# Computational Validation of a Clinical Decision Support Algorithm for LAI-PrEP Bridge Period Navigation at UNAIDS PrEP Target Scale (21.2 Million Individuals)

## TL;DR

A new tool helps improve the success rate of HIV prevention injections by predicting and addressing barriers during the critical initial period.

## Contribution

A clinical decision support tool with external configuration architecture for LAI-PrEP bridge period navigation is developed and validated at UNAIDS scale.

## Key findings

- The tool improved bridge period success rates from 23.96% to 43.50% at UNAIDS target scale.
- Evidence-based interventions could prevent approximately 100,000 HIV infections annually.
- Vulnerable populations saw disproportionately higher relative improvements in success rates.

## Abstract

Long-acting injectable HIV pre-exposure prophylaxis (LAI-PrEP) demonstrates superior efficacy to oral PrEP but faces a critical implementation challenge: 47% of patients fail to receive their first injection during the “bridge period” between prescription and initiation. We developed a clinical decision support tool with an external configuration architecture synthesizing evidence from major LAI-PrEP trials (HPTN 083, HPTN 084, PURPOSE) and implementation studies. The tool provides population-specific risk stratification, barrier identification, and evidence-based intervention recommendations from a library of 21 interventions with mechanism diversity scoring to prevent redundant recommendations. We conducted progressive validation on four scales: 1000 (functional), 1,000,000 (large-scale), 10,000,000 (ultra-large-scale) and 21,200,000 patients (UNAIDS PrEP target), with comprehensive unit testing achieving a test pass rate of 100% (18/18 edge cases). Progressive validation demonstrated convergence and increasing precision: 1K (±2.6 pp), 1M (±0.09 pp), 10M (±0.028 pp), and 21.2M (±0.018 pp). At UNAIDS 2025 PrEP target (21.2 million) scale, the tool predicted baseline bridge period success rate of 23.96% (95% CI: 23.94–23.98%), with evidence-based interventions improving success to 43.50% (95% CI: 43.48–43.52%)—an absolute improvement of 19.54 pp (or 81.6% relative improvement), representing 4.1 million additional successful transitions globally. Population disparities were substantial: People who inject drugs (PWID) showed 10.36% baseline success versus 33.11% for men who have sex with men (MSM)—a 22.75 pp gap. Regional disparities were equally significant: Sub-Saharan Africa (serving 62% of global patients) achieved 21.69% baseline versus 29.33% in Europe/Central Asia—a 7.64 pp gap. However, evidence-based interventions disproportionately benefited vulnerable populations. PWID experienced +265% relative improvement, and adolescents experienced +147% relative improvement, demonstrating that systematic implementation support can narrow rather than widen health equity gaps at UNAIDS 2025 PrEP target (21.2 million) scale. The tool demonstrates predictive validity with policy-grade statistical precision. Using published epidemiologic parameters (HIV incidence 2–5% among indicated users, LAI-PrEP efficacy 96%), our model translates the 4.1 million additional successful transitions into approximately 80,000–100,000 prevented HIV infections annually (midpoint: 100,000), corresponding to an estimated USD 40 billion in averted lifetime treatment costs.