A discrete-choice experiment to determine patient preferences for injectable multiple sclerosis treatments in Germany

Introduction

Multiple sclerosis (MS) is a chronic, autoimmune, inflammatory disorder of the central nervous system characterized by axonal demyelination and eventual axon loss [Kamm et al. 2014; Saidha et al. 2012]. First-line treatment for MS in Germany has evolved over the past 3 years to now include oral treatments such as dimethyl fumarate and teriflunomide [Warnke et al. 2013; Weber et al. 2012]. However, historically, the mainstay of disease-modifying treatments (DMTs) for MS are injectable immunomodulatory drugs [Patti, 2010; Wingerchuk and Carter, 2014]. The injectable DMTs are well-established therapies with a wealth of real-world safety and efficacy data that span two decades [McGraw and Lublin, 2013]. In 2014, a pegylated interferon (peginterferon) beta-1a administered subcutaneously was approved in the European Union (EU) and United States (US) for relapsing MS (RMS) [Biogen 2014a, 2014b]. This new, injectable DMT has the longest dose interval compared with alternative injectable treatments: administered only once every 2 weeks [Biogen 2014a, 2014b], allowing patients the option of having injection-free weeks. All injectable DMTs are associated with a number of adverse events (AEs) that differ in severity depending on the treatment regimen. The most common, nonserious AEs include flu-like symptoms (FLS) and injection-site reactions (ISR) (including pain, erythema, pruritus) [Patti, 2010; Wingerchuk and Carter, 2014].

Adherence to injectable DMTs is an ongoing issue for patients and physicians [Patti, 2010]. A phase IV study of DMT therapy in approximately 2500 patients with relapsing–remitting MS (RRMS) reported that 25% of patients were nonadherent [Devonshire et al. 2011]. A total of 50% of these nonadherent patients stated that they forgot to administer the injection, and 32% stated injection-related reasons for their nonadherence [Devonshire et al. 2011]. Adherence rates have been shown to decline as more frequent injections are needed. Indeed, an analysis of more than 6600 patients with MS showed greater treatment adherence in those needing once-weekly injections than those with a more frequent injection regimen [Halpern et al. 2011]. Consequently, poor treatment adherence due to frequent injections is an important factor for physicians, healthcare authorities, and the pharmaceutical industry to consider [Devonshire et al. 2011; Steinberg et al. 2010].

Injection-related reasons for nonadherence include physical or cognitive difficulties, perceived lack of treatment efficacy, treatment-related AEs occurring with each injection, injection anxiety, needle phobia, and injection fatigue [Lugaresi, 2009; Patti, 2010; Menzin et al. 2013]. New treatments with prolonged efficacy that allow for increased time between doses have now been developed; these require fewer overall injections when compared with older treatments, thereby reducing injection-related factors experienced by patients [Menzin et al. 2013; Patti, 2010; Shingler et al. 2013].

Patient preferences, and consequent trade-offs for specific features and attributes of injectable DMTs, are not captured by health-related quality of life (HRQoL) instruments, such as the EQ-5D, or quality-adjusted life year (QALY) models [Shingler et al. 2013; Whitehead and Ali, 2010]. A discrete-choice experiment (DCE) is an established methodology for quantifying patient preference for attributes of medical treatments. DCEs are based on the concept that a treatment can be described as a combination of features, or attributes, and that choice of alternative treatments depends on the relative preferences for the different attributes of those treatments. Attributes are described as a series of levels; for instance, the attribute of dosing frequency could include the levels of weekly, bi-weekly or monthly dosing. In a DCE study, participants are presented with a series of treatment pairs with different attributes and levels. Statistical analysis of patients’ treatment choices across a series of paired comparisons reveals the importance of individual attributes and the trade-offs between attributes that patients are willing to make (for example, the reduction in efficacy that patients would accept in exchange for an improvement in dosing frequency).

