Pill Identification using a Mobile Phone App for Assessing Medication Adherence and Post-Market Drug Surveillance

david Prokop,Joseph Babigumira,Ashleigh Lewis

DOI: https://doi.org/10.48550/arXiv.2004.11479

2020-04-23

Computers and Society

Abstract:Objectives: Medication non-adherence is an important factor in clinical practice and research methodology. There have been many methods of measuring adherence yet no recognized standard for adherence. Here we conduct a software study of the usefulness and efficacy of a mobile phone app to measure medication adherence using photographs taken by a phone app of medications and self-reported health measures. Results: The participants were asked by the app 'would help to keep track of your medication', their response indicated 92.9% felt the app 'would you use this app every day' to improve their medication adherence. The subjects were also asked by the app if they 'would photograph their pills on a daily basis'. Subject responses indicated 63% would use the app on a daily basis. By using the data collected, we determined that subjects who used the app on daily basis were more likely to adhere to the prescribed regimen. Conclusions: Pill photographs are a useful measure of adherence, allowing more accurate time measures and more frequent adherence assessment. Given the ubiquity of mobile telephone use, and the relative ease of this adherence measurement method, we believe it is a useful and cost-effective approach. However we feel the 'manual' nature of using the phone for taking a photograph of a pill has individual variability and an 'automatic' method is needed to reduce data inconsistency.