Big data and patient reported outcome measures (PROMs)

Amy Abernethy1

1Flatiron Health, New York, NY, USA

Abstract

Leveraging the emerging wealth of “big data” from diverse healthcare sources is critical to developing the evidence base that informs decisions made by patients, providers, and policy makers.  While healthcare today is generating a vast volume of data, there are certain ingredients to which the entire industry must commit in order for the data to be facile and meaningful.  Solutions are needed to pull out key data points that are fundamental to cancer-focused analyses from unstructured documents (e.g., case notes, biomarker reports) at scale.  Even the structured data that are routinely available in clinical datasets are messy, requiring normalization and harmonization, such that all data points are merged into a common analyzable format.  And, continuous amalgamation of longitudinal data is necessary to depict the full healthcare story.   The increasingly central role of the electronic health record (EHR), coupled with a renewed focus on the importance of the voice and experience of the patient, creates a clear opportunity to generate meaningful patient-centric data that contribute to the big data vision.  PROM data include everything from personal reporting of demographic characteristics to personal reflections on satisfaction with health care and assessments of symptoms and quality of life.  Collection of complete, high quality PROM data at scale is most practical when real-time collection of PROMs is useful by clinicians at point of care, then these data are stored in the EHR and available for a variety of secondary analytic purposes.  Today, there are many PROM efforts occurring at the institutional level; these efforts are demonstrating the value of engaging with patients to generate key components of a clinical data set, which facilitates patient-centered and individualized care, leading to improved outcomes for society as a whole.