Study Design
This cross-sectional study employing quantitative methods was part of a pilot quasi-experimental study seeking to evaluate the acceptability, appropriateness, and feasibility of group therapy for the management of stigma among pregnant adolescents living with HIV/AIDS in Northern Uganda. The study was conducted between August and November 2023. We report the baseline burden and factors associated with stigma in this population.
Study Setting
The study was conducted at Gulu Regional Referral Hospital (GRRH) and St. Mary’s Hospital, Lacor (SMHL), both located in Gulu City, Northern Uganda. GRRH is the largest public health facility in the district and serves as the referral hospital for all 11 districts in the Acholi sub-region. SMHL is the largest church-based not-for-profit hospital in Northern Uganda. Both hospitals provide healthcare for persons living in Gulu and the nearby districts of Omoro, Pader, Amuru, West Nile, and some parts of Southern Sudan. These hospitals offer a comprehensive range of services typical of regional referral hospitals within the Ugandan healthcare system. This includes general and specialized medical care, pediatric services, surgery, obstetrics, and gynecology. Additionally, each hospital houses a functional HIV clinic and delivers services for the prevention of mother-to-child transmission (PMTCT) of HIV/AIDS, catering to pregnant adolescents and adults. SMHL antiretroviral therapy (ART) clinic offers care to over 8,000 clients, 1000 of whom are adolescent girls. The infectious disease clinic (IDC) at GRRH has over 10,000 registered clinics, 1000 of whom are adolescent girls.
Study Population
The study was conducted among pregnant adolescents living with HIV, attending ART or antenatal clinics, or seeking other healthcare services at GRRH and SMHL.
Selection Criteria
We included female adolescents aged 10–19 years of age, with a confirmed diagnosis of HIV/AIDS according to the Ministry of Health guidelines, and a confirmed pregnancy either by urine or serum HCG, or ultrasonography, seeking care or attending ART/antenatal clinic at Gulu Regional Referral Hospital or St. Mary’s Hospital Lacor, with a written informed consent to participate in the study. Pregnant ALWHIV with critical illnesses requiring immediate medical attention or those with severe mental illnesses and unable to respond to the interviews were excluded.
Sample Size and Sampling Criteria
The sample size for this pilot study was determined using the methodology outlined by Whitehead et al. [63], specifically designed for pilot studies with continuous numerical outcomes. This approach, described in detail in the publication, considers several factors, including the anticipated standardized differences within the study groups, the power of the study, and the projected sample size for the subsequent main trial [63]. At 80% power, the required sample size for each arm of a pilot study using the non-centrality t-distribution approach is 20 [63]. Therefore, the sample size required for the two arms in this pilot study was 40, with 20 from each hospital. Study participants were recruited consecutively until the target sample size was reached.
Study Variables
The primary outcome variables included HIV-related stigma and stigma to adolescent pregnancy. The secondary outcome variables were the sub-construct of the stigma experienced and the impact of stigma on HIV/AIDS and pregnancy-related outcomes. The independent variables included the participants’ sociodemographic data (age, marital status, parental history, social support, schooling status, level of education, employment status, household income, nature of residence, and transportation). We also collected obstetric history (parity, age at first birth, gestational age, and antenatal care visits) and HIV/AIDS history (age at diagnosis of HIV/AIDS, mode acquisition, ART initiation, adherence to ART, last viral load and CD4 counts)
Data Collection Tools and Procedure
A semi-structured questionnaire consisting of 70 questions with sections on demographics, HIV-related history, pregnancy-related history, HIV stigma, and pregnancy-related stigma was used to collect data. HIV stigma was assessed using the HIV-Stigma scale for children (HSS-10) [64]. This scale, consisting of 10 Likert-style items, is a concise version of the original 40-item Bergen HIV Stigma Scale, which has undergone validation among adolescents living with HIV/AIDS in Kenya. The HSS-10 provides a four-point response scale (strongly disagree, disagree, agree, strongly agree), with individual question scores ranging from 1 to 4, leading to an aggregate score between 10 and 40. The questionnaire had a high internal consistency/reliability level in assessing HIV stigma in our study (Cronbach alpha = 0.82).
For measuring pregnancy-related stigma, the Adolescent Pregnancy-Related Stigma Scale (APSS-15) was used. This 15-item Likert scale was developed through an extensive literature review, consultation with experts, and input from the nursing/midwifery staff at participating hospitals. The APSS-15 provides a five-point response scale for each item (strongly disagree, disagree, neutral, agree, strongly agree), with scores ranging from 1 to 5, leading to a total range of 15 to 75. The questionnaire had a high internal consistency/reliability level in assessing teenage pregnancy stigma in our study (Cronbach alpha = 0.88).
For linguistic accuracy and cultural relevance, the questionnaire was translated into Acholi and then back into English by two independent language experts, ensuring consistency and comprehensibility. Two midwives, experienced in working with pregnant adolescents and with sufficient training in conducting surveys, were recruited as research assistants for this study. Before data collection, the research assistants received training to ensure they were well-prepared for data collection. The interviews were primarily conducted in Luo, the most spoken language in the district. However, in special circumstances where participants spoke different languages, their preferred language was used to facilitate clear and effective communication.
Data collection was carried out using a digital format for efficiency and accuracy. The questionnaire was developed on Enketo webforms, a versatile tool for creating user-friendly digital surveys. This digital questionnaire was then deployed on the Kobo Toolbox Platform, a robust and secure data collection software developed by the Kobo Organization based in Cambridge, Massachusetts, USA.
Data analysis
Data was exported from the Kobo Toolbox cloud to STATA 18.0 Statistical Software (StataCorp LLC, College Station, Texas, USA) and R Software for cleaning, coding, and analysis. The responses from the Likert scales were converted into numerical scores (ranging from 1 for 'strongly disagree' to 5 for 'strongly agree'), and the total stigma scores for each participant were calculated by summing these scores for both the HSS-10 and APSS-15. categorical variables were presented as frequencies and percentages, while continuous variables were tested for normal distribution using the Shapiro-Wilk test. Depending on their distribution, they were then presented as means with standard deviations or medians with interquartile ranges. Notably, HSS-10 scores exhibited a non-normal distribution, while APSS-15 scores were normally distributed. The median score and interquartile range (IQR) for each of the two scores were then summarized. For comparative analysis between HIV stigma and adolescent pregnancy stigma, both scores were normalized to a scale of 0 to 100% using a standardized formula below.
$$Normalized score = \frac{Original score-Min}{Max-Min}$$
where Min was the minimum value (10 for HSS-10, and 15 for APSS-15, whereas Max was the maximum value (40 for HSS-10 and 75 for APSS-15).
Simple and multivariable linear regression models were used to assess factors associated with HIV and adolescent pregnancy stigma. Participants' level of education, living with a spouse, rural vs. urban residence, having a child, delayed ART initiation, history of skipping ART, and mode of HIV acquisition were included to construct the multivariable linear regression models. The multivariable linear regression models for pregnancy stigma incorporated factors like living with a spouse, nature of residence, level of education, sources of emotional and social support, parity, gestational age, delayed ART initiation, and skipping ART. The selection of variables for inclusion in the models was guided by literature and expert opinion. A p-value of less than 0.05 was considered statistically significant.