranimf. een mooie vriendschap aan overgehouden. Je bent zo’n bijzonder mens, zo lief, warm en grappig. Wat fijn dat ik je heb leren kennen!
135
DANK
Wij hebben elkaar leren kennen in een moeilijke en emotionele periode, maar we hebben er
Dankwoord
En dan de mannelijke collega’s niet te vergeten, Joram en Prejaas. Prejaas, doordat we nu zo dicht bij elkaar zitten heb ik je na 3 jaar eindelijk leren kennen. Een bijzonder chaotische, niet plannende en afspraken vergetende collega, maar ondertussen een lief familiemens en een attente man waar we allemaal stiekem een beetje verliefd op zijn geworden! De data-unit verdient een speciaal woord van dank: Sylvie, Monica, Manon, Natasja, Jos, Laura en nu ook Josca, bedankt voor jullie inzet en hulp met het organiseren van afspraken, de data invoer en de grote hoeveelheden telefonische interviews! Jullie zijn onmisbaar! Ook veel andere medewerkers van de afdelingen neurologie en epidemiologie & biostatistiek ben ik dank verschuldigd, voor de ondersteuning van het onderzoek en de hulp bij patiënten visites. Verpleegkundigen (in het bijzonder Daniëlle, Marijke, Ina en Siny), balie medewerkers (Claudia, Marion, Anita) en het secretariaat van de poli/afdeling. Karin en Brenda, de begintijd als student bij jullie op het secretariaat heeft me broodnodige afleiding en gezelligheid gegeven! Anne Marie (alias de mama van E&B) en Hülya, wat leuk dat we zo’n gezellig en succesvol (organisatie)team zijn gebleken en naast jullie wil ik ook Elles en Karin & Annette (bureau MS centrum) bedanken voor veel regelzaken en attente hulp. Lieve vrienden, in het bijzonder lieve Arlette, Bas, Frederieke, Jeroen, Martijn en Susan, wat heerlijk dat ik altijd op jullie terug kan vallen. Volgens mij hebben jullie geen idee wat ik al die jaren uitspookte, maar bedankt voor de momentjes dat ik even met jullie kon sparren over kleine dingen rond mijn proefschrift. Vooral bedankt dat ik bij jullie nooit aan mijn werk hoef te denken door al onze mooie, emotionele, gezellige en feestelijke momenten met elkaar. In het kader van ‘MEER IS BETER’, laten we onze vriendschap koesteren en op deze manier voortzetten tot we oud en grijs zijn! Jullie zijn me heel dierbaar! Lieve Arlette, ik vind het heel bijzonder dat je vandaag naast me wilt staan als mijn paranimf. Ik ben trots op je, zorgzame vriendin, stoere mama, ambitieuze vrouw en altijd even sterk en vrolijk. We zijn de afgelopen jaren steeds meer naar elkaar toe gegroeid en ik hoop dat we (met onze gezinnen) nog lang zo intens van elkaar mogen blijven genieten! Lieve Ellen, ook al zijn we een slecht onderzoeksteam samen (waar zijn onze publicaties uit Suriname?!), we gaan toch allebei promoveren. Je blijft voor mij een hele bijzondere vriendin!
136
Dankwoord
Lieve vriend Daan, ik wil jou bedanken voor je interesse in mijn proefschrift, je goede ‘oefenvragen’ over MS en je inspiratie met betrekking tot de klassieke muziek. Daardoor heb jij ook een bijzondere bijdrage geleverd aan deze promotie. Lieve familie en schoonfamilie, dank voor jullie liefde en interesse in mijn werk en onderzoek. We zijn een bijzondere familie, waar ik veel waarde aan hecht! Bijzondere dank gaat uit naar de meest trotse mensen die ik ken, mijn oma’s en opa. Lieve broer en zus. We zijn een bijzonder trio, uit hetzelfde gezin maar alle drie zo verschillend! Lieve Danielle, ik bewonder je inzet en enthousiasme voor maatschappelijke projecten, je vrolijkheid en avontuurlijke reislust. Nu nog je grote liefde vinden Lieve Tim, ik ben blij dat je na wat omzwervingen je geluk hebt gevonden als skileraar, in je studie en natuurlijk met die schat van een Kiira! Dank dat jullie er altijd voor mij en mijn gezin zijn! Lieve papa en mama, dank voor alles wat jullie voor mij gedaan hebben, de mogelijkheid die jullie me gaven om te studeren en alle vrijheid die ik kreeg in wat ik wilde doen. Af en toe verzon/verzin ik erg veel tegelijk, maar jullie hebben altijd achter me gestaan en me overal in gesteund. Ik ben dankbaar voor zulke lieve ouders. Lieve Raymond, ik heb een proefschrift geschreven (waar je volgens mij niet heel veel last van hebt gehad), maar ondertussen ben jij van ons twee de hardste werker. En dat terwijl je graag huisman zou willen worden naast een carrièrevrouw… helaas! Met name nu streef je naar het behalen van iets bijzonders, ver weg waar een ander deel van jouw geluk en ons geluk ligt. We kennen elkaar zo lang dat we een goed en op elkaar ingespeeld team zijn, we kunnen de toekomst samen aan! Te quiero, te adoro… Lieve Sofia, je hebt nog geen idee maar ik word zo vrolijk van jou! Gedachten aan jou en jouw vrolijke foto’s rond mijn computer geven me ook tijdens mijn werk een trots en gelukkig gevoel. We verheugen ons op de toekomst met jou en je broertje(s) en/of zusje(s)!
DANK
Judith, maart 2014
137