DCEs have been used previously to study preferences for MS treatments [Johnson et al. 2009; Shingler et al. 2013; Utz et al. 2014; Wicks et al. 2015; Wilson et al. 2014, 2015]. Shingler and colleagues assessed patient preference for features of injection devices and treatment efficacy, showing that while efficacy is of primary importance, device attributes can play an important role in patient preference. They did not assess patients’ aversion to side effects [Shingler et al. 2013]. Johnson and colleagues assessed patient preferences for risk of serious AEs as the efficacy of treatment improved, indicating that patients would be willing to accept a degree of increased risk associated with their treatment if there was a corresponding increase in treatment efficacy. However, they did not assess preferences for dosing features, such as dosing frequency or mode of administration [Johnson et al. 2009]. Utz and colleagues examined preferences for dosing features of DMTs and found that pills were preferred to injections [Utz et al. 2014]. However, by varying dosing frequency independently from mode of administration and side effects, they found that injections were preferred when pills had more frequent mild side effects or more frequent dosing. However, Utz and colleagues did not assess preferences for efficacy attributes. The studies described by Wicks and colleagues and Wilson and coworkers examined preferences for a broad range of treatment attributes, including efficacy, serious and mild side effects, and mode and frequency of administration [Wicks et al. 2015; Wilson et al. 2014, 2015]. Wicks and colleagues examined preferences for oral DMTs only, while Wilson and coworkers focused on oral and parenteral DMTs. Mode of administration and dosing frequency were not varied independently, which limited their ability to assess acceptable trade-offs among dosing frequency and other treatment features and health outcomes.

The objective of this study was to assess the relative importance of a number of features of a hypothetical injectable DMT for patients with MS in Germany. This study used a DCE to quantify patients’ preferences for features of injectable DMTs, including efficacy, side effects, and features related to administration, such as dosing frequency and injection time.

Methods

Study design

The study investigators employed an online DCE, similar to the design previously used to examine US patient preference for injectable treatments for MS [Poulos et al. 2015]. The survey design, administration, and analysis followed the guidelines for DCE applications in healthcare [Bridges et al. 2011]. The choice and descriptions of each treatment attribute and its associated levels were informed by: (i) consultation with clinical experts; (ii) features of currently available injectable DMTs; (iii) selected clinical study findings; and (iv) the relevance to patients. The final six treatment attributes were: the number of years until MS symptoms get worse (i.e. disability progression), the number of relapses in the next 4 years, injection time, frequency of injections, FLS, and ISRs (Table 1). A draft survey was pretested using in-depth, in-person interviews with patients.

OPEN IN VIEWER

Table 1.

Specific attributes and the corresponding levels for the treatment-choice questions used in the study.

Attribute	Levels
Number of years until MS symptoms get worse	4 years
	2 years
	1 year
Number of relapses in the next 4 years	1 relapse
	3 relapses
	4 relapses
Injection time	3 seconds
	10 seconds
Frequency of injections	1 time each month (monthly)
	2 times each month (biweekly)
	4 times each month (weekly)
	12 times each month (3 times each week)
	30 times each month (daily)
Flu-like symptoms	No flu-like symptoms after any injection
	Symptoms for 1 day after some injections
	Symptoms for 3 days after some injections
	Symptoms for 3 days after every injection
Injection-site reactions	No reaction
	Mild reactions

MS, multiple sclerosis.

Patients were asked to choose one of two hypothetical injectable treatments for MS, defined by the attributes and levels in Table 1. An example treatment-choice question is shown in Figure 1. During the survey, patients were asked to consider the cost of all medications as equal, since cost was not chosen as an attribute for analysis.

OPEN IN VIEWER

Figure 1.

An example treatment-choice question.

SAS 9.3 (SAS Institute, Cary, NC, USA), an analytics software program, was used to generate the experimental design [Johnson et al. 2013]. The attribute levels and the pairs of profiles were optimized for design efficiency, level balance, correlations between level differences, and number of choice tasks [Huber and Zwerina, 1996; Kanninen, 2002; Dey, 1985; Kuhfeld et al. 1994; Kuhfeld, 2010]. To maximize the quality of responses and to minimize cognitive burden, SAS 9.3 generated 48 pairs of profiles that were split into six versions of eight questions. Patients were assigned randomly to one of the six versions, and the order of the questions was randomized for each patient. One question was repeated later in the sequence to assess the stability of patients’ choices (data not shown).

To ensure that all patients were considering the same level of initial disability and the same improvements in efficacy when evaluating the hypothetical treatments, they were asked to answer the questions as if their health status conformed to a hypothetical reference scenario. According to the reference scenario, patients’ ‘baseline’ or initial disability was ‘mild’ on the Hohol disease steps scale, and disability would progress to ‘moderate’ when their symptoms got worse [Hohol et al. 1995]. Data on patient demographics and socioeconomic characteristics and experiences with MS and MS treatments were also collected.

Results

Sample characteristics

Of 202 adults who responded and completed the survey, 13 (6%) were excluded from data analysis because they always chose either medication A or medication B in the treatment-choice questions; this may indicate that they did not pay close attention to the choice tasks. As a result, the final sample size for analysis was 189 patients.

