Abstract
Introduction:
In order to estimate the value of interventions in multiple sclerosis (MS) – where lifetime costs and outcomes cannot be observed – outcome data have to be combined with costs. This requires that cost data be regularly updated.
Objectives and methods:
This study is part of a cross-sectional retrospective study in 16 countries collecting data on resource consumption and work capacity, health-related quality of life (HRQoL) and prevalent symptoms for patients with MS. Descriptive analyses are presented by level of disability, from the societal perspective, in EUR (2015).
Results:
A total of 779 patients (mean age = 57 years) participated; 72% were below retirement age and of these, 36% were employed. Employment was related to disease severity, and MS affected productivity at work for 84% of patients. Overall, 96% and 72% of the patients experienced fatigue and cognition as a problem. Mean utility and annual costs were 0.735 and 11,400GBP at Expanded Disability Status Scale (EDSS) = 0–3, 0.534 and 22,700GBP at EDSS = 4–6.5, and 0.135 and 36,500GBP at EDSS = 7–9. The mean cost of a relapse was estimated at 790GBP.
Conclusion:
This study illustrates the burden of MS on UK patients and provides current data on MS that are important for development of health policies.
Introduction
The United Kingdom has been a leading force behind the integration of economic studies into decision-making in health care. The academic interest in cost of illness studies, outcome measurement and cost-effectiveness analysis dates back more than half a century and a large number of studies have been published. However, the formal integration into policy and decision-making only happened in 1999 with the creation of the National Institute for Health and Care Excellence (NICE). Since then, although its decisions are only applicable to England and Wales, NICE has been a leader in the development of methodology, and its decisions are observed with interest across Europe.
A considerable number of cost of illness studies have been performed in multiple sclerosis (MS) prior to 1999, and new studies were performed around the timing of introductions of new disease-modifying treatments (DMTs).1–5
The availability of DMTs has led to changes in patient management and a focus on earlier and better diagnosis and adjustments in the diagnostic criteria themselves. One of the consequences in this regard is that the recorded prevalence of the disease is quite different from that estimated two or three decades ago, 6 leading to an increase in prevalence to 203/100,000 population in 2014. 7 With diagnosis already possible after a clinically isolated event, 8 one must also expect a different distribution of the type of MS and the severity of the disease than 10 years ago: a larger proportion of patients with relapsing-remitting disease and thus of patients in the early stages of the disease and with less disability (a low score on the Expanded Disability Status Scale (EDSS)).
It is, therefore, important to update the information on the burden of MS, and the study presented here is part of a European-wide effort in 16 countries, endorsed by the European Platform of MS Societies (EMSP) 9 and carried out with the support of national MS societies. It uses a similar methodology as the last European survey in 2005. 10
Materials and methods
The detailed methodology for the European survey is published separately. 11 Therefore, we only provide a short summary of the general methods and issues specific to the United Kingdom.
Data
The study aimed to estimate the costs of all health care and other resource utilisation related to MS: hospitalisation, rehabilitation, consultations, diagnostic procedures and tests, medication, community care, family support and production losses (sick leave, early retirement and invalidity). In addition, information on major symptoms such as fatigue and cognition health-related quality of life (HRQoL) as well as self-assessed disability using descriptions based on the EDSS was collected.
Data were collected with a standard questionnaire, at a single point in time, for a retrospective period of time. The latter was varied depending on the question in order to minimise recall bias: 1 month for use of drugs, community services and family help; 3 months for hospitalisation, consultations, tests, sick leave and relapses; 12 months for major investments. Resource utilisation is reported for these time periods, while cost calculations are annualised.
Disease information such as the type of MS, disability (EDSS), HRQoL, utility (EQ-5D 12 ), symptoms (fatigue and cognition) and the effect of MS on work related to the current day or week was obtained.
The handling of missing data for the cost calculation is explained in more detail in the paper describing the study methods. 11 For resource use, we present actual answers without any imputation for missing answers. Also, no imputations were made for missing information on disease status, symptoms and HRQoL.
Costs
Costs were calculated from the societal perspective, including all costs regardless of who ultimately is responsible for them. Patient co-payments and patients’ out-of-pocket expenses were thus included. The cost of a relapse was calculated as the difference in quarterly costs between patients with or without a relapse and an EDSS score of 0 to 6; patients who were unsure were excluded from the estimation. Invalidity, early retirement and DMT costs are not considered in this calculation, as they are unlikely to be affected within 3 months. In addition, investments in the United Kingdom were essentially for modifications to the house or the car, which again are unlikely to be affected directly by a relapse.
