Applying health equity implementation science frameworks to population genetic screening
Institute for Public Health Genetics, University of Washington, Seattle, WA, USA
Department of Bioethics & Humanities, University of Washington School of Medicine, Seattle, WA, USA
Department of Laboratory Medicine and Pathology, University of Washington, Seattle, WA, USA
Department of Biomedical Informatics and Medical Education, University of Washington, Seattle, WA, USA
Kaiser Permanente Washington Health Research Institute, Kaiser Permanente Research Affiliates Evidence-based Practice Center, Seattle, WA, USA
*Corresponding author; email: ndrao@uw.eduABSTRACT
Implementation science frameworks with a focus on health equity have emerged to help guide the introduction of new interventions into healthcare and community settings while limiting health disparities. The purpose of this research was to explore the applicability of such frameworks to guide the equitable implementation of population genetic screening programs. We searched PubMed and reference lists for relevant frameworks and examples of their use in health settings. We then assessed if and how selected frameworks provide guidance for different stages of population genetic screening: recruitment, sample collection, result return, follow-up care and prevention, and cascade screening. Findings were synthesized into a list of health equity considerations specific to each stage. We identified 5 implementation frameworks that focus on health equity. Guidance varied by framework type: determinant (explaining what affects implementation outcomes), process (translating research into practice), or evaluation (assessing implementation). Common characteristics included focusing implementation efforts on populations who have historically experienced health inequities and adapting interventions to fit local contexts. Process models also highlighted the importance of community partnerships. Overall, frameworks offered broad recommendations applicable to population genetic screening program implementation. However, gaps still exist in guidance provided for later stages of population genetic screening. To improve the equitable implementation of genetic screening, future programs may benefit from utilizing one or more of these frameworks or by incorporating the health equity considerations and outcomes compiled in this analysis.
Article notes
Competing Interest Statement
The authors have declared no competing interest.
Funding Statement
This study was funded by the Brotman Baty Institute for Precision Medicine and the National Institutes of Health R01HG010144.
INTRODUCTION
Population genetic screening, or genetic screening of people regardless of personal or family history of disease, has been proposed to increase the reach of genetic services and identify more people at risk for preventable conditions (Grzymski et al., 2020; King et al., 2014; Limburg et al., 2011). However, population screening is complex for many reasons including the need to appropriately inform large numbers of people about the benefits and harms of genetic screening, collecting DNA samples, and following people over time to ensure they receive appropriate care based on their results. If not implemented with care and the needs of underserved people in mind, population genetic screening may perpetuate or further exacerbate already existing health disparities (Grzymski et al., 2018).
To limit harmful consequences, health equity must be a central consideration in the design and implementation of population genetic screening programs. Health equity is defined as everyone having a fair and just opportunity to be as healthy as possible (Centers for Disease Control and Prevention, n.d.). Striving for health equity requires focusing on the needs of those who are at greatest risk of poor health due to social circumstances (Braveman, 2014). It involves the elimination of health differences that are linked to social determinants historically connected to exclusion, such as race, ethnicity, socioeconomic status, gender, age, religion, disability, sexual orientation, gender identity, and geographic location (Braveman, 2014).
Implementation science frameworks with a focus on health equity have emerged to guide the introduction of new interventions into healthcare and community settings, and their principles could improve the incorporation of genomic discoveries into healthcare (Chambers et al., 2016; Roberts et al., 2017). How well such frameworks suit population genomic screening programs is not well understood.
We conducted a systematic literature search to identify and describe published equity-focused implementation science frameworks and assessed the applicability of these frameworks to population genetic screening programs.
METHODS
Population genetic screening stages
We conceptualized population genetic screening in 5 major stages (Figure 1) based on the design of existing pilot screening programs (David et al., 2021; East et al., 2021; Grzymski et al., 2018; Rao et al., 2023). Our descriptive model’s stages include recruitment, sample collection, return of results, follow-up care and prevention, and cascade screening (the notification of biological relatives about genetic risk). This model formed the basis for our analysis of framework applicability to population screening.
Framework identification
We searched PubMed for frameworks designed to promote health equity during the implementation of health interventions using the following keywords: (“health equity” or “health disparities” or “health inequalities”) and (“implementation” or “translation”) and (“framework” or “model” or “theory”). One author, NDR, screened the resulting titles and abstracts. When articles cited potentially relevant frameworks or included a review of frameworks, NDR examined the reference lists for pertinent publications. Other frameworks previously known to the authors were also considered. Article review was restricted to work published between January 2010 and December 2021, as the focus on health equity in implementation science has become more prominent relatively recently (Odeny, 2021).
