Evaluating Community Health Worker Programs for Noncommunicable Disease Management: An Integrative Review

Hannah S. Scranton, Johns Hopkins University

Angela Chang Chiu, Johns Hopkins University

Introduction

Management of chronic, or noncommunicable diseases (NCDs) is a global health priority due to their significant contribution to disability and mortality (Pan American Health Organization [PAHO], 2021). Reducing NCD-related morbidity and premature mortality is a key target of the United Nations’ 2030 Sustainable Development Goals (UN, 2015). Social determinants of health disproportionately expose low-income communities to higher NCD risk, poorer health and limited access to preventative care (Marmot & Bell, 2019). The COVID-19 pandemic underscored these longstanding health inequities and prompted public health experts to advocate for improved data systems to better address the needs of marginalized populations (Maani et al., 2021).

Globally, Guatemala is one of the nations with the highest burden of NCDs, where diabetes and hypertension affect 9.7% and 21.2% of the population respectively (PAHO, 2022). This data does not account for the morbidity and mortality of these two diseases in rural Guatemala which is comprised primarily of indigenous populations, who are largely underserved by the healthcare system. Community-based health interventions utilizing trained community health workers (CHWs) are recognized as effective models of care that can address healthcare disparities in underserved and marginalized communities (Berini et al., 2022; Scott et al., 2017).

CHW programs implicate a long-term investment of financial, human and physical resources (Berini et al, 2022). This review explores how the impact of a CHW program for NCD management in marginalized communities could be quantified to evaluate achievement of goals, return on investment, and overall value to the communities served.

Methods

Search Strategy

An integrative review was conducted to examine best practices in CHW program evaluation in rural and indigenous settings. A search of international literature was conducted using Medical Subject Headings (MeSH) and other relevant keywords, such as “community health worker*) OR as “community health aide*” OR as “health auxiliary*” AND “program evaluation” OR “impact assessment” OR “evaluation” AND “population health” OR “population health management” OR “noncommunicable disease” OR “chronic disease” OR “hypertension” OR “diabetes” OR “rural” OR “indigenous”. Four electronic databases were queried: PubMed, Embase, Web of Science, and Latin American and Health Sciences Literature (LILACS) were queried electronic databases (Appendix 1).

The literature search was performed including articles published between 2019 and 2024. A total of 770 records (133 from PubMed, 68 from LILACs, 231 from Web of Science, and 177 from Embase) were identified and imported to Covidence reference management software where title/abstract screening, full-text screening, removal of duplicates, and data extraction were performed. After removing 280 duplicates, 490 titles and abstracts were screened for relevance.

Articles met the inclusion criteria if they described CHW programs and relevant NCD outcomes. The primary outcome of interest was validated evaluation frameworks or tools for CHW-related programs, while the secondary outcome was improved NCD-related outcomes within the communities. Inclusion criteria were studies that involved CHW program evaluation in rural or indigenous settings published between 2019 and 2024. Exclusion criteria were research comprising children under 18 years old, acute care setting, urban setting, infectious disease management, or study protocols. Non-peer-reviewed publications, articles without full text, and studies conducted in languages other than English (or without English translations) and Spanish were excluded.

Data Extraction and Synthesis

The Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines and checklist for reporting were employed for reporting the results of this integrative review (Page et al., 2021). Reference management and data extraction procedures were performed using Covidence online reference software (Covidence, 2018). All the references were uploaded to Covidence, and records of the screening process were retained in Covidence throughout the review process. Two articles in an urban setting were manually included to reflect the importance of utilizing a validated framework for evaluation while addressing NCDs in underserved communities (Dannefer et al., 2024; Feinberg et al., 2019). Two authors (HS and ACC) collaboratively screened abstracts against inclusion and exclusion criteria to identify the selected papers. No conflicts arose during the review.

Quality Appraisal

Evidence was appraised using the Johns Hopkins Nursing Evidence-Based Practice (JHNEBP) model, which assigns levels based on study design and quality. The model utilizes a hierarchical scale from Level I to Level V. Level I represents the strongest evidence and includes systematic reviews and meta-analyses. Level II includes quasi-experimental studies, Level III covers non-experimental designs, qualitative studies and meta-synthesis, Level IV entails expert consensus informed by research evidence, and Level V includes expert opinion papers (Dang & Dearholt, 2017). Additionally, the quality of each evidence was also rated on a three-point scale: A for high quality, B for good quality, and C for low quality or studies with major methodological limitations (Dang & Dearholt, 2017).

