|Home | About | Journals | Submit | Contact Us | Français|
Community-based health workers (CBHWs) are frontline public health workers who are trusted members of the community they serve. Recently, considerable attention has been drawn to CBHWs in promoting healthy behaviors and health outcomes among vulnerable populations who often face health inequities.
This systematic review synthesized evidence concerning the types of CBHW interventions, the qualification and characteristics of CBHWs, and patient outcomes and cost effectiveness of such interventions in vulnerable populations with chronic, non-communicable conditions.
Four electronic database searches, including PubMed, Embase, Cumulative Index to Nursing and Allied Health Literature (CINAHL), and Cochrane, and hand searches of reference collections were undertaken to identify randomized controlled trials published in English before August 2014.
A total of 934 unique citations were screened initially for titles and abstracts. Two reviewers then independently evaluated 166 full-text articles that were passed onto review processes. Sixty-one studies and six companion articles (e.g., cost-effectiveness analysis) met eligibility criteria for inclusion.
Data were extracted by 4 trained research assistants (RA) using a standardized data extraction form developed by the authors. Subsequently, an independent RA reviewed extracted data to check accuracy. Discrepancies were resolved through discussions among the study team members. Each study was evaluated for its quality by two RAs who extracted relevant study information. Inter-rater agreement rates ranged from 61% to 91% (average 86%). Any discrepancies in terms of quality rating were resolved through team discussions.
All but 4 studies were conducted in the U.S. The two most common areas for CBHW interventions were cancer prevention (n=30) and cardiovascular disease risk reduction (n=26). The roles assumed by CBHWs included: health education (n=48), counseling (n=36), navigation assistance (n=21), case management (n=4), social services (n=7), and social support (n=18). Fifty-three studies provided information regarding CBHW training, yet CBHW competency evaluation (n=9) and supervision procedures (n=24) were largely underreported. The length and duration of CBHW training ranged from 4 hours to 240 hours with an average of 41.3 hours (median: 16.5 hours) in 24 studies that reported length of training. Eight studies reported the frequency of supervision, which ranged from weekly to monthly. There was a trend toward improvements in cancer prevention (n=21) and cardiovascular risk reduction (n=16). Eight articles documented cost effective analysis and found that integrating CBHWs into the healthcare delivery system was associated with cost-effective and sustainable care.
CBHW interventions appear to be effective when compared to alternatives and also cost-effective for certain health conditions particularly when partnering with low-income, underserved, and racial and ethnic minority communities. Future research is warranted to fully incorporate CBHWs into the health care system to promote non-communicable health outcomes among vulnerable populations.
We conducted a review of the studies in which the effect of community-based health workers—public health workers who are trusted members of the community they serve—was tested for chronic disease management and care among people who are at-risk for health disparities. We found 67 relevant studies. Most studies focused on preventing cancer and cardiovascular diseases. In these studies, community-based health workers carried out several tasks. The tasks included providing health education and counseling, helping patients navigate the healthcare system, managing care, and providing social services and support. How community-based health workers were trained or their ability to carry out certain tasks verified, or who supervised their work were mostly under-reported. Compared to no intervention or other alternatives, partnering with community-based health workers tended to result in increasing screening tests for breast, cervical, and colorectal cancers, decreasing blood pressure, blood glucose, and weight, and promoting exercise in study samples. In several studies reporting costs, community-based health workers tended to save costs as well. Our findings support the benefits of working with community-based health workers in promoting health among people who are at-risk for health disparities.
Vulnerable populations—defined as those “capable of being hurt” or “susceptible to injury or disease” 1—refer to a wide range of groups including the economically disadvantaged, the uninsured, racial and ethnic minorities, the elderly and children, or those who encounter barriers to accessing healthcare. 2 Their health problems often intersect with social factors such as housing, poverty, absence of a usual source of care, and inadequate education. 3
The needs of vulnerable populations are multifold and require extensive medical and nonmedical outreach and services. However, current health care financing and service delivery arrangements do not always address the complexity and breadth of needs. For example, since the advent of the Patient Protection and Affordable Care Act (PPACA), the rate of the uninsured dropped initially (nearly 4%) in early 2014 but there has been no substantial change in this statistic from the second to the third quarter of 2014. 4 In addition, the proportion of U.S. adults who delay medical treatment for serious conditions in the past year has risen since 2013 (from 19% to 22%).5 In 2013, more than 41 million U.S. individuals under 65 years of age did not have health insurance, due in large part to the fact that they could not afford coverage. 6 Compared with their insured counterparts, the uninsured were less likely to receive timely preventive care within the last year (33% vs. 67% of the nonelderly with Medicaid and 74% of nonelderly individuals with employer-based insurance) or have access to appropriate follow-up care after abnormal screening results.6 More than half (58%) of the uninsured with a chronic illness reported that they did not buy a prescription drug because of cost, compared with 39% of those with publicly funded insurance and 34% of those with private insurance. 7 Other vulnerable populations such as the elderly or individuals with disabilities also have high levels of unmet health care needs. For example, the State of Aging and Health in America 2013 report revealed that only about 51% of male and 53% of female older adults (65+ years) were up-to-date on certain preventive care such as flu vaccination or colorectal cancer screening.8 Similarly, inidividuals with disabilities had more than 1.5 times higher odds of delaying care due to costs compared to those without. 9
Novel approaches to address the risks and multiple needs of vulnerable populations is an important public health imperative.10,11An emerging approach is to work with community-based health workers (CBHWs). CBHWs are indigenous public health workers who not only share the same ethnicity, language, or geographic community of the patients they serve but also share the life experiences with target populations and communities. 12 Hence, they are uniquely aware of the ethnic, linguistic, socioeconomic, cultural, and experiential factors that may influence that community's use of healthcare services. 13 With their unique ability to provide ‘bridges’ between the community and healthcare services, CBHWs play a role which could address health inequities: culturally appropriate health education, individual and community capacity building, advocacy, and informal counseling and social support in diverse settings (e.g., community-based organizations, community clinics, or primary and emergency care centers).14
A number of systematic reviews were published with regard to CBHW interventions.15-21 Previous systematic reviews found that CBHW interventions are effective in promoting a wide range of healthy behaviors, such as breast cancer screening15; self-management of diabetes16-18, hypertension19, and asthma20; and medication adherence among patients with HIV/AIDS. 21 Only a few reviews highlighted the additional emphasis on the roles and training of CBHWs, however. 16-18 Furthermore, the field is rapidly evolving with greater attention to the synergistic effects of CBHWs as part of patient-centered care teams. A comprehensive systematic review on CBHW interventions to control non-communicable diseases including diabetes among vulnerable populations is needed, in order to develop a better understanding of integrating CBHWs into the delivery of care to vulnerable populations.
