Background

The World Health Organization estimates that epilepsy affects about 50 million people worldwide [1]. Individuals with epilepsy have poorer health outcomes and report a poorer quality of life [2,3,4]. Self-management is important for individuals with epilepsy in order to maintain optimal physical, cognitive, and emotional health following diagnosis [5, 6]. Engagement in self-management may increase self-efficacy, medication adherence, avoidance of seizure triggers, and improves patient and family knowledge about when to seek urgent medical care [7, 8]. For instance, education based-approaches positively impact self-management, whereas psychosocial therapy-based approach positively impact quality of life [9, 10]. A recent Cochrane review identified weak support for intervention strategies such as using epilepsy-trained nurses and educational interventions to support self-management in adults with epilepsy [11]. Yet, rates of self-management and adherence to treatment regimens remain suboptimal among individuals with epilepsy, leading to increased medical care and premature death [5, 6].

As with many chronic health conditions, challenges to self-management for adults with epilepsy include low health literacy, poor social support, low levels of education, medication and disease-related side effects, and low socioeconomic status [12,13,14,15,16,17,18]. Patients with epilepsy have additional disease-specific self-management concerns which can impact engagement in, and adherence to, self-management [19]. The effectiveness of self-management in adults with epilepsy may be influenced by commonly co-occurring medical conditions such as traumatic brain injury, mood disorders, associated cognitive impairment, and/or impulse-control issues. Further complicating self-management plans, high levels of disease-related stigma, driving or mobility restrictions, and social isolation may serve as barriers to attending intensive epilepsy self-management programs and/or medical appointments that occur in-person and/or require a sustained time commitment [20,21,22]. Thus, engagement in self-management and treatment adherence may be positively influenced when self-management programs are informed by knowledge of epilepsy-specific considerations and how these considerations may influence self-management.

A possible solution to epilepsy-specific self-management concerns may be self-management programs offered within the context of clinicians providing health care for patients with epilepsy. Implementing self-management programs in health care systems requires consideration of many multi-level factors [23]. Facilitators and barriers to implementation can occur at any level of an intervention, including the person (patient or caregiver), program, and system levels. Research has shown the importance of allocating resources, balancing context-specific adaptation and program fidelity, and involving key stakeholders during initial and sustained implementation and adoption [24,25,26,27]. Although scientific organizations recommend self-management support [28,29,30], there is little guidance about how to implement these programs, and we identified no prior systematic reviews that focused on this issue.

We conducted a systematic review of quantitative, qualitative, and mixed-methods studies to examine the barriers and facilitators to implementing and adopting epilepsy self-management programs in large health care systems. We chose qualitative evidence synthesis methods to incorporate evidence from multiple study designs that addressed the same overarching concept of implementation and adoption of epilepsy self-management programs. Our study addresses gaps in understanding facilitators and barriers at the person, program, and system levels that impact implementation and adoption of epilepsy self-management programs.

Methods

This systematic review was part of a Veterans Health Administration (VHA)–funded report available online (www.hsrd.research.va.gov/publications/esp) [9, 10]. The purpose of the systematic review was to address gaps in evidence regarding the efficacy and effectiveness of self-management programs for adults with epilepsy, synthesize the current evidence on self-management programs for patients with epilepsy, and identify potential barriers in the adoption of these programs within a health care system. We developed and followed a standard protocol for this review in collaboration with stakeholders (PROSPERO: CRD42018098604). We adhered to the Preferred Reporting Items for Systematic Reviews and Meta Analyses (PRISMA) guidelines [31] (Additional file 1).

Literature search and study selection

The conceptual model, literature search, and study selection methods for the systematic review are described in greater detail in the full monograph [9, 10]. A full list of inclusion and exclusion criteria (i.e., population, intervention, comparator, outcome, timing, and setting) is listed in Table 1. The major criteria for eligibility included randomized or quasi-experimental studies that enrolled adults with epilepsy and evaluated self-management interventions and reported a relevant clinical, process, or economic outcome. Eligible interventions were required to have a primary focus on self-management of epilepsy, could not be education-only or general medical care interventions, and could not focus primarily on a comorbid psychological diagnosis. Additionally, eligible studies focused on adults (aged ≥ 18) with new or chronic epilepsy, family members and/or caregivers of adults with epilepsy, and/or stakeholders involved in implementation of epilepsy self-management programs. For this specific analysis, we included additional observational designs and qualitative studies that addressed facilitators and barriers to intervention implementation and adoption. We operationalized self-management as programs that facilitated the provision of epilepsy-specific knowledge and skills that assisted an individual and/or caregiver in incorporating epilepsy self-management behaviors into daily life. Our definition was based upon a conceptualization of self-management by Jonkman and colleagues [32]; however, we modified the definition to only include one component beyond knowledge acquisition [9, 10]. We operationalized implementation and adoption as the uptake, and continued use, of an evidence-based intervention in health care practice [34].

