Organization: Health Economics and Epidemiology Research Office, Queen’s University, Yale University and University of Arizona
Study objectives: This study aims to identify communication strategies and delivery configurations most likely to increase HPV vaccine uptake among school-aged children in South Africa. Using a discrete choice experiment, it aims to solicit preferences from parents and caregivers for communication material for boys and preferences for delivery methods for boys and girls. It also aims to adapt communication materials with parents and caregivers, using a design thinking approach.
Study design description: The study will use a mixed-methods, sequential design approach, comprising three components: a pre-DCE survey, two discrete choice experiments (DCEs), and a human-centred design (HCD) workshop to identify optimal HPV vaccine communication and delivery strategies for South African schoolchildren aged 9–17 years.
The pre-DCE survey (Component 1), administered via REDCap immediately before DCE tasks, collects child details, HPV awareness, and attitudinal measures from 800 caregivers (400 per DCE arm). It screens eligibility (parents and caregivers of boys-only for DCE 1; parents and caregivers of boys and girls for DCE 2), characterises baseline knowledge, and enables subgroup analyses by child sex and parental characteristics, ensuring seamless transition into experimental tasks.
Within each DCE arm, participants are then assigned to one of several pre‑defined choice‑set blocks (8–10 choice sets per block, with an uptake opt‑out option and a consistency‑repeat task), ensuring balanced exposure to attribute levels across the sample. Attributes, derived from formative research, the Three Cs model, and pilot testing with parents and caregivers (n = 30), are embedded in Stata‑generated D‑efficient fractional factorial designs. The HCD workshop (Component 3; n = 15 via the Indlela B‑Hub) prototypes the top‑performing DCE profiles into culturally appropriate communication materials (e.g., messages, posters) for feasibility and acceptability testing with parents and caregivers
Study outcomes: The study will identify optimal HPV vaccine communication strategies and delivery configurations to maximise uptake among South African schoolchildren aged 9–17 years (prioritising boys), through ranked parental/caregiver preferences from two targeted discrete choice experiments (DCEs; n=800) capturing trade-offs across key attributes. These insights will yield prototype-ready, culturally adapted materials (e.g., WhatsApp messages, posters) and delivery protocols via a human-centred design (HCD) workshop, validated for acceptability, usability, and scalability within school-based vaccination programmes
Study site: This fully online study for Components 1 and 2 has no physical sites for the pre-DCE survey and DCEs, utilising digital platforms for national recruitment across South Africa. Prolific Academic screens and randomises 800 parents/caregivers of children aged 9–17 years (400 per DCE arm) based on residence, age, parental status, and approval rating (>80%), providing basic demographics. REDCap, hosted by the University of the Witwatersrand/HE2RO, captures anonymous electronic consent, surveys, and DCE instruments with branching logic and Stata export compatibility. Component 3 (HCD workshop; n=15) will be conducted on-site in Johannesburg via the B-Hub within the Indlela Behavioural Insights for Better Health unit at the Health Economics and Epidemiology Research Office (HE2RO), University of the Witwatersrand. This will enable an iterative prototyping process with local parents/caregivers.
Target population: Parents and caregivers of school going boys and girls, aged 9-17 years.
Sample size: The study plans a total sample of 800 parents/caregivers for Components 1 and 2 (pre-DCE survey and DCEs), with 400 per arm exceeding the Johnson-Orme minimum threshold of +-100 (based on 10 attributes × 4 levels max, 8–10 choice sets, 2 alternatives) and ISPOR guidance for precise utility estimates (>300/arm), enabling robust subgroup analyses by child sex and socio-demographics. For Component 3 (HCD workshop), a standard sample of 15 participants ensures optimal engagement for prototyping and refinement.
Analysis: The analysis for the discrete choice experiment (DCE) components will focus on estimating how parents’ preferences for HPV vaccination communication and delivery are shaped by different attribute levels. Using random‑utility theory, choices will first be modelled with a multinomial logit framework to estimate utility coefficients for each attribute (e.g., mode of delivery, message framing, benefits framing, and linkage to care), and then extended to mixed logit or latent‑class models if significant preference heterogeneity is detected. Relative attribute importance will be calculated to show which aspects of communication and delivery drive decisions most strongly, and, where appropriate, willingness‑to‑pay‑type trade‑offs will be derived for attributes that reflect cost or burden (e.g., private‑vaccine booking arrangements). Subgroup analyses by child sex, parental education, SES, and prior HPV knowledge will examine how preferences vary across key population segments, supported by sensitivity checks on model specification and response quality.
For the pre‑DCE survey and human‑centred design (HCD) components, analysis will be largely descriptive and qualitative. The pre‑DCE survey data will be summarised using frequencies, percentages, and basic descriptive statistics to characterise parental demographics, HPV knowledge, and vaccination attitudes, and to contextualise the DCE results. The HCD workshop will be analysed thematically, with workshop notes, feedback on prototypes, and design discussions coded to identify key themes around acceptability, usability, and feasibility of the HPV communication materials. Together, these analyses will link quantitative DCE‑derived preferences with qualitative insights from parents, providing a coherent evidence base for refining HPV vaccination communication and delivery strategies in the South African school‑based programme.
Duration: 3 months