Unit costs for the individual resources were taken from public sources and are described in the paper on the study methods. 11 Results are reported in GBP (2015).
Patients
The objective was to include a sample where all levels of disease severity (defined by EDSS) were represented in sufficient numbers to permit analysis, rather than a prevalence sample. This highlights how costs and HRQoL change as the disease progresses and provides the necessary data for cost-effectiveness analysis of treatments that are expected to change the course of the disease. Mean results may thus not be representative and should neither be extrapolated directly to national costs nor be compared directly to the results in earlier studies. Therefore, we report results by EDSS or by disease severity groups only (mild MS EDSS = 0–3, moderate MS EDSS = 4–6.5, severe MS EDSS = 7–9).
In anonymous surveys, participation will depend heavily on the methods used for the survey: collecting data in MS centres tends to overestimate the number of patients with early but severe disease and on treatment with DMTs; collecting data from members in patient organisations may lead to the opposite. Internet surveys may bias towards patients with better education, while postal surveys may include older patients.
The UK participants were contacted by the UK MS society, and the vast majority answered the questionnaire online. The data were collected during the third quarter of 2015. The UK MS society invited 5928 individuals by e-mail, but despite two reminders, the response rate reached only 13%.
Results
A total of 779 evaluable responses were received (96% online). Patients from all regions were well represented, with participation in Northern Ireland, Scotland and Wales proportional to the population. In England, participation from the South East, South West and London exceeded 40% and was thus considerably higher than expected according to population density. 13 Table 1 provides details on demographics, employment and disease.
Sample demographics.
MS: multiple sclerosis; SD: standard deviation; EDSS: Expanded Disability Status Scale; DMT: disease-modifying treatment; RRMS: relapsing-remitting multiple sclerosis.
Total sample (n = 779).
Patients of working age (N = 563).
Patients working (N = 214).
Demographics and employment
The age of respondents in the United Kingdom ranged from 26 to 89 years (mean = 56.7 years, median = 57 years and standard deviation (SD) = 10.8 years). This makes the UK sample the oldest in our study series with a clear effect on the results – more patients with severe disease, lower DMT usage, fewer patients of working age and actually working. These differences make comparisons of mean costs in the sample with those from other countries in this study neither meaningful nor informative.
Women represented 70% of the sample; 80% of the sample lived with their family, and eight patients were in a nursing home at the time of the survey. Education levels in our sample appeared to be high: 10% of the patients had basic education, 52% had a secondary or a professional degree and 38% a university degree. This compares to 36%, 37% and 27% in the general population in England and Wales in 2011 14 and would indicate a sample with high education, as would be expected on an Internet survey.
Despite the high mean age, a majority of patients in the sample were below average effective retirement age (63 years for women, 64 years for men 15 ) numbering 563 patients (72%). Of these, 200 patients (36%) were employed or self-employed. In total, 14 patients above retirement age also worked, bringing this group to 214 patients (or 28% of the full sample), with a mean age of 50.8 years. The employment rate compares to an activity rate of 74.4% in the general population aged 16–64 years. 16 The majority of patients worked part time (92%), and of these, 82% did so because of MS. In the general population, 27% of people work part time and of these, 3% indicated that it was due to illness or disability, 14 illustrating again how much MS affects patients’ activities. Average working hours in the sample were 39 hours for full-time workers, 21.4 hours for part-time employees. Sick leave during the past 3 months was reported by 22% of patients, with a mean duration of 9.2 days. This appears to be mostly due to relapses that were reported by 18% of the patients.
Overall, employment decreased rapidly with increasing disease severity, as shown in Figure 1. The reasons for the difference found between patients at EDSS = 2–3 and EDSS = 4–5 might partly be due to the skewed disease severity distribution in this sample. The number of patients with an EDSS of 0–3 was small, with, however, a relatively high age (52 years) and a long mean disease duration (14–18 years) that may explain their leaving the work force for reasons other than a physical handicap (fatigue and cognition). Of non-employed patients, 57% indicated MS as the reason.
Employment by disease severity. Proportion of patients below retirement aged employed or self-employed. In total, 563 patients (72%) were below retirement age, and of these, 36% were employed or self-employed.