Table 1 Eligibility criteria (reproduced with permission from Luedke et al. [9])
Full size table

For this systematic review, we collaborated with a medical librarian to develop a search strategy for each database (Additional file 2) [9, 10]. We searched MEDLINE® (via PubMed®), PsycINFO, Cochrane Central Register of Controlled Trials (CENTRAL), and CINAHL in April 2018, and the MEDLINE search was updated in March 2019. Citations and the full-text of potentially eligible studies were evaluated by two investigators. We resolved disagreements using consensus between two investigators, and involved a third investigator as needed. All articles that met eligibility requirements and addressed facilitators and barriers to implementation and adoption were included for data abstraction.

Data extraction strategy

Data abstraction was completed by a team of two co-investigators (AAL, AS) who had experience in qualitative methodology, under the guidance of the primary investigator (JWW). Data from published reports were entered into a Microsoft Excel spreadsheet created for this project. Abstracted data included study purpose, study design and details, participant eligibility criteria, recruitment strategy, control group, participant type (i.e., patient, caregiver, clinician, researcher), sample size and demographics, main findings salient to facilitators and barriers of implementation and adoption, and other comments. We abstracted implementation and adoption barriers (i.e., description of themes or factors that impeded the use and implementation of the intervention as reported in the study’s results and/or findings sections) and facilitators (i.e., description of themes or factors that aided the use of the intervention as reported in the study’s results and/or findings sections) to the implementation of self-management interventions (as distinct from barriers and facilitators of an individual engaging in self-management behaviors). Data were entered independently by one investigator, and then independently reviewed by a second investigator. Following the completion of data entry, the two investigators met to discuss abstracted data; disagreements were resolved by consensus.

Risk of bias

To evaluate the risk of bias (ROB) across included studies, we employed ROB instruments appropriate for each study design (e.g., qualitative, cross-sectional, mixed-methods) [35, 36] (Additional file 3). The evaluation of ROB in each study was based on an assessment of whether the study was well-designed and conducted within the research paradigm proposed by the study authors. These tools assess risk of bias in 5 to 10 domains, categorizing each domain as low, unclear, or high ROB. The tools are not intended to create summative scores [35, 36], and so we applied a more holistic approach by considering the individual domains and overall impressions of the study. First, each study was independently rated by each qualitative investigator (AAL, AS) using a ROB tool. Then, using this first ROB rating, the two investigators met and discussed each study in the context of valid and rigorous qualitative research methodology [37,38,39]. For instance, we discussed sampling technique (e.g., did the authors describe recruitment and sample selection in detail? Did the sampling align with the population described in the introduction?); analysis (e.g., Was analysis completed by one individual or more than one individual? Was analysis completed by the individual who completed data collection?); and measures (e.g., Does the data collection method align with the introduction and purpose of the study?). After discussing each study, we assigned an overall ROB using the following definitions: (1) low ROB—bias, if present, is unlikely to alter the results seriously; (2) unclear ROB—a risk of bias that raises some doubts about the results; and (3) high ROB—bias may alter the results seriously. Disagreements were resolved by consensus between the two qualitative investigators and the senior author. We did not validate each ROB tool for use in our systematic review as we did not make any real change to the intent or the goal of each instrument.

We used the 10-item Critical Appraisal Skills Programme (CASP) for Qualitative Research Studies [35] for the included qualitative studies (n = 5). Each item is rated “Yes,” “No,” or “Can’t tell”; there is no summary rating. The overall ROB in the 5 qualitative studies was low [22, 40,41,42,43]. For the quantitative descriptive studies, we used the 5 items in the Mixed Methods Appraisal Tool (MMAT) [36]. These criteria address the sampling strategy, the sample representativeness, measurements, risk of nonresponse bias, and appropriateness of the statistical analysis. The MMAT rates each item “Yes,” “No,” or “Can’t tell”; there is no summary rating. We used the 5-item MMAT specific to mixed-methods studies [36] for the mixed-methods study (n = 1). These criteria address the rationale for using mixed-methods, the integration of the study components, the interpretation of the study components, discussion of divergences or inconsistencies between the quantitative and qualitative data, and how each component of the study adheres to the quality criteria of each of the quantitative and qualitative methods. Similar to the quantitative descriptive studies, the MMAT for mixed-methods studies rates each item “Yes,” “No,” or “Can’t tell”; there is no summary rating.

