Using a hybrid implementation-effectiveness design for this feasibility study, we evaluated the implementation and effectiveness of the CHANGE program with a single arm, pre-/post-study design,(22) where participants were measured before and immediately after receiving the intervention. The University of North Carolina (UNC) Non-Biomedical Institutional Review Board (IRB) approved and monitored the study, beginning with approval in January 2016; direct interaction with study participants ended in September 2017. All participants provided written informed consent, and clinic patients consented to have study staff obtain CVD-related lab values from their medical record by signing a separate Health Insurance Portability and Accountability Act (HIPAA) consent form. For participants enrolled at the health department, no information was gathered from their medical record.
The CHANGE Intervention: Two CHWs delivered an adapted version of the evidence-based Heart-to-Health lifestyle intervention (23) and referred participants to community and clinical resources. Heart-to-Health is a low-intensity behavioral lifestyle intervention targeting CVD risk reduction through dietary and physical activity behavior changes, smoking cessation, and medication adherence. To support these behavioral changes, CHANGE also included a community ‘heart healthy’ resource guide and protocols for referring participants to and following up on their use of those resources. CHWs delivered the CHANGE intervention over 4 monthly, in-person counseling visits (45-60 minutes) in participants’ homes or at local venues selected by the participant. Between these monthly counseling visits, the CHW made short ‘booster calls’ (about 15-20 minutes) to follow up with participants on the progress made with goals set and actions taken on referrals made at the last counseling visit. Each participant received a program manual with educational materials on healthy eating, taking medicine, physical activity, stopping smoking, and a community resources directory including resources in their community related to heart health, health care, and transportation services. To maximize the potential benefits of lifestyle changes, program topics were introduced to participants based first on the participant’s selection of the behavior they most wanted to change, then on the potential CVD risk reduction expected by making the behavior change (ranked from highest to lowest). See Table 1 for more details on the CHANGE program content (4 main areas), listed in order of importance for CVD risk reduction. The total planned contact time (4 counseling visits + 3 booster calls) for this low-intensity intervention is estimated at 4 to 6 hours.
Site, CHW, and Participant Recruitment: The two sites selected for our feasibility study included a Federally Qualified Health Center (FQHC) and a local health department in Hertford, a rural NC county. We selected Hertford county because of its high rates of CVD risk factors and our prior relationships with FQHC leadership. Hertford county is located in the northeastern region of NC, with a population of about 24,000, poverty rate of 26%, and over 60% of the population self-identified as African American.(24) In 2016, Hertford County was ranked 89th for health outcomes and 93rd for health behaviors, among NC’s 100 counties.(25) The research team created subcontracts with both sites to cover the costs of staff member participation on the community-engaged research team and the salary and benefits for a full-time CHW.
The CHANGE Study’s enrollment goal for Hertford County was 150 participants. This sample size would provide a reasonable estimate of feasibility as measured by recruitment and attrition rates, while allowing for a loss to follow-up of 20%. To be eligible, participants had to: live in or receive medical care in Hertford County, North Carolina; be 18-80 years of age; and speak English. Women who reported that they were pregnant were excluded or withdrawn, as pregnancy may account for observed changes in weight and blood pressure. The CHW at the health department recruited participants through community outreach, including strategies such as word-of-mouth, flyers, local newspaper or magazine advertisements, participation in health fairs and community events, and visits to churches, beauty salons, and senior centers. Recruitment by the health department was tilted toward primary prevention of CVD which meant anyone encountered who was interested and eligible to participate was invited. In contrast, the CHW at the FQHC recruited through the electronic health record systems, with a focus on secondary prevention. Clinic nurses pre-screened existing patients for elevated risk of a cardiac event, and then created a list that the CHW used to recruit study participants, either at a clinic visit or via a phone call. they were smokers or had uncontrolled diabetes (A1c greater than 8%), hypercholesterolemia (low density lipoprotein [LDL] greater than 130 mg/dL), hypertension (systolic blood pressure >140 or diastolic >90 mm Hg) or a previous cardiovascular event. Patients who appeared in multiple risk categories were prioritized.
Staff Training for Intervention Delivery: Prior to patient recruitment, the research team conducted an intensive 6-day centralized study training with the staff responsible for participant recruitment and intervention delivery (site supervisors and CHWs). The training sessions included reviews of study protocols, informed consent and participant confidentiality, participant recruitment and study site protocols, CHANGE intervention content, community referral resources, and data collection methods. Training also included opportunities for CHWs to practice motivational interviewing skills and to role-play enrolling, counseling, and referring study participants.
Data Collection: Data collection included measures of both implementation and effectiveness outcomes; methods for each type are detailed below. The CHWs collected all participant data at counseling visits. For health center patients, eligibility screening data used to rank patients by CVD risk factors was provided to the study staff through a data sharing agreement.
Implementation Outcomes: Data were collected to assess reach and delivery fidelity. Data on reach were captured thorough tracking logs that CHWs maintained of the number of individuals they invited to participate, whether they agreed to participate, and reasons for declining. Data on fidelity were collected through an online system where CHWs documented delivery of the intervention including contact duration, content covered, goals set, referrals made to community resources, and disposition of referrals given (action taken and/or services received).
Effectiveness Outcomes: CHWs collected blood pressure, weight, self-reported dietary and physical activity data to measure program effectiveness. They collected outcome measures at the first and last intervention visits (Home visits 1 and 4). With this study’s primary focus on the effective implementation of the CHANGE program, and a secondary focus on its effectiveness in reducing CVD risk, we intentionally limited our data collection to reduce the burden on both the CHW and participants. Data collection included physical measures of weight and blood pressure, brief validated surveys of dietary and physical activity behaviors, and general demographic and health information. Weight, as the average of two measures, was assessed in pounds to the nearest tenth, by an electronic scale (Seca 874, Seca, Hanover, MD). Blood pressure (BP) measurements were taken with an automated BP machine (Omron HEM-907XL, Omron Healthcare, Lake Forest, IL). Two BP measurements (reported as an average systolic and diastolic value) were taken at 1-minute intervals after the participant was seated for 5 minutes. Self-reported dietary behaviors were measured with items from two validated brief food frequency surveys (10 total items) measuring dietary fat quality (26) and estimated intake of fruits and vegetables.(27) A single item (adapted from the 2 items used in BRFSS)(28) was used to assess usual daily consumption of sugar-sweetened beverage consumption. We gathered self-reported data on physical activity behaviors with a validated adaptation of the RESIDE survey which focuses on walking.(29, 30)
Statistical Analysis: For this pilot feasibility study, baseline sample characteristics were summarized using descriptive statistics such as means, percentages, standard deviations etc. Analyses of primary and secondary outcomes and pre-post- changes at 4 months were conducted using paired t-tests. Additionally, we assessed for group differences between males and females in pre-post- changes. To determine our rate of attrition and loss to follow-up, we included in our denominator all enrolled participants who completed the first counseling visit. Since intervention effectiveness outcomes were secondary aims in this study, we did not use any imputation methods to account for missing values but provide a description of those lost to follow-up. For weight and blood pressure measurements taken by CHWs at counseling visits, we used values from the last completed visit when the fourth visit values were missing. For 10 participants, weight and blood pressure values from their last visit after baseline served as their post-intervention values. All analyses were conducted using SAS Version 9.4 (SAS Institute, Cary, NC).