Most employed patients felt that MS affected their productivity at work (84%) and only 8% indicated that they had no problems, while 8% had not answered the question. The severity of the effect covered the entire visual analogue scale (VAS) range from 0 to 10, with a mean of 4.0 (SD = 2.7; Figure 2). Fatigue was considered the most bothersome symptom (73%), followed by difficulties thinking (43%), mobility (40%), pain (31%) and low mood (21%).
Fatigue, cognitive difficulties and effect of MS on productivity at work. Mean score on the visual analogue scales (0 = no problem; 10 = severe problems) for fatigue, cognitive difficulties and impact of MS at work (only for patients working). Patients with missing EDSS (1) or missing answers are excluded.
Disease information
Disease information is summarised in Table 1. The mean EDSS in the sample was 5.5 (SD = 2.2), with 64% of the patients at EDSS = 6 or higher. This makes it the most severe sample in the European study. However, the number of patients at all levels of EDSS was sufficient to yield a stable analysis, with the exception of EDSS = 9 (10 patients). The mild group represented 18% of the sample, the moderate group 51% and the severe group 31%.
The proportion of patients with relapsing-remitting multiple sclerosis (RRMS) was 37%, with secondary progressive multiple sclerosis (SPMS) 38% and with primary progressive multiple sclerosis (PPMS) 24% (14 missing). Thus, while one would expect a high proportion with SPMS considering the high EDSS, the proportion stating PPMS is considerably higher than normal prevalence. 6 This suggests, as we found in previous studies, that patients might be uncertain in their answers regarding the type of disease. Therefore, we did not include disease type in our analyses and focused instead on EDSS levels. DMTs were used by 28% of the sample, with usage declining with higher EDSS levels, as expected (Table 1). Among users, 13% were on their first DMT, likely again an expression of the high age and severe disease in this sample; first-generation DMTs were used by 58% of the users (Table 2).
Type of DMTs used (N = 218 a ).
DMT: disease-modifying treatment.
Missing information on type of DMT for three patients.
Relapses in the preceding 3 months were reported by 141 patients (18%) of which half occurred in the past month (Table 1). However, 176 patients (23%) were unsure whether they had a relapse or not, and we assumed that the answer was no. Thus, the mean relapse rate over a 3-month period was estimated at 0.3 (SD = 0.7). Corticosteroids were used by 25 patients (18%) with relapses.
Symptoms and HRQoL
Fatigue and cognitive difficulties were an issue for a majority of patients, and both were related to disease severity (Figure 2). Fatigue was present in 96% of the patients. The mean VAS score was 5.9 (SD = 2.5) for the sample and 4.4 for patients with mild disease and 6.2 for patients in the moderate and severe groups. Cognitive difficulties were recognised by 72% of the patients. The mean VAS score in this group was 4.8 (SD = 2.0) overall, 4.4 in the mild, 4.8 in the moderate and 5.0 in the severe group. For the full study sample (assigning 0 to the group with no problems), the mean score was 3.5 (SD = 2.9) and 2.8, 3.6 and 3.7 in the three groups, respectively.
The detailed answers to the EQ-5D indicated that both the severity and the domains affected changed with disease severity. 17 Self-care was unaffected in patients with mild disease but declined rapidly with advancing disease. Anxiety was present in around half of the patients at all levels. For the other three domains, around half of the sample indicated problems already in early disease, with a rapid decline at higher EDSS levels (Figure 3).
Proportions of patients at different levels of disease severity experiencing difficulties in the five domains of the EQ-5D, N=772. Difficulties are increasing with increasing disease severity, with the exception of anxiety/depression that appears similar at all stages of the disease.
Utility
Utility declined with increasing disability (EDSS; Figure 4). Mean utility in the sample was 0.469 (SD = 0.3), the lowest in this European study and explained mostly by the high EDSS.
Utility by EDSS level using the EQ-5D, N=772. Utility is calculated by relating the scores (1 = no problem, 2 = some problems and 3 = severe problems) of the five domains to a health state valuation system established with the general population using decision analytic methods. 15 Values decrease from levels comparable to the population to negative values in severe disease.
Resource utilisation
Resource utilisation is presented in Table 3. During the preceding 3 months, 28 patients (4%) were admitted as inpatients, most often in a general ward (25 patients), on average 2.3 (SD = 1.6) times and for a mean of 6.5 days (SD = 6.6 days). In total, 55 patients (7%) had day admissions, on average 2.5 (SD = 1.9) times. Inpatient or day admissions to rehabilitation centres occurred for 17 patients (2%).
