|Home | About | Journals | Submit | Contact Us | Français|
Asian Americans have lower colorectal cancer (CRC) screening rates than non-Hispanic Whites. Hmong Americans have limited socioeconomic resources and literacy. This randomized controlled trial (RCT) was conducted to determine if bilingual/bicultural lay health educator (LHE) education can increase CRC screening among Hmong Americans
We conducted a cluster RCT among Hmong Americans in Sacramento, California. LHEs and recruited participants were randomized to intervention or control groups. The intervention group received CRC education over 3 months delivered by a LHE. The control group received education about nutrition and physical activity delivered by a health educator. The outcomes were change in self-reported ever and up-to-date CRC screening after 6 months.
All participants (n=329) were foreign-born with mostly no formal education, limited English proficiency, and no employment. Most were insured and had a regular place for care. The intervention group had greater changes after the intervention than the control group for ever screening (p=0.068) and being up-to-date (p<0.0001). In multivariable regression analyses, the intervention group had a greater increase than the control group in reporting ever screening (AOR = 1.73, 95% CI: 1.07–2.79) and being up-to-date (AOR = 1.71, 95% CI: 1.26–2.32). Having health insurance had > 4 times the odds for receiving screening, both ever and up-to-date. A higher CRC knowledge score mediated the intervention effect for both screening outcomes.
A culturally and linguistically appropriate educational intervention delivered by trained LHEs increases colorectal cancer screening in an immigrant population with low levels of education, employment, English proficiency, and literacy.
Clinicaltrials.gov # NCT01904890
Colorectal cancer (CRC) incidence and mortality can be reduced through screening tests but screening participation is suboptimal for racial/ethnic minorities and even more for those with limited English proficiency.1 In national surveys, fewer Asians have reported CRC screening than non-Hispanic whites, even after adjustment for socioeconomic status, access, and language barriers.2 In California, Asian Americans are less likely to be screened for colorectal cancer compared with non-Hispanic Whites, despite narrowing disparities over 2003–2009.3 Only 64.4% of eligible Californians reported being up-to-date for CRC screening in the 2013 Behavioral Risk Factor Surveillance System.4
To overcome cultural and language barriers, bilingual/bicultural lay health educators (LHEs) who are community members receiving training in health topics have been effective in delivering community-based interventions.5 LHE interventions have been effective among Vietnamese Americans and Chinese Americans for increasing screening rates for CRC6 and other cancers in randomized controlled trials.7,8 However, such trials have rarely been conducted among Hmong Americans who have some of the highest rates of poverty (27.4%),9 low levels of education,11 and limited literacy even in the Hmong language.10,11 Our team members previously conducted a LHE intervention of in-home education and patient navigation that increased Hepatitis B screening among 260 Hmong adults,12 and a LHE intervention that increased breast cancer screening among 434 Hmong women.13
The objective of this study was to conduct a LHE-delivered educational intervention to promote CRC screening among Hmong Americans and evaluate the intervention’s impact on CRC screening, compared to an attention control with nutrition and physical activity (NPA) education. We hypothesized that participants in the intervention group would report higher increases in ever screening and being up-to-date for CRC screening than those in the control group.
Using a community-based participatory research (CBPR) approach,14 we partnered with Hmong Women’s Heritage Association (HWHA), a community-based organization that has provided services for Hmong families since 1994. Since 2000, HWHA and the academic research team have been engaged in a CBPR partnership through the Asian American Network for Cancer Research Awareness, Research, and Training (AANCART), a national CBPR network funded by the National Cancer Institute’s Center to Reduce Cancer Health Disparities.12–14 HWHA was a full partner in the implementation, evaluation, and dissemination of the study.
We used a two-arm cluster randomized controlled trial (RCT), with clustering at the level of the LHEs, who were recruited through Hmong radio and HWHA clients. After receiving training on participant recruitment, LHEs recruited participants through their own social networks. Some participants were recruited through radio announcements and HWHA clients. LHEs were randomized by a computer program to either the intervention or control arm after completing recruitment. The LHEs assigned to the intervention were trained to deliver CRC prevention information, while the LHEs assigned to the control arm delivered no intervention. The control group participants received nutrition and physical activity (NPA) education from a health educator. This design enabled a comparison of a LHE intervention versus usual care with attention control while ensuring that the control group benefited from research participation. All participants attended two small group educational sessions lasting approximately 90 minutes and separated by 2 months, received 2 follow-up calls about 1 month after each session, and completed surveys at baseline and 6 months. The study was conducted from March 2012- August 2015 in Sacramento, CA. Human subject approval was obtained by the Institutional Review Board of UC San Francisco.
