Developing a Methodology for Real Time Tracking of Heat Related Morbidity Counts in Maricopa County

Background/Objective: Heat related morbidities (HRM) are a major concern for the Maricopa County Department of Public Health (MCDPH). Environmental temperatures in Maricopa County (MC) are more concerning during summer months (June, July and August), where temperatures can reach an excess of 115^›F. Consequently, high temperatures lead to an increase of HRM. The objective of this study was to identify the baseline levels and epidemic thresholds of heat-related morbidities in MC.

Methods: Total weekly hospital visits and HRM-associated visits were extracted from hospital discharge data (HDD) for January 2006 through December 2012.  The R statistics package was used to determine the expected levels of HRM visits to total visits and establish a baseline and epidemic threshold.  Data from 2006-2007 were used as a reference data set and the model was applied to 2008-2012.  Alarms were triggered when the proportion of HRM visits exceeded the expected baseline and the epidemic threshold.

Results: A total of 11,698,199 hospital visits occurred during the study period. The HRM rate was 84.13/100,000 visits; 95%CI (82.47-85.79), and seasonally ranging from 5.68; 95% CI (4.22-7.14) in January to 300.37; 95%CI (289.26-311.48) in summer months. Numbers of HRM remained below the epidemic threshold during the study period but several alarms were produced when the numbers of HRM exceeded the expected baseline, even during years with lower rates.

Conclusions: Time-series methods are useful and sometimes necessary for aberration detection.  These methods are especially useful for seasonal data and to identify aberration during low-rate periods often missed by methods that ignore the temporal dimension.