The purpose of this article is to provide a critical review of the evidence on CBHW interventions. Specifically, we examined the types of interventions in which CBHWs were employed, the qualification and characteristics of CBHWs, and the patient outcomes and cost effectiveness of such interventions in vulnerable populations with non-communicable chronic conditions. We also considered the integration of CBHWs into the mainstream healthcare work force for both the prevention and management of non-communicable chronic diseases that overburden vulnerable populations. Our review systematically extends the previous efforts by providing an understanding of: 1) how CBHWs are trained prior to the delivery of an intervention; 2) how CBHWs implement an assigned intervention; 3) how CBHW interventions achieve desired effects; and 4) how CBHWs are integrated into the current healthcare system.
The authors searched four electronic databases, including PubMed, Embase, Cumulative Index to Nursing and Allied Health Literature (CINAHL), and Cochrane, and conducted hand searches of reference collections to search for potential studies. Following consultation with a health science librarian, the authors used a combination of keywords that contained MeSH terms: “vulnerable populations”, “community health worker”, and “randomized controlled trials.” More detailed information about search terms is described in Appendix 1. The searches were restricted to articles published in peer reviewed journals in English before August 2014 (for studies focused on individuals with diabetes since 2011). The hand searches involved review of reference lists from articles obtained from the four electronic databases.
Figure 1 summarizes the results of the literature search. Initially, 922 studies were retrieved from four electronic databases after 575 duplicates were discarded. Twelve additional studies were obtained from hand searches of reference collections. Two reviewers independently conducted an initial screening of titles and abstracts with relevance to non-communicable chronic diseases. After screening the initial titles and abstracts, 397 abstracts were passed onto a second review process to exclude (1) studies focused on children, (2) non-data based articles (e.g., editorial, commentary), and (3) studies focused on non-vulnerable populations. Of 397 abstracts, 166 abstracts were included in a full-text review. Two reviewers (KK and BA) independently evaluated full-text articles to determine whether studies met the following inclusion criteria: (1) randomized controlled trials published in English in peer-reviewed journal; (2) studies testing CBHW-led interventions; (3) studies focused on adults; and (4) studies focused on chronic conditions. We excluded 89 articles for the following reasons: (1) full-texts were unavailable (conference abstracts; n = 16), (2) were non-randomized controlled trials (n=55), (3) studies did not include a CBHW-led intervention (n=6), (4) studies focused on diabetes that were published before 2011, given their inclusion in a recent systematic review on CBHWs for individuals with diabetes (n=8), (5) studies that tested the effectiveness of an intervention to change behaviors among CBHWs (n=1), and (6) articles that reported preliminary or intermittent findings and reported the long-term findings of other articles (n=3). Articles that included a cost-effectiveness analysis only were merged into the main outcome studies. Discrepancies regarding the extracted data (see data extraction selection below) between two reviewers were reconciled based on a series of team discussion. A total of 67 articles met criteria for inclusion. Figure 1 provides a detailed outline of the article selection process.
Relevant data were extracted by 4 trained research assistants (RA) using a standardized data extraction form developed by the authors. The following data were extracted from the selected studies: author, year, country, randomization, intervention unit, setting, sample (% non-Caucasian), the method of outcome ascertainment, time to outcome measure, theory use, CBHW selection criteria, type of training, training frequency, training intensity, duration of training, participant satisfaction, delivery approach, control group, types CBHW of intervention, measurability of the CBHWs' effect, intervention dose, intervention intensity, intervention duration, number of subjects in the study groups, mean age and sex proportion of the study sample, proportion of target condition or behavior for the treatment and control groups at baseline and follow-up, fidelity, and study quality. Subsequently, an independent RA reviewed extracted data to check accuracy. Discrepancies were resolved through discussions among all RAs and authors.
