Prevalence of hospitalizations due to chronic diseases at the state and county level: The implementation of the Chronic Condition Indicator tool with Mississippi hospital discharge data, 2011

Background: Across the United States and in Mississippi, healthcare providers and public health agencies face the challenge of controlling the escalating financial and health-related cost of chronic diseases. This study aimed to determine the prevalence of hospitalizations due to chronic conditions by implementing a free clinical grouper that clusters medical diagnoses as chronic and non-chronic.  In addition to the geographical distribution of hospitalizations due to chronic conditions, the demographics, average charges, and length of stay were also determined. Finally, a comparative analysis of hospitalizations due to chronic and non-chronic conditions was conducted.

Methods: Conditions were classified as chronic by the implementation of the Chronic Condition Indicator, an algorithm that classifies ICD-9-CM diagnoses into two mutually exclusive groups, chronic and non-chronic. A chronic condition in this classification system is defined as any medical condition that lasts for 12 months or more, and might result in limitations on self-care and/or ongoing medical intervention. Mississippi hospital discharge data for 2011 included only Mississippi residents and data were analyzed with SAS 9.3. Means were compared with t-tests and proportions with chi-square tests.

Results: There were 153,040 (40.6%) discharges due to chronic conditions in 2011. Among them, mood disorders were the most prevalent group, followed by congestive heart failure and chronic obstructive pulmonary disease. Males (45.9%) and Caucasians (41.0%) were more likely to be hospitalized with a chronic condition than females (40.0%) and other races (40.0%). On average, patients hospitalized for chronic conditions were 12.5 years older (57.4 years versus 44.9 years); stayed 1.3 days longer (6.1 days versus 4.8 days); and their average charges were $7,418 ($31,259 versus $23,841) higher than patients hospitalized with non-chronic conditions. All of the above-mentioned differences were statistically significant at p = 0.001.While there was no statistically significant difference in the hospitalizations of patients residing in metro areas compared to patients residing in rural areas, there were considerable county-level variations in hospitalization rates due to chronic conditions.

Conclusion:  The implementation of the Chronic Condition Indicator algorithm demonstrated a cost effective and comprehensive method for clustering and studying the epidemiology of chronic conditions.  Furthermore, this study demonstrated the usefulness of administrative health care data for evaluating the resource utilization of hospital admissions due to chronic conditions. Finally, this study pinpointed the geographical distribution of hospitalization rates due to chronic conditions, an approach that can be used to target the most affected areas and populations in an evidence-based way.