The study was powered to detect a net effect size of 0.20 between the intervention and control group in the proportion ever screened for CRC, assuming a 0.25 in the intervention group and 0.05 in the control group, LHE cluster size of 12 participants, intracluster correlation coefficient (ICC) of 0.05, and an attrition rate of 0.05. We determined the number of participants per LHE based on our prior LHE studies.7,15
The study was implemented over 3 time periods (waves). Each LHE participated in only one wave. Inclusion criteria initially were for LHEs to be Hmong and be over 50 years, similar to trial participants, but due to recruitment problems, the lower age cutoff was changed to 18 years, starting in wave 2. The LHE were native Hmong speakers who could also speak English, and their educational background ranged from some high school to college graduates; some had prior experience as LHEs in previous HWHA projects. Intervention and control LHEs received an identical first training session by HWHA but in separate groups to minimize contamination. The first session included a description of the LHE program, roles, and responsibilities, and training in participant recruitment. The LHEs were trained on protection of human subjects in recruitment and participation but did not administer consent. Following the training, each LHE recruited 12–15 participants using a script describing the purpose of the project and scope of participant involvement. After completing recruitment and being randomized, the intervention LHEs received a second training session to conduct small group sessions and deliver CRC information. The control LHEs did not receive a second training session as the HWHA staff delivered the NPA information. Pre- and post-training surveys assessed the effect of training on LHE knowledge and confidence, and additional training was administered as needed. Each LHE was paid $1,200 at the conclusion of the last data collection to reflect the value of their cultural expertise and the time involved in the research.
Eligibility criteria for participants included being 50–75 years old, self-identifying as Hmong, speaking Hmong or English, living and intending to stay in the area for at least 6 months, having no personal history of CRC, having no medical problems preventing them from attending sessions, and being willing to participate in a study about CRC screening or NPA. Participants were recruited regardless of prior CRC screening history to reflect the general community and minimize selection bias. Only one person per household could be a participant. At the first educational session, bilingual research staff obtained consent by reading from a written document outlining study activities, risks, and benefits. Surveys were administered verbally by research staff given participants’ low literacy level. Participants were paid $60 for their research participation, with $20 after the first session (baseline survey) and $40 after the third and final session (6-month survey).
The CBPR team developed a LHE training program, manual, and flipchart for CRC based on prior interventions.6,16 The development of the CRC materials was guided by the Social Cognitive Theory17,18 and the Transtheoretical Model,19 specifically addressing: 1) knowledge of CRC risk and prevention; 2) expectations about CRC screening (positive anticipatory outcomes of screening);17,18 3) self-efficacy (confidence that one can obtain screening);17–19 and 4) intention (motivation and readiness to obtain screening).19 The CRC flipchart promoted the goal of obtaining any CRC screening at the time interval recommended by the U.S. Preventive Services Task Force (USPSTF).20 The flipchart described the needs and benefits for screening, the different screening tests and USPSTF recommendations for screening frequency, and barriers to screening through brief educational and culturally appropriate messages. System barriers, such as access to care, were addressed with a list of locally available services, but the study was not designed to offer patient navigation services. HWHA provided appropriate cultural images and used double simultaneous forward translation21 for the flipchart into Hmong from English, with review by 2 focus groups and the research team. For the control group, the content of the NPA PowerPoint presentation was on healthy nutrition (food types and portion and serving sizes) for cardiovascular health and diabetes prevention, based on prior interventions,6,16 and did not include any CRC information. CRC LHEs were trained to deliver brief follow-up telephone calls to check in on participants regarding questions about CRC and stages of change (readiness to obtain screening). The follow-up telephone calls for the control group were conducted by NPA LHEs who asked participants about their diet.
