The Role of Community Health Advisors’ Cancer History in Implementation and Ecacy of a Cancer Control Intervention

Purpose: Community health advisors (CHAs) play a key role in promoting health in medically underserved communities, including in addressing cancer disparities. There is a need to expand the research on what criteria makes for an effective CHA. We examined the relationship between CHAs’ personal and family history of cancer, and implementation and ecacy outcomes in a cancer control intervention trial. Methods: Twenty-eight trained CHAs implemented a series of three cancer educational group workshops for N=375 workshop participants across 14 churches. Implementation was operationalized as participant attendance at the educational workshops and ecacy as workshop participants’ cancer knowledge scores at 12-month follow-up, controlling for baseline scores. Results: CHA’s personal history of cancer was not signicantly associated with implementation, nor knowledge outcomes. However, CHAs with family history of cancer had signicantly greater participant attendance at the workshops than CHAs without family history of cancer (p=.03). In addition, there was a signicant, positive association between male CHAs’ family history of cancer and male workshop participants’ prostate cancer knowledge scores at 12 months (estimated beta coecient=0.49, p<.01) after adjusting for workshop participant baseline knowledge scores, the CHAs’ competence score, and the CHAs’ education levels. Conclusions: Findings suggest CHAs with family history of cancer may be particularly suitable for cancer peer education, though further research is needed to conrm this and identify other factors conducive to CHAs’ success.

ompetence, and demographic characteristics such as age, education, and gender [6,7].Often, the CHA is an individual indigenous to the community, who is already recognized as a community leader in some capacity, and who has racial/ethnic concordance with the intervention's target population [6,7].Albeit limited, there is also evidence that those with health-related experience are more com etent as CHAs [6,7].

Studies have drawn different conclusions on whether CHAs' sociodemographic characteristics are important factors in predicting CHAs' performance [8].A study in Uganda found that CHAs' age, sex, and education had no effect on their ability to classify and treat pneumonia [9].A cancer control intervention reported that CHA sociodemographic characteristics did not predict the CHAs' success in recruiting intervention participants [10].On the other hand, a study in Kenya found that CHAs' sociodemographic characteristics such as age, gender, education level, and experience were signi cantly associated with CHA performance and program e cacy in a pregnancy home visit program [11].Speci cally, it was found that CHAs aged 40-50 years performed more effectively on most performance metrics than their younger or older counterparts.CHAs' age was also signi cantly, positively associated with program e cacy (i.e.client enablement).Males were also more effective in certain roles such as record-keeping, while females were more effective in other roles such as counseling their clients.Finally, CHAs with higher education levels were reported to be more effective on some performance metrics such as record-keeping and appropriate use of job aids [11].

While there is limited research on CHA sociodemographic characteristics, albeit with mixed results, the role of CHA attributes relevant to the particular public health issue of focus is largely unexplored.Individuals with a personal or family history of cancer often become cancer advocates and may be particularly suitable as CHAs for cancer control interventions for medically underserved communities.As CHAs, they can address their community members' particular barriers to screening [12].There is evidence that barriers to cancer screening, including fear can be particularly salient for some communities [13,14].

In the African American community, fears of cancer and of the medical establishment have been linked with lower screening rates [13].Many community members also continue to think of cancer as a "death sentence" after hearing of family or loved ones that were diagnosed and shortly thereafter were lost to the disease [14].

There is also a considerable literature base demonstrating that cancer survivor stories are a compelling tool in a health promotion context [15].CHAs with a cancer history may be akin to a role model and community members may learn from t eir experiences [15].Community members may also perceive such CHAs as more credible due to their cancer experience [16].Finally, CHAs with a personal and/or family history of cancer may be especially drawn to and passionate about their role due to impact of this experience.

