Using an Intervention Mapping Approach to Improve Adherence to Disease-Modifying Treatment in Multiple Sclerosis

Multiple sclerosis (MS) is the most prevalent chronic inflammatory demyelinating disease of the central nervous system and is currently incurable.¹  Although no treatment fully prevents or reverses the progressive neurologic deterioration, disease-modifying treatments (DMTs) have been shown to limit the number of relapses and prevent new lesion formation, as well as reduce the accumulation of disability^1,2  among individuals with relapsing remitting MS (RRMS). For individuals with RRMS, these outcomes translate into lower risk of relapse, better quality of life, and higher treatment satisfaction. More than a dozen DMTs are now available.¹ 

Many of the DMTs require consistent self-administration across time; however, a meta-analysis on adherence to oral DMTs found that by the end of the first year of treatment, 20% of individuals with RRMS were not taking their DMTs as prescribed,³  known as nonadherence,⁴  and 25% discontinued treatment, known as nonpersistence.⁵  Studies with a 2-year follow-up period reported that only 60% of individuals with RRMS persisted in their treatment.⁶  Both nonadherence and nonpersistence are associated with lower quality of life as well as with increases in relapse risk, hospitalizations, emergency department visits, neuropsychological issues, work loss, and MS-specific medical costs.^7–14  Notably, adherence rates among individuals with MS are higher than among those with other conditions such as epilepsy, Parkinson disease, and rheumatoid arthritis.¹⁵ 

Adherence to DMT poses special challenges. Despite the chronic nature of RRMS and the necessity of long-term treatment, the need for DMTs may be less obvious during periods of disease inactivity. Disease progression can occur with no noticeable symptoms, potentially making individuals with RRMS less concerned about appropriate adherence. Benefits of treatment accrue only in the long term, yet adverse effects (AEs) of medication that reduce quality of life are immediate.¹⁶  These are the challenges described in the therapy-related factor in the World Health Organization (WHO) adherence model⁴ ; the other factors are social and economic (eg, age, ethnicity, education), health care system (eg, type of insurance), condition related (eg, duration, comorbidity), and patient related. The patient factor in particular is often studied because its components—such as perceptions of illness, medication beliefs, habits in medication, and affective states—are considered potentially modifiable in interventions. In MS, most studies on adherence and persistence have focused on the socioeconomic factor, data on which are easily retrieved from administrative records, and only a few studies have investigated patient-related^17–21  or therapy-related factors.^11,22–24 

Implementing interventions to increase adherence is essential to improve the care of individuals with RRMS. However, despite avid interest in assessing adherence and unravelling its predictors/correlates, few studies have used previous findings to develop interventions. This paper focuses on the development of an evidence-based and theory-driven motivational interviewing intervention—a technique focused on exploring and resolving ambivalent feelings and perceptions—to support individuals with RRMS who are taking DMTs and to improve their adherence rates.

We designed the intervention described here based on theory and prior empirical evidence to ensure the possibility of replication and adoption, if successful. We used the Intervention Mapping (IM) approach²⁵  to structure the process of design and implementation. The study follows recent guidelines requesting more comprehensive and transparent reporting²⁶ ; this affords a better understanding of the considerations that led to choosing specific intervention components, as well as potential pitfalls in the process. One model for such reporting is a self-management intervention for medication adherence among survivors of breast cancer.²⁷ 

The IM framework is anchored in 2 principles: First, include stakeholders to ensure relevance, acceptability, implementability, and contextual and cultural appropriateness; second, use theories and evidence. The framework includes 6 stages: (1) assessing needs, (2) identifying intervention objectives and behavioral determinants, (3) designing and/or selecting theory-based methods for behavior change, (4) organizing and producing the intervention; (5) implementing the intervention, and (6) evaluating the intervention (see FIGURE S1). The first 4 stages are described in detail in the rest of this paper.

The first stage, which included data collection, was approved by the Institutional Review Board of Carmel Medical Center (#0061-14-CMC). All participants signed written informed consent forms confirming that they were free to leave the study at any time.

