The electronic health record company Epic offers financial incentives to healthcare systems that use its proprietary algorithms — and those algorithms can deliver inaccurate predictions, a Stat News investigation found.

Epic is the largest electronic health record company in the United States, and it holds the health records for around 250 million people. The company has around 20 algorithms designed to predict things like how long a patient might stay in the hospital, which patients might become seriously ill, and which might develop a deadly condition called sepsis.

Like many other groups that build health algorithms, Epic doesn’t publicly share details of how the algorithms are built. Researchers at hospitals that use Epic are able to scrutinize the tools, but any investigations are challenging: they can’t disclose proprietary information and may worry about complicating their employer’s relationships with the company if they publish negative results, Stat News reported.

The studies that have been done on the algorithms show that they often don’t work as well as advertised. One analysis of the sepsis algorithm, which Epic says works 76 percent of the time, found it was only right 63 percent of the time.

Despite the problems with the algorithms, Epic incentivizes hospitals to adopt them. The company’s “honor roll” program has awarded hundreds of thousands of dollars and, in some cases, nearly $1 million to healthcare systems that implement technical upgrades, including the algorithms. “It would be a terrible world where Epic is giving people a million dollars, and the end result is the patients’ health gets worse,” Glenn Cohen, director of the Petrie-Flom Center for Health Law Policy, Biotechnology, and Bioethics at Harvard University, told Stat News. The financial incentives could create a conflict of interest.

“Epic’s Honor Roll is a voluntary program which encourages the adoption of features that help save lives, increase data exchange, and improve satisfaction for patients, physicians, and health systems,” Epic said in a statement to Stat News.

Some healthcare systems with financial resources and robust information technology departments are able to analyze Epic’s off-the-shelf algorithms and adjust them to fit to their needs. Others, though, may not be able to examine the tools as closely. “The least sophisticated hospitals who don’t have the large data teams are also the ones for whom $1 million is likely to make the most difference,” Cohen said.