The purpose of the current study was to evaluate the association of CHAs' personal and family history of cancer on intervention implementation and e cacy outcomes in a cancer educational workshop series delivered by trained and certi ed lay peer CHAs in N = 14 African American churches.The study addressed the research question of whether CHAs with a personal [family] history of cancer have greater intervention implementation and e cacy outcomes compared to CHAs without a personal history of cancer.We hypothesized that CHAs with a personal [family] history of c ncer would have greater intervention implementation outcomes (operationalized as intervention participant attendance at all three cancer educational workshops) than CHAs without a personal history of cancer.Additionally, we expected that study participants who attended the cancer educational workshops taught by CHAs with a personal

[family] history of cancer, would report greater 12-month cancer knowledge scores than participants who attended workshops taught by CHAs without personal [family] history of cancer.


Methods


The Intervention

This work presents a secondary analysis of data collected from the Project HEAL (Health through Early Awareness and Learning) cancer control implementation trial, conducted from 2012-2016 in 14 African American churches in Prince George's County, MD, USA.The study was approved by the University of Maryland Institutional Review Board (protocol # 10-0691) nd is registered with clinicaltrials.gov(NCT02076958).Community partners recruited churches that were predominantly African American, midsize (e.g., 200-800 members), and had not host

s similar cancer
ducational intervention.Two lay persons in each church were trained and certi ed to conduct a series of evidence-based cancer educational workshops for breast, prostate, and colorectal cancer, with an e phasis on screening.Female participants attended the cancer introduction, and the breast and colorectal cancer workshops, while male workshop participants attended the cancer introduction, and the prostate and colorectal cancer workshops.Additionally, while both the female and male CHAs led the colorectal cancer workshop, only the female CHA led the breast cancer workshop and only the male CHA led the prostate cancer workshop.Detailed information about the Project HEAL intervention, recruitment strategies, and outcomes are reported elsewhere [3,4,17].

The current analysis examined data from Project HEAL CHAs and workshop participants.Each of the 14 church pastors identi ed one male and one female member to be trained, resulting in a total of 28 CHAs.Eligible CHAs were at least 21 years of age, a member of the church, had regular access to the internet, were able to complete web-based training activities, and self-identi ed as African American.CHA recruitment is discussed in detail elsewhere [17].The cancer educational workshop participants were members of the 14 participating churches who self-identi ed as African Ame ican, between 40 and 75 years of age, able to complete project surveys and attend all three workshops, and had no personal history of breast, prostate or colorectal cancer.


Measures Independent variables

The CHA personal cancer history data were collected from questions on an evaluation survey administered upon completion of their training, which asked "Are you a cancer survivor?"and "Do you have a family history of cancer?".These responses were used to derive two indicators for CHA personal and family history of cancer, respectively, coded as 0 (no CHAs had history), 1 (1 CHA with history) and 2 (both CHAs had history) at each church.


CHA competence and CHA education

The CHA's competence measure was generated from study team-reported rati

lity Checklist during
orkshop observations.The study team rated CHAs on 18 competence items that measured the CHA's skills (e.g., "Delivered with con dence"), knowledge (e.g., "Evidenced mastery of the content"), presentation style (e.g., "Spoke in appropriate volume"), and engagement with workshop participants (e.g., "Was able to keep audience attention").Each item was rated on a scale of 1-4 representing poor, fair, good, and excellent, respectively.For each

HA, an average score was calculat
d across the 18 items.CHAs' education was measured as the CHA having four or more years of college education or less than four years of college.


Implementation Outcome

A church-level measure of number of participants that attended all three workshops was captured through a project management tracking database.This metric, ranging from 1 to 22, is a measure of the CHAs' success in engaging participants in the intervention.As the intervention was designed to have participants attend all three workshops, the CHAs ability to increase attendance is an indicator of successful implementation.