Each study was evaluated for its quality, based on published quality rating scales (see Table 1). 22-25 Specifically, we used the quality rating scales published by Jadad et al. 22 and Sackett and Haynes. 25 Our quality rating scale also incorporated additional evaluation items addressing intervention setting and outcome assessment methods used in published systematic reviews. 23 The total quality rating scale score ranged from 0 to 12 with 0 being the lowest quality to 12 indicating the highest quality. Based on the possible range of scores, studies with quality ratings of 0-4, 5-8, and 9+ were categorized as low, medium, and high quality studies, respectively. Two RAs who extracted relevant study information rated each study for its quality independently. Inter-rater agreement statistics using percent agreement ranged from 61% to 91% (average 86%). Any discrepancies in terms of quality rating were resolved through team discussions.
Table 2 summarizes the key characteristics of studies included in this review. There were 67 publications including 61 studies26-86 with 6 companion articles. 87-92 The companion articles presented cost analysis of the main studies. All but four of the 67 studies were conducted in the U.S.; two studies were conducted in India, 47,83 one68 in Pakistan, and one86 in Taiwan. Ethnic minorities were the focus of all but four studies, which included predominantly low-income non-Hispanic white participants (61% to 95%) at risk for experiencing inequality in healthcare access (e.g., Appalachians). 37,48,49 Across the 67 articles, sample sizes varied widely from 2562 to 167,915. 47 Participants generally ranged in age from 32 years42 to 71 years82 and 28%71 to 100% of participants were women.
Of 67 articles, thirty studies involved CBHWs for cancer prevention for specific cancer types (cervical, 28,32,33,35-39,41-44,49-52 breast, 26,27,31,33,36,37,39-41,46,48,50,53-55 colorectal, 29,30,34,36,45 and oral47). Twenty six studies focused on cardiovascular disease (CVD), 56-59,62,64,67,73,80, and key risk factors such as diabetes, 63,69,70,76-79,81 and hypertension. 61,65,68,71,72,74,75 Two studies tested an intervention on other CVD-related topics including promotion of physical activity in women who have 1+ CVD risk factors60 and chronic disease screening (e.g., annual exam). 66 Three studies tested CBHW interventions on cognitive functioning, 82 and mental disorders (depression, 84 and schizophrenia83). Finally, two studies involved CBHWs for asthma85 control and medication safety among rural elders with chronic diseases. 86
Forty-six (75%) studies employed individual randomization and 14 (23%) studies used cluster randomization. 29,33,34,36,39,47,48,54,55,58,62,63,68,82 The two most common types of comparison groups were less intensive intervention (n=17 [28%]),27,35,36,38,40,46,53,57,58,62,65-67,73,74,78,85 and usual care (n=16 [26%])31,32,37,41,44,47,51,55,60,64,71,72,76,77,83,86 or enhanced usual care (n=4 [7%])56,59,69,84 followed by wait-list control (n=7 [11%]),42,45,48-50,63,77 and attention control (n=3 [5%]).39,52,82 Eight studies involved more than one comparison group (e.g., usual care and minimal intervention). 26,28-30,61,68,75,80 Study sites commonly involved participants' homes (n=31 [51%]), community health clinics (n=15 [25%]), community-based organizations (n=11 [18%],35,36,38,40,57,59,67,69,73,77,82 and faith-based organizations (n=4 [7%]).34,36,42,62 Some studies relied solely on telephone contact for CBHW interventions.27,37,53,74
Thirty-nine studies fell under the high quality category (i.e., quality scores of 9 or higher with a maximum possible score of 12) (see Table 1). Most studies clearly described the research questions, study design, sample characteristics, sample inclusion and exclusion criteria, study setting, study outcomes, and data collection timepoints. None of the studies were considered to be low quality. However, there were several notable methodological limitations. For example, only about half of the studies (n=34) discussed using a theoretical framework to develop the intervention or from which to select study outcomes. 26-29,31-35,39,41-44,46,49,53-55,62-65,73-80,84,85 In addition, less than one third of the selected studies discussed conducting a power analysis a priori (n=21) and less than half of studies (n=30) clearly described how they randomized study participants. Similarly, less than one in five studies discussed any type of blinding (i.e., the outcome assessor was aware of the status of the participant's group assignment) (n=13).
In the context of cancer screening, half of the studies measured primary cancer screening behavior using self-report only29,34-36,38-41,48-50,53-55 as opposed to objective chart review. Of the seven cancer screening studies which verified self-reported screening behavior with chart review, discrepancies were noted in all studies (sensitivities from 59%52 to 83%33,42 and specificities from 81%33 to 100%28). Finally, less than one third of the studies (n=21) described how they maintained and monitored CBHW intervention fidelity. 28,32-34,37,42-44,46,47,49,51-56,60,67,70,78
CBHWs delivered a wide range of interventions including education, counseling, navigation assistance, case management, social services, and social support. These interventions were often delivered in addition to traditional outreach responsibilities of CBHWs which included participant recruitment and data collection. CBHW interventions were performed in collaboration with health care professionals. CBHWs were supervised by research staff, clinic staff, and study psychologisits (Table 2). Fifty out of 61 interventions involved CBHWs alone or the effect of the work of CBHWs was tested separately, whereas in 11 studies, CBHWs patnered with other professionals such as primary care providers, 26,27,48,56,84,93 nurse case managers, 63 dietitians, 69,73 and social workers 85 to deliver the study intervention.
