Adopting P4P early vs. late didn't affect hospitals' mortality rates

The observational study used Medicare claims and looked at clinical process scores and 30-day mortality in patients cared for at hospitals that voluntarily participated in pay-for-performance programs or those that later participated in the Hospital Value-Based Purchasing program.


A recent study found no difference in clinical process scores or 30-day mortality between U.S. hospitals that voluntarily participated in pay-for-performance programs (early adopters) and hospitals that later participated in the Hospital Value-Based Purchasing program (late adopters).

Researchers used Hospital Compare data from 2003 through 2013 to compare clinical process scores and 30-day mortality among 214 early adopters and 975 matched late adopters. The observational study used Medicare claims and included more than 1.3 million patients ages 65 years and older (n=263,088 admitted to early adopters; n=1,108,276 admitted to late adopters). Results were published online on Jan. 4 by The BMJ.

Even after matching, there were some differences between early and late adopters. Compared to late adopters, early adopters cared for a higher proportion of white patients (87.4% vs. 84.9%) and started from a higher average clinical process score for the combined target conditions (91.5 vs. 89.9) during the baseline period of early participation in the Premier Hospital Quality Incentive Demonstration (fourth quarter of 2003 to fourth quarter of 2009).

In the period leading up to the current Hospital Value-Based Purchasing program (first quarter of 2010 to second quarter of 2011), early adopters performed at a slightly higher level than late adopters, although there was no significant difference (97.3 vs. 96.7; difference, −0.55; 95% CI, −1.01 to −0.10).

During the Hospital Value-Based Purchasing program period (third quarter of 2011 to fourth quarter of 2013), there was no significant difference in clinical process scores between early and late adopters. During this period, both groups reached a ceiling of about the same level: 98.5 for early adopters versus 98.2 for late adopters (difference, −0.27; 95% CI, −0.77 to 0.22). The difference in the trend differences for the program period versus the pre-program period was not significant (P=0.19).

At baseline in the fourth quarter of 2003, the groups had similar 30-day mortality (14.9% for early adopters vs. 14.8% for late adopters). In the fourth quarter of 2013, both groups ended at the same rate of 9.9%. Mortality trends were not significantly different between groups for either targeted conditions (0.05% point difference quarterly) or nontargeted conditions (−0.02% point difference quarterly).

“These findings provide evidence that having additional time is not likely to turn these programs into a success, at least as far as patient outcomes are concerned,” the study authors wrote. They noted some limitations, such as the potential for uncaptured differences in patient characteristics between early and late adopters, as well as a lack of generalizability to a younger patient population.