EHR Data Accurately Identify Joint Implant Parts

Nicholas Giori, M.D., Ph.D.

Automated analysis of the electronic health record (EHR) can be an alternate way to assess implant use and performance in patients with total hip arthroplasty.

The automated approach was low cost and low burden for the health system.

Joint replacement is a huge cost area that has been of concern to Medicare. Major joint replacement or reattachment is one of Medicare’s top volume Medicare Severity Diagnosis Related Groups. In an effort to avoid denial of claims, medical records should contain enough detailed information to support the determination that major joint replacement surgery was reasonable and necessary.

Such surgeries are costly to health systems, too. In 2014, there were more than 400,000 hip and knee replacement surgeries, costing more than $7 billion for the hospitalizations alone. The average Medicare expenditure for surgery, hospitalization, and recovery ranges from $16,500 to $33,000.

Nicholas Giori, M.D., Ph.D., and a team of investigators quantified the extractability and accuracy of registry-relevant data from the EHR and assessed the ability of such data to track trends in implant use and the durability of implants using data from the largest integrated healthcare system in the U.S., Veterans Health Administration (VHA). The team included more than 37,000 patients receiving total hip arthroplasty at any VHA medical center from 2000 to 2017. The main data source was the VHA Corporate Data Warehouse database created in 2006 to collect information stored in the VHA’s distributed data repositories. Structured records were used to obtain data on date of surgery and numeric patient identifier.

Free text called “ScheduledProcedure” or “PrincipalPostOpDiagnosis” populating the Corporate Data Warehouse helped identified surgery side. Prosthesis part numbers were in a free-text field named “ProsthesisModel.” The cleaned part number was put onto the U.S. Food and Drug Administration Global Unique Device Identification Database (GUDID). The investigators identified company name, model, size, and whether the part was a major component of the total hip arthroplasty or not.

Giori and the investigative team determined whether patients had a subsequent total hip arthroplasty or revision anywhere in the VHA system on the same side. They calculated Kaplan-Meier curves for prosthesis used at least 100 times.

The team determined the annual number of primary and revision total hip arthroplasty procedures and the number and percentage of major parts identified by part number each year. For accuracy assessment, a surgeon reviewed 100 randomly selected total hip arthroplasty operations.

There were 45,351 total hip arthroplasty procedures identified with 192,805 implant parts and data completeness improved over the time. Following 2014, 85% of prosthetic heads, 91% of shells, 81% of stems, and 85% of liners used in the VHA healthcare system were identified by part number.

The study, “Assessment of Extractability and Accuracy of Electronic Health Record Data for Joint Implant Registries,” was published online in JAMA Network Open.