Criteria for inclusion were frameworks focused on health equity and implementation of health services, and that were developed for high-resource settings, as these are the most relevant to population genetic screening. Frameworks were excluded if they were specific to a certain health condition or intervention, provided little guidance for implementation or if the article was not available in English. Discussion papers, or those that only described a need for considering health equity during program implementation or provided no explicit framework or model, were also excluded.
Data extraction and evaluation
For each of the selected frameworks, the following data was extracted: name, author, year of publication, type, audience, development, and description. Framework type was determined according to Nilsen’s categorizations of implementation science theories, models, and frameworks: determinant frameworks, process models, or evaluation frameworks (Nilsen, 2015). Determinant frameworks are designed to assist with understanding barriers or facilitators that influence implementation outcomes; process models to guide the process of translating research into practice; and evaluation frameworks to specify implementation outcomes (Nilsen, 2015).
Data synthesis
We assessed the applicability of each framework to population genetic screening by evaluating if and how the framework provided guidance for the 5 major stages of population genetic screening we identified. To further our understanding of potential framework application, we looked for examples of how each framework may have been used in other settings by examining articles published through December 2021 that cited our selected frameworks. We also searched for empiric evidence of framework validation.
We then compared the selected frameworks and discussed their strengths and weaknesses with respect to guiding the implementation of population genetic screening. Using findings from our applicability assessment, we also compiled a list of health equity considerations and outcomes specific to each stage of population genetic screening.
RESULTS
The initial PubMed search yielded 1,013 results. An additional 37 articles were identified through reference lists or because they were previously known to authors. Records were screened by title and abstract followed by full text review (Figure 2). We identified five frameworks designed to reduce or prevent health disparities during the implementation of health interventions (Table 1). One framework was described in two of the selected articles.
We categorized the frameworks as one determinant framework (HEIF, Woodward et al., 2019), three process frameworks (Proctor reframed, Baumann & Cabassa, 2020; Transcreation, Nápoles & Stewart, 2018; EquIR, Eslava-Schmalbach et al., 2019), and one evaluation framework (RE-AIM extension, Shelton et al., 2020). Researchers were the primary intended audience for each framework. Frameworks were largely conceptually developed by the authors, though one, EquIR (Eslava-Schmalbach et al., 2019), was developed using stakeholder engagement (Table 2). Our search for evidence of framework validation yielded no information about if and how any of the five frameworks had been validated.
The Health Equity Implementation Framework (HEIF)
Description
Woodward and colleagues developed HEIF by integrating the i-PARIHS implementation science framework (Harvey & Kitson, 2016) and the Health Care Disparities Framework (Kilbourne et al., 2006). HEIF is designed to help researchers determine factors related to innovation uptake and disparities in healthcare to improve outcomes for marginalized populations (Woodward et al., 2019). Health equity domains include culturally relevant factors, the clinical encounter, and societal context (Table 2). Culturally relevant factors are specific to intervention recipients based on their lived experience and can include characteristics such as socioeconomic status, implicit bias, health literacy, trust in providers, language, race and ethnicity. The clinical encounter encompasses interactions between providers and patients, which influence if an intervention is offered by a provider or accepted by a patient. These encounters are influenced by inner context at the local (e.g., clinic) and organizational (e.g., hospital) levels, and outer context (e.g., the healthcare system). Finally, the societal context includes economies, physical structures (how environments are built or arranged), and sociopolitical forces (social norms or political forces). These impact health disparities by influencing the inner and outer context, the clinical encounter, and culturally relevant factors. The HEIF has previously been applied to design an interview guide and direct content analysis to identify implementation factors and best practices for social needs screening in primary care settings (Drake et al., 2021).
Application to population genetic screening
The HEIF is well suited to provide guidance for anticipating possible barriers or facilitators to implementation across all stages of population genetic screening (Table 3). For example, during recruitment, attention to cultural factors can help researchers anticipate how language and cultural beliefs influence informed consent and enrollment. HEIF’s physical structures domain can inform how in-person sample collection and return of results may facilitate or impede screening depending on access to reliable transportation and the location of facilities. Similarly, during follow-up care and prevention, understanding potentially inequitable physical spaces can inform implementation. In the cascade screening stage, reflecting on sociopolitical forces, such genetic privacy laws, may illuminate barriers to information sharing among relatives.
