PATIENT CHARACTERISTICS AND RESOURCE ALLOCATION FOR INDIVIDUALS WITH CHRONIC KIDNEY DISEASE

Purpose: We hypothesize that Chronic Kidney Disease (CKD) patients with different laboratory results and different comorbidities experience different disease progressions, and therefore require different resources. We believe that grouping patients based solely on the characteristics of their health state will translate into clusters of individuals with different appointment and hospitalization frequencies.

Method: We obtained de-identified, administrative data for 29,735 CKD patients at eleven VA facilities. We used clustering and a linear discriminant analysis (LDA) models to group the data and to identify significant predictors. After grouping the patients for a second time using only the significant LDA predictors, we compare clusters based on a set of characteristics that have not been introduced in any of the previous models. Those characteristics measure the inpatient and outpatient resources used by individuals. Using all eleven sites, we checked for significant differences in appointment frequencies across patient groups.  Additionally, we analyzed each facility individually, and compared the results to those obtained when using the entire dataset.

Result: We discovered that the significant inpatient length of stay and outpatient appointment frequency differences were primarily associated with group comorbidity differences. Significant differences in resource usage across CKD patient groups are primarily associated with individuals’ diabetes and CKD stage. At the individual facility level, groups’ resource rankings depend on the proportion of individuals with diabetes, as well as peripheral vascular disease (PVD), cardiovascular disease (CVD), dialysis states, and the average group time since the onset of dialysis. There appear to be regional differences among the facilities.  For hospitals primarily located in Southern California, laboratory value results play a more important role. 

Conclusion: Our results suggest that there exists a significant association between resource utilization and patient characteristics, primarily patient comorbidities. Facilities may wish to consider special outpatient appointment tracks, or targeted case management, for CKD patients who have been diagnosed with PVD/CVD or diabetes, or who are on dialysis, because of their expected heavy utilization of resources. A diagnosis of heart failure or cirrhosis for a CKD patient does not appear to be associated with an increased utilization of either outpatient or inpatient resources.