Abstract: FR-PO514
Thirty-Day Readmissions in the Peritoneal Dialysis Population Have High Clinical Variability
Session Information
- Peritoneal Dialysis: Modality, Catheter, Infections
November 08, 2019 | Location: Exhibit Hall, Walter E. Washington Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Dialysis
- 703 Dialysis: Peritoneal Dialysis
Authors
- Kim, Jackson, LAC USC Medical Center, Los Angeles, California, United States
- Lin, Eugene, LAC USC Medical Center, Los Angeles, California, United States
Background
Dialysis facilities are evaluated on their all-cause 30-day readmission rate. We investigated the clinical heterogeneity of 30-day readmissions in the peritoneal dialysis (PD) population and the relatedness of readmissions to index hospitalizations.
Methods
In adults with Medicare receiving PD in the United States from 1/1/2013-12/31/2014, we classified index hospitalization and 30-day readmission pairs as “related” if the principal diagnoses were of the same organ system. Using multinomial logistic regression and adjusting for the patient, facility, and geographic factors, we studied whether prior hospitalization burden was associated with a higher likelihood of related readmission. For each of the most common index hospitalization diagnoses, we summarized the most likely reason for 30-day readmission.
Results
The adjusted probability of an unrelated 30-day readmission was 19.2% (95% CI: 18.7%, 19.7%) in patients with 0-1 hospitalizations in the prior year and 26.3% (95% CI: 25.0%, 27.2%) in patients with 4 or more hospitalizations (relative increase 1.4, 95% CI: 1.3, 1.5). Cardiovascular index hospitalizations were most likely followed by a related 30-day readmission (11.8%, 95% CI: 11.3, 12.4%), while renal index hospitalizations were least likely (3.4%, 95% CI: 2.4, 4.3%). For the most common index hospitalizations (peritonitis, sepsis, hypertension, and pneumonia), the most common reason for 30-day readmission was the same as the index hospitalization. Rates of these readmissions varied widely: 43% of peritonitis admissions were followed by peritonitis, sepsis, or peritoneal catheter complications; 34% of sepsis admissions by infections; 21% of hypertension admissions by hypertension or volume overload; and 21% of pneumonia admissions by pneumonia, sepsis, or pulmonary complications.
Conclusion
High background hospitalization rates were associated with increased unrelated 30-day readmissions. When stratifying by index hospitalization diagnosis, there was substantial variability in the rate of and reason for 30-day readmission. Medicare could improve 30-day readmission metrics in PD by accounting for clinical heterogeneity and background hospitalization burden.