Proctor reframed
Description
Baumann and Cabassa (2020) reframed the Proctor implementation science framework to provide an example of how to apply an existing framework to address healthcare inequities. The original Proctor framework posits that interventions differ from their implementation strategies and requires the involvement of various stakeholders at multiple levels (Proctor et al., 2009). The original Proctor also proposes outcomes in three interrelated but distinct domains: implementation (e.g., feasibility, fidelity, acceptance), service (e.g., efficiency, safety, effectiveness), and client (e.g., satisfaction, function). The reframed Proctor framework emphasizes collaborating with stakeholders and community members throughout intervention planning, design, and implementation in order to understand and meet the needs of historically underserved communities (Table 2) (Baumann & Cabassa, 2020). It proposes continually adapting programs based on the needs of populations with the goal of reducing inequities through systematic changes to intervention and implementation strategies. Finally, Proctor reframed suggests conducting descriptive and explanatory studies to identify factors that contribute to inequities in implementation outcomes. We did not identify any published applications of Proctor-reframed in health settings.
Application to population genetic screening
Proctor-reframed specifies guidance most relevant to the recruitment stage, including ensuring that populations that have previously experienced inequities in genetic services are included in population genetic screening programs. Suggested strategies for enhancing inclusion are conducting programs in non-traditional settings such as in faith communities or community centers. This framework also discusses how face-to-face presentations with community members and person-to-person recruitment can assist with enrolling people who would otherwise not participate.
Transcreation
Description
Transcreation is defined as the process of planning and delivering interventions to reduce health disparities that resonate with the intended community (Nápoles & Stewart, 2018). Nápoles and colleagues created this framework to address the differences that occur between original intervention implementation settings (often among mainstream populations or in academic settings) and when interventions are adopted among a population facing health disparities.
Collaboration is a central principle of Transcreation, which proposes stakeholder and community involvement through the entire process of intervention design, implementation, and adaptation (Table 2) (Nápoles & Stewart, 2018). This framework assumes the presence of an established partnership between researchers and community members and a shared understanding of the disparity to be addressed. As part of the framework’s proposed collaboration, Transcreation suggests involving community workers in implementation by training them in intervention delivery. Fitting interventions to context and population needs is also prominent.
Transcreation has previously been applied in other health settings; for example, it has been used to adapt a stress management intervention for Latina breast cancer survivors living in rural settings (Santoyo-Olsson et al., 2019).
Application to population genetic screening
Transcreation provides guidance most relevant for the initial recruitment stage of population genetic screening by suggesting focusing attention on populations who experience disparities in access to and utilization of genetic services (Table 3). Through the incorporation of scientific evidence, programs can also adopt recruitment strategies that have been proven to work in similar settings.
Unique to Transcreation is training community members in intervention delivery. This idea is relevant for all population screening stages, as members can be trained to provide cultural, informational and logistical support to specific communities within a general population. During recruitment, this can promote informed decision-making. In the follow-up and prevention stage, community members acting as patient navigators can assist individuals with information about insurance or recommended medical interventions.
EquIR
Description
Eslava-Schmalbach and colleagues developed EquIR for researchers and decision makers to reduce or prevent health inequities during the implementation of health programs or policies (Eslava-Schmalbach et al., 2019). This conceptual framework is cyclic, with social determinants of health considered throughout. The cycle begins with identifying disadvantaged populations and quantifying current health inequalities (Table 2). It then suggests developing and implementing recommendations to meet the needs of disadvantaged populations with key players such as health professionals, patients, community members, and stakeholders. It finishes by recommending the monitoring of implementation outcomes (Table 2) and identifying how the intervention has impacted the health status of populations receiving the intervention. From here the cycle continues and the new population health status becomes the starting point of the intervention. Beyond its initial development, we identified no examples of the use of EquIR to guide implementation of health interventions.
Application to population genetic screening
Of the guidance proposed by EquIR, the described outcomes are most readily applied to population genetic screening and can be used to understand how programs impact disadvantaged populations at each stage. For example, measures of acceptability and appropriateness can be applied during recruitment, sample collection, and return of results stages to understand stakeholder perceptions of fit and usefulness of program procedures (Table 3). Measures of fidelity and coverage may be useful during follow-up care and cascade screening stages for understanding how often people receiving positive screening results are able to act on these results and how often risk information is shared with biological relatives. In addition, the cyclical nature of EquIR promotes ongoing program adjustments informed by these outcome measures.