CBHWs fulfilled the role of an educator in 48 articles. CBHWs provided education via individual sessions or group education sessions34-36,38-42,48,52,57,58,62,67,69,73,77,79,81of varying sizes, from 338,41 to 2057participants (median: 3.5 to 11), with a duration of intervention up to 30 months72 and each session lasting 5-10 minutes75to 3 hours (average: 93 minutes). 42,52,67 Education sessions took place at participating clinics, community locations, or participant's home or work. Varying educational materials were used, including standardized powerpoint presentations, 34 videos, 28,50,52 print education materials, 50,52 and monthly newsletters. 61 In addition, role playing was also adopted for interactive education sessions. 62 Some studies provided different types of CBHW-led education based on the individual's risk profile after baseline assessment. In a study delivering a nurse-CBHW team intervention to support diabetes self-management in American Samoa, participants assigned to a higher risk group attended weekly group sessions while participants assigned to moderate or lower risk group were seen individually by CBHWs monthly or at a lower frequency. 63 Only one study reported initial testing and validation of the educational materials. 28 In a study promoting cervical cancer screening among Mexican American women, Byrd et al. 28 validated the educational materials and lesson plans at two half-day workshops with bilingual/bicultural CBHWs who had experiences working with Mexican American women. CBHWs reported that the materials were easy to use and successfully demonstrated their ability to use lesson plans. 28
In 36 articles, CBHWs delivered counseling sessions to address barriers in adopting target behaviors and to reinforce benefits of behavior change. 26,27,30-32,37,43-46,49,51-55,57-64,66,67,72,75-77,79,81,83,85,86 CBHWs communicated with participants via telephone calls, home visits, or regular meetings to assess and problem-solve personal and environmental barriers throughout the intervention. In one study, 36 CBHWs provided theory-based scripted messages for each barrier, including personal belief, fear, healthcare provider, personal need, and management barriers. 46 In another study, 65 CBHWs delivered weekly 5 to 10 minutes of counseling to reinforce patient lifestyle, medication-taking, and appointment-keeping behaviors.75
In 21 articles, CBHWs provided navigation assistance for their study participants in obtaining preventive care services and managing chronic diseases. 27,31,33,37,38,40,41,43,45,46,49,51,62,63,70-72,77,79,81 As navigators, CBHWs provided information on how to access medical services and helped with scheduling appointments at health centers. Information on access to medical services included availability of low-cost or free medical services in the community, local providers, and health insurance. In addition to appointment scheduling, CBHWs facilitated participants' attendance at scheduled health services by arranging transportation and accompanying participants to appointments. In a study addressing cervical cancer screening among Chinese women, CBHWs provided transportation assistance through taxicabs or bus passes and medical interpreter services during clinic visits for Pap testing.51
CBHWs were involved in case management in 4 studies by planning and coordinating appropriate healthcare services. 77 Studies in which CBHWs provided case management services usually addressed chronic conditions often involving care from multiple health professionals. In a study delivering a diabetes education and management program for uninsured Mexican Americans, 77 CBHWs facilitated immediate physician contact to address acute problems, assisted with pharmacy refills, and arranged specialty visits, such as dental care and dilated retinal exam. These CBHWs were state-certified health workers and they delivered management services in the setting of an urban community health services clinic.
In 7 articles, CBHWs assisted participants in assessing social services in addition to medical services. 56,65,71,72,76,83,84 In these studies, investigators attempted to address systematic barriers preventing study participants from adopting target behaviors by connecting them to existing social services. The social services provided included referrals to community transportation, 71 child care, 71 housing, 65,76,84 legal benefits, 83 and employment opportunities.65,83,84
Eighteen studies assigned CBHWs to provide social support to promote targeted health behaviors. 27,29,31,36,48,49,51,52,55,58,60,63,72,80,81,86 CBHWs directly provided support for behavior change by encouraging the study participants through multiple conversations and offering emotional support. Social support was also offered indirectly by educating family and friends on how to be supportive. For example, in a study delivering an intervention to decrease blood pressure in an urban African American population, 72 CBHWs taught family members or friends how to provide daily support to the patient, and also assist with appointment keeping or with behaviors related to blood pressure control.