Data synthesis and presentation

Data synthesis was completed by the same co-investigators under the guidance of the primary investigator. We analyzed the abstracted data using thematic synthesis and the framework method [44, 45]. Thematic synthesis utilizes a three-stage method for data synthesis: (1) free coding of findings in the included studies, (2) organizing data into descriptive themes, and (3) generating analytical themes that combine the findings of individual studies into interpretations that cross the findings of the studies [45]. The framework method is a complementary approach to management and analysis of qualitative data, which relies on charting the data into a matrix based on descriptive and analytic codes [44]. Using the research question and our full-text review and familiarization with these data as a guide, we created a framework based on three levels from the social ecological framework [46] that included barriers and facilitators as reported for a category (i.e., patient with epilepsy or caregiver, program or intervention, and site or health system). Then, we described each piece of data either as being a facilitator or a barrier to implementation of a self-management program. Each piece of data was subsequently grouped into an ecological framework level (i.e., patient/caregiver, program/intervention, site/health system). We then completed first-level analysis of these data and confirmed the validity of our interpretations by referencing the original texts. After the data were independently coded and discussed among the two qualitative researchers, we conducted thematic synthesis by identifying and grouping related codes within each category (e.g., patient/caregiver, program/intervention, site/health system). Then, each researcher independently organized related codes into themes. We reviewed the themes and then identified overarching themes that applied to both facilitators and barriers. The creation and identification of codes and themes was iterative. To ensure rigor and validity of these findings, we independently coded and developed themes. We then discussed theme development and identification until we reached agreement between the two researchers. Next, we conducted a narrative synthesis of the data and displayed its summary in a crosstab of themes by study.

Ecological levels

For each facilitator and barrier, we first identified the respondent (e.g., the patient with epilepsy, caregiver, or health care clinician) associated with each theme. Then we examined each theme at one of three levels, adapted from ecological models of health behavior. The ecological models of health emphasize that determinants of behavior derive from individuals and characteristics of their environments that influence behavior directly and in interaction with one another [46]. Table 2 provides operationalized definitions and exemplars for each of the three ecological levels. These categories were mutually exclusive; thus, each facilitator and barrier could only be placed in one category (i.e., patient/caregiver, program/intervention, site/health system).

Table 2 Definitions and examples of the ecological levels
Full size table

Results

We identified 13 studies addressing facilitators and barriers to implementing self-management interventions for persons with epilepsy (Fig. 1) [22, 40,41,42,43, 47,48,49,50,51,52,53,54]. Study designs in this analysis included semi-structured interview (n = 5) [22, 40,41,42,43]; cross-sectional survey (n = 5) [47, 48, 50, 53, 54]; longitudinal survey (n = 1) [52], mixed-methods including group semi-structured interview, cross-sectional survey, and records review (n = 1) [49], and a discrete choice experiment (n = 1) [51]. Study respondents included patients with epilepsy, caregivers, and health care clinicians together [42, 49, 52,53,54], patients and caregivers only [22, 40, 41, 43, 47, 48, 51], and health care clinicians [50]. We describe study characteristics including study sample, epilepsy type, and intervention in Table 3.

Fig. 1
figure 1

PRISMA literature. OECD, Organization for Economic Cooperation and Development. Asterisk means reference did not report a relevant clinical, process, or economic outcome

Full size image
Table 3 Study characteristics
Full size table

Thematic synthesis identified 5 themes that could be applied conceptually to facilitators and barriers. These themes, described in Table 4, included relevance, personalization, intervention components, technology considerations, and clinician interventionist. The presence of themes varied by study as shown in Table 5.

Table 4 Themes across studies of self-management of epilepsy
Full size table
Table 5 Presence of themes by study
Full size table

Facilitators

The presence of facilitators of epilepsy self-management interventions at any level (i.e., person, program, site/system) was noted in 11 studies [22, 40,41,42,43, 47,48,49,50, 52, 54]. Two studies did not include any facilitators [51, 53].

