Scope and Design
This cross-sectional study was conducted in the informal settlements of Kampala, the largest urban centre and capital city of Uganda. Kampala’s population is estimated at 1.5 million, 27.5% of whom are aged between 15-24 years (28). The city has five metropolitan administrative divisions -Nakawa, Rubaga, Makindye, Central and Kawempe, and is home to Uganda’s national referral hospitals, including Mulago national specialized hospital, Kiruddu hospital, Kawempe hospital and Butabika national mental referral hospital.
Sampling
The sample size was calculated using the Kish Leslie formula for cross-sectional studies (29). Since there was limited evidence on the prevalence of Hepatitis B screening or completion of the Hepatitis B vaccination schedule among young psychoactive substance users in informal settlements, we chose a conservative prevalence of 50% (30), a 95% level of confidence, a margin of error (d) of 0.05 and a design effect of 2.0 (30, 31) was used to determine the sample size. This yielded a sample size of 768 young psychoactive substance users. A total of 12 informal settlements were purposively selected for geographical representation of the informal settlements in the city. The settlements included Kinawataka; Luzira; Luzira-Kirombe; Kitintale, Nalukolongo; Wankulukuku-Kabowa; Kamwokya; Bwaise; Katanga; Katwe-Kinyoro; Namuwongo-Soweto and Kyebando.
After the purposive selection of informal settlements, respondent driven sampling was used in the selection of study participants. For each of the informal settlements, we used community leaders who had participated in a previous study in the informal settlements to identify four individuals who acted as primary seeds (32). At the time of enrolment of primary seeds, research assistants made sure that the selected individuals were not under the influence of psychoactive substances. The selected seeds were first interviewed by research assistants prior to being given coupons to enroll secondary seeds. The secondary seeds were then requested by the primary seeds to report at an agreed venue to be interviewed after providing informed consent. The detailed procedure of undertaking respondent driven sampling is documented by (31).
Eligibility criteria
Only young psychoactive substance users aged 18 years and above were interviewed. These must have stayed in the informal settlement for at least 6 months. All young psychoactive substance users who were sick or under the influence of psychoactive substances were not interviewed.
Variable measurement
The main outcomes of interest in this study were having undergone hepatitis B screening in the last 12 months and completion of the Hepatitis B vaccination schedule based on the WHO recommendation of 3 vaccine doses (33). However, the completion was irrespective of the timing of the vaccinations. Self-reports were used to measure completion of the vaccination schedule.
The independent variables included the respondents’ socio-demographic characteristics such as sex, age, history of substance use, level of education, and their knowledge and attitude towards Hepatitis B prevention strategies. Knowledge was assessed using questions on the recommended Hepatitis B vaccine dose and the duration the vaccine protects someone against the Hepatitis B infection. Attitude was measured using a question on the perceived efficacy of Hepatitis B vaccine. History of substance use was classified as “ever used’ which referred to lifetime use of a psychoactive substance; ‘recent use’ which referred to having used a psychoactive substance in the last 12 months and ‘current use’ referring to the use of a psychoactive substance in the last 30 days.
Data management and analysis
A structured questionnaire was designed using the kobo tool box online platform and later preloaded onto the kobo collect mobile application. The data collection tools were designed with skips to reduce errors by research assistants. Prior to data collection, all research assistants received training on the study protocol and data collection. All study tools were translated into the local language and thereafter pretested. Data were collected using smart phones and tablets, and later uploaded to an online server at; https://kobo.humanitarianresponse.info. Upon submission, the data were reviewed on a daily basis by the principal investigators as a means of ensuring quality control. Prior to data analysis, data were downloaded in a Microsoft Excel format and further cleaned to reduce any possible errors. Measures of central tendency such as means, median and mode were particularly used to identify errors in the continuous variables.
Data analysis was done using STATA version 14.0. Descriptive statistics were performed to summarize both continuous and categorical variables (background characteristics of respondents, history of substance use, prevalence of hepatitis B screening and completion of the Hepatitis B vaccination schedule.
Inferential statistics were used to determine the factors associated with Hepatitis B screening, and completion of the Hepatitis B vaccination schedule. Modified Poisson regression analysis was used to determine the factors associated with Hepatitis B screening since the prevalence of screening for Hepatitis B was greater than 10% (34, 35). Bivariate analysis was done first to establish the association between predictor variables and screening for Hepatitis B. A cut off p-value of less than 0.2 was set for variables eligible to be included in the multivariable model (36). Prevalence ratios (PR) and their corresponding 95% confidence intervals were used as the measure of risk.
Given that the prevalence of completion of the Hepatitis B vaccination schedule was a rare occurrence (less than 10%) among young psychoactive substance users, we used logistic regression to determine the predictors. Initially, bivariate logistic regression was used to determine the predictors of completion of the hepatitis B vaccination schedule. Predictors that had a p-value of less than 0.2 were included in the multivariable model. Odds Ratios (OR) were used as the appropriate measure of risk.