The effects of CBHW interventions reviewed are summarized in Table 2. The findings presented in this section highlight the effects of CBHWs as an intervention component as long as the effect was measured separately. Due to the heterogeneity of settings, sample characteristics, and types of interventions, it was not possible to conduct a quantitative meta-analysis. Overall, most studies reviewed reported positive outcomes for the targeted health behavior. Eight publications including 6 companion articles also demonstrated that the use of trained, culturally competent CBHWs resulted in cost savings.36,43,87-92
Of the 30 studies which tested the effect of a CBHW-led intervention on cancer control, 21 studies (70%) found improvements in cancer screening behaviors. 26,28,30-33,35,37,38,40-51 Positive changes in mammogram uptake (6% to 33% increase) 26,31,37,40,41,43,46,48,50 were noted in nine of the sixteen studies focusing on breast cancer. The trial with the largest increase in mammogram screening (33%) employed a multi-faceted intervention designed for African American women that included 4 monthly CBHW-led culturally-tailored counseling sessions and mailing of a postcard message tailored to the participant's barriers . 46 Similarly, significant improvements in Pap tests, ranging from 7% to 29%28,32,35,37,38,42,49,50,52, were reported in nine of the sixteen studies targeting cervical cancer. Of the three studies focused on colorectal cancer, only one yielded a significant increase in colorectal cancer screening in the CBHW-led education group compared with usual care (27% vs. 12%, p<0.001).45
The studies without significant changes in mammogram, 27,36,53,55 Pap test,27,30,32,34 or colorectal cancer screening21,27 tended to compare one type of CBHW intervention to another (e.g., education vs social support group)27 or to a less intensive intervention (as opposed to no intervention). 27,53 In addition, some of these studies included a high proportion of participants who were up-to-date for screening41 or had significantly different demographic characteristics between the two groups at baseline. 55 One study44 reported significant improvement in self-reported Pap test use among Appalachian women (N=286) (71% vs. 54%, p=0.008); however, the result ceased to be significant when using chart review to ascertain the outcome (51% vs. 42%, p=0.135). Holt et al. 34 found a significant negative effect of spiritually-based CBHW-led intervention on fecal occult blood testing among African Americans compared to the non-spiritual group (2%↓ vs. 9%↑, respectively, p=0.03).
Sixteen studies (62%) included in the review found a significant effect of CBHW intervention on CVD risk reduction. 56,58-60,62-65,68-72,77-79 Of the nine studies that tested the effect of CBHW-led intervention on global CVD prevention, five (56%) studies found significantly greater improvements in lipid profile (total cholesterol, LDL, HDL, or triglycerides), 56,59,62 blood pressure, 56,58,62,64 HbA1C, 56 and global CVD risk59 for the CBHW intervention group as compared to the comparison group. Mixed or non significant results were noted in the three remaining studies, which might have been due to a small sample size (48 to 61), 67,73 low statistical power, 80 low follow-up rates (67% to 73%),67,80 or variability in fidelity of intervention implementation. 80
Of the two studies that focused on other CVD-related topics, 60,66 only one study60 found a significant improvement in self-reported moderate (71% to 84%, p<0.001) and vigorous (13% to 33%, p<0.001) physical activity from baseline to 6-month follow-up in the intervention group. No significant increase was noted in the comparison group.
Of the eight studies63,69,70,76-79,81 that exclusively focused on HbA1C or fasting glucose as a primary outcome, all but two76,81 found significant improvements in diabetes control. Tang et al. 81 compared the effect of peer leaders (bilingual residents in the target community with diabetes and aged 21+ years) vs. CBHWs on diabetes management. In the study, 81 the peer leader group had a significant reduction in HbA1C at 18-month follow-up (-0.6% from baseline, p=0.009). In contrast, the CBHW group failed to maintain a HbA1C reduction (-0.3% from baseline, p=0.234).
Eight studies examined CBHW interventions for better blood pressure control. Significant improvements in blood pressure control were seen in four studies. 65,68,71,72 Of the four studies that found non-significant results, 57,61,74,75 two lacked statistical power. 57,61
Three studies involved CBHWs to address issues related to cognitive functioning and mental disorders, such as depression and schizophrenia; 82-84 the study results were mixed. Using data collected from a cluster randomized trial94 designed to test the effect of a weight-loss intervention for obese older adults (60+ years) using cognitive training as an attention control, Beck et al. 82 compared a cognitive training intervention to a weight-loss intervention. Participants in the intervention group had about 3 times higher odds of achieving better cognitive functioning compared with those in the attention control (weight-loss) group (OR, 2.7; 95% CI, 1.3-5.6, p=.011). 82 Two studies that focused on mental disorders yielded partially significant83 or non-significant findings. 84 Chatterjee et al. 83 found a a significant decrease in disability from schizophrenia (p=0.01) but not in symptom severity. In the study84 that tested a CBHW intervention on depression, the authors argued that non-significant findings might have been associated with fidelity issues, instead of an ineffective interevention.
One study tested the effect of CBHW intervention on asthma control and found that the intervention was effective in promoting self-efficacy; however, there was no significant difference in clinical outcomes (e.g., symptomatic days and night over the past 14 days) between groups. 85 However, the authors reported that this study was underpowered to detect self-management in asthma control and clinical outcomes. In a study86 that tested the effect of a volunteer coaching on medication safety in community-dwelling elders with 2 or more chronic illnesses, the volunteer coaching program was effective in promoting medication safety knowledge as well as 3 (out of 6) medication safety behaviors, compared with the usual care.
Eight out of 61 studies (13%) included cost analyses. Of the 8 articles with cost analyses, three studies focused on diabetes control, 88 high blood pressure control, 91 and cardiovascular risk reduction. 87 Lawlor et al. 88 found that a lifestyle intervention delivered by a registered dietitian/CBHW group could be cost effective. In the study, estimated direct medical costs per capita were $850 and $142 for the registered dietitian/CBHW and the registered dietitian groups, respectively. When total costs were calculated, however, it was higher for the registered dietitian only group than for the registered dietitian/CBHW group ($7,596 vs. $6,027, respectively). In a study to lower blood pressure in 12 randomly selected communities in Pakistan, 91 Jafar et al. found that a ‘home health education by CBHWs plus trained general practitioner' intervention was the most cost-effective intervention compared to ‘home health education only’, ‘general practitioner only’, and ‘usual care’ interventions. The combined intervention resulted in an incremental cost effectiveness ratio (ICER) of $23 (95% CI 6%-99%) per mmHg systolic blood pressure reduction compared with the usual care group. 91 In a CVD risk reduction program delivered by NP/CBHW teams in urban community health centers, 87 Allen et al. also reported estimated savings of $157 and $190 per 1% reduction in systolic and diastolic blood pressure, respectively. They also reported an ICER of $149 for 1% reduction in HbA1C and $40 for 1% reduction in low-density lipoprotein cholesterol.