E cacy Outcomes

Workshop participant-level measures of breast, prostate, and colon cancer knowledge were collected through self-administ

ed surveys completed at
baseline and 12-month follow-up.Breast cancer knowledge was assessed using an established instrument previously used with African American women [18,19].The measure is a composite of six items such as "Are older women more likely to get breast cancer than younger women?".The items used a yes/no/not sure response scale and correct responses were summed to form the knowledge scale score, which has a range of 0-6.The scale had

modest alpha re
iability of α = .32[3].Prostate cancer knowledge was also assessed using items from an established instrument [20] which was used in our previous research [3].The measure is a composite of ve items such as, "More African American men are diagnosed with prostate cancer than Whites".The items had a yes/no/not sure response scale and correct responses were summed to form the knowledge scale score, which has a range of 0-5.Internal reliability for the scale was α = .63[3].The colon cancer knowledge measure was assessed with a validated instrument previously used with African Americans [21].The measure is a composite of seven knowledge items such as, "Colorectal cancer screening is not necessary if there are no symptoms".The response options are agree/disagree/don't know, and the nal score has a range of 0-7.Internal reliability for the scale was α = .72[3].For all knowledge measures, higher scores indicate greater knowledge levels.


Analyses

All analyses were conducted using SPSS Version 26.0.First, the sample characteristics were summarized through descriptive statistics.Spearman's correlation coe cients were used to examine the bivariate associations between the CHA cancer history measures and the implementation and e cacy outcomes.One-way analysis of covariance (ANCOVA) tests with Tukey's post-hoc procedures were used to determine if there was any signi cant difference by the church's number of CHAs with personal/family history of cancer on the continuous intervention implementation outcome (i.e., the number of participants

hat atten
ed all three workshops), after controlling for the CHAs' competence scores and education levels.Linear regressions were used to examine if there was a signi cant association between the CHA cancer history measures and the workshop participant cancer knowledge scores at 12-months after controlling for workshop participant baseline knowledge scores, the CHAs' competence scores, and the CHAs' education levels.Mixed models were not used to examine the association between the outcomes and CHA cancer history because the ICC scores indicated that outcomes were very minimally explained by between-church and between study-group differences.


Results


Sample Demographics

The average CHA age was 51.4 (SD= 13.3) years old, and 50% were female (see Table 1).While less than 20% of the CHAs had a personal history of cancer (N=5), over half (53.6%, N=15) reported a family history of cancer.There was a total of N=375 church members who were enrolled across the 14 churches and attended at least one of the three cancer educational workshops (not shown).Among workshop participants, 67.8% were female and the average age was 55.3 years old (SD=9.3)(not shown).


Bivariate correlations

CHAs' personal history of cancer was not signi cantly associated wi

e outcomes (Table 2)
However, CHA family history of cancer was positively associated with implementation and one of the e cacy outcomes (Table 2).The number of CHAs with a family history of cancer was positively associated with the number of participants who attended all three workshops (r s =.77, p=.001), while the male CHA's family history of cancer was positively associated with the workshop participants' average prostate cancer knowledge score (r s =.26, p=0.01).


CHA Cancer History and Intervention Im

ementation

There was n
signi cant difference between CHAs with a personal history of cancer and CHAs without a personal history of cancer on the median number of participants that attended all three workshops (see Table 3), after controlling for the CHAs' competence.However, there was a signi cant difference between CHAs with a family history of cancer and CHAs without, on the median number of participants that attended all three workshops [F= 5.4, p = .05],after controlling for the CHAs' competence.Tukey's posthoc tests indicated the churches where both CHAs had a fami

history of cancer had a signi cantly higher median
number of participants who attended all three workshops compared to the churches where neither or only one CHA had family history of cancer.


CHA Cancer History and Intervention E cacy

Table 4 reports the results from the linear regression model that examined the adjusted associations between CHA cancer history and the e cacy outcomes.There was a signi cant and positive association between the male CHAs' family history of cancer and the workshop participants' prostate cancer knowledge scores at 12-months (estimated beta coe cient=0.49,p<0.01) after controlling for the baseline knowledge scores and CHA's competence measures.No other cancer knowledge outcomes were associated with CHA personal or family history.


Discussion

The purpose of the current study was to evaluate the role of two CHA characteristics -CHAs' personal and family history of cancer -in implementation and e cacy outcomes in a CHA-led cancer educational intervention.Although limited in scope, previous literature led to the hypothesis that CHAs with personal or family history of cancer may be particularly suitable for facilitating cancer control intervention outcomes given their expertise in the subject matter, perceived credibility by the target audience, and role model status for community members.This hypothes

ndings, whi
h we will discuss in greater detail below.This research makes a unique contribution to the literature, as there is little previous study on what characteristics make for an effective CHA, outside of select sociodemographic characteristics such as age, gender, or education level, and health experience.

