Health worker perceptions of stigma towards Zambian adolescent girls and young women: a qualitative study

Background: The high prevalence of HIV among adolescent girls and young women aged 15–24 in Eastern and Southern Africa indicates a substantial need for accessible HIV prevention and treatment services in this population. Amidst this need, Zambia has yet to meet global testing and treatment targets among adolescent girls and young women living with HIV. Increasing access to timely, high-quality HIV services in this population requires addressing the intensified anticipated and experienced stigma that adolescent girls and young women often face when seeking HIV care, particularly stigma in the health facility setting. To better understand the multi-level drivers and manifestations of health facility stigma, we explored health workers’ perceptions of clinic- and community-level stigma against adolescent girls and young women seeking sexual and reproductive health, including HIV, services in Lusaka, Zambia. Methods: We conducted 18 in-depth interviews in August 2020 with clinical and non-clinical health workers across six health facilities in urban and peri-urban Lusaka. Data were coded in Dedoose and thematically analyzed. Results: Health workers reported observing manifestations of stigma driven by attitudes, awareness, and institutional environment. Clinic-level stigma often mirrored community-level stigma. Health workers clearly described the negative impacts of stigma for adolescent girls and young women and seemed to generally express a desire to avoid stigmatization. Despite this lack of intent to stigmatize, results suggest that community influence perpetuates a lingering presence of stigma, although often unrecognized and unintended, in health workers and clinics. Conclusions: These findings demonstrate the overlap in health workers’ clinic and community roles and suggest the need for multi-level stigma-reduction approaches that address the influence of community norms on health facility stigma. Stigma-reduction interventions should aim to move beyond fostering basic knowledge about stigma to encouraging critical thinking about internal beliefs and community influence and how these may manifest, often unconsciously, in service delivery to adolescent girls and young women. Data were collected from health workers across six public health facilities participating in the parent study [23]. All facilities are in the densely populated, primarily low-income urban and peri-urban areas of Lusaka, Zambia, with an estimated population of 2 million people and 190 public or private health facilities. We conducted 18 in-depth interviews (IDIs) (15 females/three males), ranging from around 30–90 min, with 14 clinical and four non-clinical staff across the six facilities. Eligibility criteria included being aged 18 or older and working in the study health facilities for at least 6 months in a position, whether clinical or non-clinical, that interacts with AGYW clients. The study team and a Community Advisory Board member identified potential participants by liaising with the leadership from each facility to nominate three clinic staff from different departments who they assessed would be knowledgeable and actively engaged informants. Selected participants included clinic in-charges (heads of departments), community health workers, and adolescent focal point persons, as well as employees from three departments where AGYW are commonly seen—Maternal and Child Health, labor ward, and registration. Duration of respondents’ service in the health sector ranged from 2.5 to 20 years. IDIs were conducted in August 2020 by three experienced Zambian female qualitative interviewers in a private setting using a semi-structured guide developed for the study (Supplementary File 1). IDI topics focused on eliciting health workers’ perceptions of the drivers, manifestations, and impacts of stigma towards AGYW in both the clinic and the community. For example, interviewers asked questions such as ‘Do you feel stigma is an issue for adolescent girls and unmarried young women who are having sex?’ and further probed about what this stigma looked like, who it was coming from, and how this stigma manifested when seeking care at the clinic. Interviewers participated in a 3-day training that included content related to stigma, practice with the interview guides, a refresher on qualitative methods, and research ethics. Within 1 day of each IDI, the interviewer summarized the main findings in a debrief report, which is a short, structured form designed to highlight key concepts [24]. Each debrief report was reviewed by the senior co-investigator leading this sub-study to provide rapid feedback on interviewing techniques and probing approaches. Interviewers engaged in debriefing sessions regularly throughout the data collection period to share and incorporate this feedback. All interviews were conducted in English, audio recorded and transcribed verbatim, and then checked for quality by the field research team lead. After close reading of transcripts and debrief reports, the research team developed a codebook capturing deductive (i.e., from topics probed in the semi-structured guides) and inductive (i.e., emerging from participant narratives) themes. Two trained analysts applied codes to text segments from IDI transcripts using the Dedoose (2019) web application. At the beginning of coding, each team member independently coded one transcript using the preliminary codebook, and the coding results were compared to reach consensus on how codes should be applied, the adequacy of code definitions, and the completeness of the codebook. The codebook was revised, and this process was repeated with one additional transcript. Coding discrepancies were reviewed and resolved during weekly quality control meetings, where the team also discussed any questions and emerging themes. After each meeting, the codebook was revised to reflect any changes, and coding of previous transcripts was updated as needed to reflect these changes. An inter-rater reliability test of several key concepts resulted in a Cohen’s Kappa score of 0.75, which is considered to indicate substantial agreement between coders [25]. After coding all transcripts, analysts reviewed coded text segments, identifying salient themes and patterns to aid data synthesis and interpretation. Ethical clearance was provided by three review boards: the ERES Converge Research Ethics Committee in Zambia as well as the Institutional Review Boards at the Population Council and the University of North Carolina-Chapel Hill. Interviewers provided study details and obtained written consent from all participants before initiating the interviews.