In 2008, the World Health Organization published a report titled: Closing the gap in a generation: Health equity through action on the social determinants of health outlining social determinants as the primary driver of preventable health inequity. Despite this, commonly used risk scores like Medicare’s and Medicaid’s still neglect to incorporate clinical biological disease severity and socio-behavioral risk factors. This is due in large part to the inaccessibility of incorporating socio-behavioral data.
To better understand how socio-behavioral data exists in EHRs, we took a look at about 1 million patient EHR records across various EHRs and states. We quickly found that the overwhelming majority of social determinant and behavioral health data are housed in free-text fields. Additionally, in recent years, documentation of socio-behavioral factors has improved alongside the adoption of EHRs in FQHCs. As seen in the figure below, the prevalence of socio-behavioral data in EHRs grew from 7.3% to 24.7% between 2014 and 2016.
Linking socio-behavioral data to patient care can guide targeted intervention. Because of this, we investigated how socio-behavioral data is stored in EHRs and discovered it was primarily stored in free-text fields. From analyzing key words and phrases within non-discrete data, anxiety and stress emerged as the most common factors. Substance abuse, socioeconomic circumstances and incarceration each affect less than 2% of the patients we analyzed, but 1% of 100,000 patients is 1,000 people.
Using socio-behavioral data can guide providers with more informed decisions about the best course of care, but so much is trapped in EHRs. Without a socio-behavioral companion diagnostic guide, incoporating these factors into care will remain challenging. As the healthcare industry works to improve quality while reducing cost, integrating socio-behavioral data may be the missing puzzle piece we have been looking for.