Relevance

At the person level, facilitators included desire of the patient with epilepsy to acquire self-management skills [40, 54] and the desire to have content that was highly applicable to living with epilepsy (e.g., through eliciting concerns about self-managing and daily living from the patient or caregiver) (41, 52, 54). At the program level, facilitators included content designed to: enable the patient’s acquisition of skills for living with epilepsy, learn how to apply self-management skills, develop coping strategies for daily life, and communicate with family, caregivers, and health care clinicians about epilepsy [41, 42, 48, 54]. No relevance facilitators were identified at the site/system level.

Personalization

At the person level, facilitators included identifying whether the patient owned the necessary technology for the intervention (e.g., computer, mobile telephone) [47], had an identified source of social support (e.g., parents, significant others, friends) [43], and whether the intervention was congruent with the patient’s preference for individual peer-support or group interaction [41, 42]. At the program level, facilitators included developing the intervention and tailoring its components to build on the current self-management practices of the patient [47, 54]. No personalization facilitators were identified at the site/system level.

Intervention components

At the person level, facilitators included patient or caregiver receiving written materials (e.g., educational content) to refer to during and after the intervention [42]. At the program level, facilitators included involving family members in the intervention [49], using an empowerment approach to help the patient develop self-management skills [41, 54], and the format of the intervention (e.g., group format that included both the patient and caregiver) [42]. Additional program level factors include the ability to personalize materials to each patient [42, 48, 54], the availability of written materials [41, 42, 52], the ability to interact with a group [22, 49], the provision of peer support [22, 49], and the length and duration of the intervention sessions [22, 49]. At the site/system level, facilitators included developing intervention materials using uniform program standards to ensure program fidelity across intervention sites [42]. One study indicated that the site of the intervention (e.g., medical center) was unimportant, as patients with epilepsy indicated no preference of one site over another [48].

Technology considerations

At the program level, facilitators included the high level of usability of intervention components located on the internet, mobile applications, or phones [47, 52, 54]. No technology facilitators were identified at the person or site/system levels.

Clinician interventionist

At the program level, facilitators included creating an intervention team consisting of a patient in tandem with an expert health care clinician who could deliver the intervention content [48]. No clinician interventionist facilitators were identified at the person or site/system levels.

Barriers

The presence of barriers to epilepsy self-management interventions at any level was noted in 11 studies [22, 41,42,43, 47, 49,50,51,52,53,54]. Two studies did not include any relevant barriers [40, 48]. We noted that in the studies that addressed barriers, stakeholders included clinicians, social service providers, and researchers [50], and patients and clinicians [53].

Relevance

At the program level, barriers included incongruent responses between patients and clinicians about the patient’s problems and needs [53], incongruent opinions by clinicians, researchers, and social service providers on the necessary intervention content [50], and incongruent responses between patients and clinicians on who should lead the intervention and provide epilepsy self-management education and support [53]. Additional barriers at the program level included educational content that was too general or lacking in personalization or tailoring to the patient, his or her disease state and relevant comorbidities [22, 53, 54], and not identifying what the patient views as important in self-management and living with epilepsy [49]. No relevance barriers were identified at the person or site/system levels.

Personalization

At the person level, barriers included the patient’s memory and/or cognitive impairments [22, 42, 47, 50, 51, 53], the patient’s disinterest in participating in a self-management intervention [51], not identifying the patient’s preference or desire for self-management support [43], and no current use of the technology by the patient [47]. At the program level, barriers included not accounting for the cognitive limitations of the patient [22]. At the site/system level, barriers included not accounting for the characteristics of the patient population such as the patients’ access to health care [50] or transportation concerns [49].

Intervention components

At the program level, barriers included having the patient incur a cost to participate in the intervention [51], not identifying how demographics (e.g., country of origin, burden of disease, socioeconomic status, level of activation) influence the patient’s participation and views of the intervention [51], not identifying the ideal time for follow-up by the clinician after the intervention [49], and not having clear instructions for the role of caregivers participating in or affected by the intervention [49]. Additional program level barriers include not having written materials (e.g., program manuals, handouts, website resources) the patient can refer to during and after the intervention [41, 49], having groups heterogeneously composed of individuals with disparate experiences of living with epilepsy [22], experiencing challenges to scheduling group intervention sessions because of calendar conflicts for participants and clinicians [42], and not identifying the optimal duration and length of the intervention for patients [22, 41, 51]. Barriers at the site/system level included having different levels of attrition at study sites [52] and challenges to using a participatory approach to intervention development and content identification (e.g., lengthy time to complete, need to obtain ethical approval, and efforts to ensure participant engagement) [42]. No intervention component barriers were identified at the person level.

