Hospital Spending and Quality

A lot of time and effort is being expended in trying to determine if there is a relationship between how much is spent on health care, or how much it costs a provider to deliver, and patient outcomes, particularly in hospitals.  There appears to be great variation in hospital spending for the same conditions and the Dartmouth research, for example, has suggested that there are few or no better care outcomes in high-spending regions.  This research encourages policymakers to think maybe they can cut payments to high-spending regions or hospitals without affecting the quality of care.  More recent research, however, has challenged this conclusion and it remains a subject for virulent controversy.  A new study published under the auspices of the National Bureau of Economic Research takes a new approach by examining patients who were delivered to a hospital by ambulance in an emergency, since these deliveries often have some element of randomness to them, and they also used data from New York State on admissions for patients on either side of an ambulance territory boundary.    (NBER Study)

The researchers looked at the average costs of the hospitals and compared that with mortality outcomes for the patients delivered to the hospital.  Higher cost hospitals had lower mortality.  A ten percent increase in hospital costs, for example, was associated with a 4% decline in mortality.  In trying to ascertain what characteristics of higher cost hospitals might account for the gain, the researchers found it was not associated with performance on “appropriate care” measures, for example, but seemed most correlated with treatment and procedure intensity.  In other words, it was the quantity of care more than the cost of the items of care that seems linked to the lower mortality.  Being taken to a teaching hospital or to a “high-tech” hospital was clearly associated with lower mortality.  These tend to be very high-cost institutions.  Overall these results are not good news for those who think that spending and quality are not related and that reimbursement can be cut without an effect on outcomes.