Study design
This study was conducted in Guangzhou, China. Guangzhou is the capital city of Guangdong Province with 14.5 million residents in 2017 [47]. As of 2019, about 4% in MSM were infected with HIV in Guangzhou [48]. This study was conducted after the National ART Guideline started to recommend free ART to all PLWH regardless of their CD4 counts. The participants were recruited between June, 2016 and June, 2017, and were followed up till December 31, 2017. Dates of initiating ART for the first time were extracted from the national HIV/AIDS comprehensive information system.
Participants and data collection
The inclusion criteria included: 1) aged at least 18 years, 2) received confirmatory HIV diagnosis, 3) without contradiction of ART initiation, 4) had never received ART before the date of baseline survey, and 5) willing to have the research team access their ART-related information in the national HIV/AIDS comprehensive information system. Exclusion criteria included presence of: 1) major psychiatric illness (schizophrenia and bipolar disorder), and 2) not able to communicate with the interviewers.
Participants were recruited from 6 out of 11 administrative districts of Guangzhou with relatively high HIV prevalence for the survey. There are six district Center for Disease Control and Prevention (CDC) and 29 community healthcare centers (CHC) providing HIV treatment and care for PLWH. To avoid selection bias, all HIV-positive MSM on the service record of these CDC/CHC were invited to join the study. Trained staff of these CDC/CHC contacted all HIV-positive MSM on their service records through telephone and/or during their regular follow-up, screened their eligibility, briefed them about the study, and assured them refusal would not affect their right to use any services and they could quit at any time without being questioned. Participants were asked to leave their ID number for extracting the date of ART initiation from the National HIV/AIDS Comprehensive Information System. Guarantee was made that their ID number would be kept strictly confidential, and would not appear in the questionnaire or dataset for analysis. Those who showed interest in the study were asked to visit one of these CDC/CHC. Out of 461 eligible HIV-positive MSM approached, 102 (22.1%) declined to participate in the study, and 359 (77.9%) provided written informed consent to join the study, and 303 (84.4%) completed the baseline survey. No name or personal contact was collected during the interview.
Measurements
Design of the questionnaire
A panel consisting of epidemiologists, health psychologists, and CDC/CHC workers was formed to design the questionnaire. The questionnaire was tested among 15 HIV-positive MSM. Based on their feedback, discussion was made by the panel to finalize the questionnaire.
ART initiation
Using their ID number as identifiers, date of initiating ART for the first time was extracted from the national HIV/AIDS comprehensive information system. Therefore, we could access all participants’ status of ART initiation within the follow-up period. For participants who had initiated ART within the 12-month follow-up period, the observational time was the time interval between the date of baseline survey and the date for initiating ART for the first time. For those who had not initiated ART within the follow-up period, their observational time was the entire follow-up time interval.
Background characteristics of the participants measured at baseline
Participants’ background information was collected, including demographic characteristics (age, marital status, education level, income level, medical insurance, household registration) and disease-related characteristics (Time since diagnosis of HIV, CD4 cell counts in the most recent episode of testing, and self-rated their health status).
Perceptions related to ART measured at baseline
Five scales were constructed for this study to assess perceptions related to ART. They were based on the HBM.
Perceived severity of delayed ART initiation was measured by four items (e.g., ‘Failure to participate in treatment in time could lead to poor treatment effect’). The Perceived Severity Scale was formed by summing up individual item scores (1=strongly disagree to 5=strongly agree). Higher scores on the scale indicated that consequences of delayed ART initiation were perceived to be more severe.
Six items (from 1=strongly disagree to 5=strongly agree) were used to measure perceived benefit of immediate ART initiation (e.g., ‘Participate in ART in time can effectively improve CD4 cell counts’). The Perceived Benefit Scale was formed by summing up individual item scores. Higher score indicated that participants perceived more benefit for initiating ART immediately.
Perceived barriers of immediate ART initiation was measured by six items (e.g., ‘Early treatment means you have to suffer from the side-effects for a longer time’), with the following response categories (from 1=strongly disagree to 5=strongly agree). The Perceived Barrier Scale was formed by summing up individual item scores, higher scores indicated participants perceived more barriers to initiate ART immediately.