Technology considerations

At the person level, barriers included the patient’s lack of knowledge about eHealth tools and functions, having concerns about the privacy of eHealth tools, and varying individual preferences for using technology for epilepsy self-management [47]. At the program level, barriers included difficulty developing eHealth tools with high usability, and a lack of help for users encountering technical difficulties [47, 49, 54]. At the site/system level, barriers included not acknowledging or addressing the inequity of access to eHealth tools within the sample or the person with epilepsy’s concerns about the privacy of eHealth tools [47].

Clinician interventionist

At the program level, barriers included not incorporating the duties of the intervention into the clinician interventionist’s normal job duties (i.e., a collateral duty) [49], not adequately preparing the clinician interventionist to deliver the intervention content [49], developing a complex intervention protocol that is difficult to deliver [49], and not identifying the optimal professional role and educational training of the clinician interventionist [50, 53]. At the site/system level, barriers included a lack of opportunity for the clinician interventionist to engage in continuity of care for the person with epilepsy following the conclusion of the intervention [49] and not accounting for the limited time allotted for medical visits [50, 54]. No clinician interventionist barriers were identified at the person level.

Risk of bias

The tools used to assess risk of bias (ROB) (Table 3, Additional file 3) for the descriptive quantitative, mixed-methods, and qualitative studies did not provide for the calculation of summary scores for individual papers. Among the 7 descriptive quantitative studies, [47, 48, 50,51,52,53,54] ROB was heterogeneous. Patterns that led to judgments of higher ROB included unclear representativeness of the sample (n = 6) [47, 48, 50, 51, 53, 54] high (n = 2) [50, 53] or unclear (n = 4) [48, 51, 53, 54] ROB from non-response, unclear risk of bias in sampling strategy (n = 3) [48, 50, 54] and unclear appropriateness of measures (n = 2) [50, 54].

Discussion

This study is the first known systematic review of facilitators and barriers to implementation and adoption of epilepsy self-management interventions. This review aimed to identify facilitators and barriers by examining evidence from diverse studies. We identified five themes: relevance of intervention content, personalization of intervention content and methods, intervention components, technology considerations, and clinician interventionist considerations. Within these themes, facilitators and barriers were identified at the person, program, and system-levels. Our findings add to the literature as recent systematic reviews on self-management interventions for adults with epilepsy [11, 57, 58] have not focused on facilitators and barriers to intervention implementation and adoption (Table 6).

Table 6 Highest priority evidence gaps
Full size table

Our findings underscore a number of key disease-specific considerations for the implementation of epilepsy self-management interventions. Self-management practices that align with an individual’s needs, values, and preferences are considered an important component in delivering patient-centered care [59, 60]. Data in the present review indicates that interventions should be flexible in order to accommodate a person with epilepsy’s preferences about peer support, group interactions, existing self-management strategies, existing self-management social support, and technology preferences. Intervention materials should include written and portable self-management information that is personalized and tailored to the unique situation of each person with epilepsy. Notably, self-management information should be available prior to, during, and after intervention sessions so that individuals with epilepsy, and their caregivers can review and reference self-management content when engaging in self-management behaviors.

Epilepsy providers should incorporate delivery of self-management education and skills into their regular clinical practice to enhance successful delivery of relevant self-management programmatic content. Incorporating the intervention duties into the epilepsy provider’s role and duties will help ensure successful implementation of the intervention in practice and that the interventionist is appropriately trained in the intervention protocol. Overall, our findings indicate that the number of clinical personnel and the time requirements for participation in the intervention are important factors to consider when implementing and adopting interventions. However, no studies in our review addressed the costs related to implementing and adopting self-management programs for epilepsy, nor did any studies address clinician-related costs. Given the personnel and time required for epilepsy self-management programs, new models of care delivery (e.g., telemedicine, mHealth) should be considered as a means to facilitate implementing and adopting epilepsy self-management programs in a cost-effective manner. Virtual care delivery models may be less costly than face-to-face programs and may be more scalable, thus increasing the reach and impact of self-management programs for persons with epilepsy.