Surveys were conducted immediately before and 6 months after the first session. Sociodemographic measures included age, gender, birthplace, years in the U.S., education, employment, marital status, English language proficiency, and household income. Access was measured by participants reporting whether they had health insurance, had a primary doctor, had a regular place of care, and whether they had seen a doctor in the past year. Health status was measured by self-rated health (excellent, very good, good, fair, poor) and whether a physician had told the participant that he/she had cancer. CRC-related measures included awareness of CRC and screening (colon cancer, colonoscopy, sigmoidoscopy, FOBT), knowledge about CRC screening, and self-report of test receipt and when the test was obtained. For the multivariable regression analyses, a knowledge score from 5 questions about CRC screening was created (score 0–5 total): 1) heard of colon polyps, 2–4) frequency of testing for FOBT (yearly), sigmoidoscopy (every 5 years), and colonoscopy (every 10 years), and 5) age of screening starts at 50.
The main outcome measures were self-reports of: 1) CRC ever screening (ever had fecal occult blood test (FOBT), sigmoidoscopy, or colonoscopy) and 2) self-reported up-to-date CRC screening (FOBT at 1 year, sigmoidoscopy at 5 years, or colonoscopy at 10 years). Comparison of study arms with respect to participant characteristics and pre-post change in proportion screened and awareness were evaluated with generalized linear models. Multivariable logistic regression models were created for reporting ever receiving any CRC screening and being up-to-date with CRC screening as a function of time (post vs. pre), study arm, and their interactions, adjusted for participant characteristics. In order to assess the mediating effects of knowledge, we also created multivariable models of knowledge score and then added knowledge score as a covariate to the models of ever and up-to-date CRC screening. All participants were included in analyses regardless of prior CRC screening history. Generalized estimating equations (GEE) were used in all models to account for clustering by LHE. Analyses were conducted on an intention-to-treat basis, with baseline values carried forward for dropouts. All analyses were conducted with SAS software, version 9.3 (SAS Institute, Inc., Cary, NC); statistical significance was assessed at the 0.05 level (2-sided).
Figure 1 shows the CONSORT diagram for participant flow. The study had 29 Hmong LHEs (aged 21–55, 82.7% women, 14 in the intervention group) recruited by the community partner. One LHE in the control group dropped out before study activities began, and that LHE’s 2 participants were assigned to another control group LHE. Out of 429 eligible participants, 93 (21.7%) refused to participate, and 329 participants were randomized to the intervention (n=161) and control arms (n=168). The retention rate at 6-month follow-up was 98%, with 5 participants who could not be contacted.
Participants had a mean age of 60.4 years and were mostly women and married (Table 1). All were born in Laos with 83.6% having lived in the U.S. more than 10 years. Over 89% spoke only Hmong at home and 70.5% reported speaking English poorly or not at all. Participants also had low socioeconomic status: 53.8% reported annual household incomes of less than $20,000, 89.6% reported no formal education, and 90.9% reported no employment. Healthcare access was high with 95.1% having health insurance, 94.2% having a regular place for healthcare, 92.1% having a primary care doctor, and 84.8% having seen a doctor in the past year. Almost half rated their general health as fair or poor. No significant differences appeared between the intervention and control arm participants.
Figure 2a shows that at baseline the majority of participants had heard about FOBT, but the intervention group had significantly greater increases post-intervention in awareness than the control group (intervention: 69.6% to 90.7%; control 74.4% to 79.8%, p=0.0017). Fewer were aware of other terms at baseline, but the intervention group also had significantly greater increases post-intervention in awareness than the control group for colon cancer (intervention 44.1% to 87.0%; control 39.3% to 58.9%, p<0.0001), colonoscopy (intervention: 36.0% to 83.9%; control 32.1% to 45.8%, p<0.0001), and sigmoidoscopy (intervention: 38.5% to 81.4%; control 32.1% to 43.5%, p<0.0001).
As shown in Figure 2b, at baseline very few participants knew anything about colorectal cancer screening guidelines, such as starting at age 50 or knowing the frequency of tests. By post-intervention, the intervention group had significantly greater increases in knowledge compared to the control group for colon polyps (intervention: 23.6% to 78.3%; control 19.6% to 37.5%, p<0.0001), starting at age 50 (intervention: 14.3% to 36.0%; control 11.9% to 14.3%, p=0.0056), getting a FOBT yearly (intervention: 10.6% to 38.5%; control 11.9% to 17.3%, p=0.0001), getting a sigmoidoscopy every 5 years (intervention: 3.7% to 24.2%; control 1.2% to 4.2%, p<0.0001), and getting a colonoscopy every 10 years (intervention: 2.5% to 20.5%; control 3.6% to 6.5%, p=0.012).
