Development of smart phone tools for printed diagnostics: Challenges and solutions

摘要

Among the various forms of low-cost diagnostic tests (e.g. lateral flow, fluorescent), colorimetric tests are powerful tools for performing a wide range of point-of-care health diagnostics and screening using biological samples (blood, urine, or saliva). However, due to human color perception, colorimetric test results are difficult to quantify, particularly in the case of multi-parameter tests, which pose an even greater challenge. In order to meet this need, we propose the use of a mobile app with augmented reality (AR) that can automatically track the relative position and orientation of a printed test, which includes a reference color pattern that enables automatic color calibration. This solution is highly scalable, does not require any hardware (e.g. phone stand), and can be easily adapted to any colorimetric test geometry. For quantitative evaluation, we conducted a field study with 23 Indian health workers who had never previously used a smart phone. With only 10 minutes of instruction, participants were asked to perform 50 randomized consecutive measurements manually, then repeat 50 additional measurements with our mobile app. Results with mobile app (R2=0.944, duration=32 sec/test) compared favorably to the manual data capture (R2=0.819, duration=37 sec/test). This work demonstrates that mobile phones can be used successfully to read colorimetric tests without external hardware, and also demonstrates that augmented reality can be a valuable ingredient for creating mobile tools that are accessible to novice users in low-resource communities.