Evaluating the impact of chronic diseases and healthcare interventions on post-discharge patient outcomes
MetadataShow full item record
As nearly half of all Americans suffer from one or more chronic diseases, resulting in inefficient healthcare utilization and increased costs of care, effectively treating and preventing chronic disease has become a public health priority. This is particularly true for vulnerable patient populations, such as the aged and individuals with lower socioeconomic status struggling with additional barriers to care. Improving the treatment of chronic disease begins with understanding the risk factors associated with high rates of in-patient, emergency, and out-patient services. Examining healthcare utilization from a population health management perspective showed that individuals with chronic disease(s) had hospital readmission rates that were 10 times as high and revisit rates (in-patient or emergency care) that were 12 times as high as rates for individuals without chronic disease. A Cox regression model analysis showed that female patients, using managed care, and longer lengths of stay during the index hospitalization decrease risk of readmission and revisit. Level of chronic disease complexity had the greatest impact on utilization as individuals with complex chronic disease are 4.5 times more likely to be readmitted and 2.8 times more likely to have a revisit than individuals without chronic disease. Treatment in an integrated model Accountable Care Organization was found to reduce the likelihood of readmission but not revisit. While it is known that chronic diseases place patients at a heightened health risk, further study is needed to understand the impact of multiple chronic diseases. Using observational causal inference methods disease combinations in the renal and cardiac pathways were analyzed for a large sample of New York State Medicaid patients. In the renal pathway, disease combinations that included a coronary artery disease diagnosis had the highest baseline readmission rates and causal effect estimates, therefore are the most influential to hospital readmissions. In the cardiac pathway the most influential diagnoses were heart failure and tobacco abuse (smoking). In addition, behavioral health conditions and substance abuse disorder diagnoses were highly prevalent in the study population and presented a substantial opportunity for improving patient outcomes. Reducing hospital readmissions through the increasing adoption of intervention programs has created a need for new modeling approaches capable of representing the temporal nature of these initiatives. Three models were assessed for their ability to provide meaningful insights from limited patient and program outcome data. The Cox Proportional Hazard model was found to be useful for quantifying the effects of covariates on outcomes while the Extended Cox Model provided an understanding of the effects of influential variables as they change over time. Further, the Discrete Time Markov Chain model was particularly useful for comparing alternative program designs. This work explored the impact of chronic disease and care transition programs on hospital readmissions, identifying the key risk factors influencing healthcare utilization and quantifying the effects of chronic disease(s) on patient outcomes in the low-income chronically ill population. These developments will improve the ability of public health professionals to identify the highest-risk individuals and design more effective and personalized care interventions. Further, the flexible analytic methodology for assessing the impact of care transition intervention programs using available program and patient data will guide analysts and healthcare administrators in evaluating existing programs and efficiently using limited healthcare resources.