The first stage of IM involves carrying out a needs assessment to establish the extent of the health problem and to identify the determinants associated with the problem behavior. This needs assessment was informed by a review of the literature, using search terms around “DMT,” “MS,” “adherence,” and “persistence,” as well as a search for systematic reviews on medication adherence in general and in MS.^{3,15,22,28,29}  In addition, we carried out a longitudinal observational study pertaining to the patient factors in medication adherence in the target population of individuals with RRMS. This study assessed medication adherence and persistence among individuals with RRMS taking oral and injectable DMTs in multiple modalities (eg, self-reported; administrative records) and examined the prospective associations among demographic, clinical, and perceptual predictors to adherence and persistence.^20,30  As results from the longitudinal study began accumulating, we adopted the Necessity-Concerns Framework, which evolved into the Perceptions and Practicalities Approach (PAPA).^31,32  This approach to the analysis of patients’ actions in regard to their medication use focuses both on perceptions that shape an individual’s motivation (ie, intentional non-adherence) and on practical issues sometimes beyond a patient’s control, regardless of intention (ie, unintentional nonadherence).

The intervention’s desired results were determined in this stage. Specific goals for enhancing adherence were developed via theory-based PAPA constructs and guided by the specific needs of individuals with RRMS. These goals were dissected to determine particular activities or techniques that could improve adherence. At this point in the process, practitioners can be explicit about the intervention goals and the targeted determinants.

In stage 3, we identified theory-based behavior change techniques and methods that relate to the determinants found in the needs assessment (ie, stage 1) by scanning the literature for studies that offered insights into the determinants specified in the intervention objectives.

In this stage, the identified behavioral strategies are combined to develop the intervention. The development was an iterative process and was informed by feedback from the research team, individuals with RRMS, and clinicians. Two psychologists in the clinic reviewed the content of the intervention. In addition, 5 female individuals with RRMS taking a DMT gave their opinions at 2 decision points. Their primary language was Hebrew (n = 3) or Arabic (n = 2). They were recruited either via their physician’s recommendation or after filling out a survey.

Nonadherence to oral DMT among individuals with RRMS was found to be prevalent. A recent systematic review and meta-analysis evaluated adherence using either the medication possession ratio or the proportion of days covered, which are based on administrative records of persons being supplied with prescribed medication within a given time period. The mean medication possession ratio of >80%— which is considered adherent—was present in 63.5% to 88.5% of individuals with RRMS, and it decreased over time.³  Moreover, discontinuing 1 medication or switching medication in the first year of treatment was common (25%).³ 

We conducted a prospective cohort study to examine adherence and persistence, and their predictors, among individuals with RRMS (n = 186). The study included demographic and clinical characteristics as well as patient-reported medication beliefs, illness perceptions, medication habits, perceived health, and affect (ie, depression and anxiety). The patients were prospectively assessed at 3 time points: baseline, 6 months later, and 12 months later. Using multiple measures (multiple self-reports, military personnel records, and others²⁰ ), we found that, at different time points, 69.9% and 71% of individuals with RRMS were adherent (with slight variation among medications^20,30 ) and 64.5% were persistent. This prospective analysis also identified a consistent barrier: the concerns and beliefs held by individuals with RRMS about their prescribed medications,²⁰  specifically those about potential overtreatment and about harm the medications could cause.

Perceptions regarding medication were assessed within the Necessity-Concerns Framework.³¹  The framework posits that a person weighs the perceived need for treatment (necessity) against the potential harm of the treatment (concerns) prior to deciding whether to take the treatment. Indeed, recent work has identified reasons for intentional nonadherence to include AEs, mild course of disease, and cost²¹  (however, cost had little relevance for our sample because all were enrollees in a public health system). Another patient-related variable is satisfaction with treatment, which was associated with adherence both in a meta-analysis³³  and among individuals with MS.¹⁷ 

Other studies among individuals with MS have found that unintentional nonadherence is also common, often associated with memory difficulties, affective states, and personality.¹⁹  Other statistically significant (or marginally significant) predictors of nonadherence or nonpersistence at some measurement point were socioeconomic status, anxiety, habits, and depression.²⁰  These findings, concomitant with variability within the same individual over time³⁰  and across treatments,³²  indicate that ideally, an intervention for an individual with RRMS needs to be flexible to best address different barriers to adherence. TABLE 1 lists principal determinants of nonadherence at the patient level identified at this stage.

