| Background: Since 1999 WHO has tracked the procurement of rapid diagnostic tests (RDT) for common infections, such as HIV, and has reported a significant increase in usage. In 1999 less than 40% of all HIV diagnoses around the world were performed with rapid tests whereas the remaining 60% were conducted with laboratory-based tests such as ELISAs. By 2009 over 80% of HIV diagnoses had transitioned to using RDTs. Many of these tests abide by the ASSURED criteria, which outline standards for the evaluation of point-of-care diagnostics. Although these criteria endorse important characteristics such as low cost and rapid time-to-results, they do not provide quantitative measures to compare similar tests or guide RDT developers. In this analysis we aim to determine how different countries select HIV RDT(s), and to parse out which characteristics are responsible for their decisions. Understanding what drives the purchase and use of current HIV RDTs can help inform future designs and specifications for test developers.;Methods: We analyzed data from The Global Fund's Price and Quality Reporting (PQR) database from January 2010 to January 2014. The sample consisted of 945 purchase records of HIV rapid diagnostic tests, including order date, purchasing country, and the number of tests purchased per order. The main independent variable was unit cost per test (in USD). Linear regression with random effects on country was used to examine the relationship between test purchase, a proxy for test usage, and test cost. Models were adjusted for country random effects and country and test-level covariates.;Findings: This study included 945 purchase records of 38 different HIV RDTs from 93 countries. Our exploratory analysis showed that the characteristics that drive test purchase differ across a variety of settings. As cost is increased by just one $1.00 overall purchase across all settings falls by 18.7% (95 % CI: -25.4 to -16.0 %). Other important driving characteristics were test time, maximum test storage temperature, and whether or not the test required additional equipment to run. When observations are stratified by country income level, the major driving factor is test cost and the degree of influence changes across strata. For low-income countries, every $1.00 increase in test cost decreases test purchases by 62.1 % (95 % CI: -155 to -39.4 %). For high-income countries this decrease is only 10.4 % (95 % CI: -24.7 to -2.7 %). Stratification by country HIV prevalence, high vs. low, again showed cost was a major driving factor, while test time, maximum test storage temperature, and whether or not the test required additional equipment to run were also significant indicators. For all of the analyses, test accuracy --- sensitivity and specificity --- was not found to significantly influence test purchase within this dataset.;Findings: The ASSURED criteria are a good first step toward evaluating diagnostics for a variety of settings but this work presents the first quantitative assessment of which characteristics influence the purchase and usage of HIV RDTs. Overall, cost was a major driving factor, but the degree of its influence, and whether or not other factors were also important, varies between settings. This exploratory analysis can serve as a step toward better understanding what drives HIV RDT usage and help researchers and test developers design future tests that focus on these important characteristics. |