Findings have implications for CHA effectiveness in the area of lay or peer health advisor interventions.


CHA's personal history of cancer

The expectation that CHAs with a personal history of cancer would have greater intervention implementation and e cacy outcomes compared to CHAs who had not had cancer, was not supported by the current analysis.While an intuitive hypothesis, it is possible that the limited number of CHAs in the current study who had had cancer may have attenuated our ability to detect signi cance due to a lack of variability.The current study's nding are consistent with those of Maxwell and colleagues [10], who reported that CHA's personal history of cancer was not associated with the CHA's success in

cruiting study participants.The s
udy by Maxwell et al. was similarly limited by a low number of CHAs who had had cancer.

Future studies in this area may bene t from more intentional recruitment to ensure greater representation of individuals with a personal cancer history and survivorship experience.Indeed, the current study did not specify cancer history as a CHA eligibility criteria, as this was not an a priori part of the larger study' research question.Future research could also be strengthened by considering the CHA's physical and emotional functioning in the context of their cancer survivorship experience.It is possible that a CHA who is a cancer survivor may face health issues related to survivorship such as fatigue, pain, or limited cognitive function [22], which may limit their ability to ful ll all aspects of the CHA role and research protocol, such as intervention participant recruitment and intervention implementation.These factors related to functioning may outweigh the expectations around the bene ts of source credibility and content area knowledge, which in part drove the hypothesis that CHAs who had had cancer would be more effective in their role than those who had not.


CHA's family history of cancer

We anticipated that CHAs with a family history of cancer would have greater intervention implementation and e cacy outcomes compared to CHAs without family history of cancer, which was partially supported by the current data.The intervention was designed to have participants attend three workshops, and this intention was better met when both CHAs had a family history of cancer compared to when neither or only one CHA had family history.There was a greater median number of participants that attended all three workshops when both CHAs had family history of cancer compared to when neither or only one CHA ha

family history.The signi cant n
ings despite the small sample size of churches indicates, arguably, a particularly robust association.Due to having a personal connection to the intervention's goal, CHAs with family history of cancer may be more motivated to deliver the project as intended than CHAs without such family history.If so, CHAs with family history may have engaged more proactively and regularly with workshop participants about the workshops and to attend the workshops.Concomitantly, workshop participants may have responded to the CHAs' potential greater engagement and credibility by attending the workshops.

Additionally, the male CHA having a family history of cancer also appeared to be associated with an improved understanding of the cancer educational content as evidenced by the higher workshop participants' prostate cancer knowledge scores at follow-up.Because CHA family history of cancer was not associated with intervention participant gains in breast or colorectal cancer knowledge, the current ndings may re ect the complexity of the cancer educational content being presented by the CHA.In speci c, this characteristic could be more relevant for cancer information that is considered complex, as is the case with prostate cancer.There is consid rable controversy about the screening recommendations for prostate cancer [23] involving a lack of research evidence that the prostate speci c antigen test reduces mortality and questions on whether screening bene ts outweigh the risks of follow-up tests and cancer treatment in cases of indolent disease.The Project HEAL prostate cancer knowledge survey items re ected this complexity and the controversy surrounding prostate speci c antigen screening.

Alternatively, it could be that CHA cancer family history played a particularly powerful role either when the CHA, or the audience, were men.The prostate cancer workshops were men-only and taught by the male CHAs.Unfortunately, the current study was not able to disentangle whether the male CHA cancer family history led to prostate cancer knowledge gains in male workshop participants was due to the information complexity, audience and/or participant sex, or some other factor.The male and female CHAs worked together to present the colorectal cancer workshop to a mixed-sex audience.Had the colorectal cancer workshops been separated by sex and taught by same-sex CHAs like the prostate and breast cancer workshops, we would have been able to determine if the observed association for prostate cancer would have been observed for colorectal cancer.Further study, perhaps informed by qualitative insights by both CHAs and workshop audience members, would be required to explain the current nding around family cancer history.

