By Ken Kleinberg, Point-of-Care Partners, Practice Lead, Innovative Technologies
From an IT perspective, SDoH can be considered in terms of what data can be obtained, how it can be analyzed and, perhaps most importantly, how it can influence the action and workflow of payers, providers, and patients. Electronic health records (EHRs) were designed primarily to capture medical information and submit claims, and payer systems were oriented around benefits and claims adjudication. Both need to be substantially expanded to consider these new data types. Population Health Management (PHM) and case/care management systems are particularly well suited to benefit from SDoH data and analysis. The ability to determine a patient’s/member’s need and then refer them to the proper community-based services is key.
Additional Data to Inform
SDoH introduces many new data sources including zip code demographics, financial and credit information, self-reported information from patient and family, direct observation, Internet-of-things data (air quality, temperature), etc. Accessing/managing these new data sources often have the same problems as any data: sourcing (quality, cost, completeness, timeliness/refresh, structured/unstructured), definition (standards, terminology), storage (data model/architecture), and access (shareability, interoperability, governance, privacy, consent). While there are many and extensive sources of various SDoH data (e.g., from various agencies), this type of data, in terms of IT usability, is still in the early stages. For example, most SDoH data is neither included in the first draft of the United States Core Data for Interoperability (USCDI), nor required to be managed as part of current EHR certification, although progress is being made (e.g., via the USCDI update process and the Interoperability Standards Advisory initiative.
Analysis
Once SDoH data has been collected, the harder– how to make sense of it – must be addressed. The challenges can seemingly be simplified by focusing on one major or overarching factor, like zip code or a few of the big factors such as food, transportation, housing. But the complexities and interdependencies of these could vary. For example, they could change according to the day of the week, weather or season. What if a key family member that provides care loses a job or is only around sometimes? What if a patient has no one to care for their dog during a needed hospital stay? Expanding the analysis models to 5, 10, 20 or even 100 factors or more would potentially lead to more accurate insights, especially if volumes are high. However, while that approach can be enticing – especially given ever-more powerful machine learning/artificial intelligence (ML/AI) technology to analyze the additional factors – it can also lead to false confidence and could even be dangerous (e.g., are the more “accurate” results going to be repeatable? Will they transfer to other locations? Is an organization willing to bet money or lives on that?).
Action
If getting and normalizing the data was a challenge and analyzing it even harder, perhaps the most difficult and important step is how to then ensure the best actions by the various stakeholders are taken. Here’s where big strides can and are being made with SDoH, giving the providers, payers, pharmacists, and other healthcare professionals better insight into what social and community services to recommend and being able to incorporate action more readily into treatment plans (e.g., eCare plans) to refer the patient appropriately. More advanced solutions can help the stakeholders follow the progress of the referral (did the patient visit the foodbank?).
SDoH can also have effects at more of a population level. For example, understanding specific population needs for a payer may influence which markets they want to enter or risk models to participate in. For providers, it may influence which patients they feel they can help the most, or whether they want to participate in various VBC programs. Even more exciting will be to see if and to what degree better outcomes, lower costs, and higher stakeholder satisfaction can be achieved as SDoH data is increasingly utilized across the care continuum.
SDoH Initiatives and Resources
The Office of the National Coordinator for Health Information Technology (ONC) has been involved in SDoH since 2015 and is helping to lead the SDoH charge. The ONC recently held a half-day workshop on SDoH - which highlighted many initiatives and resources, including:
Other SDoH resources (of many) include:
In addition to the significant efforts by EHR, PHM, interoperability, and analytics vendors, there are many commercial solutions now available with specific focus on sourcing, analyzing, and acting on SDoH, with many more likely to surface including. Examples include:
Conclusion
The COVID pandemic has highlighted how different socioeconomic and additional factors can disproportionately effect health and outcomes – it raises the bar for us to understand these differences and take the right action. Telehealth is now giving clinicians a view inside a patient’s home in a way we’ve not seen in a few generations (when house calls were common). Social media, financial and demographic data, and more are being collected like never before. Finding ways to analyze and use this information efficiently, compassionately and ethically will lead to a healthier country. IT is key to that endeavor!
We’re here at Point-of-Care Partners to help you through how to leverage SDoH with payer and provider systems regarding standards, FHIR-based API interoperability and targeting the most impactful use cases. If you have questions or just want to talk through your challenges, reach out to me at ken.kleinberg@pocp.com.