Five studies assessed cost effectiveness of CBHW intervention in the context of cancer screening. A CBHW intervention designed to promote cervical cancer screening among Vietnamese-American women resulted in an ICER of $30,015 per quality-adjusted life year. 89 In a study which tested the effect of a CBHW-facilitated AMIGAS (Ayudando a Las Mujeres con Información, Guía y Amor para su Salud: Helping Women with Information, Guidance, and Love for Their Health in English) program (video and/or flipchart) among non adherent Mexican origin women 21+ years, Lairson et al. 92 reported an ICER of $980 per additional women screened compared to a video-only intervention and wait-list control. Wagner et al. 90 conducted a CBHW outreach program for low-income women with abnormal Pap test results in Alameda, California and obtained an ICER of $959 per follow-up for the intervention when compared with usual care. While Larkey et al. 36 found no difference in cancer screening outcomes between intervention and comparison groups, the cost per participant screened was approximately three times greater in the individually delivered group than in the social support group ($1,716.22 vs. $516.53, respectively). Paskett et al. 43 estimated a cost associated with CBHW intervention to promote mammography screening among low-income, ethnically diverse female patients aged 40 years or older, which equated to a cost saving of $4,986 per each mammogram in the CBHW group.43
The characteristics, training, and roles of CBHWs are summarized in Table 3. Studies widely varied in their approaches to identifying CBHWs. Approaches included identification by community leaders, 29,34,86 use of existing CBHWs in the community, 30,45 use of participating churches49 or a community self-help organization, 50 or community members who demonstrated the positive behavior targeted within the study population. 26
Studies also used different selection criteria. Twenty-five (%) studies used living in the same residential area with the study participants as a CBHW selection criterion. 27,31,35,40,44,47-50,53-55,60,62,64,66,69,70,77-81,84,86 All studies targeting racial/ethnic minorities included bilingual CBHWs. 28,36,45,52,60,64,66,67,77,78,81,84 Three (%) studies selected CBHWs based on their similar background to the study population in terms of marital status, age, socioeconomic status, occupation, or having children. 35,49,52 Some studies selected CBHWs based on their educational level31,45,47,60 or having previous experiences working with the community. 32,36,50,56,66,80,86 Studies addressing women's health-related diseases or recruiting only female participants used sex as one of the selection criteria (n=16). 27,28,31,32,35,37,39,40,42,44,49,52,53,60,64,66 Two studies addressing women's cancer screening specified age selection criteria as 55 years or older37 and 40 years or older. 40
All but 8 studies30,42,44,48,51,61,65,76 reported information regarding training of CBHWs, yet the extent and breadth of such information varied. Occupation of trainers varied from health center staff to study team members including study coordinators and investigators. One study utilized previously trained CBHWs from community-based organizations as trainers. 57 The length and duration of training ranged from 454 to 240 hours81 with an average of 41.3 hours (median: 16.5 hours) in 24 studies that reported length of training. CBHWs with shorter training tended to serve relatively simple roles such as recruitment and education, 38,40,73 whereas CBHWs with longer training tended to take on additional roles including data collection, care management/ coordination, and navigation assistance. Longer training also encompassed both knowledge-based and competency-based contents such as motivational interviewing techniques and computer/Internet skills. 67,79,81,85,95 For example, in Tang et al. study,81 CBHWs received 240 hours of training covering both intervention-specific (e.g., diabetes education and home visit) and general contents (e.g., human subjects and computer skills). Training was delivered via didactics and interactive sessions encompassing relevant health information as well as interviewing and teaching skills. Competency evaluation was only reported in 9 studies. Studies reported that the competency of CBHWs were evaluated after training by using mock educational sessions, 28,34 role playing scenarios, 85 or written assessments. 29,55 CBHWs received continual training after initial training sessions through monthly skill building sessions in three studies.32,53,78
Supervision of CBHWs were largely underreported. Twenty-nine (48%) studies reported details of CBHW supervision for quality control. In these studies, CBHWs were supervised by the study team members including study coordinators (n=13), clinic staff (n=7; e.g., community health directors, nurse case managers), CBHW coordinators (n=2), and study psychologists (n=2). Supervision was imparted by weekly or monthly meetings with the study team members, 27,32,37,43,50,73,82 and/or direct observations by the supervisors. 28,44,53,63 In one study, the fidelity of the intervention was maintained by having CBHWs document details of intervention implementation, including outcome of home visit attempts, types of study materials used, and follow-up phone calls. 52 No studies reported the amount of time spent in supervision, limiting an understanding of the resources (personnel and cost) needed to support CBHWs.
