A survey from Black Book Research found when healthcare organizations don’t have a master patient database across various departments, match rates when exchanging records with other health systems are much lower.

Using survey responses from nearly 1,400 health technology managers, Black Book said hospitals without enterprise master patient index (EMPI) tools have a 24 percent match rate when exchanging records between systems. In contrast, hospitals with EMPI have reported “consistently correct patient identification” since 2016, averaging 93 percent for registrations and 85 percent for records shared with non-networked providers.

“Despite the increases in record sharing among providers, increased risk and cost from redundant medical tests and procedures because of fragmented data trapped in siloes makes tracking patients especially difficult,” said Doug Brown, managing partner at Black Book.

Patient matching has been identified as a major problem for years by health information management groups and one of the many obstacles which has held back interoperability. As consolidation sweeps across the healthcare space, larger hospitals and systems have need to bring newly acquired practices and facilities up to their level on patient identification and EMPI tools.

Without those tools, there will be financial consequences for systems and hospitals. Inaccurate patient information is to blame for one-third of all denied claims, costing a hospital an average of $1.5 million annually and $6 billion for the U.S. healthcare system as a whole. Survey respondents said that prior to implementing an EMPI tool, an average of 18 percent of their organizations’ patient records were duplicates, which could result in extra costs for the patient—an average of $1,950 per inpatient stay and $800 per emergency room visit—as well as forcing hospitals to expend resources for months to clean up their records.

“Ultimately, the real challenge of identity management and parsing together a longitudinal health record has to do with integration and interoperability,” Brown said. “Many systems still do not communicate and store data in disjointed architectures and an upsurge of identifiers continue to be created.”