Abstract: FR-PO1066
The Medical Evidence Report (CMS-2728) Misidentifies Dialysis Start Dates and Inpatient Dialysis Starts
Session Information
- Home Dialysis: Modalities, Barriers, Utilization, and Clinical Outcomes
October 23, 2026 | Location: Exhibit Hall A, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Dialysis
- 802 Dialysis: Home Dialysis and Peritoneal Dialysis
Authors
- Ji, Grace, University of Southern California, Los Angeles, California, United States
- Lung, Khristina I., University of Southern California, Los Angeles, California, United States
- Mohan, Sumit, Columbia University Vagelos College of Physicians and Surgeons, New York, New York, United States
- Cervantes, Lilia, University of Colorado Anschutz Medical Campus School of Medicine, Aurora, Colorado, United States
- Erickson, Kevin F., Baylor College of Medicine, Houston, Texas, United States
- Lakdawalla, Darius, University of Southern California, Los Angeles, California, United States
- Lin, Eugene, University of Southern California Keck School of Medicine, Los Angeles, California, United States
Background
Dialysis facilities report the dialysis start date on the Medical Evidence Report (Form 2728), which is widely used in dialysis research (e.g., in the United States Renal Data System [USRDS]). Policymakers use the start date to assess optimal dialysis starts and the presence of an inpatient (IP) start (colloquially, "crash starts").
Methods
From the USRDS, we identified patients initiating chronic dialysis 2012-2021 with 60+ days of fee-for-service (FFS) Medicare before dialysis start. We compared 3 start date definitions: (i) USRDS' first dialysis date ("FIRSTDIAL"), (ii) Form 2728 facility-reported date, and (iii) a FFS claims-derived start date. We identified IP starts (start date coinciding with IP dialysis) under definitions ii and iii. We assessed the method that better predicts home dialysis use, an outcome downstream of optimal starts.
Results
For 98% of 339,612 patients initiating dialysis, Form 2728 and USRDS start dates were identical. Only 31% of patients had identical Form 2728 and claims-based start dates, which differed by ≥14 days in 14% of patients and ≥30 days in 4%. Both methods classified 55% of starts as IP. The claims method reclassified another 16% of starts as IP (non-IP under Form 2728). For starts concordant under both methods, IP starts were associated with less home dialysis use than non-IP starts (Figure, p<0.001). Dialysis starts reclassified from non-IP to IP (using claims) exhibited home dialysis use patterns closer to IP starts than non-IP starts (Figure, p<0.0001, log-hazard scale).
Conclusion
Facility-reported Form 2728 dialysis start dates underidentify IP ("crash") starts by 16%. Claims are better at identifying suboptimal IP starts, which are associated with less home dialysis. Researchers and policymakers should consider data sources other than Form 2728 to identify dialysis start dates, especially when assessing the quality of dialysis starts.
Acknowledgment
Data from this study came from the United States Renal Data System. The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy or interpretation of the U.S. government.
Time to Home Dialysis Use, by Inpatient Start Definition
Both methods = using Form 2728 and Claims
Funding
- NIDDK Support