Figure 2c shows that from pre- and post-intervention, the intervention group had significantly greater increases than the control group in ever having had an FOBT (intervention: 67.7% to 79.5%; control 68.5 to 70.8%, p=0.039) and sigmoidoscopy or colonoscopy (intervention: 26.1% to 36.0%; control 18.5% to 17.9%; p=0.0052). The increase in the intervention group compared to the control group in ever screening for CRC was borderline significant (intervention: 72.1% to 83.2%; control: 72.0% to 75.0%, p=0.068).
The intervention had significantly greater increases in being up-to-date with CRC screening (intervention 44.1% to 57.1%; control 43.5% to 43.5%, p<0.0001), FOBT (intervention 32.3%% to 41.6%; control 35.1% to 34.5%, p=0.001), and sigmoidoscopy or colonoscopy (intervention 19.9% to 26.7%; control 16.1% to 14.3%, p=0.0053) (Figure 2d).
Table 2 shows the multivariable regression analyses for the intervention effect. From pre- to post-intervention, the LHE intervention group had significant increases in ever screening (Adjusted Odds Ratio [AOR] 1.95, 95% Confidence Interval [CI] 1.40–2.72) and being up-to-date for screening (OR 1.73, 95% CI 1.34–2.21). The LHE intervention was superior to the control group, which did not have significant changes in the screening outcomes, for ever screening (OR 1.73, 95% CI 1.07–2.79) and being up-to-date (OR 1.71, 95% CI 1.26–2.32). When knowledge score was added to the models, the intervention effect became non-significant. For every point increase on the knowledge score on a 0–5 point scale, the odds of ever screening (OR 1.29, 95% CI: 1.08–1.55) and being up-to-date (OR 1.49, 95% CI 1.24–1.79) were significantly increased, which supported knowledge being a mediator of the intervention effect. Having health insurance was highly associated with screening outcomes in all models with OR greater than 4, while having seen a physician in the past year was associated with being up-to-date.
The results of this study showed that a culturally and linguistically appropriate educational intervention delivered by trained LHEs is effective in increasing both ever screening and being up-to-date with CRC screening among Hmong Americans. The up-to-date screening rate (57.1%) at post-intervention with Hmong Americans is comparable to what has been reported post-intervention for similar interventions with Vietnamese Americans (56%)6 and Chinese Americans (55.7% FOBT)16 and approaching the rate reported for the general population who are fluent in English (59.6%).2,22 Taken together, these studies show that LHE interventions that are culturally and linguistically appropriate are effective in increasing CRC screening among Asian Americans with limited English proficiency. The intervention effect sizes in this study are comparable to clinic-based interventions to increase CRC screening among disadvantaged populations with direct offering of FOBT kits.23–25 However, the post-intervention CRC up-to-date screening rate of 57% falls below the National Colorectal Cancer Roundtable’s screening goal of 80%.22 While over 70% of the participants at baseline reported ever screening (mostly with FOBT), it is possible that the 6-month time period of the study is too short for participants to schedule and complete a doctor’s appointment or obtain a sigmoidoscopy or colonoscopy.
The finding of knowledge as a mediator of LHE education in this population is significant in that it provides evidence of a pathway for increasing CRC screening. Previous research26,27 has demonstrated the association between knowledge change and intention to obtain screening, but few have shown the direct association between knowledge change and screening outcomes. The LHEs’ verbal and visual flipchart delivery of education about CRC screening helps to address the fact that Hmong Americans in general have low education and limited literacy even in their own language.11 Visual tools have been previously described as effective teaching methods for the Hmong community to promote breast and cervical cancer screening.28,29 In addition, the sociocultural relationship between LHEs and their participants may be helpful in encouraging the participants to participate in questions or discussions about screening.30 For example, LHEs might help address any conflicting cultural beliefs and norms about cancer screening, such as any uncertainty about Western treatments11 for the Hmong participant.
