April’s Electronic Health Records Association (EHRA) white paper “Recommendations for Determinants Capture” importantly calls out the variety of ways in which health care teams learn about patient’s Social Determinants of Health (SDOH) or Health Related Social Needs (HRSN). The Association’s Social Determinants of Health and Health Equity Task Force is tasked with ensuring that EHR technology can be used to reduce health disparities. In doing so, it took a close look at recent recommendations and guidelines around identifying and categorizing SDOH and mapped out how these can be incorporated into electronic health records in ways that meets its defined Characteristics of an Effective Solution, which are critical for widespread technology change and adoption as well as interoperability and support of data use for research.
In discussing standardization of SDOH in EHR systems, they note that “…an EHR should be able to indicate whether a patient was assessed for a domain risk (e.g. housing instability or food insecurity), whether that risk is present, and the method of assessment if a standardized instrument or questionnaire was used.” The Task Force specifically note that while this may come through an embedded electronic screening tool, “A user might also evaluate risk through informal methods such as a conversation or a paper form, subsequently coding the identified risk using Z codes.”
Their final recommendation is that “Risk assessment methods should remain flexible for now.” Clearly, this is not just for the convenience of the EHR companies. Instead, it meets the myriad needs of patients and their healthcare teams to address whole person care in a variety of ways, and at any time, not just at specified intervals for recommended screenings.
The Task Force further notes that the source of the data can be represented with an “Optional Coded Value: Corresponding to the instrument or question used for risk assessment.” This will provide the appropriate provenance or metadata for research purposes, but may also be used to vary clinical decision support around how follow-up assessments are approached.
Taken together, these statements point the way for additional assisted technology. In this case natural language processing (NLP) can help both clinicians and other members of the health care team recognize when their conversations have covered an assessment of a SDOH domain, and then to code it, making that assessment available for representation in the EHR without adding to staff burden. These “conversations” are not new to healthcare and are often spontaneous and not scripted. What is new is the recognized need for more systematic capture as “data” and more importantly action based on the social stressors that come up in these conversations.
.png)
.jpg)