Resource utilisation, health care and community services.
SD: standard deviation; MS: multiple sclerosis; MRI: magnetic resonance imaging; DMT: disease-modifying treatment; OTC: over-the-counter.
Two-thirds of patients (524, 67%) had a consultation during the past 3 months, most often with a general practitioner (GP; 262, 34%), an MS nurse (209, 27%) or a neurologist (195, 25%). Investigations and tests were needed by 120 patients (15%) and medications for MS and MS-related symptoms were used by 73% of the patients during the past month. Drugs other than DMTs and corticosteroids were used by 284 patients (37%), predominantly treatments for walking, spasticity and pain (199 patients, 26%); incontinence (126 patients, 16%); and depression (109 patients, 14%). Non-prescription drugs were purchased by 420 (54%) patients.
Investments in equipment and devices to aid patients’ mobility were made during the past 12 months for or by 380 patients (49%), most often for walking aids, wheel chairs and modifications to the car or the house.
Community and social services were used by 22% of the patients, most frequently home help and transportation. Help from family was used by 69% of the patients, on average 19.3 days/month and 4.9 hours/day. Both community services and informal care were related to disease severity (Figure 5).
Use of informal care during the past month (days/month and hours/day, per user). In total, 69% of the sample use help from the family, but use is clearly concentrated in the severe group: 26% of the patients in the mild, 74% in the moderate and 86% in the severe group are relying on family support.
Costs
Total mean annual costs per patient for patients with mild, moderate and severe disease and by EDSS score are presented in Figure 6 and Table 4.
Total mean annual cost per patient by disease severity (N = 779). Total costs increase with disease severity, but the type of resources changes. Overall, health-care costs dominate in mild disease; production losses, informal care, investments and community services dominate in more severe disease. However, due to the underrepresentation of younger patients with a short disease duration, production losses are already high at low EDSS levels, while the cost of DMTs is lower than expected.
Total mean annual cost per patient by disease severity (mild, moderate and severe), N = 779 (GBP 2015).
GBP: British Pound; EDSS: Expanded Disability Status Scale; SD: standard deviation; DMT: disease modifying treatment.
The average cost of a relapse for patients with an EDSS of 0 to 6 was estimated at 792GBP. All types of costs, but particularly informal care, increased for patients with relapses (Figure 7).
Mean cost of a relapse (3 months), GBP 2015. The cost is estimated as the difference of costs of patients below EDSS = 6.5 with and without a relapse (N = 83 and N = 367, respectively). Patients who were unsure (N = 176, 23%) were excluded from the estimation. Invalidity, early retirement and DMT costs are not considered in this calculation, as they are unlikely to be affected within 3 months. In addition, we excluded investment costs as they related for their vast majority to transformations of the car or house unlikely to be affected by a relapse. The cost of a relapse is thus estimated at 792GBP. The use of all types of resources increases during a relapse, with the biggest increase seen for inpatient care (22%) and informal care (24%).
Discussion
This study provides an update to current understanding of the burden of MS on patients, the health-care system and society in general, based on a cross-sectional survey conducted in 2015.
The resource utilisation patterns reported here reflect the clinical needs and disease experience of patients in the United Kingdom. As was observed in previous studies of this type, we report relatively high levels of utilisation of MS nursing services and reliance on over the counter medication and informal care among study respondents. DMT use is lowest among the countries participating in this study but is partly explained by the underrepresentation of people with RRMS and a short disease duration. These findings should be evaluated with reference to local clinical guidelines, such as those published by the Association of British Neurologists (ABN) in 2015 18 or by benchmarking versus the findings from other countries in this research programme.
It is also important to note that in addition to clinical needs, these patterns reflect the structure of health-care provision and delivery of various services through the UK National Health Service (NHS). That is, patients utilise the particular set of services and treatments that are commissioned by the NHS, which is not necessarily the same as that provided to equivalent patients in other countries. Similarly, cultural norms also impact on how often a person will seek to visit their GP, and their willingness to present with symptoms of early disease prior to a confirmed diagnosis. Thus, comparisons of these findings with those from other countries should be interpreted with caution.