Perhaps one of the most intriguing questions raised by the current research is why family history, but not personal history of cancer was related to better workshop attendance overall, and prostate cancer knowledge gains among men.It may be that learning about cancer from a peer who has had cancer themselves may raise feelings of discomfort or elevated cancer worry in the audience, whereas a peer with a family history may not evoke the same level of fear.Learning about cancer from a person who has had it may result in psychological reactance, whereby an individual feels threatened or as if someone is taking away or limiting their options [24,25].Exper -driven health information, often delivered through nurse-led health education interventions, can be interpreted as a threat for participants [25].It follows that a CHA who has had cancer themselves may be perceived as imposing their health message on the workshop participant whereas, perhaps, the CHA who has had family history of cancer is perceived more so as a peer than an authority gure.Future research should examine this further and consider qualitative insights from both CHAs and audience members in an effort to shed light on this unexpected nding.


Strengths Limitations

This study has a number of strengths and implications for strengthening cancer control interventions through identifying people most likely to be effective CHAs.We had the opportunity to investigate whether the relationship between CHA cancer history and the intervention outcomes differed by type of cancer history (i.e.personal and family).We examined the association with cancer knowledge across three different cancer sites to understand if there are unique implications.The current ndings indicate that there is more to be learned in future research.

The current ndings are also limited by several factors.First, there were a limited

mber of CHAs with a person
l cancer history in the current sample.This may have attenuated our ability to detect a signi cant role of CHA personal history of cancer on intervention outcomes.Second, there may be other more salient factors in predicting participant workshop attendance and knowledge gains than CHA-level characteristics, such as participants' own family history of cancer.In these analyses, we focused on assessing potential CHA-level characteristics of interest and did not measure such participant-level factors.Future research would bene t from a mixed model approach a sessing the role of a more comprehensive set of participant, CHA, and church-level covariates on the outcomes of interest.Third, the study was not able to evaluate the role of CHAs cancer history on a wider range of intervention outcomes.

It would be bene cial to understand if the CHAs' cancer history had an association with one of the primary evaluation metrics of a cancer control intervention, participants' actual screening behaviors.


Conclusion & Implications

While the present ndings suggest that CHAs who have a family history of cancer may yield greater engagement in and e cacy of cancer educational interventions, additional research is warranted to determine conditions under which this is true and why this may be the case.Conveying prostate cancer information is complicated due to shifting, risk-based screening recommendations, screening test-related limitations and controversies, and continued c allenges in distinguishing indolent prostate cancers suitable for active surveillance from aggressive cancers that need treatment.It appears that African American male CHAs with a family history of cancer m

be particularly effective
for conveying this complex information to other men in their community.Our ndings have implications for the design of future CHA-led health promotion interventions, CHA workforce development, and reduction of cancer disparities, particularly those disproportionately impacting African Americans.Note.To predict the breast cancer knowledge score, only the female CHAs cancer history data were used.


Tables

Additionally, to predict the prostate cancer knowledge score, only the male CHAs cancer history data were used.

+ The adjusted model includes covariates for participants' baseline knowledge scores, CHA competence, and CHA education measures (not reported).The CHA competence measures are not statistically significant at α = .05.

Table 1
1
Project HEAL Community Health Advisor Sample Characteristics (N=28)
Baseline characteristicn%, Mean (SD)GenderFemale1450.0Male1450.0Age28 51.4 (13.3)Education<4 years of college1242.9=<4 years of college1657.1Marital StatusNot Married/living with partner1657.1Married/living with partner1242.9Work statusRetired621.4Disable13.6Part-time27 1Full-time1553.6Missing414.3CHAs with family history of cancerNo1035.7Not sure310.7Yes1553.6CHAs with personal history of cancerNo2175Yes517.9Missing27.1CHA's competence level/1-4Female CHA's average competence level133.54 (.40)Male CHA's average competence level123.50 (.61)Both CHAs' average competence level253.51 (.47)Table 2