Share Abstract
How have policies influenced PrEP uptake? An analysis of 139 countries and 33 African countries
Abstract Content:
BACKGROUND: Oral PrEP is at varying stages of scale-up globally, as longer-acting PrEP products enter the market in limited supply. This analysis evaluates which policies significantly enable PrEP uptake to understand how to expand access to PrEP options.
METHODS: Data on cumulative PrEP initiations (dependent variable) for 139 countries, as of September 2024, was collected from AVAC's Global PrEP Tracker. Data on the status of 54 policies (independent variable (IV)) related to HIV testing and prevention, structural changes, and health systems across all 139 countries, as of July 2024, was collected from the HIV Policy Lab. Multiple and individual regressions, using RStudio (2024.12.0+467), assessed the significance of policies altogether and individually (p=0.05). A separate analysis was done for 33 African countries. Correlation analyses of IVs assessed multicollinearity using the Pearson coefficient. Ridge regression analysis assessed if multicorrelated variables should be eliminated from the model to remove confounding in multiple regressions.
RESULTS: The global analysis showed that eight policies on task-shifting (1), decriminalizing same-sex sex (3), decriminalizing HIV exposure (3), and HIV self-testing (1) were significantly associated with PrEP uptake in individual regressions. Multiple regression showed only one policy on task-shifting is significantly associated with PrEP uptake.
The analysis of African countries showed that four policies decriminalizing HIV exposure (2), prohibiting compulsory HIV testing (1), and loosening age restrictions on testing and treatment (1) were significantly associated with PrEP uptake in individual regressions. None were significantly associated with PrEP uptake in multiple regression.
The discrepancy in the individual and multiple regressions indicates multicollinear IVs. Correlation analyses confirmed that at least four pairs of policies are highly correlated in enabling PrEP uptake. Ridge regression results showed that these same variables were very strongly contributing to the predictive power of the model and cannot be eliminated.
CONCLUSIONS: Policies found to significantly enable PrEP uptake making access easier, reduce discrimination, and align with user preferences. While the effect of individual policies is reduced when considering several altogether, they influence PrEP uptake both individually and systemically. Differences in policies that significantly enable PrEP uptake at global and regional levels imply that context affects which policies should be implemented.
METHODS: Data on cumulative PrEP initiations (dependent variable) for 139 countries, as of September 2024, was collected from AVAC's Global PrEP Tracker. Data on the status of 54 policies (independent variable (IV)) related to HIV testing and prevention, structural changes, and health systems across all 139 countries, as of July 2024, was collected from the HIV Policy Lab. Multiple and individual regressions, using RStudio (2024.12.0+467), assessed the significance of policies altogether and individually (p=0.05). A separate analysis was done for 33 African countries. Correlation analyses of IVs assessed multicollinearity using the Pearson coefficient. Ridge regression analysis assessed if multicorrelated variables should be eliminated from the model to remove confounding in multiple regressions.
RESULTS: The global analysis showed that eight policies on task-shifting (1), decriminalizing same-sex sex (3), decriminalizing HIV exposure (3), and HIV self-testing (1) were significantly associated with PrEP uptake in individual regressions. Multiple regression showed only one policy on task-shifting is significantly associated with PrEP uptake.
The analysis of African countries showed that four policies decriminalizing HIV exposure (2), prohibiting compulsory HIV testing (1), and loosening age restrictions on testing and treatment (1) were significantly associated with PrEP uptake in individual regressions. None were significantly associated with PrEP uptake in multiple regression.
The discrepancy in the individual and multiple regressions indicates multicollinear IVs. Correlation analyses confirmed that at least four pairs of policies are highly correlated in enabling PrEP uptake. Ridge regression results showed that these same variables were very strongly contributing to the predictive power of the model and cannot be eliminated.
CONCLUSIONS: Policies found to significantly enable PrEP uptake making access easier, reduce discrimination, and align with user preferences. While the effect of individual policies is reduced when considering several altogether, they influence PrEP uptake both individually and systemically. Differences in policies that significantly enable PrEP uptake at global and regional levels imply that context affects which policies should be implemented.