Findings

After screening based on the inclusion criteria, 393 studies were found to be irrelevant. Of the 97 articles, 83 were excluded from the database search results, leaving 14 articles after full-text review. A hand search of the article reference list was performed, identifying five additional articles that met the eligibility criteria; two out of the five articles were excluded due to their design. A total of 17 articles were selected and included for data extraction based on the inclusion and exclusion criteria, shown in the PRISMA flowchart (Figure 1).

Figure 1. PRISMA Flow Diagram

Seventeen articles were selected and included in the full-text review. Articles were published between 2019-2024, eight of which were set in rural, underserved communities in the US, one each in Indonesia, Zimbabwe, India, Guatemala, and British Columbia. One study was conducted in urban Harlem, NYC (Dannefer et al., 2024); another publication about the same program was manually added (Feinberg et al., 2019).

Of the 17 articles appraised, only one study was rated Level I Quality B, representing the highest level of evidence and good quality, providing strong support for the findings it reports. Six articles were Level II with Quality A or B ratings, including quasi-experimental studies and one randomized control trial (RCT), offering moderate to strong evidence for program effectiveness. Most of the articles (eight) were Level III, Quality A or B, consisting of non-experimental designs, qualitative studies, and mixed-method approaches. While these studies provided valuable insights into implementation processes and contextual factors, they are limited in their ability to establish causational relationships. Two articles were rated Level III Quality C, indicating lower quality evidence with methodological limitations. A comprehensive table of the literature appraisal is outlined in Table 1.

The six mixed-methods studies provide both quantitative outcome data and qualitative perspectives on how programs function (Cheadle et al., 2019; Dannefer et al., 2024; Fullman et al., 2023; Kidder at al., 2024; Rodela et al., 2021; Szelest et al., 2021). The four quasi-experimental studies (Daniels et al., 2024; Duffy et al., 2020; Feinberg et al, 2019; Lukitasari et al., 2021) and single RCT (Joshi et al., 2019) offer evidence of program impact, though the limited number of controlled studies reflects the challenges of conducting rigorous experimental research in community-based settings. The qualitative studies, secondary analysis and single-arm study do not confirm whether observed outcomes are directly attributable to CHW interventions (Chimberengwa & Naidoo, 2019; Glenn et al., 2019; Lalla et al., 2020; Samuel-Hodge et al., 2020). Two evaluation frameworks developed in the U.S. were manually added to inform the discussion on standardized program evaluation (Kidder et al., 2024; Rodela et al., 2021). While these frameworks were not developed specifically for rural Indigenous settings like Guatemala, they are designed to be adaptable to local contexts (Kidder at al., 2024; Rodela et al., 2021).
Through an iterative review of the included articles, recurring themes across the 17 articles were identified. Patterns were identified inductively during the review process rather than through formal thematic analysis. Through this process, four main themes were identified across the literature; three relevant to measures of impact: (1) improved health outcomes; (2) increased, more appropriate, or more cost-effective utilization of healthcare services; and (3) perceived value to the community served, and one theme addressing barriers of CHW program evaluation.

Improved health outcomes

Improved biometrics relating to NCD management, especially blood pressure and HgA1C were primary outcomes. Four articles reported statistically significant improvements, as in one large study in Sonoma County, California which reported that of 1,751 participants, those with blood pressure control increased from 57% in 2014 to 67% in 2016 (P<.001) (Cheadle et al., 2019). Similarly, by the second year of CHW interventions in Harlem, blood pressure control in hypertensive participants improved from 57% to 73.9% (P=0.002) (Feinberg et al., 2019). In four studies, changes in biometrics were not statistically significant, but still trended towards improvement (Dannefer et al., 2024; Fullman et al., 2023). Many articles lacked reporting on statistically significant outcomes, which is consistent with observations from other literature reviews about CHW outcomes in rural US populations (Berini et al.).