Approximately 50% of all patients reported a diagnosis of RRMS, 31% reported secondary progressive MS (SPMS), 11% reported primary progressive MS (PPMS), and 3% reported progressive relapsing MS (PRMS) (Table 2). On most days, 10% had no limitations due to MS, 22% had mild symptoms (according to the Hohol scale), 30% had moderate symptoms on the Hohol scale, 18% of patients needed a cane for walking long distances, 9% needed a cane for short and long distances, and 4% needed bilateral support. Approximately 53% of patients had experienced FLS caused by their MS medication (Table 2), and 47% had experienced a mild ISR to their medication. Approximately 71% of patients had current or prior experience with injectable MS medication, and 12% had never been prescribed medication to take on a regular basis to manage their MS (Table 2).

OPEN IN VIEWER

Table 2.

Characteristics of the patients enrolled in the survey.

Characteristic	(N = 189)
Age, mean (SD), years	43.8 (10.5)
Female	68.8%
Employed full time	25.4%
Type of MS:
Relapsing–remitting	50.3%
Secondary-progressive	31.2%
Primary-progressive	10.6%
Progressive–relapsing	2.6%
Number of years since MS diagnosis, mean (SD)	9.2 (6.9)
Number of relapses in last 2 years, mean (SD)	3.1 (5.9)
Currently manages MS symptoms through:
Exercise or have physical therapy regularly	59.3%
Over the counter medicines	24.3%
Dietary supplements, such as vitamins and herbs	25.9%
Alternative medicine, such as acupuncture	8.5%
Oral medicine (pills) prescribed by doctor	45.5%
Receive regular IV infusions of medicine	12.7%
Receive regular injections of medicine	34.9%
None of the above	8.5%
MS symptoms, on most days:
No limitations	10.1%
Mild symptoms	22.2%
Moderate symptoms	29.6%
Need cane for walking long distances	18.0%
Need cane for walking short and long distances	8.5%
Need bilateral support	4.2%
Need wheelchair/scooter	7.4%
Patients that have never regularly been prescribed a medicine to manage their MS	12.2%
Patients with current or prior experience with injectable MS medication	70.9%
Patient experience of device: Seconds it takes for medicine to come out of injection device, mean (SD)
Among those currently receiving regular injections	10.9 (6.5)
Among those who have received regular injections in the past	11.2 (12.1)
Ever had flu-like symptoms caused by your MS medicines	53.4%
Ever had mild injection-site reactions caused by your MS medicines	46.6%

IV, intravenous; MS, multiple sclerosis; SD, standard deviation.

Preferences for selected attributes of injectable disease-modifying therapies

The results of the RPL estimation indicate that patient preferences were consistent with the natural ordering of each of the attributes, for example, one relapse in the next 4 years was preferred to two relapses in the next 4 years (p ⩽ 0.05) (Figure 2). Further, there was a statistically significant difference between at least two preference weights within each attribute, except injection time. The relative importance of an improvement in the number of years until MS symptoms get worse from 1 to 4 years (measured as the vertical distance between preference weights corresponding to these levels) was 2.1 (Figure 2). The relative importance of this change was the greatest, when compared with the relative importance of any other change amongst other attributes included in the study. Furthermore, it is approximately twice as important as a change from four relapses to one relapse in the next 4 years (2.1 versus 1.1), a change of 30 injections to 12 injections per month (2.1 versus 1.0), or a change in FLS presenting 3 days after every injection to 1 day after some injections (2.1 versus 1.0) (Figure 2).

OPEN IN VIEWER

Figure 2.

Mean preference weights for the total study population (N = 189).

The relative importance of a change in the frequency of injections per month from 12 to 2 was 0.7 (Figure 2). The relative importance of this change was approximately equivalent to the relative importance of the number of relapses in the next 4 years from three to one (0.6), or an improvement from 1 to 2 years in the number of years until MS symptoms get worse (0.6).

Data from the three additional RPL models explored interactions between frequency of injections per month with injection time, FLS and ISR. None of these interactions were found to be statistically significant (p > 0.05), suggesting that preferences for injection time, FLS, and ISR do not vary systematically with frequency of injections per month (data not shown).

Discussion

This study used DCE methodology to evaluate preferences of patients in Germany for different attributes of a hypothetical injectable MS treatment. As expected, patients preferred an improvement in outcomes for each attribute, which is consistent with the results of previous DCE studies [Johnson et al. 2009; Shingler et al. 2013; Poulos et al. 2015; Utz et al. 2014; Wicks et al. 2015; Wilson et al. 2014, 2015].