Twenty-three out of 61 studies (38%) reported details of payment to CBHWs. Hourly rates of CBHWs from those reported ranged from $12.11 per hour31 to $22.26 per hour. 89 Two studies reported that the CBHWs were paid $1,500 after recruiting 20-22 participants from their social networks and providing group sessions. 35,40 Funding sources of CBHWs were mostly from study grants. CBHWs were also employed by a managed care organization26 or compensated by a community-based organization. 57 In some studies, CBHWs were entirely volunteer-based without pay. 29,86 The inclusion of CBHWs into the healthcare system as a way to improve the care of vulnerable populations was rarely discussed. However, some studies pointed out that sustaining CBHWs beyond the funding period was made possible through establishment of long-term, non-federal funds (philanthropic and local agency funding) for CBHWs. 57,77 Other studies maintained and expanded CBHW interventions following the cessation of grants by exporting the interventions into other settings such as outpatient clinics, 75 community sites, 77 and rural or other urban communities.48
To our knowledge, this is the first systematic review that provides a critical appraisal of CBHW interventions targeting vulnerable populations with or at risk for non-communicable chronic diseases. Overall, we found that CBHW interventions were effective in promoting CVD risk reduction, cancer screening, and cognitive functioning, although mixed results were also noted by studies. The 2015 Community Preventive Services Task Force report also revealed that CBHW interventions are effective in controlling blood pressure and cholesterol among patients who are at risk for CVD. 96 There was insufficient evidence to support CBHWs in addressing mental disorders. In addition, there was insufficient evidence concerning the cost-effectiveness of CBHW interventions.
There was no consistency in terms of the duration and intensity of CBHW training in the included studies. In fact, more than half of the studies lacked full descriptions of CBHW training and fidelity monitoring; many failed to describe the characteristics of CBHW and criteria for their selection. When CBHWs received rigorous training, patient outcomes related to cancer prevention and cardiovascular risk reduction were significantly improved. For example, CBHWs in Prezio et al. study77 were required to obtain a state-level certificate. In the study, the CBHW intervention group had a significant decrease in HbA1C compared to the usual care group (1.6% vs. 0.9%; p<0.05, respectively). Staten et al. 80 argued that questionable competency levels of CBHWs prior to studies and variability in fidelity of CBHW intervention implementation could be possible explanations for non-significant effects found in some studies. Previous systematic reviews17,19 underscored the importance of required training and competency levels in relation to assigned responsibilities. Limited yet growing research has focused on the degree to which CBHWs can achieve their competency levels to serve successfully as an interventionist in vulnerable populations. There is a strong need for studies to clearly elabote the contents and processes of CBHW training such as competency evaluation and supervision to optimize the use of this approach.
Our findings offer implications for the successful delivery of CBHW interventions as part of patient-centered and community-oriented care teams. As ‘natural helpers’, CBHWs play an essential role in bridging between the healthcare services and the communities they serve. 97 In particular, their natural helper roles97,98 building on ‘trust’, ‘rapport’, and ‘an ability to communicate with the community’ draw much attention to CBHWs as part of patient-centered care teams. Indeed, CBHWs delivered both medical (e.g., culturally tailored health education) and non-medical services (e.g., social support, social services) in the studies included in this review. Nevertheless, their natural helper roles were not easily quantifiable and created difficulty in evaluating the quality of the work of CBHWs. A recent diabetes management study99 developed and used a comprehensive CBHW encounter form which seems to offer a promising avenue for monitoring and evaluating CBHW work in a naturalistic setting. In the study, Lemay et al. 99 argued that CBHWs' daily activities with patients as captured and documented in a standardized encounter form need to be incorporated into a patient's medical record. This information may help estabilish a foundation for proper payment for CBHWs' services.
Cherrington et al. 100 acknowledged that paid CBHWs tend to cover a wider scope of work, be more flexibile in terms of scheduling, and produce the full impact of which they are capable. In countries such as U.S. and some Asian countries such as Pakistan where the included studies were conducted, failure to secure sustainable funding sources for CBHWs appears to be a major barrier to the full integration and maintenance of this model into healthcare delivery systems. A few recent examples present promising avenues for working with CBHWs as a reimbursable, alternative model of care for vulnerable populations. For example, two states in the U.S., Minnesota and Massachusetts, initiated policies to cultivate CBHWs. In response to a lack of a healthcare workforce and an increase in diverse populations, Massachusetts recognized CBHWs as an essential component of the state healthcare workforce; however, many CBHWs in Massachusetts still rely on federally funded or philanthropic programs, rather than a state-wide funding program. 101 Minnesota made diagnosis-related (e.g., patient education), not social service-related, CBHW services reimbursable under Medicaid if the CBHWs completed a 14-credit certificate program and worked under the supervision of Medicaid-approved healthcare professionals such as physicians, advanced practice nurses, dentists, public health nurses, and mental health providers. This was the first state in the U.S. to establish a potentially sustainable funding source to maintain CBHWs.102
Only eight articles (13%) documented cost-benefit analyses associated with the integration of CBHWs into the healthcare delivery system for prevention and management of chronic conditions that most often overburden vulnerable populations. The U.S. PPACA—also called Obamacare or Affordable Care Act—aims to increase the quality and affordability of health insurance, lower the uninsured rate by expanding public and private insurance coverage, and reduce the costs of healthcare for individuals and the government (PPACA Public Law 111-148, 42 USC 280g-11, §399v). The PPACA acknowledges the essential role of CBHWs in improving health behaviors and outcomes by indicating CBHWs as an important part of healthcare teams for the delivery of care, particularly among medically underserved populations and communities. The PPACA emphasizes the need for CBHWs in communities with a high rate of uninsured but eligible individuals with a high percentage of chronic diseases or infant mortality; PPACA calls for more attention to be focused on tailored interventions responsive to multi-faceted, underlying challenges threatening communities. 103 The PPACA presents unprecedented opportunities to include CBHWs as a core component of medical teams, promoting health behaviors and outcomes as a sustainable part of the healthcare system. Clearly, more systematic cost evaluations of collaborating with CBHWs as an alternate care model is warranted to expedite the translation of research into evidence-based guidelines and recommendations for clinical practice in vulnerable populations.