It is nevertheless possible to comment on some notable aspects of the data presented here. This is a relatively old sample of MS patients, and the distribution of EDSS scores does not represent current prevalence. Patients with severe disease and long disease duration are overrepresented and thus introduce a bias in the mean analysis for the total sample. We partly overcome this bias by presenting all results by disease severity rather than for the sample but acknowledge that the survey underrepresents people with shorter disease duration, of whom there will be many with RRMS and mild disability.
The mean time to diagnosis in this UK sample was 8 years. This is likely a reflection of the high age of the sample where patients received their diagnosis some 15 years ago, at which time diagnostic criteria were used that did not allow as early a diagnosis as is nowadays possible, and when a ‘wait-and-see’ approach to clinical care of patients with MS was more likely. More recent and emerging evidence suggests longer term benefits of DMTs, 19 potentially more so with early treatment. 20 Although DMTs are a major cost item in mild MS, the observation of the large increase in costs, alongside the severe drop in utility in patients with severe disease as demonstrated in this work, suggests the potential for such treatments to ultimately result in fewer costs and better health outcomes for the NHS.
In addition to characterising various aspects of the burden of MS on patients, and capturing patterns of resource utilisation that could inform future policy recommendations, this work can provide valuable inputs into cost-effectiveness analyses for appraisals by NICE and similar bodies. The data generated by this study provide a comprehensive and contemporary source to describe the HRQoL and resource utilisation patterns of patients in the United Kingdom with MS. The level of detail in which the cost items are captured and described is a strength of this study; the methodology underlying the calculations is clearly laid out. The cross-sectional approach to data collection is mandated by the fact that the disease process of MS and related resource consumption cannot be observed over its entire duration, and covering the entire disease spectrum is key. This is seldom, if ever, possible even in randomised controlled trials (RCTS). Also, the extent to which pivotal studies capture health-care resource utilisation is patchy; furthermore, RCT populations may not reflect the true complexity of real-world patients with MS, in terms of their clinical profiles or their experience of the health and social care system outside the context of a clinical study.
The United Kingdom has seen a number of cost of illness assessment studies in MS, and a number of data sets have been used in NICE appraisals. However, the most recent of these studies suffers from a much smaller data set, 5 while the larger research efforts were conducted in 2005 against the background of a different NHS.4,21,22 Simply inflating the values extracted from that earlier study to suit today’s needs is no longer satisfactory, and this update is timely.
In summary, these data provide the latest and most comprehensive source of information on the burden of MS among UK patients. The data provided can help direct future design of health-care provision and policy changes, but decision makers should interpret the findings carefully and be aware of the limitations of this approach. Moreover, a range of factors make inter-country comparisons of data from this research initiative difficult. In particular, the patient group represented is the oldest of all included countries, suggesting that younger patients with mild MS and short disease duration are underrepresented in this survey.
Footnotes
Acknowledgements
The authors are grateful to Diane Redfern-Tofts (MS Society) for support with data collection, Marie Braisher (University College London) for support with ethics submission, Mia Gannedahl and Johan Dalén (MAPI Group) for organisational and statistical support, Barbara Rosengren for language editing and Cara Harley for formatting.
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship and/or publication of this article: A.T. has received honoraria/support for travel for consultancy from Esai, Biogen (MAPI), MedDay, Novartis, Teva, Remedia and Excemed; he has received support for travel for consultancy from the International Progressive MS Alliance, National MS Society USA and the Multiple Sclerosis International Federation. He receives an honorarium from SAGE Publishing as Editor-in-Chief of the Multiple Sclerosis Journal. G.K. has received consulting fees from Almirall, Bayer, Biogen, Merck Serono, Novartis, Oxford PharmaGenesis, Sanofi Genzyme and Teva. J.B., D.C. and J.E. were at the time of the study employed by the MAPI Group, a research company acting as consultants to pharmaceutical companies. D.M. has received honoraria for Advisory Committee and/or Consultancy work from Biogen Idec, Novartis, Mitsubishi Pharma Europe & Bayer Schering Pharma and grants for performing central MRI analysis of MS trials from Biogen Idec, Novartis, Apitope and Merck. The Queen Square MS Centre at UCL Institute of Neurology is supported by the UK MS Society and UCL-UCLH Biomedical Research Centre.
Funding
The author(s) declared receipt of the following financial support for the research, authorship and/or publication of this article: European Health Economics and MAPI have received funding for the full research and production of the manuscripts from Biogen.