EHR Data Sharing Shows Promise for Public Health Monitoring of Drug Use

EHR data sharing can support near-real-time information access on drug use-related healthcare use to guide clinical and public health responses, according to a study published in Health Affairs.

Researchers leveraged the Minnesota EHR Consortium, a data sharing collaboration across health systems in Minnesota. The study used hospital and emergency department (ED) visit data from May 2012 to June 2023.

Based on monthly data from six health systems, the researchers found that hospital and ED visits involving methamphetamine had the most significant increase over time (302 percent), and opioid-related visits increased across the countywide sample, consistent with patterns at the national level.

The study found that in Hennepin County, Minnesota, Native American, multiple-race, and Black people had substantially higher rates of drug-involved hospital and ED visits than other racial and ethnic groups. Nationwide research found similarly higher rates of drug-involved hospital and ED visits among Black people compared to White people but did not report rates among Native American populations.

“Our findings suggest that the data infrastructure we developed can be used to evaluate the effectiveness of targeted interventions for Native American and Black communities in Minnesota and inform modifications, when needed, to maximize the impact of programs aimed at reducing overdose deaths,” the study authors wrote.

The EHR is a promising source of real-time data for the identification of public health priorities, the researchers noted. Monitoring drug-involved healthcare use through EHR data may help public health officials inform early intervention and detect trends before mortality spikes.

Notably, EHR data contains sociodemographic information, including residence, race, ethnicity, and preferred language. Monitoring drug-involved hospital and ED visits across these different categories can ensure a health equity-driven response to substance use challenges, the study authors said.

For instance, the study found that in 2022 and 2023, more than 14 of every 100 hospital and ED visits among Native American people were drug-involved, compared to 4 of 100 hospital and ED visits among White people.

“Targeted prevention and treatment efforts are needed in communities with high rates of drug-involved hospital and ED visits and could be evaluated using the surveillance methods we developed,” the authors wrote.

Stakeholders can also merge EHR data with social service data to provide insight into service use by groups at higher risk for overdose or other drug-involved healthcare use. The researchers said that they are actively expanding their work to monitor substance use disorder-related ED visits among people experiencing criminal legal involvement or housing instability.

“We are integrating flags for recent criminal legal involvement or housing services use so that public health officials can stratify trends in drug-involved visits by specific groups,” they explained. “Such data integration can facilitate timely evaluations of social service programs that may improve patient health and reduce drug-involved health care use (for example, low-barrier housing efforts).”

Although the study demonstrated the utility of a countywide EHR collaboration, stakeholders can scale the model to different geographic levels.

For instance, the Minnesota EHR Consortium has used statewide EHR data to track trends in COVID-19 vaccination inequities to guide targeted outreach. Researchers could leverage a similar approach to monitor trends in drug-involved healthcare use statewide.

“This information can provide insights to support an effective response to the drug overdose epidemic that decreases mortality and reduces disparities across racial and ethnic groups,” the authors emphasized.

They noted that the addition of emergency medical services data would further improve EHR-based surveillance efforts by including data on nonfatal overdoses that do not require hospital and ED care.