Our study also demonstrates that even among a population with high healthcare access, having health insurance remains a significant factor associated with CRC screening receipt. Still, economic barriers may extend beyond health insurance, as publicly insured adults report barriers such as cost and coverage.31 One Hmong study about barriers to care reports that, more so than medical mistrust or discrimination, lack of health insurance, making co-payments, language, and issues related to scheduling appointments were important.32 LHEs in our study provided mostly informational and social support, not logistical support. Patient navigators33 could lead to a larger intervention effect, but that would have required more resources. With the Centers for Medicaid and Medicare Services 2014 rule (CMS-2334-F) opening up payment opportunities for preventive services by nonlicensed individuals,34 this might be more feasible in future. A future enhanced intervention might incorporate this LHE-delivered intervention to address attitude and knowledge changes, plus a patient navigation component to address logistical support for increased CRC screening rates.
There are several limitations to this study. First, the results from Hmong living in one county in California may not be generalizable to all Hmong Americans, although Sacramento has the third largest Hmong population in the U.S.10 Second, the small community could have created contamination of the control group; however, that would reduce the intervention effect size and thus if present, would strengthen the conclusions. Third, it is possible that LHEs may choose participants who may be more likely to get screening, but we attempted to deal with this selection bias by blinding LHEs and participants to study arm assignment until after recruitment was completed. Fourth, this study includes individuals who were up-to-date at baseline, but the increase in being up-to-date was significantly greater in the intervention group in the multivariate regression analyses. Finally, the study outcomes are self-reported, which may lead to social desirability bias among intervention participants to report CRC screening. Validation of self-reports would have been ideal but would have been difficult given study budget limitations due to the large number of healthcare providers in this community. However, a CRC education intervention for Filipino Americans found that the intervention effect for CRC screening was upheld when self-report bias was taken into account.35
The significant contribution of this study is that it shows the effectiveness of LHE in increasing CRC screening behavior through greater knowledge among Hmong Americans, a population with significant socioeconomic and health disparities. Strengths of this study include the cluster randomized controlled trial study design, the high participation and retention rates, and the diverse participant sample. In order to meet the national goal of 80% CRC screening among immigrant populations such as the Hmong, there is an urgent need for increased training of bilingual, bicultural LHEs to deliver culturally and linguistically appropriate CRC education that can help improve community engagement with healthcare services. Further research could help identify what is needed to reduce additional barriers to CRC screening in the Hmong community and describe elements of what makes LHEs work and their cost-effectiveness.5,36 Future studies may also consider studying the use of LHEs for management of other non-cancer health issues5,37–39 for the Hmong population.
This study was funded by the National Cancer Institute (U54 CA153499). However the content of this paper reflect those of the authors, and not necessarily those of the National Cancer Institute. We acknowledge Youa Lo and May Chee Lo for their research staff assistance with data collection and the rest of Hmong Women’s Heritage Association for their community partnership in support of this study.
Conflict of interest: None to declare from any authors.
Author Contributions:Elisa K. Tong: Conceptualization, Methodology, Formal analysis, Writing (original draft), Visualization, Project administration
Tung T. Nguyen: Conceptualization, Methodology, Formal analysis, Writing (original draft), Supervision, Project administration, Funding acquisition
Penny Lo: Conceptualization, Methodology, Validation, Investigation, Writing (original draft), Project administration
Susan Stewart: Conceptualization, Methodology, Software, Validation, Formal analysis, Writing (original draft), Visualization, Supervision, Funding acquisition
Ginny Gildengorin: Conceptualization, Methodology, Software, Validation, Formal analysis, Writing (review and editing)
Janice Y. Tsoh: Conceptualization, Methodology, Formal analysis, Writing (original draft)
Angela Jo: Conceptualization, Methodology, Writing (review and editing)
Marjorie Kagawa-Singer: Conceptualization, Methodology, Writing (review and editing)
Angela Sy: Conceptualization, Methodology, Writing (review and editing)
Charlene Cuaresma: Conceptualization, Methodology, Writing (review and editing)
Hy Lam: Conceptualization, Methodology, Software, Validation, Resources, Data curation, Writing (review and editing), Project administration
Ching Wong: Conceptualization, Methodology, Resources, Writing (review and editing), Project administration
Mi T. Tran: Validation, Data curation, Writing (review and editing), Project administration
Moon S. Chen, Jr.: Writing (original draft), Supervision, Funding acquisition