RE-AIM extension
Description
The extension to the RE-AIM framework authored by Shelton and colleagues is designed to promote sustainability and health equity. The original RE-AIM framework focuses on evaluation and includes both individual and staff/setting level domains: Reach and effectiveness (individual), adoption and implementation (staff/setting), and maintenance (individual and staff/setting) (Glasgow et al., 1999, 2019). While the extension to RE-AIM discusses these same domains and previously described indicators, Shelton et al. provide additional guidance to consider health equity during the measurement of these indicators (Table 2). This guidance focuses on assessing indicators over time across different populations of focus (defined by age, race, ethnicity, disability, insurance status, literacy level or other social determinants of health), to identify and address health inequities (Shelton et al., 2020). The extension to RE-AIM also considers the link between health equity and costs or resources and suggests incorporating cost estimates and resource requirements into planning discussions with stakeholders. This framework has previously been used to evaluate the implementation of a COVID-19 vaccine program seeking to facilitate equitable vaccine access and uptake among Latinx community members (Marquez et al., 2021).
Application to population genetic screening
The outcome indicators and health equity considerations listed by RE-AIM extension give measures that can be monitored at each stage of population genetic screening (Table 3). During recruitment, relevant indicators include the number of people who are offered screening and the number of people who agree to screening. Taking into account social determinants of health when interpreting these indicators can determine if all populations are offered and enroll in screening similarly and reveal which populations are not reached. Reach can also be ascertained for sample collection, return of results, and cascade screening to find inequities that may be emerging during these stages.
Measures of effectiveness across social determinants of health are also relevant for the return of results, follow-up care, and cascade screening stages. For return of results, indicators include the number of people experiencing psychosocial harms upon learning results. For follow-up care and prevention, relevant indicators are the number of people who are able to engage in preventive interventions who desire it and the number of people who experience psychosocial harms because of difficulties accessing care. During cascade screening, indicators include the number of biological relatives who receive testing.
Comparing frameworks
A number of characteristics were shared across the analyzed frameworks. The first was to focus implementation efforts on populations who have historically experienced health inequities. This is a crucial consideration as placing specific emphasis on underserved populations at the beginning of implementation planning can reorient design and procedures to better prioritize the needs of such communities. Another common element across frameworks was to adapt interventions to fit local context and meet the needs of marginalized communities. Doing so can limit the implementation gap, which occurs when the context where interventions are designed and developed does not align with realities of implementation settings. Constraining this gap can increase the appropriateness of an intervention (Baumann & Cabassa, 2020).
The identified frameworks were conceptually focused, rather than validated theories, and the guidance provided varied by framework type, as expected. The process models, Proctor reframed (Baumann & Cabassa, 2020), Transcreation (Nápoles & Stewart, 2018), and EquIR (Eslava-Schmalbach et al., 2019), for example, tended to be high-level, and provided overarching considerations and recommendations for program design, implementation, and evaluation rather than specific guidance that lends itself to individual stages of an intervention like population genetic screening. Regarding screening, Proctor reframed (Baumann & Cabassa, 2020) and Transcreation (Nápoles & Stewart, 2018) recommendations applied most directly to the recruitment stage. While all three process models described evaluating implementation outcomes keeping social determinants and differences in outcomes across populations in mind, they varied in the specificity with which they described and defined these outcomes.
In contrast, the determinant framework, HEIF (Woodward et al., 2019), provides an explicit means to identify barriers to program implementation throughout all stages of population genetic screening. Similarly, the evaluation framework, RE-AIM extension (Shelton et al., 2020), detailed indicators and health equity considerations for monitoring program outcomes relevant to all program stages.
Among the process models, Proctor reframed (Baumann & Cabassa, 2020), Transcreation (Nápoles & Stewart, 2018), and EquIR (Eslava-Schmalbach et al., 2019), another main concept was the importance of involving community partners and other stakeholders throughout implementation. Such collaboration allows researchers to learn more about local customs and build trust with community members (Cabassa & Baumann, 2013; Wallerstein & Duran, 2010). Interventions can better be tailored to a specific population and integrate relevant perspectives, norms, and social and cultural values. As a result, this may improve intervention acceptability and effectiveness and prevent health disparities from emerging.
DISCUSSION
In this analysis we outline relevant equity considerations for population genetic screening program implementation guided by five selected frameworks: HEIF (Woodward et al., 2019), Proctor reframed (Baumann & Cabassa, 2020), Transcreation (Nápoles & Stewart, 2018), EquIR (Eslava-Schmalbach et al., 2019), and RE-AIM extension (Shelton et al., 2020). The HEIF (Woodward et al., 2019), RE-AIM extension (Shelton et al., 2020), and outcome measures provided in EquIR (Eslava-Schmalbach et al., 2019) were applicable to all stages of population screening. Remaining guidance from EquIR (Eslava-Schmalbach et al., 2019) and ideas proposed by Proctor reframed (Baumann & Cabassa, 2020) and Transcreation (Nápoles & Stewart, 2018) tended to be broad and less applicable to each individual screening stage.
