Is Customer Data Enough to Optimize EMPI Patient Match Rates?
The ramifications of incorrectly matched patient records or unmatched patient records can be life threatening. Combining medical and medication history for two different people can lead to adverse drug events, incorrect diagnoses, unnecessary diagnostic tests and even death. Conversely, missing medical and medication history for an individual if multiple medical records for that individual are not linked together, results in a fragmented, incomplete medical record which can also lead to dire consequences.
Without a nationwide unique patient identifier, accurately matching multiple records for the same patient from disparate sources is a great challenge. We have been talking about the issues associated with patient identifiers for some time. The industry’s answer to this challenge is the Enterprise Master Patient Index (EMPI) which uses a variety of statistical algorithms to match patient records while simultaneously endeavoring to minimize the number of false positive and false negative matches.
Health information exchanges (HIEs) and integrated delivery networks (IDNs) prefer to err on the side of caution — they typically have a low to zero tolerance level for patient record mismatches. This means that they tend to implement systems that have a high match accuracy but a large number of false negative matches. This also results in quite a bit of time consuming, costly manual remediation of questionable matches.
Most HIEs and IDNs naively believe that their own data sources are sufficient for EMPIs to deliver the highest patient record match rates. However, individual data records are not designed to handle certain types of changes and updates. One simple example is an address change. Over time people move, change addresses and phone numbers, or potentially change names due to marriage or divorce. Many IDNs find that if a patient returns for care after a number of years and has a new address, they are unable to match that patient to his historic medical record that has the old address. Even a “fat fingers” typing error can prevent matching old and new records.
By looking beyond internal only date, EMPI match rates can rise dramatically. External data helps link together old and new addresses, or appropriately separate information about a divorced couple or adult child who has moved away from their parents’ house. Bob Smith, Jr. when approaching for care, does not want his father’s records (Bob Smith, Sr.) – showing potential heart disease, drug allergies, or other conditions – mixed up with his own.
Equifax has over 310M detailed consumer records in the U.S., including current demographic data (address, phone, etc.) and a robust archive of historical demographics and name changes. We have seen that the addition of external data results in higher match rates and fewer false negative matches for our customers. It also increases the productivity of those performing manual remediation of questionable record matches. An in-depth analysis we conducted against a full month’s worth of transactions across three randomly-selected customers — each representing a different market segment — showed match rates of up to 98% when they included data outside their own internal records or basic credit records.
Once accurate matches across various medical records have been determined, those records can be permanently linked within the EMPI. By incorporating 3rd party Big Data as a best practice to optimize patient record matching, care providers prevent adverse medical events and improve patient outcomes.
Recommended For You
Chalcy Raja leads the Big Data Platform team at Equifax. She has been a trailblazer in the area of big […]
What’s the best way to get someone to pay attention to marketing and advertising? Make it meaningful to them on […]
You have a lot of customer data — housed in a lot of different locations. You’re working hard to make […]
The CERCA Spring Conference, held on May 16, capped a broadly successful 2018 filing season that saw tax identity theft reduced by […]