ONC launching Patient Matching Algorithm Challenge
Recognizing that the misidentification of patients remains a difficult problem for healthcare organizations, the Office of the National Coordinator for Health Information Technology is planning to launch its Patient Matching Algorithm Challenge early next month.
“There’s a lot of work going on with patient matching in the industry,” says Steve Posnack, director of the ONC Office of Standards and Technology. “But with all the matching that’s gone on, there are very few benchmarks that are publicly available … about how well the algorithms that people are using to do patient matching should perform.”
Posnack defines patient matching as the process of comparing different demographic elements from different health IT systems to determine if they refer to the same patient. He also acknowledges the potentially negative effect that mismatching can have on patient safety.
The aim of ONC’s Patient Matching Algorithm Challenge is to shine “a little bit of sunlight and transparency around what the benchmarks should be and how well the current tools are performing (and to) see if there are other tools and algorithms out there that could do a better job potentially than what’s currently in use,” Posnack contends.
The ultimate goal of the challenge is to “spur the development of innovative new algorithms, benchmark current performance and help the industry coalesce around common metrics for success,” according to Posnack.
Participants in the challenge, which officially launches in early June, will be provided a dataset and will have their answers evaluated and scored against a master key. These participants will be given as many as 100 run-throughs to see how well they can match the patients in the dataset.
ONC will award as many as six cash prizes totaling $75,000. The major prize category will involve three prizes for the highest “F-Score”—a combination of best precision and recall. In addition, best in category prizes will be awarded for “best precision” (least mismatched patients), “best recall” (least missed matches) and “best first F-Score run.”
Earlier this year, the ECRI Institute ranked the top 10 patient safety concerns in 2017 for healthcare organizations, with patient identification errors ranking sixth overall.
Also See: 10 top patient safety concerns in 2017
Posnack notes that the College of Health Information Management Executives has launched and is currently conducting a $1 million National Patient ID Challenge designed to develop a solution that ensures 100 percent accuracy of every patient’s health data to reduce preventable medical errors.
He describes ONC’s Patient Matching Algorithm Challenge as being complementary to the CHIME’s National Patient ID Challenge, with finalists in the innovation round to be announced on May 12 and an ultimate winner announced in November.
“We have a great relationship with CHIME and have certainly been in touch with them related to their challenge,” adds Posnack. “But, we’re approaching patient matching from a different angle. We’re really trying to spotlight how the algorithms perform and the benchmarks that we can establish from an industry perspective.”
ONC will conduct three informational webinars—May 10, 17 and 24—to provide additional details about the challenge and participant requirements.
“We hope our challenge will get the attention of experts from other industries that deal with individual matching, such as the financial and airline industries,” Posnack concludes.