Cues to action were measured by 2 items (e.g., ‘People who are important to you (relatives or spouses) support you to participate in the treatment in a timely manner’). The Cue to Action Scale was formed by summing up individual item scores (1=strongly disagree to 5=strongly agree). Higher scores on the scale indicated more events or information from close others promoting ART initiation were perceived.
Five items were used to measure perceived self-efficacy of immediate ART initiation (e.g., ‘How confident you are about participating in ART’) (response categories: 1=none to 5=a great deal). The Perceived Self-Efficacy Scale was formed by summing up individual item scores, with higher score indicated perceived higher self-efficacy in initiating ART immediately. The Cronbach's α of these five scales ranged from 0.80 to 0.91.
Weighting of pros versus cons related to immediate ART initiation measured at baseline
Regarding weighting of pros versus cons related to immediate ART initiation, participants were asked to rank the overall advantages versus overall disadvantages for immediate ART (1= disadvantages prevail, 2= disadvantages and advantages are similar, 3=advantages prevail), and its benefit for themselves versus others (1=more benefits for oneself, 2=more benefits for others, 3=benefits for oneself and for others are similar, 4=not sure).
Psychosocial variables measured at baseline
Probable depression was measured by validated Chinese version of the Patient Health Questionnaire-9 (PHQ-9) [49], which has been widely used in studies targeting PLWH and MSM. Responses were reported by using a four-point Likert Scale ranging from 0 (rarely or none of the time) to 3 (most or all of the time). Scores of 0-4, 5-9, and ≥10 corresponded to minimal, mild and moderate-severe depressive symptoms. In this study, Cronbach’s alpha of the PHQ-9 was 0.901 [50].
Generalized anxiety disorder was measured by validated Chinese version of the 7-item Generalized Anxiety Disorder Scale (GAD-7) [51] (Response categories: from 0= rarely or none of the time to 3= most or all of the time). Scores of 0-4, 5-9, and ≥10 were defined as minimal, mild, and moderate-severe anxiety symptoms. In this study, Cronbach’s alpha of the GAD-7 was 0.947 [51].
Perceived social support was measured by the 12-item multidimensional scale of perceived social support (MPSSS) from three 4-item subscales (i.e., family (FA), friends (FR) and significant others (SO)) [52]. Each item is rated on a 7-point Likert-type response format, the possible scores range from 4 to 28 for each subscale [52]. Cronbach’s alpha of FA, FR and SO were 0.883, 0.844 and 0.834 respectively.
Public HIV-related stigma was measured by Chinese Courtesy Stigma Subscale (CCSSs) of Public HIV-related Stigma Scale [53, 54]. The subscale contains 13 items, with a 4-point ordinal response format ranging from 1 to 4. A lower score indicates the higher perceived public stigma [53]. Cronbach’s alpha of CSSSs was 0.925.
Self-stigma was measured by the 9-item short version of the Self-Stigma Scale (SSS-S) [55]. Each item is rated on a 5-point Likert-type response format ranging from 1 to 5. A higher score indicates a higher level of self-perceived stigma [55]. Cronbach’s alpha of SSS-S was 0.891.
Statistical analyses
Using background variables measured at baseline as independent variables, univariate Cox regression models were used to estimate hazard ratios (HRu) and respective 95% confidence interval (CI). Baseline background variables with p<0.05 in such univariate analysis were used as candidates for fitting a forward stepwise Cox regression. Adjusting for significant background variables in multivariate analysis, associations between independent variables of interest (i.e., perceptions based on the HBM, weighting of pros and cons of immediate ART, and psychosocial variables) and the dependent variable were then assessed by adjusted hazard ratios (HRa). Each of such HRa was obtained by a single Cox regression model, which involved one of the independent variables of interest and the significant background variables. Finally, a summary forward stepwise multivariate Cox regression model was fit, using all variables with P<0.2 in adjusted analysis as candidates. Multivariate hazard ratios (HRm) were obtained. Data were analyzed with SPSS 25.0, P-values smaller than 0.05 were regarded as statistically significance; P-values between 0.05 and 0.2 were considered as of marginal significance. The figures of the survival curves were performed using R (version 3.4.4).