The needs assessment identified that while the intervention should address both intentional and unintentional nonadherence, it was more important to focus on intentional nonadherence (ie, medication beliefs). Reducing intentional nonadherence was found to be consistently predictive of adherence in our sample. Retaining flexibility, in case other barriers emerged, was set as a priority. Several determinants of adherence were identified as modifiable targets for the intervention objectives. The intervention needed to help the individual with RRMS:

• (1) develop an accurate perception of MS as an illness;

• (2) increase knowledge of and accept the necessity of DMTs;

• (3) reduce concerns about DMTs and accept the increased importance of DMTs in their life;

• (4) reduce the impact of AEs;

• (5) increase self-efficacy regarding DMT adherence; and

• (6) reduce the likelihood that they will forget to take DMTs.

Although knowledge about MS and DMTs was not found to be directly associated with DMT adherence, we reasoned that increased understanding of MS and DMT could affect beliefs on necessity, harm, overtreatment, and concerns about the medication. We also reasoned that addressing the issue of DMT-associated AEs could increase confidence in managing AEs and also affect concerns and beliefs (eg, around DMTs’ necessity and possible associated harms).

Many adherence interventions are not grounded in theory,²⁸  although use of theory was found to have a consistent (albeit modest) effect.²⁸  Basing interventions in theory has the potential to increase their effectiveness; this is accomplished by identifying constructs on which to focus, then targeting effective behavior change techniques to address these constructs.³⁴  Our findings in stage 1—ie, that beliefs about medications were prospectively associated with adherence—directed our search for a theoretical framework. This led us to adopt PAPA^31,32  and the aligned behavior change techniques.

Moderator analyses on interventions in medication adherence^35,36 —examining whether intervention effects vary depending on intervention characteristics—indicated that delivery mattered: that is, face-to-face interventions administered directly to patients were more effective.²⁹  Content mattered as well: Intervention content that resulted in higher adherence included behaviorally focused and habit-based interventions, but cognitive-focused interventions designed to change knowledge, beliefs, or attitudes resulted in less adherence. Finally, standardized (vs individualized) interventions resulted in higher adherence. Accordingly, we decided on a face-to-face administration, a flexible yet structured modality, and inclusion of information and behavioral content.

Despite the seeming ineffectiveness of information provision in improving adherence,³⁶  we chose to include the information component in the intervention for several reasons. First, providing information on the clinical benefits of DMTs is particularly important for this class of preventive medications. DMTs are beneficial in that they reduce disease activity (eg, annual relapse rate, appearance of new lesions in MRI, accumulation of disability1) and these benefits accrue over the long term. However, daily symptoms are not relieved by taking DMT medications (ie, there is no explicit positive reinforcement), and medication AEs are immediate; it is important for patients to understand this scenario. Secondly, good communication and provision of information about pharmacological treatment and diseases are aligned with the principles of patient-centered care and shared decision-making. Lastly, when education or information is integrated with other interventions, there is some evidence that they have a positive impact on adherence, although results are not consistent.³⁶  Accordingly, the educational part was followed by other components.

For our sample, beliefs about medications were consistently identified as determinants of nonadherence in the Needs Assessment stage. As described above, providing information to ameliorate inaccurate perceptions was an initial intervention component; however, in order to confront these unhelpful beliefs, a nondirective style was employed, using strategies such as reflective listening and open-ended questions to draw out the patients’ motivations, difficulties, and decisions related to adherence.³⁷  This was accomplished by using motivational interviewing, which is a patient-centered, evidence-based psychological counseling method that is aimed at exploring and resolving ambivalent feelings and perceptions regarding behavior change. This technique has grown in use across diverse health settings in order to support behavior change,³⁸  including medication adherence,³⁸  and among individuals with MS.³⁷  This intervention component was delivered by a psychologist specializing in health.

Individuals with MS (n = 5) provided feedback on the format (eg, group vs one-on-one), scope (ie, number of meetings/sessions), and content of the intervention. Initially, the intervention was designed as a support group that would meet for 6 to 8 sessions; however, patients pointed out that this commitment was too demanding for working adults, and they questioned whether individuals with MS would complete such a schedule. Accordingly, the format was changed to 3 one-on-one meetings, either integrated into regular clinic visits or conducted over the phone. Individuals with MS agreed that a longer meeting with the attending physician, integrated into a planned clinic visit and focusing on MS and DMT information, would be valued and that personal sessions with a health psychologist could be helpful. At a later timepoint, other individuals with RRMS provided feedback on the general content of the intervention.