There are methodological issues to be taken into consideration when interpreting the findings in this review. While 39 of 67 studies (58%) were of high quality, many studies lacked methodological rigor, which might have led to false negative results (no effects of CBHW interventions). For example, studies without a priori power analysis failed to find significant effects for CBHWs. 62,66,73 In addition, approximately one in four studies did not report the number and/or reasons for participant withdrawals or drop-outs, and 31 studies used per-protocol analysis instead of intent-to-treat analysis. About half of the studies were conducted without the guidance of a theorical framework, which might have resulted in mixed results in some selected studies in this review. A theoretically grounded CBHW intervention can strengthen the theoretical underpinnings of CBHW practice. 104 In some studies CBHWs took on both traditional outreach and recruitment responsibilities, as well as the delivery of the intervention. 42 This dual role is likely to have led to the disclosure of group allocation, hence, threatening the internal validity of the results. Future studies should address these issues by calculating proper sample size a priori, conducting intent-to-treat analysis, and concealing group assignments.
A few limitations of this review should be noted. First, because many terms are used to describe CBHWs and front-line outreach public health workers, it is possible that we did not extract all relevant articles in the existing literature. However, to avoid this, in addition to hand searches of reference collections, we conducted a systematic electronic search using a comprehsive list of MeSH terms as well as similar keywords, such as lay health advisor or lay health counselor, after a consultation with a trained health science librarian. Nonetheless, given the diversity in the CBHW literature –including grey literature such as research findings outside of academia or reports from organizations—publication bias may exist. The inclusion of grey literature might have offered a more comprehensive understanding of CBHW characteristics and roles. Second, the CBHW workforce was developed to predominantly serve vulnerable populations, though it is possible that some skills can be used for other populations. Thus, our findings may not be applicable to other populations such as mid- or high-income populations. Third, we included only articles written in English; therefore, we limited the generalizability of the findings concerning studies published in non-English languages. Fourth, some studies included CBHWs as one part of a multifaceted intervention and did not test the effectiveness of CBHWs separately. Therefore, caution should be taken when interpreting the effects of CBHWs. Finally, we arbitually categorized studies with quality ratings of 0-4, 5-8, and 9+ as low, medium, and high quality studies, respectively. We considered that each item is an equally weighed factor that constitutes study quality. However, one might argue that certain factors (e.g., how a study is randomized) may contribute to its quality rating more so than others.
Working with CBHWs to deliver important health-related interventions is a growing trend. As CBHWs are typically trusted members of their communities with whom they share the same cultural and linguistic backgrounds and life experiences, they are ideally positioned to provide tailored, culturally responsive interventions. Thus, CBHWs have a unique role in facilitating community health promotion and may be the mechanism by which to establish close ties between healthcare providers and community members. CBHW models also support the movement from a healthcare system that focuses only on “sickness care” to one that is also “prevention-focused.” As Rosenthal et al. 102 pointed out, integration of CBHW models into the healthcare system appears to be an effective strategy for restructuring primary care delivery and, focuse on accessible, continuous, comprehensive, compassionate, and culturally effective care. In conclusion, our review of 67 articles shows that CBHWs can be an effective intervention model that is also cost-effective for certain health conditions (e.g., high blood pressure, diabetes) or behaviors (mammogram and Pap test use) for low-income, underserved, and racial and ethnic minority communities. Our findings support the use of CBHW as an intervention model and suggest as well the need for more rigorous and continued evaluations of this approach for a wide range of conditions and populations.
This study was supported, in part, by a grant from the National Cancer Institute (R01CA129060). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Additional resources were provided by Center for Cardiovascular and Chronic Care and Center for Innovative Care in Aging at the Johns Hopkins University School of Nursing. We would like to express our appreciation to our research assistants, Betsega Awelachew and Judy Liu, for their work in article search and data extraction.
Contributors: K. Kim reviewed records and articles for eligibility, retrieved and coded the relevant articles, and co-wrote the article. J. S. Choi retrieved and coded the relevant articles, and co-wrote the article. E. Choi reviewed records and articles for eligibility and provided critical comments on the article. C. L. Nieman, J. H. Joo, F. R. Lin, and L. N. Gitlin provided substantive contributions, including advice on the coding scheme and directions, and made substantive edits to the article. H. R. Han originated the project, designed the coding scheme, led the writing, along with K. Kim, J. S. Choi, and E. Choi, reviewed studies, and conducted the analysis.
Human Participant Protection Statement: This is a systematic review of published articles. The Institutional Review Board approval was not needed.