Results of our analysis may offer insights for researchers designing new population genetic screening programs and assist with identification and selection of relevant frameworks to direct implementation. To make the best use of the variety of recommendations brought up by the different frameworks, these frameworks may best be used in tandem. For instance, determinant domains can be used when process model steps suggest identifying implementation barriers and specific indicators can be drawn from evaluation frameworks when steps call for assessing implementation outcomes.
We found that guidance for later stages of population genetic screening programs, such as follow-up care and cascade screening, was limited beyond central framework characteristics. For true public health impact, individuals receiving positive screening results must have access to services to delay or prevent disease onset. Health benefits may also be seen if genetic risk information is communicated to relatives. However, frameworks lacked specific guidance about how to ensure equitable referrals to follow-up care or promote adherence to recommended medical interventions. Discussion about sharing health insights and how to engage biological relatives who may be implicated by an individual’s risk results was also missing. This is not entirely surprising given that these frameworks were not specifically developed with genetic services in mind, but demonstrates that some considerations specific to genetic testing are not addressed by the current literature.
Additionally, implementation frameworks emphasizing health equity have limited guidance for ongoing programs. While the cyclical nature of EquIR (Eslava-Schmalbach et al., 2019) and ongoing evaluation measures provided by RE-AIM (Shelton et al., 2020) can be used to some extent, guidance particularly relevant to genetics again appears to be absent. For instance, frameworks provide little assistance about how to incorporate new risk information over time or ensure that providers are up to date on genetic recommendations so that they can best advise their patients. Again, as the frameworks analyzed here were not designed to be genetics specific, gaps in guidance are expected and suggest a need for frameworks tailored for genetic programs.
Population genetic screening health equity considerations
Synthesizing findings from the included frameworks, we have compiled a list of relevant health equity questions and outcomes that warrant consideration during the implementation of population genetic screening programs in order to limit health disparities (Table 4). Though not exhaustive, questions may be useful throughout the design and implementation of future screening programs and spur further discussion related to pursing health equity. Broadly, considerations include the accessibility and cultural sensitivity of different population screening processes. Outcomes focus on understanding the distribution of benefits and harms from genetic screening, and the acceptability of program procedures across various demographic groups.
One of the overall considerations for pursuing health equity is involving community partners (Table 4). As members of the community are likely more in-tune with local settings compared to researchers, they may be better equipped to understand and identify drivers behind complex inequities (Leveraging Community Expertise to Advance Health Equity, 2021). Through community engagement, researchers and public health professionals can ascertain what communities identify as problems to be addressed and what community health priorities are. This can inform if population genetic screening is a suitable intervention in a particular setting and truly meeting community needs. Investment by communities in population genetic screening programs can also promote sustainability of such programs.
Even with these health equity considerations identified, challenges may emerge when incorporating these ideas into practice. For example, answers to these questions may vary by communities included in a single screening program. Resource constraints may also prevent the adoption of more equitable practices. Additionally, outcome measures may be difficult to ascertain if they involve time-intensive data collection and the continued engagement of people who have taken part in genetic screening. As such, researchers and health professionals looking to implement screening programs may benefit from using these considerations to appropriately plan and allocate resources.
Limitations
Our study may have been limited by the frameworks considered for analysis. Due to stringent inclusion criteria, we may not have identified all relevant frameworks. In addition, though we give a broad overview of implementation science frameworks that center health equity, we did not assess the quality of these frameworks. However, our analysis of how frameworks have been used in other settings provides an indirect measure of utility and quality. Our assessment of framework applicability to population genetic screening programs may also have been limited. Relevant stages, such as sample lab testing, were not included and some of our analysis may be applicable to multiple stages though not detailed here. Despite these limitations, our study is a first step in describing the current state of implementation science frameworks that explicitly focus on health equity and how they can be applied to improve the equitable implementation of population genetic screening programs.
CONCLUSION
Current implementation science frameworks that emphasize health equity offer broad recommendations applicable to the implementation of population genetic screening programs. However, gaps still exist in guidance provided for stages of screening that are ongoing, such as follow-up care and cascade screening. Through our application of frameworks to population genetic screening, we have created a list of considerations and outcomes that may assist with more equitable implementation. Researchers planning to implement screening programs may benefit from consulting these considerations or following guidance from analyzed frameworks.
Data Availability
All data produced in the present work are contained in the manuscript