Materials and Methods
Study Design and Participants
From February 2019 to May 2020, a cross-sectional survey was administered by telephone to patients throughout the United States with current or prior HCV infection who self-identified as Black or African American. The UNC-Chapel Hill IRB approved the study procedures. All participants provided verbal or written consent prior to participation.
Inclusion and Exclusion Criteria
Participants were eligible for inclusion whether they had participated in a national prospective cohort study (PROP UP: NCT02601820) or attended an outpatient liver clinic at UNC-Chapel Hill; were at least 21 years of age; had a current or prior diagnosis of HCV; and self-identified as Black or African American.
Recruitment and Consent
We recruited participants through two methods. Participants who participated in PROP UP and self-reported being Black or African American were contacted via phone or a thank you letter and invited to participate. To meet enrolment numbers, we also recruited patients from the UNC outpatient liver clinics. Participants were reimbursed $20 for survey completion.
After consent was given, participants completed surveys over the phone with trained research staff. Coordinators entered survey responses, including open-ended responses, directly into a web-based Research Electronic Data Capture (REDCap).[26,27] Participants self-reported all data with the exception of cirrhosis status, which was extracted from the PROP UP study database or electronic health records.
To develop the quantitative measures below, the investigators reviewed literature and selected pre-existing surveys assessing factors associated with WTP in research. Pre-existing survey items were combined into a survey to evaluate all relevant domains. This preliminary survey was shared with an HCV patient engagement group (2 Black, 1 Biracial, 1 Latino, 3 White participants) during a focus-pilot group. The investigators and patients worked to reduce item redundancy, minimize survey burden and select items with the most breadth, relevance and coverage of the patients' perspective. A second draft was piloted by the patient engagement group to ensure item comprehension and clarity. Pre-existing surveys had similar formats and response options based on a five-point Likert scale as follows: 1 (strongly disagree), 2 (disagree), 3 (unsure/neutral), 4 (agree) and 5 (strongly agree), unless otherwise indicated. For each scale, an overall score was calculated for each participant by taking the mean of all items; overall scores were not calculated if a participant had missing data on a scale item. All multi-item scales demonstrated good internal consistency assessed through Cronbach's α: WTP (α = 0.84), perceived benefits (α = 0.92), perceived barriers (α = 0.91), trust in doctors (α = 0.84) and health literacy (α = 0.77).
Willingness to Participate. Thirteen items were used to assess participants' WTP in different types and circumstances of health/medical research studies. Of these thirteen, three items were selected from a pre-existing survey. The investigators then developed ten WTP items to assess willingness to participate in a broader array of research activities and piloted these items with the patient engagement group. Responses were based on a five-point Likert scale from 1 (not at all willing) to 5 (very willing), with 3 connoting 'unsure/neutral' 2 and 4 were not labelled.
Perceived Benefits of Research Participation. A 16 item scale was created from pre-existing surveys[29–32] to assess participants' perceived benefits of participation in research. A high score indicates high perceived benefits.
Perceived Barriers to Research Participation. A 27 item scale was created from pre-existing surveys[17,29–31,33] to assess participants' perceived barriers to participating in research. A high score indicates high perceived barriers.
Trust in Doctors. A twelve item scale was created from pre-existing surveys[17,24,31] to assess participants' trust in doctors or doctors who conduct research. Negative items were reverse coded prior to analysis. A high score indicates high trust in doctors.
Health Literacy. A three item scale assessed the health literacy of participants with a five-point response scale as follows: 1 (Always), 2 (Often), 3 (Sometimes), 4 (Rarely) and 5 (Never). Participants were asked how frequently they asked others to help them read medical materials, how often they felt uncomfortable filling out medical forms and how often they did not understand their medical condition due to difficulty reading medical materials. A high score indicates high health literacy.
Subjective Knowledge of General Health/Medical Research. Participants were asked a single item about how well they felt they understood what health/medical research involved. This item was assessed on a five-point response scale as follows: 1 (Not at all), 2 (Not too well), 3 (Somewhat well), 4 (Pretty well) 5 and (Very well). A high value indicates high subjective knowledge.
Objective Knowledge of Clinical Trials Scale. Knowledge of clinical trials was measured with ten true/false questions compiled from pre-existing surveys[17,35] about aspects of clinical trials. Responses were coded as correct or incorrect. Missing responses and 'I don't know' were coded as incorrect. The percentage of correct responses for each question and the overall mean percentage correct was calculated.
Sociodemographic Variables. Participants self-reported age, sex, marital status, education level, annual household income (<$40,000 vs. ≥$40,000), employment status, health insurance and prior research participation.
Strategies to Increase Participation in Health/Medical Research. Interviewers asked participants the following question: 'What recommendations do you have for researchers about strategies or ways to increase participation from the Black community in health and medical research?' Probes maximized the quantity and quality of responses.
Descriptive statistics were used to conduct an initial data analysis. Prior to constructing a regression model, we assessed the distribution of the WTP score variable by reviewing summary statistics and a histogram. Ten a priori predictor variables were selected based on an extensive review of the literature concerning WTP in research. A multivariable median regression model was used to assess the relationship between these 10 a priori hypothesized predictor variables and overall WTP in research. The robustness of the main results to variations in statistical assumptions and methods was assessed by conducting several sensitivity analyses. Statistical analysis was performed with SAS 9.4 (SAS Institute, Inc., Cary, NC). A single coder used rapid qualitative content analysis to inductively code open-ended responses and summarize results of emerging themes.
J Viral Hepat. 2021;28(7):982-993. © 2021 Blackwell Publishing