Arizona county fights heat-related illnesses with analytics
- By Amanda Ziadeh
- Oct 03, 2016
An Arizona county is using analytics software to monitor risk factors associated with heat-related illnesses caused by the state’s extremely high summer temperatures.
“Heat can be deadly down here in Arizona,” Graham Briggs, administrator for the Pinal County disease investigation program, told GCN. "We knew that people were getting sick related to the heat in our specific county, but we didn’t know anything more than that."
Officials knew they needed real data to take a more rigorous approach to heat morbidity. The Pinal County Public Health Services District turned to SAS, a provider of analytics software, to help disease investigators monitor and analyze heat-related illnesses.
The county wanted to measure its heat morbidity rate and the populations affected as well as determine how to mitigate and identify the risks. Program staff gathered years of data from the statewide mortality databases and hospital discharge sheets hoping to find patterns and risk factors for heat stroke, hyperthermia and heat shock. The software ran through a million rows and 200 columns of data, which revealed spikes and clusters of heat illnesses in certain populations and geographies.
In public health, epidemiologists use statistical methodology to analyze and manipulate large amounts of data that may be too large for staff and Microsoft Excel to manage, Briggs said. That’s where SAS Analytics came into play.
Integrating the data was relatively seamless, Public Health Data Analyst Sammy Packard said. SAS uses the same files as Excel, and it is capable of reading, detecting and identifying the various diagnostic codes used by hospital staff. Packard simply merged the list of codes with discharge data so that the system automatically matched codes to diagnosis descriptions, which makes searching for certain heat-related information easier.
Discharge data can also include a patient’s demographics and location. SAS geocoding tools work with existing geographic information systems software to help users map at-risk areas.
For example, Briggs said preliminary analysis uncovered a surprising at-risk population among younger people. By combining geospatial and socioeconomic data, investigators also identified clusters of heat illnesses in poorer parts of the county.
Once occupational and population risk factors are discovered, staff can create data visualizations and send tailored alerts, resources and educational tools to county leaders, health providers, emergency responders, the media and other agencies.
In addition to tracking heat-related illnesses, the county also uses the software for case management for tuberculosis patients, who often face six months of therapy, various medications, daily observations and tests. According to Briggs, a TB investigator is using SAS to combine lab results, treatment dates and follow-up plans to generate automated reports, including daily or weekly “to-do” lists, reminders, treatment verifications and so on.
Additionally, SAS allows staff to produce long-term trend reports through biosurveillance of hundreds of infectious and sexually transmitted diseases. For example, the software can help spot trends in diseases, like salmonella, that are harder to identify causes for or be used for outbreak detection and analysis.
The county has been using SAS for heat-related illness monitoring since the beginning of summer 2016. It plans to spend the winter analyzing all the data to characterize the at-risk populations and find ways to mitigate risk by spring 2017.
Amanda Ziadeh is a former reporter/producer for GCN.