There were no significant differences in preferences among patients reporting a diagnosis of RRMS versus patients reporting other types of MS (including those who did not know what type of MS they had). There was also no statistically significant difference between the preferences of patients who were treatment-experienced and the patients who were treatment-naïve across all attribute levels (online supplementary material, Figure S1). Nor was there a statistically significant difference between preferences of patients who were injection-experienced versus patients who were injection-naïve, across all attribute levels (online supplementary material, Figure S2). These data suggest that although patients have different actual experiences with respect to injections, treatments, and their MS type, all patients included in the sample have similar preferences regarding features of hypothetical injectable DMTs.

Overall, frequency of injections and disability progression were the two key drivers for patient preference. Specifically, the relative importance of a change in frequency of injections from 30 injections per month to one injection per month and an improvement in the number of years until MS symptoms get worse from 1 to 4 years, were almost twice as important as an improvement in number of relapses from four relapses to one relapse or an improvement in FLS from 3 days after every injection to none. These results suggest, that patients do differentiate between injectable treatments based on injection frequency. This is consistent with data presented in previous studies [Poulos et al. 2015; Shingler et al. 2013; Utz et al. 2014; Wicks et al. 2015]. The change in injection time was not shown as being important to patients compared with the other attributes studied. Regarding the tolerability of treatments, patients showed a strong negative preference towards FLS, which would be experienced for 3 days, after every injection.

Together, these data support the evidence that a perceived lack of efficacy and treatment-related AEs remain a concern for patients with MS [Treadaway et al. 2009; Patti, 2010; Shingler et al. 2013].

Disease severity is classified by the Hohol disease step scale [Hohol et al. 1995].

Of the patients who reported that their ‘current’ symptoms were moderate or worse (n = 128 [67.7%]) and the patients who reported that their ‘current’ symptoms were mild or better (n = 61 [32.3%]), there were no significant differences in overall preferences for treatment attributes. However, patients with mild or better symptoms placed greater weight on having only one relapse over the next 4 years than patients with moderate or worse symptoms (p < 0.01). Interactions between frequency of injections per month and injection time, FLS, and ISR were not statistically significant. We therefore did not find any evidence that preferences for these three attributes were dependent on the level of injection frequency.

There are limitations to DCE which should be taken into consideration when evaluating these data. Not all of the attributes related to injectable DMTs which would influence or affect patients’ treatment choice were incorporated in the study. The set of attributes selected was chosen to be sufficient to cover only the fundamental aspects of injectable DMTs.

Patients were aware that the questions were regarding a hypothetical device and actual device choice may be influenced by factors not assessed in this study (e.g. financial factors such as treatment cost).

With regard, to the study populations assessed, the patients had a ‘self-reported physician diagnosis of MS’; therefore, it is difficult to evaluate how representative this sample is of a typical patient population with MS. There is a small possibility that some of the survey respondents were not patients with MS, and participated in the survey to earn modest compensation (€5). Also, given the sample size of the study population, we acknowledge that the data may not be representative of all individuals with MS in Germany. No evidence was found that preferences among patients with and without injection experience or those patients who were treatment-experienced and treatment-naïve were statistically different from one another.

The survey instrument was translated from English into German by a professional translation service, and the survey was adapted to the local context. The instrument was pretested in German and, based on the pretest findings, revised by native German speakers with expertise in DCE studies. Despite the care taken in translation and adaptation of the survey instrument, the survey was not back-translated and reconciled with the English survey to ensure that the survey content was equivalent. It is possible that patients’ perceptions of the meaning of the descriptions and survey questions were somewhat different than we have described in this manuscript due to possible discrepancies in the translation.

Finally, when interpreting these results, it should be noted that they do not indicate the overall importance of treatment attributes, but rather, the importance of changes in attribute levels that lie within the ranges included in this study. Also, despite several of the attributes that are used in this study having been previously shown to influence adherence, these results do not describe actual adherence, or even intended adherence.

Conclusions

Understanding which attributes of injectable DMTs influence patient preference could help improve wellbeing and, potentially, treatment adherence in patients with MS. This study provides evidence that patients place a significant value on improvements in the frequency of dosing and disability progression. The results also suggest that changes in injection frequency may be as important as some changes in efficacy or safety attributes of injectable treatments for MS.

The majority of the attributes examined did have an influence on preference, suggesting that DMT attributes are carefully considered by patients and may influence patient adherence. This should be taken into consideration by physicians when selecting appropriate treatments for patients with MS.