One key finding of our study is that no studies directly addressed implementation and adoption issues for epilepsy self-management programs in large health systems. Our included studies addressed factors primarily at the patient level or program level. Important themes that could inform the development, implementation, and/or adoption of future epilepsy self-management interventions included the desire of patients with epilepsy to be involved in the development of intervention content and recognition that cognitive limitations may affect engagement and adherence. Stakeholder engagement is particularly important in developing self-management programs for epilepsy as differences exist among individuals with epilepsy and clinicians in regard to reasons for nonadherence and factors that impact epilepsy self-management [30, 61, 62]. Involving key stakeholders (e.g., patients, caregivers, providers, administrators) may help promote facilitators, and attenuate barriers, to implementing and adopting an epilepsy self-management program. Of note, no studies addressed facilitators and barriers to implementing and adopting epilepsy self-management interventions in relation to cost-effectiveness. Therefore, future research should focus on identifying cost-effective strategies that promote facilitators and reduce barriers to implementation and adoption of epilepsy self-management programs.

Persons with epilepsy expressed a desire for a team composed of an individual with epilepsy and a clinician interventionist to deliver self-management education and support. Thus, future research should further examine the composition of this dyad and identify who the interventionist should be (e.g., registered nurse, advanced practice registered nurse, physician, physician assistant, etc.). One study identified that persons with epilepsy did not have a preference to intervention site such as a medical center or other location. Future research should determine the preferred location for a self-management program for persons with epilepsy and caregivers of persons with epilepsy, with an awareness that access to public transportation may vary widely within a community. Across the studies, persons with epilepsy stated several factors such as access to health care, familiarity and ownership of technology, finances, current self-management support, and transportation that influenced intervention engagement. Future research should focus on the extent to which these factors either individually or in combination influence the person with epilepsy’s initial and sustained engagement in an epilepsy self-management program. We did not identify any study that focused explicitly on the facilitators and barriers to the implementation and adoption of self-management programs for persons with epilepsy for a large health system. Future research is needed that specifically addresses the implementation and adoption of epilepsy self-management programs, as there may be additional personal, program, and site/system level barriers that need to be identified and addressed.

Limitations and strengths

This is one of the first known systematic reviews to apply rigorous qualitative methods to synthesize data from studies with a wide range of designs (e.g., cross-sectional and longitudinal surveys, discrete-choice experiments, semi-structured interviews, records review, and mixed-methods designs). Using these qualitative evidence synthesis methods, we were able to obtain a more complete understanding of the phenomenon of facilitators and barriers to implementation of epilepsy self-management programs than by comparative methods alone. Yet, several limitations of our study need to be acknowledged. First, while we made every attempt to identify and locate interventions and manuscripts that report on the implementation of epilepsy self-management programs, we may have inadvertently missed relevant articles. Additionally, we limited our search to Organization for Economic Co-operation and Development countries; this search limitation may have inadvertently excluded relevant self-management interventions completed in other countries that could have provided valuable insight into the implementation of epilepsy self-management programs in large health systems. Second, the tools used to assess the risk of bias (ROB) for the studies included in the present review did not allow for the calculation of summary scores. We assessed all studies for ROB. Of the 13 studies, only 1 exhibited a high ROB due to insufficient information about ethical concerns, lack of rigorous analysis of study findings, and no clear value of the research. The remaining studies exhibited either low or unclear ROB. Additionally, we did not contact authors of the studies included in this review to clarify questions pertaining to methods or data collection. Contacting authors to obtain additional information on methods may have influenced our ROB assessment. For the studies in the present review, we addressed the inherent diversity by identifying the respondent (e.g., patient with epilepsy, caregiver, or clinician) and then synthesizing emerging themes within ecological levels.

Our study has a number of strengths. Our review benefited from leveraging input from content experts, being protocol driven, and using rigorous qualitative methods for analyzing barriers and facilitators to implementation and adoption. The use of a theory-based, a priori framework, and a thorough, multiple-investigator review process ensured rigor and validity of our process, findings, and interpretation. By not limiting our inclusion criteria to either quantitative or qualitative designs, we were able to obtain a more comprehensive picture of relevant facilitators and barriers to implementation and adoption of epilepsy self-management programs.

Conclusion

Findings on facilitators and barriers to implementation underscore key considerations for the design, implementation, and adoption of self-management interventions, including factors of patient personalization, information delivery, use of technology, and intervention personnel. Future research should be designed to address these implementation issues. Future research also should focus on the extent to which intervention components (e.g., peer support), use of technology, and other identified barriers/facilitators influence the person with epilepsy’s initial and sustained engagement in an epilepsy self-management program. In conclusion, our data indicate the importance of incorporating disease-specific considerations when designing, implementing, and adopting an epilepsy self-management program.