Study design and participants
The study was conducted in Guangxi Zhuang Autonomous Region in southwest China (Guangxi). Guangxi has been one of the regions with the highest HIV disease burden in China [19]. This cross-sectional study was conducted between July and August 2021. HIV-infected children were recruited from the pediatric HIV clinic of the Guangxi Center for Disease Control and Prevention (Guangxi CDC). Children who met the following eligibility criteria were recruited for this study: (1) HIV positive; (2) being 11 and 17 years old at the study recruitment; (3) having initiated ART for at least 6 months; and (4) being willing to participant in the study and providing written informed consent from themselves and their parents or guardians. This study did not include HIV-infected children who were admitted to an inpatient care or with serious physical injury.
Data Collection
Baseline characteristics: Trained HIV clinic staff provided guidance to participants who completed a 10-minute questionnaire for data collection. The questionnaire consisted of three parts: BEDs, physical health, and social support. Upon completion of the questionnaire, blood sample was collected by a nurse for viral load and CD4 cell count testing. Transportation compensation of 30 RMB (approximately $5 USD) was provided to each participant. The clinic staff informed all participants and their parents or guardians the blood test results when the results became available.
All HIV positive cases were required to report through a national surveillance portal in China, the National Free Antiretroviral Treatment Program (NFATP) database. We linked questionnaire data to NFATP database to collect the following information: sociodemographic data (age, sex, and route of HIV infection), HIV clinical characteristics (WHO clinical stage before ART), laboratory data (CD4 cell count before ART), HIV care information (age at ART initiation, year of ART initiation, initial ART regimen, and time on ART).
Behavioral and emotional difficulties (BEDs): BEDs were assessed using the Chinese version of the self-reported Strengths and Difficulties Questionnaire (SDQ-C), which was publicly available on the official website of SDQ (http://www.sdqinfo.com). This tool has been validated in China and other countries and has been used in HIV-infected children in previous studies [18, 20]. The SDQ-C contains 25 items, which were categorized into five subscales: emotional symptoms, conduct problems, hyperactivity/inattention, peer problems, and prosocial behavior. The severity level of each item was rated on a 3-point scale (0 = not true, 1 = somewhat true, and 2 = certainly true). A questionnaire with one or more unanswered questions was considered invalid, therefore, were not included in the analysis. The total difficulties score was calculated by summing up the four subscale scores (emotional symptoms, conduct problems, hyperactivity/inattention, and peer problems). In general, a higher score represents greater difficulties for the total difficulty score and the four subscales. While for the prosocial subscale, a lower score indicates greater difficulties. We also calculated Cronbach’s alpha to measure the internal consistency of subscales. In our sample, Cronbach’s alpha scores for the total difficulties was 0.75 and were 0.50 and 0.72, respectively for emotional symptoms and hyperactivity/inattention subscales. The Cronbach’s alpha was comparably low for conduct problems (a = 0.41) and peer problems (a = 0.11) subscales. In addition, the proportions of HIV-infected children in the ‘normal’, ‘borderline’, and ‘abnormal’ categories were calculated respectively using the norms for Chinese children (Shanghai norms) [21].
Physical health
The Chinese version of the Medical Outcomes Study HIV Health Survey (MOS-HIV), which was translated by Yu et al. [22], showed good reliability and validity in Chinese people living with HIV [23]. In this study, physical health was measured using the sections of the MOS-HIV on physical function and general health perceptions. The scores of all domains were obtained by an established calculation method and then transformed into a 0-100 scale [24]. A higher score represents better physical functioning and well-being. For multi-item scales, mean substitution was generally used for missing items if the missing items were ≤ 50%. In our sample, Cronbach’s alpha scores were good for the physical functioning score (a = 0.81) and the general health score (a = 0.57).
Social support
The questionnaire collected the information of social support participants have received. Participants reported the frequency of having received assistance and support in the past 3 months from parents and other various sources including siblings, relatives, neighbors, friends, medical and nursing staff/CDC staff, teachers, fellow students, net friends, and other HIV-infected patients. The frequency of receiving support from each source was rated on a 5-point scale (1 = never, 2 = occasionally, 3 = sometimes, 4 = often, and 5 = always).
HIV treatment outcomes
ART medication adherence and virological indicators were the treatment outcomes. Adherence was measured by children’s answers on a question about whether they had missed doses in the past month. Virological failure was defined as HIV-1 RNA viral load of ≥ 50 copies/ml using the blood test result for this study [25].
Laboratory Tests
All viral load and CD4 cell count tests were conducted in the Guangxi CDC laboratory. Viral load was tested using the COBAS TaqMan 96 system (Roche Molecular Systems, Branchburg, NJ, USA) and CD4 cell count was assessed using the Alere PimaTM Analyzer (Abbott Laboratories, Germany).
Statistical analysis
All paper-based survey data were entered into EpiData software version 3.1 (The EpiData Association, Odense, Denmark) and two research staff cross-checked for errors. Statistical analyses were conducted using SAS V9.4 (SAS Institute Inc., Cary, NC, USA). The internal consistency of the questionnaire was assessed using Cronbach’s α coefficient. We reported the means and standard deviations (SD) for continuous variables. Comparisons were conducted using independent T-tests for continuous variables and Chi-square tests for categorical variables. Univariate and multivariate logistic regression models were used to investigate potential factors that were associated with missed doses in the past month and virological failure. Further, we constructed a multivariate logistic regression model stepwise to select variables independently associated with missed doses in the past month and virological failure. All tests were two-tailed, and a P-value of < 0.05 was considered statistically significant.