Photo credit: pandpstock001
Using social determinants of health to guide population initiatives will require effectively mapping conditions to electronic health record codes such as ICD-10, researchers with the University of California, San Francisco believe.
There’s a short window to do so, they say, in an article published in the November edition of Health Affairs. Amid the proliferation of disparate front-end screening tools, they write, back-end systems must be able to meaningfully extract data for initiatives such as improving quality and payment reform.
For low-income populations, for instance, whose hypoglycemia-related admissions increase predictably at month’s end as they run out of money for food, efforts to decrease hospital readmissions could focus on providing access to food.
To explore possibilities for aggregation of data using existing databases, the researchers mapped several existing social screening tools to ICD-10 codes.
Some social issues mapped well, such as unemployment to code Z56.0 and social isolation to the ICD-10 category Z60, which includes specific codes for problems related to living alone. Refugee status could be linked directly to category Z65.5, which includes exposure to disaster and war.
Problems arose, however, when there was no ICD-10 code equivalent, there was more than one possible code equivalent or the condition’s meaning was lost in the conversion to code. For instance, most existing social screening tools include questions related to housing, but the ICD-10 category Z59, which includes specific codes for homelessness (Z59.0) and inadequate housing (Z59.1) doesn’t make clear the difference between the two.
The authors call on stakeholders such as the National Quality Forum, the Office of the National Coordinator for Health IT and researchers in health services social intervention to create an effective crosswalk between these systems.