Imagine 1,000 people dying every single day.
That’s equivalent to the number of people who lose their lives to medical errors or identity issues annually. This devastating outcome is largely caused by the difficulties healthcare providers face in accurately identifying and matching patient records. According to another report, the number of duplicate records soars up to 12% in U.S. healthcare organizations. For large systems, the number goes up to 16%.
The complications with patient matching
Healthcare is transitioning- there are multiple data sources in a single healthcare system. With over 95% healthcare organizations using EHRs, matching records to the correct patient is even more complicated. Plus, as patients receive care across multiple settings and as organizations use different systems to store and share records, identification again become a challenge. The problems are only multiplying as value-based care contracts and population health management strategies encourage disparate systems to share patient data.
Typos, missing information, data entry errors, workaround- anything can result in duplicated or wrong records. Such data integrity issues don’t just put a patient’s health at risk. Duplicate data impacts everything from patient satisfaction to reimbursement- an average healthcare facility loses a whopping $17.4 million every year in denied claims and patient misidentification.
Why is patient matching important in this day and age?
With hundreds of data systems existing in healthcare, patient matching has become one of the critical pillars of interoperability. It has become extremely difficult for healthcare organizations to effectively exchange information without knowing what records should be linked together and which shouldn’t.
Out of everything, patient-centric care takes a hit. If the providers don’t have all of the patient’s information, or worse, have ‘wrong’ information, that could result in redundant tests, delayed care, or worse, patients losing their lives. On the financial front, providers need to track utilization and expenditure closely to deliver cost-effective care. Unknown costs cannot be identified with distributed data, let alone managed. The effectiveness of interventions and care programs is hard to determine if they can’t be compared in the context of other services provided.
Beating the complexities of patient identification
Advancing data-driven patient care, one of the largest health systems in North Dakota was determined to create an one-stop access to patient records and improve the accuracy of matching its patients. With multiple entities under their umbrella, the health systems experienced problems keeping a track of their patients and their utilization across numerous data sources inside their own organization. Additionally, they were wrestling with creating an accurate and comprehensive picture of the population that could detail the care timeline for each patient.
As part of its overall strategy, the health system developed a single repository of clinical, financial, claims, and operational data. This repository, a Hadoop-based Big data lake, aggregate data incoming from three different facilities. The health system dealt with patient demographics, appointments, and transaction data along with regular ADT feeds and CCDA documents and needed to integrate the records together to implement an analytic structure. Without a method to link the data coming from these systems using a common patient identifier, the health system couldn’t accurately identify and link patient outcomes.
The health system established an enterprise master patient index, EMPI, to accurately correlate data from various sources. The algorithms used a combination of fuzzy logic and exact matching to identify records. Fields such as first name, middle name, and last name were matched using fuzzy logic. For example, a single patient had his name recorded as Robert Woods, Rob Woods, and Robert J. Woods across different settings. The fuzzy logic could parse the patient identity through such variations. The exact logic, on the other hand, was run on fields which had to have a unanimous value- such as date of birth, or gender. The algorithms ran behind the scenes, and without the need for manual intervention.
Post EMPI assignment and identification, the health system could create holistic, 360-degree patient records. The records had complete patient information from associated provider and episode history to prescribed medications and risk scores. These Patient 360 records were used by the health system to identify the gaps in care, proactively manage population health, and ensure cost-effective care delivery across the network.
The health system established streamlined records for more than 27,000 patients. Over 160,000 patient documents were extracted and integrated to support data-driven care management.
The road ahead
Enabling a single source-of-truth has become an essential prerequisite as healthcare goes digital. Data quality has to be paramount- otherwise, we’re just integrating and moving ‘wrong data’ faster and faster through the system. At later stages, it becomes hard and expensive to clean up databases. Improving the quality and safety of patients should come first, and the sooner the errors are found, the better. As data becomes intertwined with healthcare, there has to be something that powers insights. It’s time for a trusted, safe source that smoothens the fragments, keeps information secure, and above all, keeps patients safe. One patient, one record, all empowered insights.
To learn how robust patient matching can enhance population health management, get a demo.
For more updates, subscribe.
Join Team Innovaccer at booth #649 this HIMSS ’18, from March 5 to 9, at Venetian Palazzo, Sands Expo Center in Las Vegas, and learn how we can assist you in delivering an efficient, data-driven healthcare.