Financial penalties designed to encourage hospitals to avoid repeat hospitalizations are working, but they’re also associated with higher mortality rates for patients with heart failure, a U.S. study suggests.

The Affordable Care Act (ACA) of 2010 created punishments for hospitals with high rates of readmissions within 30 days of discharge for Medicare patients with three common problems: heart failure, pneumonia and heart attacks. Readmissions are considered a benchmark for the quality of care, and in theory hospitals that do a better job shouldn’t have as many patients returning soon after they’re sent home.

In one respect, the law worked as intended. Researchers examined data on 115,245 heart failure patients hospitalized from 2006 to 2014 and found readmission rates dropped from 20% before the law took effect to 18.4% after penalties kicked in.

But the proportion of patients who died within 30 days of going home increased from 7.2% to 8.6% over that same period. One-year mortality rates climbed from 31.3% to 36.3%.

“Nationwide, there may have been thousands to tens of thousands of extra deaths in patients with heart failure resulting from this policy,” said senior study author Dr. Gregg Fonarow of the David Geffen School of Medicine at the University of California Los Angeles.

“No level of reduction in readmissions or cost savings should be considered adequate justification for this level of potential harm,” Dr. Fonarow said by email.

The study wasn’t a controlled experiment designed to prove whether or how reducing readmission rates might influence survival odds for patients with heart failure.

It’s possible, however, that physicians may have made treatment decisions designed to avoid readmissions rather than to give patients the best possible care, Dr. Fonarow said.

Physicians might, for example, have postponed sending patients back to the hospital until after the 30-day window for readmission penalties had passed, allowing heart failure to worsen and decreasing survival odds, Dr. Fonarow said.

When patients did return to the hospital within that 30-day penalty window, they might have been kept in the emergency department or a general care unit for observation instead of being formally admitted to the hospital, Dr. Fonarow added. This might mean some people who needed intensive care or a specialized cardiac unit didn’t end up there.

To calculate shifts in readmission and mortality rates, researchers looked at three distinct time periods: before the ACA, from January 1, 2006 to March 31, 2010; during an ACA implementation period from April 1, 2010 to September 30, 2012; and after readmission penalties kicked in, from October 1, 2012 to December 31, 2014.