Indiana Network for Population Health tackles SDOH integration
In HIMSS20 Digital, health informatics expert Brian Dixon links the power of health information exchange networks to actionable data for hospitals.
The challenge of population health is the ability to integrate and share the data for health systems to act upon the information.
Brian Dixon, director of Public Health Informatics for the Regenstrief Institute and associate professor Department of Epidemiology, IU Fairbanks School of Public Health, shares the success of integration efforts in Indiana, with HIMSS20 Digital, in Integrating the Social Determinants into an HIE Network.
Regenstrief created the Indiana Network for Population Health. The health information exchange repository is the Indiana Network for Patient Care, started in the 1990s as an experiment to determine whether the ERs of two competing hospitals shared many of the same patients. It turns out they did.
The first major project set out to serve healthcare systems in the context of addictions and opioid overdoses, Dixon said.
Health experts already knew that, though overdose deaths were declining, the number of first-responder calls that involved Naloxone for overdoses was on the rise. Often a call was made to treat the same individual several times.
"We're able to keep people alive and discharge them home, but they're bouncing back with an overdose event," Nixon said. "Our goal is to address the crisis by making publicly available data available."
What health professionals needed to know was the factors that could turn the tide of high-overdose events.
For this, Dixon and the Indiana Network for Population Health needed to look at the social determinants of health.
For over 20 years, healthcare professionals have known that what goes on in the clinical office accounts for a small fraction of life expectancy or quality of life. The bigger factors include education, employment, income, food, the physical environment and access to care through an insurance program, all within the category of the social determinants of health.
Upstream at the community level, policy affecting transportation and healthcare administration rules could be factors.
When a patient visits his or her physician, or the ER, the clinician needs to know whether the social determinants are a factor in treatment. But this information is rarely in the electronic health record. Sometimes it's in the clinical notes, Dixon said.
"Part of our work in informatics is to mine text for the social determinants," Dixon said.
GETTING THE DATA
Data is publicly available from places such as the Census Bureau, which collects data on the social determinants, as do the Social Security Administration and the Supplemental Nutrition Assistance Program, more commonly known as the food stamp program.
Then there's information from labs and payers, which can be added to the providers' clinical data.
Federal data comes from Data.gov, healthdata.gov, the Eviction Lab and, for discharge data, the Indiana State Department of Health.
For population health, the integration of census and clinical data can tell the proportion of individuals in a particular area who have less than a high school education, or the median income in a neighborhood.
In the last three years, third-party companies have emerged to create clearinghouses for SDOH data.
Dixon and the team brought the data together to make it useful to healthcare.
They found that, overall, 25% of patients treated for an overdose had one dose of Naloxone; that the ages were 18 to 44; that the number of males, compared to females, treated was 150% higher; and that white patients had double the likelihood of having Naloxone during an EMS run than patients of other races.
The Network for Population Health has APIs that allow access to the external data and a data lake to store raw information. It has a data commons of integrated data and tools to allow access and build analytical data sets.
The network is a way for all of the work done by various state and federal agencies to come together as part of a larger system to track population health, give clinical disease support, and track disease-prevalence trends and new diseases such as the coronavirus.
"(It can) enhance health surveillance in the community, tracking new diseases, COVID-19," Dixon said, "and finally clinical and population health decision support, so it's providing situational awareness to leaders and connecting patients themselves to community-based service to help them address their social determinants of health. So this is our vision for the network."
Twitter: @SusanJMorse
Email the writer: susan.morse@himssmedia.com
Population health management enters an uncharted new era
This month, we look at how approaches to treating COVID-19 and other illnesses are shifting in this new era.