Hospital Response to Payment Reform: The Effect of Financial Incentives for Reducing Hospital Acquired Conditions

Hospital reimbursement schemes that link payments to activity on the basis of diagnosis related groups (DRGs) specify higher payments for patients with comorbidities or complications. The additional payment is due to the higher expected costs of care for complex patients, but the payment schemes do not differentiate between complications that were present on admission, and those that were inflicted by the hospital during admission.

A U.S. Medicare policy introduced in 2008 removed the additional compensation for secondary conditions acquired while the patient was admitted to hospital. The policy change aimed to incentivise hospitals to take precautions to reduce the occurrence of the following hospital acquired conditions (HACs):  Foreign object retained after surgery, air embolism, transfusion with the wrong type of blood, severe pressure ulcers, falls and trauma, catheter-associated urinary tract infection (UTI), vascular catheter-associated infection, manifestations of poor control of blood sugar levels, surgical site infections following certain procedures, and deep vein thrombosis following certain procedures.

We aimed to assess whether the Medicare payment reform had the intended effect of reducing hospital acquired complications, and assess the effects of the reform on hospital resource use (LOS and intensive care unit days) and patient outcomes (30 day readmissions and mortality) and gaming (coding of secondary diagnoses as “present on admission” and the number of co-morbidities coded per claim).

We used Medicare claims data (MedPAR RIF) for all Medicare patients admitted to an acute-care short term Inpatient Prospective Payment System hospital between 2007 and 2013 and estimated difference in differences analyses of prevalence rates, resource use, patient outcomes, and gaming for patients with hospital acquired conditions targeted by the policy compared to a group of control conditions that have been considered for inclusion in the policy but are not currently included, from before to after the reform.

The initial analysis indicates that the occurrence of HACs targeted by the policy decreased slightly over time for 6 of the 12 conditions targeted by the policy, but increased or remained unchanged for 6 other targeted conditions. In the post policy period, the share of conditions coded as acquired in the hospital increased for half of the targeted conditions, but remained unchanged or decreased slightly for the other half of targeted conditions. The occurrence of secondary diagnoses in the control group increased slightly over time for the majority of control conditions, but the rate of conditions coded as hospital acquired decreased or remained stable.

The preliminary results do not suggest that the payment reform was associated with a decrease in the rate of hospital acquired conditions compared to conditions not targeted by the policy. In ongoing work we consider whether hospitals responded strategically to avoid the financial consequences of the reform.