Overview of Trial Design
Our trial was conducted in two west coast cities with sizeable Chinese communities: Seattle (Washington) and Vancouver (British Columbia). The researchers collaborated with Chinese community coalitions in both cities. summarizes the study design. Trial participants were individuals of Chinese descent who participated in baseline, community-based surveys (conducted in Seattle and Vancouver) during 2005. Chinese men and women were eligible for baseline survey participation if they were aged 20-64 years and able to speak Cantonese, Mandarin, or English. Six months after baseline survey completion, respondents who reported they had never been tested for hepatitis B were randomly assigned to an experimental group or a control group. Seattle and Vancouver participants were randomly assigned separately.
After random assignment, individuals in the experimental group received a hepatitis B lay health worker intervention. Control group participants received a mailing of physical activity print materials (pamphlet and fact sheet), as well as a pedometer with instructions for use. Our primary trial outcome was hepatitis B testing completion within six months of randomization. Secondary outcomes included the following knowledge variables: Chinese are more likely to be infected with hepatitis B than whites; hepatitis B can be spread during childbirth, during sexual intercourse, and by sharing razors; and hepatitis B can cause liver cancer. Trial participants completed a follow-up survey six months after receiving the hepatitis B lay health worker or physical activity direct mail intervention. Outcome ascertainment was based on follow-up survey responses. Medical record review was also performed for the primary outcome.
Translation and Personnel
All study materials that were read by participants (e.g., letters, consent forms, and pamphlets) were translated into simplified and traditional Chinese using standard double-forward methods. Similarly, study materials that were read to participants (e.g., study questionnaires) were translated into Cantonese and Mandarin.22,23
All project personnel with direct participant contact (i.e., survey interviewers and lay health workers) were bicultural, trilingual (Cantonese, Mandarin, and English) Chinese Americans/Canadians. Study personnel and study participants were matched by gender.
Baseline Survey and Trial Recruitment
Previous articles provide detailed descriptions of our baseline survey methods.15-17,19
To identify Chinese households, a previously validated list of 50 Chinese last names was applied to electronic versions of the metropolitan Seattle and Vancouver telephone directories. All identified households in geographic areas of Seattle with a relatively high proportion of Chinese residents were included in the US baseline survey sample. In Canada, a random sample of identified households in East Vancouver (an area with a high proportion of Chinese residents) was selected.
Introductory mailings were sent to households selected for inclusion in the survey. Subsequently, interviewers made multiple household contact attempts (including weekday, weekend, and evening attempts). If a household included two or more eligible Chinese adults, the nearest birthday method was used to select one study participant from the household. Interviews were conducted face-to-face in participants' homes. Respondents received a small financial incentive for baseline survey completion.
Baseline survey participants specified their age, marital status, and educational level. They were also asked how many years they had lived in North America and how well they spoke English. Respondents were read the following statement: “Hepatitis B is an inflammation of the liver caused by a viral infection. It sometimes makes the skin and eyes go yellow. People with hepatitis sometimes lose their appetite and experience nausea as well as vomiting.” They were then asked if they had ever had a blood test to see if they currently have hepatitis B or have had it in the past. Finally, respondents were asked whether they thought Chinese or white residents of North America are more likely to be infected with hepatitis B; whether they thought hepatitis B can be spread during sexual intercourse, during childbirth, and by sharing razors; and whether they thought hepatitis B disease causes liver cancer.
The survey was completed by a total of 969 Chinese men and women (436 in Seattle and 533 in Vancouver). The cooperation rates for the baseline survey in Seattle and Vancouver were 58% and 59%, respectively. Individuals who completed the baseline survey were eligible for randomization into the trial if they had no history of hepatitis B testing (52% of Seattle respondents and 44% of Vancouver respondents). The sample for the randomized controlled trial included 460 individuals (226 in Seattle and 234 in Vancouver).
We used findings from an earlier qualitative study to develop culturally and linguistically appropriate materials for use in the hepatitis B lay health worker intervention.24
These materials included a video (available in Cantonese, Mandarin, and with English sub-titles) and a pamphlet (with simplified Chinese, traditional Chinese, and English text). Our audio-visual and print materials emphasized the importance of hepatitis B serologic testing for all individuals of Chinese descent, and also addressed key hepatitis B facts. For example, the materials included information about the high rate of hepatitis B infection among Chinese, routes of person-to-person hepatitis B virus transmission, and the association between chronic hepatitis B infection and the development of liver cancer. Two visual aids were also developed by the project to emphasize key educational points: A world map showing rates of chronic hepatitis B infection by country and a graph showing liver cancer rates in North America by race/ethnicity.
Lay Health Worker Intervention
Lay health workers made up to 11 attempts to complete an educational and motivational home visit with each experimental group participant. Individuals who refused a home visit were offered the educational materials (video and pamphlet). If a lay health worker was unable to contact a participant, the educational materials were mailed to his/her home. The lay health workers were trained to act as role models, give social support, and provide tailored responses to each individual's barriers to hepatitis B testing (e.g., believing that testing is unnecessary for asymptomatic people). During home visits, lay health workers systematically asked participants if they could watch the video together, offered participants a copy of the video and pamphlet, and showed participants the two visual aids.
Follow-up Survey and Medical Record Review
To trace individuals who had recently moved, we used contact information for friends and relatives, provided at the time of the baseline survey, and the most recent telephone books for Seattle and Vancouver. The follow-up survey implementation procedures were identical to those used at baseline. Specifically, multiple contact attempts were made, the interviews were completed face-to-face, and a small financial incentive was provided. Follow-up survey interviewers were unaware of each participant's trial randomization assignment. Our follow-up questionnaire included the same hepatitis B testing and knowledge items as the baseline questionnaire. Additionally, two follow-up survey questions assessed the use of hepatitis B audio-visual and print educational materials among individuals randomized to the experimental arm.
Follow-up survey respondents who reported they had received hepatitis B testing in the six- month interval since their random assignment were asked why they had been tested. They were then asked to provide information about the date of testing, as well as the location of the clinic or doctor's office where testing was performed. Each of these participants was also asked to sign a medical release form giving project staff permission to request medical record verification of his/her self-reported hepatitis B test. A copy of the hepatitis B test result was then requested (from the relevant clinic or doctor's office) using a form that provided the participant's name, age, and self-reported date of testing. The project contacted each health care facility up to three times (twice by mail and once by telephone).
Process data were collected to document the implementation and content of our hepatitis B lay health worker intervention. Specifically, lay health workers routinely completed forms addressing the outcome of home visit attempts (e.g., agreed to participate in a home visit, refused a home visit but accepted the educational materials, or refused a home visit and the health education materials). They also documented use of the project video, pamphlet, and visual aids.
We conducted an “intent-to-treat” analysis and included all randomly assigned individuals with follow-up data. Chi-square tests and, when necessary, Fisher's exact tests, were used to evaluate statistical significance with respect to differences in proportions. Unconditional logistic regression techniques were used to adjust for the following potential confounders: City (Seattle versus Vancouver), age-group (<45 years versus ≥45 years), educational level (<12 years versus ≥12 years), marital status (currently married versus not currently married), proportion of life spent in North America (<50% versus ≥50%), English language proficiency (spoke very well or fluently versus did not speak well or at all). Regression analysis of knowledge at follow-up also included baseline knowledge (for the knowledge variable under consideration) as a covariate.