eHealth and mHealth Pipelines for Clinical Decision Support to Improve Medication Selection and Safety

Abstract

Although much work has been done over the past decade on developing personalized and evidence-based medicine, such as diagnostic tests based on genetics to better predict patients' responses to therapy, stumbling blocks remain that have prevented knowledge, tests, and other pertinent patient-centric data from becoming routine in clinical practice. The main problem with uptake and implementation centres on an absence of computational infrastructures to deliver an individualized approach at the point of care. This study focused on the creation of re-usable computer and mobile-based components of clinical decision support systems for patient-centric data. The goal was to develop eHealth and mHealth pipelines to not only create interactive delivery systems for medication selection and medication safety, but also to provide templates that can be re-used on provincial, pan-Canadian, and international scales. Through this research, a novel database structure for storing and analyzing data (CASTOR), augmented fast and frugal decision trees (FFTs), and mobile and web-based applications (AntiC), were developed to improve medication selection and medication safety. These projects were initially based on rationalizing, storing, and interpreting large genomic data sets, filtering the ensuing pharmacogenomically relevant data into automated decision trees for medication selection on a country-specific basis, and developing a work-frame template (BOW) to build clinical decision applications (apps) for smartphones and tablets. As a result of this project, ultimately, community and hospital pharmacists in Canada will be better informed about the safe handling and use of oral chemotherapeutics, preventing adverse drug reactions, and Ministries of Health in developing countries will have access to curated, pharmacogenomics-based decision trees specific to their countries to rationalize medication selection at no extra cost to local healthcare systems. The models we developed (CASTOR, augmented FFTs, and BOW) were deliberately designed and built with re-usable components that can be linked to future projects to generate targeted and focused clinical decision support and mHealth applications wherever needed. This will eventually allow for the implementation of rationalized clinical decision support applications and evidence-based knowledge translation at the point of care.