Epic, an electronic health record company, offers financial incentives for healthcare systems that use its algorithms. These algorithms can make inaccurate predictions.

Epic, the largest US electronic health record company, holds data on over 250 million Americans. There are around 20 algorithms that can predict what length of stay a patient will need in hospital. They also predict which patients may become seriously ill and whether they might succumb to sepsis.

Epic, like many other companies that create health algorithms, doesn’t make public details about how they are made. Epic tools are available for hospital researchers to examine, but they are not allowed to reveal proprietary information. They also have the risk of causing problems with their employer’s relationships if they publish negative results.

Studies done on algorithms have shown that they don’t always work 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.

Epic encourages hospitals to adopt the algorithm, despite the shortcomings. The “honor roll”, a program funded by Epic, has given hundreds of thousands of dollars to hospitals that have implemented technical upgrades. In some cases, it even gave nearly $1 million. Glenn Cohen, director at the Petrie-Flom Center for Health Law Policy and Biotechnology at Harvard University, said that Epic would give people a million dollars and that patients’ health will suffer. Financial incentives can lead to conflicts of interest.

Also Read: This is the most unhinged Trump rant about the 2020 election ye

Epic’s Honor Roll encourages the adoption features that save lives, increase data interchange, and improve patient satisfaction for physicians and patients,” Epic stated in a statement.

Epic’s pre-made algorithms can be modified by healthcare organizations with the right financial resources and information technology departments. However, others may not be as able to analyze the tools as carefully. Cohen stated that the hospitals with the least advanced data teams are the ones where $1 million will make the biggest difference.