The intervention had 2 major components: educational, delivered by a neurologist, and behavioral, delivered by a psychologist, and both were largely carried out face-to face. The intervention targeted individuals with RRMS who had been identified as suboptimally adherent, and it was customized to each specific patient. The intervention (including the needs assessment of stage 1) was registered (ClinicalTrials.gov; IDCMC-14-0061-CTIL) and received institutional review board approval (IRB #NCT02488343).

* Component 1: Education. The neurologist first briefly reviewed what MS is, including its types, the definition of active disease, and the risks to the patient of remaining untreated. This information was communicated both orally and visually, reviewing the personal records of the individual with RRMS (eg, a recent MRI) if available. This section addressed types of treatment—that is, those treatments that can help address exacerbations, alleviate symptoms, improve quality of life, and prevent deterioration. The last category, which includes DMTs, was the focus of the conversation. The neurologist first addressed the rationale for DMT, showed an illustration of the DMT’s mode of action, and talked about treatment modalities. Possible AEs and how to manage common AEs were then addressed. Finally, the neurologist reviewed how outcomes would be monitored, the importance of adherence, and what would happen if doses were missed. The information about AEs was tailored to the individual patient’s specific history and current medication. The behavioral change techniques employed were: information about health consequences (eg, information provided about how missing doses can affect MS progression), salience of consequences (eg, using visuals to show what happens to the brain when medication is missed), reattribution (eg, discuss AEs and nonspecific symptoms and explore alternative explanations), and problem solving (eg, discuss how to decrease and manage AEs).

* Component 2: Behavioral. Individuals with RRMS who were identified as suboptimally adherent received 3 one-on-one meetings with a health psychologist; 2 were in-person and 1 was via telephone. The first in-person meeting with the psychologist immediately followed the education visit with the neurologist. It was followed by a telephone call 2 weeks later from the same psychologist, then followed by another in-person meeting with the psychologist 6 months after the initial meeting.

The purpose of this component was to normalize the ambivalence toward DMTs, while still communicating their importance. It began with an open-ended question about having MS—how the disease affects one’s hopes and goals for life—then moved on to medication-taking behavior. This afforded an opportunity to review the pros and cons of DMT use and the patient’s specific barriers and facilitators, if they chose to share them with the psychologist. TABLE 2 specifies the potential personalized behavior changes³⁹  that the MI techniques can encourage. The MI techniques used were also recorded by the psychologist at the end of each in-person session.

In discussing nonadherence, possible reasons were examined, as were insights into how nonadherence can be tied to other personal values and how patients envision their future selves. Then, they talked through potential ways to resolve concerns and barriers. The psychologist and the individual with RRMS also discussed ways of overcoming any drug-related AEs. When unintended reasons for nonadherence predominated, the discussion veered toward practical tips: These include conscious planning; forming habits that pair medication habits with self-identified key daily activities or times (by preference); use of reminder apps; and/or use of a daily marked table of adherence (by paper and pencil), sent weekly to the psychologist. The researchers explored including a public medication app with clinician access, but this option was not approved due to privacy issues.

Upon completion of the one-on-one meetings, the psychologist recorded how the patient perceived the importance of medication adherence as well as their self-efficacy (ie, its importance to them and their confidence in their ability to advocate for themselves) and the specific MI techniques used during the meeting. In order to monitor fidelity (ie, the extent to which an intervention was delivered as conceived and planned), 6 of the in-person meetings were recorded at random, contingent on receiving permission from the person with MS.

Adherence to DMT treatment was assessed by the patient-reported outcome measure of the Probabilistic Medication Adherence Scale³⁰  (an 18-item questionnaire assessing adherence behaviors) because it yielded the most variance among individuals with MS during the observational stage.⁴⁰ 

The intervention was initially tested for acceptability (ie, willingness to enroll in the intervention), feasibility (ie, completion of intervention), and initial efficacy; a before-after design was employed for this phase. The design was chosen due to expected acceptability and because the expected dropout rate of a between-participants design would be too low in power.

