How Representative is the AAOS Joint Replacement Registry?

Garbage in, Garbage out.  That is the #1 risk of any dataset, data stream or registry.  Now that the AAOS Joint Replacement Registry (AJRR) has become the nation’s largest such repository of joint surgery data, how representative is it in terms of national demographics, inpatient utilization, access, cost, quality, and outcomes?

A new study titled “Is American Joint Replacement Registry Data Representative of National Data? A Comparative Analysis,” has just been published in the August 25, 2021, edition of The Journal of the American Academy of Orthopaedic Surgeons.

OTW spoke with the study authors, and they explained that “generalizability (or external validity) is crucial for measuring how applicable findings and conclusions from a sample are to the cohort at large.”  To determine the external validity of the American Joint Registry data, the authors collected AJRR data for all patients undergoing total hip arthroplasty (THA) or total knee arthroplasty (TKA) between 2012 to 2018 and then compared that to similar data between 2012 to 2016 in the National Inpatient Sample dataset.

The team looked for differences between the two databases and measured the magnitude of those differences in terms of demographics, hospital volume, and geographic characteristics.

Co-author James I. Huddleston, III, M.D., vice chair of the AJRR Steering Committee and an orthopedic surgeon at Stanford Health Care, explained to OTW that the team “needed to allow the AJRR to mature to the point where there were enough data to represent a ‘national’ sample.”  And mature it did. According to the study authors, as of October 1, 2020, the AJRR includes over 2.1 million hip and knee arthroplasty procedures (more than 1,400 medical institutions and 11,000 surgeons.)

For this work, the authors analyzed:

• 2,316,345 primary THA’s from the National Inpatient Sample
• 557,684 THA’s from the AJRR
• 3,417,700 TKAs from the National Inpatient Sample
• 809,494 TKAs from the AJRR

They found that the proportions of AJRR and National Inpatient Sample patients were similar based on overall sex, small differences were identified between databases considering hospital volume and geography.

“The AJRR was underrepresented in Southern regions and hospitals with low procedure volume and overrepresented in Northern hospitals and those with larger volume. Both the National Inpatient Sample and the AJRR followed a similar overall trend, with most procedures performed at hospitals with <50 cases per year,” wrote the authors.

“When compared to the Nationwide Inpatient Sample,” said Dr. Huddleston to OTW, “the AJRR showed similar hospital volumes, patient ages, and geography. This comparison supports the generalizability of AJRR data to a larger national sample.”