Medicare Payment to Skilled Nursing Facilities: The Consequences of the Three-Day Rule

Monday, June 11, 2018: 1:30 PM
1051 - First Floor (Rollins School of Public Health)

Presenter: Ajin Lee

Co-Authors: Ginger Jin; Susan Lu

Discussant: John R. Bowblis


Medicare does not pay for the fee-for-service care provided in a skilled nursing facility (SNF) unless the patient has stayed in the hospital for at least three days. We study how this three-day rule affects inpatient discharge to SNFs and the subsequent 30- and 60-day hospital readmission rates.

Comparing regular Medicare patients to non-Medicare patients (60 and above), we find that Medicare patients discharged from a hospital 0-24 hours after the three-day-cutoff are 6.5 percentage points more likely to be discharged to SNF than those discharged 0-24 hours before the cutoff, after we control for demographics, insurance type, and medical conditions. When we use the three-day rule as an instrument (IV) for SNF discharge, SNF discharge is found to have a negative effect on hospital readmission rate in the whole sample, but this beneficial impact is entirely driven by sicker patients with an above-median Elixhauser Comorbidity Index. If we focus on patients with no or fewer comorbidities, the IV results find no significant readmission difference between those discharged to SNF and those discharged somewhere else.

The effect of the three-day rule on SNF discharge is the largest (15 percentage points) for knee and hip replacement (KHR). Though KHR patients are on average healthier than patients with heart disease, stroke and other conditions, KHR is the only category for which we find a significant increase in hospital readmission rate as a result of the three-day rule. This perverse effect is larger for patients with no comorbidities, and larger in the zip codes where the average SNFs have a lower-than-state-median occupancy rate. One explanation is that KHR patients are relatively healthy and on the margin of going to SNF or going home. As a result, most of them are discharged to SNFs simply because Medicare will pay for it and there is bed available in nearby SNFs. Not only does the over-consumption of SNF service increase Medicare expenditure in post-acute care, but it also creates inefficiencies in terms of higher hospital readmission rate and potentially longer hospital stay in order to qualify for SNF discharge upon the three-day rule.

Back of envelope calculation suggests that, pooling all conditions together, the three-day rule may have generated an extra payment from Medicare to SNFs by $57-265 million per year, without obvious reduction in hospital readmission rate. And for knee and hip replacement alone, the three-day rule may have cost Medicare $109 million per year including extra payment to SNFs and the subsequent increase in hospital readmission.