|Article: Electronic||Free Download|
The prevalence of individuals with multiple chronic conditions will increase during the coming decades along with increased proportion of older persons in the population. Persons with multiple chronic conditions present a major challenge to clinical research and to the delivery of health care. A better understanding of multiple chronic condition prevalence and patterns is needed. This study examines multimorbidity patterns using diagnoses and conditions reported to be present at the time of hospital admission among adults discharged from California hospitals in 2007. Correlation analysis, hierarchical clustering algorithms, and multivariable logistic regression are used to identify empirical patterns among diagnoses and conditions that occur together with high probability. Hierarchical cluster analysis is used to classify frequently occurring diagnoses according to their probability of occurring in the same patient discharge record, based on their polychoric correlation. Multiscale nonparametric bootstrap resampling is used to calculate ‘approximately unbiased’ probability values for each diagnosis cluster. The results demonstrate that strong patterns occur among the diagnoses and conditions reported for hospital patients. Almost all hospital patients had multiple chronic conditions, however, the relative frequency and pattern of these disease combinations varies greatly by cause of hospitalization. Detailed results are presented for clusters of diagnoses and conditions related to renal, heart, atherosclerotic, and respiratory disease, bacterial infection, and other disease groups that occur together with high probability. This study demonstrates that underlying correlation patterns among diagnoses provide a structure that can be used to improve the measurement of multimorbidity. Statistical models can be developed that incorporate this complexity to provide highly detailed measurement of multimorbidity patterns in the aging population.
|Keywords:||Multimorbidity, Chronic Disease, Hierarchical Clusters|
Associate Professor, Division of Patient Outcomes, Policy, and Epidemiologic Research, University of Virginia School of Medicine, Charlottesville, Virginia, USA