Secondary measures of NCD management included healthy behaviors such as self-reported diet and exercise habits (Daniels et al., 2024), medication adherence (Berini et al., 2021; Chimberengwa & Naidoo, 2019; Joshi et al., 2019) and reduced tobacco use (Joshi et al., 2019). Most articles in this review reported at least one improved biometric or healthy behavior after a CHW-based intervention lasting one month (Lukitasari et al., 2021) to several years (Cheadle et al., 2019; Dannefer et al., 2024; Fullman et al., 2023).

Three articles specified temporary, incomplete or inconsistent achievements of health outcome goals, which may have been due to limited time of intervention, limited power of study, or inability to overcome social determinants of health (Daniels et al., 2024; Duffy et al., 2020; Joshi et al., 2019). Overall, there was a lack of data on duration of effect post-intervention.

Improved engagement with healthcare services

Researchers reported on the frequency and type of utilization of healthcare services as a measure of return on investment in CHW programs, such as an increased utilization of primary care by 2.5% as reported by Franz et al. (2020). Szalest et al. (2021) observed that after one year of CHW intervention, the number of participant visits to primary care providers increased by 6% while the number of emergency department visits decreased by 41%; additionally, substance use and mental health related issues were no longer amongst the top reasons for visiting the emergency department, suggesting a shift towards more appropriate utilization of healthcare services.

Engagement with healthcare services was also measured by number of individuals screened for NCDs. One study reported 1,751 individuals screened for hypertension, identifying 441 individuals with high blood pressure who were contacted for follow up (Cheadle et al., 2019). As Samuel-Hodge describes, “making these community-clinical linkages is particularly important in communities … that are medically underserved and negatively impacted by many social factors that drive health” (Samuel-Hodge et al., 2020, p.8). Bringing healthcare services to the previously underserved is a crucial first step towards reducing disparities.

Building health literacy and empowerment

CHWs add value to communities by increasing health-related knowledge and building trust (Chimberengwa & Naidoo, 2019; Lalla et al., 2020; Samuel-Hodge et al., 2020). CHWs provided culturally tailored, holistic support that improved the quality of life for community members in ways not found in other areas of healthcare (Lalla et al., 2020). CHWs fostered self-efficacy amongst community members; one study reported increased Patient Activation Measures (PAM) reflecting increased engagement with their own health after CHW interventions (Glenn et al., 2019).

This theme was especially prominent in the few studies that referred to frameworks for CHW interventions. Dannefer et al. (2024), which referenced the Consolidated Framework for Implementation Research and the CHW Common Indicators, reported high participant satisfaction and development of self-advocacy skills. Chimberengwa & Naidoo (2019) referred to WHO’s Chronic Care Model in their study design and found that participants experienced increasing trust in their village health workers as they also increased their health literacy.

Barriers to impact evaluation

Studies highlighted ways in which impact evaluation may be confounded by other factors in underserved communities. Outcomes such as medication compliance may be positively impacted by CHW interventions, but still subject to larger forces such as medication availability, cost, insurance coverage, adverse effects or cultural beliefs, as described by Duffy et al.(2020) and Joshi et al. (2019). Other studies report insufficient funding to allow programs enough time to create a measurable impact in the communities served; it took two years for participants in the Harlem program to show significant improvements in blood pressure, while the program only had funding for a comparison group and cash incentives for participants for the first year (Feinberg et al., 2019).

Table 1. Table of Evidence

Table 2. Themes identified among selected studies

Discussion

This integrative review examined evidence-based methods for evaluating CHW programs to determine program impact, with a specific focus on its potential application in rural and low-resourced settings like rural Guatemala. Overall, the evidence base supporting CHW programs for NCD management is growing but remains limited by a lack of high-level experimental studies. The findings of this review should be interpreted as promising rather than definitive, as future research using more rigorous designs and standardized evaluation methods is needed to strengthen the evidence for these programs. This lack of standardization aligns with prior literature on CHW; as noted by a systematic review of 122 reviews about CHW programs published from 2005 to 2017, CHW programs are extraordinarily specific: “there are no standard blueprints that can be used to design and implement a CHW program” (Scott et al., 2018, p.13). Nonetheless, this review found several commonly used measures of impact: improved biometrics and/or healthy behaviors over time, improved engagement with healthcare services, and building health literacy and empowerment in communities.

