Person-generated health data (PGHD) are valuable for studying outcomes relevant to everyday living, for obtaining information not otherwise available, for long-term follow-up, and in situations where decisions cannot wait for traditional clinical research to be completed. While there is no dispute that these data are subject to bias, insights gained may be better than having an information void, provided the biases are understood and addressed. People will share information known uniquely to them about exposures that may affect drug tolerance, safety, and effectiveness (eg, nonprescription and complementary medications, alcohol, tobacco, illicit drugs, exercise, etc). Patients may be the best source of safety information when long-term follow-up is needed (eg, the 5- to 15-year follow-up required for some gene therapies). Validation studies must be performed to evaluate what people can accurately report and when supplementary confirmation information is needed. However, PGHD has already proven valuable in quantifying and contrasting COVID-19 vaccine benefits and risks and for evaluating disease transmission and the accuracy of COVID-19 testing. Going forward, PGHD will be used for patient-measured and patient-relevant outcomes, including for regulatory purposes, and will be linked to broader health data networks using tokenization, becoming a mainstay for signals about risks and benefits for diverse populations. This article is part of a Special Collection on Pharmacoepidemiology.
Read full abstract