For centuries, we’ve been aware that where we live and work has a significant effect on our health and wellbeing. Recent studies from Massachusetts General Hospital and Health Affairs estimate these factors affect as much as 80 and 90 percent of a patient’s health status, respectively. Even the father of medicine, Hippocrates, noted that a person’s geographical origin determined how healthy they were as far back as 4000 BCE.
Many elements comprise the building blocks of health: income, nutrition, available transportation, home life, pollution, and more. In an effort to address the needs of at-risk populations, some hospitals have launched initiatives to help improve these underlying conditions for patients. Mass General not only tries to help patients with language barriers and low health literacy to manage their medications, they also have a special program to assist patients who are domestic violence victims.
While these are people-centered solutions that require staff members for personal interactions, there are ways technology can assist as well. More healthcare organizations are using ride-share services to assist patients with transportation challenges, sometimes allowing a provider to order a ride for a patient directly from their workstation. And in some cases, AI can be used to review a patient’s record for clues that they might benefit from referrals to other community resources. Incorporating the social determinants of health (SDOH) into a more comprehensive care plan in this way not only gives the patient more of the assistance they need for healthy living, it also aids doctors by providing alerts and referrals in their workflow in real time, allowing them to give recommendations on the spot rather than taking time to gather the information later.
There are some challenges that come with using technology for SDOH, however, and one of them is interoperability. Not only are most EHRs currently unable to record these determinants effectively, it is nearly impossible to do so in a uniform way that allows providers to extract the data and analyze it. A lack of uniform coding mechanisms for problems like food insecurity or behavioral health challenges means that it is more complicated to represent those details in a patient’s record. However, it is still possible to record and use that information, according to Varun Gupta of Mount Sinai Health System: “It’s all there in unstructured format. And there’s a huge opportunity to use that data to find out insights. That’s where natural language processing comes in.”
There are still many gaps to bridge before EHRs, legacy clinical records, and other communications can rely heavily on these technologies to add value to the patient experience. But machine learning and structured data analysis are evolving at a fast pace. With major healthcare systems leading the way with pilot programs and studies using these tools, the reality may be that they will be refined and used by providers across the country in the near future.