Gaps in literature

Notably absent from the studies found through database searches were preexisting frameworks created for program evaluations, except for two studies (Chimberengwa & Naidoo, 2019; Dannefer et al., 2024) that referenced the Chronic Care Model (RHIhub, n.d.), the Consolidated Framework for Implementation Research (CFIR) (Damschroder et al., 2022) and the CHW Common Indicators (Rodela et al., 2021). Of these frameworks, the first two are nonspecific to CHW programs or program evaluation, and the third is highly complex and not designed for low-income settings. Another framework not mentioned in the articles found through database searching, but specific to public health program evaluation for underserved populations, is the updated CDC Program Evaluation Framework (Kidder et al., 2024). This framework is iterative and highly adaptable to CHW programs in rural, low-resource settings, but was not referenced in any of the literature reviewed. All these frameworks were elaborated through expert consensus; all seek to discover and address the interdependent barriers and facilitators of equitable, effective healthcare in the context of real-world settings. Evaluating CHW programs is complex due to their diverse designs, varying scales and durations, and the frequent integration within larger programs. The CHW Common Indicators and the CDC Program Evaluation Framework both acknowledge this complex reality and are designed to be multifaceted and adaptable to different settings and phases of implementation. One reason why frameworks have not been utilized in most studies found in this integrative review is that these two most relevant frameworks were not published until after 2020, therefore have not had much opportunity to be used in research yet.

Another identified gap is the measurement of monetary return on investment, which was largely absent from the studies included in this integrative review except for two articles that calculated an estimated cost-effectiveness (Dannefer et al., (2024); Szalest et al., 2021). It should be noted that the search strategy for this review did not include the phrase “return on investment”, and so relevant articles may not have been captured. Nonetheless, healthcare utilization costs are too location-specific to be considered a universal measure of impact (Rodela et al., 2021). Additionally, calculations of cost-effectiveness may not be easily calculated for CHW programs that connect healthcare services to underserved populations that otherwise do not receive any care and therefore would not have a baseline or point of comparison. Another significant limitation found in this integrative review is a lack of data on long-term outcomes; this may be related to underfunding for CHW research and evaluation, or due to more recent studies of shorter duration (Rodela et al., 2021).

Opportunities for greater inclusion of CHWs

Like the communities they serve, CHWs and their programs often face challenges and underinvestment from larger health systems. The CHW Common Indicators Project report elucidates that CHWs are “predominantly members of marginalized communities where health inequities are greatest” and as such, “experience the same oppression and denial of power experienced by their broader communities” (Rodela et al., 2021). Programs are often developed without deeper involvement of the CHWs themselves, who are most knowledgeable about the work and their communities. This leads to process and outcome measures that don’t adequately capture CHW achievements, making it challenging to report outcomes according to standardized indicators, which can in turn lead to chronic underfunding, making it harder to reach goals (Rodela et al., 2021). The CDC has come to similar conclusions in their newly updated Program Evaluation Framework, recommending collaborative, intentional and equitable approaches to program evaluation, ensuring all interest holder voices are heard, to better understand how to address drivers of health inequities (Kidder et al., 2024).

Qualitative research provides meaningful evidence for the ability of CHWs to make a positive impact in their communities through building health literacy, self-advocacy and a sense of support not found elsewhere. One area of opportunity for researchers and program evaluators is to continue to seek and incorporate from CHWs themselves to inform ongoing or forthcoming CHW efforts. This is essential to the third core tenet of the CDC Program Evaluation Framework, to “learn from and use insights” (Kidder et al., 2024, p.1). One multiyear program included in this review, the Harlem Health Advocacy Partners, can serve as an example of how a program can adapt over time to better understand and meet their community’s needs, in part by including community members and CHWs at every step of the process.

Finally, understanding the true impact of CHW programs on NCDs requires longitudinal research that follows participants over the course of many years. This type of long-term evaluation is often not realistic due to limited funding, staff turnover, and time constraints that many community-based programs and researchers face. As a result, the available literature may only capture short-term outcomes, which may not fully reflect the lasting benefits or challenges of these programs. These findings underscore the importance of longitudinal program evaluation strategies to examine program benefits and increase program sustainability.

