Expanding the use of observational data to evaluate the impact of DREAMS, a non-randomised complex intervention
Professor Sian Floyd at LSHTM
Dr Isolde Birdthistle at LSHTM
Persistently high rates of HIV infection among adolescent girls and young women (AGYW) have led to large investments in targeted HIV prevention in eastern and southern Africa. This includes the ‘DREAMS Partnership’ funded by PEPFAR during 2016-2023 to reduce HIV incidence through a complex intervention addressing individual, family, and community-based drivers of young women’s risk. LSHTM, APHRC, and KEMRI have co-led an impact evaluation of DREAMS during 2017-2023 to evaluate the impact of DREAMS in one rural and one urban setting in Kenya. Using data from two population-based cohort studies, each one enrolling over 1000 AGYW during 2017-2018, this project will use causal inference methods to evaluate the impact of DREAMS on sexual and reproductive health, and social and economic outcomes, by 2022. Existing work has used “intention-to treat” analyses and inverse-probability-of-treatment-weighting (IPTW) to compare outcomes between AGYW who were, or were not, invited to participate in DREAMS. This PhD project will extend previous work to include per-protocol analyses, “before-after DREAMS” analyses of population-wide change by comparing outcomes at cohort enrolment among an older birth cohort with 2022 outcomes among a younger birth cohort, and the application of alternative methods of analysis to control for time-dependent confounding. The project will contribute to the growing use of observational data to evaluate the impact of non-randomised interventions like DREAMS.
Project Key Words
DREAMS partnership, impact evaluation, causal inference
MRC LID Themes
- Global Health = Yes
- Health Data Science = Yes
- Infectious Disease = Yes
- Translational and Implementation Research = No
MRC Core Skills
- Quantitative skills = Yes
- Interdisciplinary skills = Yes
- Whole organism physiology = No
Skills we expect a student to develop/acquire whilst pursuing this project
Causal inference methods for evaluating the impact of an intervention using observational longitudinal data, including methods to control for time-dependent confounding
Which route/s is this project available for?
- 1+4 = Yes
- +4 = Yes
Possible Master’s programme options identified by supervisory team for 1+4 applicants:
- LSHTM – MSc Health Data Science
- LSHTM – MSc Medical Statistics
Is this project available for full-time study? Yes
Is this project available for part-time study? Yes
Particular prior educational requirements for a student undertaking this project
- LSHTM’s standard institutional eligibility criteria for doctoral study.
- MSc Medical Statistics or closely-related discipline
Other useful information
- Potential CASE conversion? = No
PROJECT IN MORE DETAIL
Scientific description of this research project
These are to evaluate the impact of DREAMS (Determined-Resilient-Empowered-AIDS-free-Mentored-Safe) interventions on sexual and reproductive health, and social and economic outcomes, among adolescent girls and young women (AGYW) in Kenya, after 6 years of implementation (2016-2022). Specifically, the project will use a range of analytical approaches to estimate the impact of a non-randomised programme. DREAMS is a PEPFAR-funded package of interventions delivered in >100 high-HIV-prevalence districts in 15 countries, with an overarching goal to reduce HIV incidence. The impact of DREAMS will be measured among two evaluation cohorts that were established from randomly-selected samples of AGYW in urban (informal settlements in Nairobi) and rural (Gem, in Siaya county) settings. In Nairobi, 1,081 AGYW aged 15-22 years were enrolled to a cohort study in 2017, and followed up in 2018, 2019, and 2022. In Gem, 1,171 AGYW aged 13-22 years were enrolled to a cohort study in 2018, and followed up in 2019 and 2022. Among those enrolled, 88% in Gem and 84% in Nairobi were included in the 2022 follow-up.
Techniques to be used
To date, analyses of the impact of DREAMS on outcomes measured in 2022 have used propensity scores and inverse-probability-of-treatment-weighting (IPTW) to control for time-dependent confounding, and inverse-probability-of-censoring weights to account for loss to follow-up. These analyses have used an “intention-to-treat” approach, comparing AGYW who were invited to participate in DREAMS by 2022 with those who were not invited to participate in DREAMS by 2022. DREAMS invitees have been distinguished as early invitees (invited during 2016-2018) or later invitees (invited during 2020-2022). This project will build on existing work, extending it to include per-protocol analyses (categorising DREAMS invitees according to whether or not they received interventions as intended), “before-after DREAMS” analyses of population-wide change by comparing outcomes at cohort enrolment among an older birth cohort with 2022 outcomes among a younger birth cohort, and using alternative methods of analysis (beyond IPTW) to control for time-dependent confounding.
Confirmed availability of any required databases or specialist materials
Cohort study data, collected during 2017-2022, are available from an impact evaluation of DREAMS conducted during 2016-2023 in a collaboration between LSHTM, the African Population and Health Research Centre (APHRC), and the Kenya Medical Research Institute (KEMRI). LSHTM has a non-exclusive shared license, together with APHRC and KEMRI, to analyse these data.
Potential risks to the project and plans for their mitigation
Data collection (2017-2022) has been completed; there is no risk that analyses cannot be done. Cleaned and documented datasets are available at LSHTM. Although LSHTM shares a non-exclusive licence for the datasets, APHRC and KEMRI are the primary owners so the principal investigators at these two institutions will be asked to permit use of the datasets by a PhD student. Based on the collaboration and previous PhD projects, permission is expected to be given. Also, the funder of the impact evaluation (Bill and Melinda Gates Foundation) requires datasets to be made available to other researchers within 2 years of databases being locked. Other hurdles are not expected.
(Relevant preprints and/or open access articles)
Additional information from the supervisory team
- The supervisory team has provided a recording for prospective applicants who are interested in their project. This recording should be watched before any discussions begin with the supervisory team.