In their recent paper, Beaulieu et al. attempt to estimate the effects of hospital acquisitions on measures of hospital quality at the acquired hospitals. To do so, the authors use four measures of hospital quality and compare the changes in these measures at acquired hospitals relative to changes at a set of non-acquired “control” hospitals. The four quality measures used in the study include a measure of patient experience (a composite score reflecting patients’ satisfaction with their hospital experience, e.g., whether they would recommend the hospital), a clinical process measure (a composite score reflecting how care is provided in the hospital, e.g., how many heart attack patients were given aspirin when discharged), and two clinical outcome measures reflecting hospital-wide readmission rates and mortality rates.
The authors conclude that acquired hospitals (relative to a comparison group of non-acquired hospitals) experienced relative declines in patient experience measures, no relative changes in 30-day readmission or mortality rates, and relative improvement in clinical process measures. However, they note that improvements in clinical process measures could not be attributed definitively to the acquisitions because these measures began to improve prior to the acquisitions.
Beyond the ambiguity resulting from the inconsistencies in the authors’ findings, we believe that important limitations in their analyses caution against relying on their research.
Assessing umbrella pricing incentives
When collusive agreements involve a subset of firms in an industry, they may create the incentive and ability for firms that are not participants in the cartel...