Limitations

Several limitations were identified in this review. While the PRISMA guidelines help guide the reporting process, it was a model originally designed for systematic review of interventions rather than integrative reviews. As a result, it may not adequately capture the decision-making process involved in selecting qualitative or mixed-method studies. Additionally, while following the PRISMA reporting guidelines can ensure transparency, it does guarantee methodological quality of the review itself. The JHNEBP appraisal tool provides a helpful structure to evaluate evidence quality, but some subjectivity in rating the studies is unavoidable. Notably, only one of the included studies were classified as Level I evidence (i.e., randomized controlled trials or systematic reviews), limiting the overall strength of the evidence base. This gap may reflect the nature of CHW projects, which are often carried out in in real-world community settings rather than controlled research environments where experimental designs can be applied. Synthesizing literature particularly regarding rural and indigenous communities may also pose challenges related to research as indigenous populations and low-resource settings have historically been underrepresented in published health literature; many programs serving these communities are documented in tribal reports, government publications, or other grey literature sources that are not captured through traditional database searches and were omitted in this review. Additionally, it is important to note that the relevant frameworks described in this review represent expert consensus rather than empirical research, which is appropriate for guiding evaluation practice but does not provide direct evidence of program effectiveness. Lastly, because narrative synthesis was used rather than statistical methods, author bias may also influence how themes are identified and organized throughout this review. Despite these limitations, this review offers valuable insights that can help guide future research and practice for CHW programs serving rural indigenous populations.

Implications for Policy and Practice

The absence of standardized evaluation frameworks for CHW-led programs in NCD management creates a significant gap in ensuring consistent outcome measurement and accountability, particularly in low-resource settings like rural Guatemala. Adopting evidence-based program evaluation models could address this gap in several important ways. Robust evaluation methods are essential not only for assessing functionality, but also for guiding improvements, demonstrating impact to stakeholders, and informing how limited resources are allocated. From a policy standpoint, clear evaluation frameworks can support the formal recognition of CHWs within Guatemala's health system, help justify sustainable funding and facilitate integration into national health strategies.

On a practical level, standardized evaluation allows programs to stay true to evidence-based practices while still adapting to the unique cultural, geographic, and social contexts of rural Indigenous communities. This approach also empowers CHWs and the communities they serve by validating their contributions and ensuring that interventions are responsive to local needs and priorities. Without consistent evaluation utilizing both quantitative and qualitative data, the ability to scale and sustain CHW programs remains limited, which undermines their potential to reduce NCD-related health disparities in underserved populations. Investing in comprehensive yet adaptable evaluation frameworks is therefore vital for strengthening the role of CHWs in NCD prevention and care in rural Guatemala and similar settings.

Conclusion

The findings from this integrative review identify strategies for impact evaluation of CHW programs for NCD management while also highlighting the lack of standardized evaluation frameworks. CHWs typically come from the communities they serve, which are often marginalized and underrepresented in research, and programs addressing these inequities are similarly underserved in the literature (Rodela et al., 2021). This is beginning to change, as seen with the CDC’s recent updates to the Program Evaluation Framework emphasizing equity and inclusion (Kidder et al., 2024). The 2030 UN Sustainability Goals focusing on reduction of morbidity and mortality from NCDs will continue to highlight a need for solutions. In rural Guatemala, where indigenous communities face significant barriers to healthcare access and carry a growing NCD burden, CHW programs offer a culturally grounded approach to reaching populations that the formal health system often overlooks. However, without consistent evaluation methods, it is difficult to demonstrate the true value of these programs to policymakers and funders who control resources. This review contributes to a growing body of evidence supporting community-based interventions as an important strategy to meet the needs of the world’s most vulnerable populations like those in rural Guatemala. Moving forward, research using standardized indicators and culturally appropriate evaluation frameworks will be key to building the support needed to sustain and expand these valuable programs in Guatemala and in similar settings.

Funding

The authors did not receive funding for this work and have no disclosures related to this publication.

Conflict of interest

The authors declare they have no conflicts of interest.

Generative AI statement

I, Scranton, Hannah, hereby confirm that this manuscript is the result of my own (and my coauthor’s) independent scholarly work, and that all material from the work of others has been fully acknowledged. No content generated by AI technologies has been used in this assessment.

Acknowledgement

The authors acknowledge Nursing Heart, Inc and their community health worker program in Guatemala for inspiring this review.

References

Appendix

Appendix 1. Search Strategies