This report describes the development of an intervention to promote adherence in individuals with RRMS taking DMTs. The intervention was created through a series of iterations that included input from researchers, practitioners, and individuals with MS. The intervention aims to support those individuals with RRMS who have been identified as suboptimally adherent with their DMT treatment; it does so by helping them to learn more about the DMT’s mechanism of action, to manage their AEs, to deal with ambivalence surrounding medication taking, and to remember to take the medication. Feedback from individuals with MS was informative, mainly shaping the intervention in terms of schedule.

This report fills a gap in the literature by utilizing the IM framework, which guides a clear and transparent delineation of theoretical, empirical, and practical considerations. The emphasis in stage 3 on theory centers this stage, guiding both the interpretation of the needs assessment (stage 1) and the intervention development (stage 4). The inclusion of theory assures neither the efficacy nor effectiveness of the intervention, yet it increases the odds of effectiveness.²⁸ 

The findings of stage 1 (needs assessment), indicating no difference in adherence among different DMTs or administration modalities,³⁰  guided intervention development. This mirrored the findings of other studies among individuals with MS and other neurological conditions^{11,22,41–45}  and steered the intervention away from administration modality, a therapy-related factor in the WHO 5-factor adherence model.⁴  The finding that adherence and persistence were not consistently associated with demographic or clinical characteristics is similar to some research findings^42,45  and different from others; it steered the intervention away from social, economic, and condition-related factors. The finding that adherence and persistence were consistently prospectively associated with specific patient perceptions about DMTs²⁰ —beliefs about overuse and potential harm, as well as general concerns—was congruent with previous work in other conditions^46,47 ; this finding steered the intervention toward MI, a technique aimed at exploring and resolving ambivalent feelings and perceptions, which also allowed for flexibility for other reasons for nonadherence.²¹ 

The acceptability, feasibility, and efficacy of the developed intervention were assessed because of lessons learned from literature reports (reviews and meta-analyses) that interventions to increase medication adherence across conditions have a moderate effect size,²⁹  if any,⁴⁸  and because of the predictors of adherence uncovered during stage 1 (needs assessment). Interventions by health care professionals and patient-support programs⁴⁹  run by pharmaceutical companies use similar behavioral change techniques,⁵⁰  including self-monitoring, shaping knowledge, framing/reframing, action planning, habit formation, prompts and cues, and problem solving. Some of these techniques were used in the few medication adherence interventions carried out among individuals with MS.^38,51,52  Similarly,^37,52  using the flexible technique of MI, the current study applied these techniques in a personalized manner, dependent on the reasons underpinning suboptimal adherence in the individual. It did so with a larger sample than previously reported³⁷  and among individuals with MS taking various DMTs, rather than a specific DMT.⁵² 

The development process had several strengths and limitations. A strength was using both general (eg, meta-analyses, reviews) and patient-specific information on medication adherence in the needs assessment. A second strength was that the development relied on evidence-based strategies for behavior change, employing MI and incorporating relevant and tested behavior change techniques. Third, researchers consulted individuals with MS and other clinicians. Fourth, the intervention targeted both oral and injectable DMTs. Fifth, the intervention targeted individuals with MS who were identified as suboptimally adherent and not a more general population with MS.

The developed intervention also harbors some weaknesses. First, the advantage of targeting individuals with RRMS who are already identified as suboptimally adherent can also present a limitation. Not all clinics routinely monitor patient adherence, and the identification is not fully valid, as there are many ways of assessing adherence.³⁰  Second, the developed intervention involved a behavioral specialist (a health psychologist), a professional who is unavailable in many clinics. Third, intervention developers consulted with only 5 individuals with MS, all women, who clearly did not represent all individuals with MS served by the clinic. Their recommendations and opinions cannot be considered generalizable, although they provided an additional perspective to those of health care professionals.

This intervention was developed in a systematic manner, following the IM guidelines, assuring incorporation of both theory and practical considerations. The intervention has been tested in a feasibility study that showed promise. As many individuals with RRMS have suboptimal adherence, this intervention has the potential to be applied widely: both for identified suboptimal adherents and as a preventive course for individuals with RRMS who are prescribed DMTs. In the latter case, materials would have to be structured to allow wide rollout and would then require separate feasibility and efficacy studies. Before upscaling, the current intervention also needs an experimental design to test efficacy and effectiveness.