A study to be presented today sought to gauge the predictive power of patient-reported outcome (PROM) data from the American Joint Replacement Registry (AJRR) on disease-specific outcomes one year after total hip arthroplasty (THA) and to assess the patient, clinical, and surgical factors predictive of disease-specific PROMs following THA.
AJRR contains a substantial volume of data before and after primary THA, including demographic, clinical, and surgical information, as well as PROM data. Little is known, however, of the ability of the data to help predict patient-specific outcomes, including both clinical improvement and likelihood of revision surgery.
In this study, conducted by Chancellor Gray, MD, FAAOS, and colleagues, the primary endpoint was the proportion of patients achieving minimal clinically important difference (MCID) and substantial clinical benefit (SCB) in hip disability and osteoarthritis outcome scores for joint replacement (HOOS-JR) following THA. Secondarily, researchers assessed factors associated with revision surgery at one year postoperatively.
Dr. Gray and colleagues analyzed data for all patients undergoing THA who had both preoperative and one-year postoperative PROM data submitted to AJRR from January 2012 through March 2020. A multivariable logistic regression was performed to determine independent associations between baseline (preoperative) PROMs; demographic, clinical, and surgical factors; and one-year postoperative HOOS-JR scores. Preoperative measures included both a general health measure (VR-12 or Patient-reported Outcomes Measurement Information System survey [PROMIS-10]) and a disease-specific measure (HOOS-JR); postoperative scores were focused on disease-specific outcomes. Further regression analyses were performed for baseline factors associated with achieving MCID or SCB and requiring revision surgery. MCID was determined with a distribution method—half of the pooled standard deviation of score change—and set to 8.2. SCB was determined with anchor-based parameters in the existing literature (22 for HOOS-JR).
During this period, 608,000 THAs were logged into the AJRR. Of that total, 23,059 patients (3.8 percent) had preoperative PROMs recorded; of those patients, 4,287 (18.6 percent) had a linked one-year postoperative PROM. The majority of THA patients with linked surveys had a PROMIS survey (69.8 percent) compared to a VR-12 survey (30.2 percent). Of the patients with baseline preoperative PROMs, 180 had revision surgery at one year (0.8 percent).
At one-year follow-up, 93 percent of patients achieved MCID, and 79 percent reached SCB via the HOOS-JR. For baseline PROM data, higher preoperative scores for physical health were associated with lower likelihood of achieving MCID and SCB.
Two specific patient factors independently predicted achievement of MCID and SCB: gender and hospital size. The procedure occurring in a medium-size hospital (100–399 beds) when compared to large hospital (>400 beds) suggested significant improved likelihood of achieving both MCID and SCB. Patients who were treated at nonteaching institutions were more likely to reach MCID compared to those treated at teaching institutions, although this finding was not mirrored for SCB.
Among institutions collecting VR-12, small- and medium-sized hospitals were less likely to have patients undergo linked revision THA surgery compared to large hospitals. Revision THA was more common, however, when an index surgery was performed at nonteaching versus teaching institutions.
Based on these data, AJRR represents a valuable opportunity to leverage a large repository of data, including longitudinally collected PROMs, to provide powerful preoperative predictors of patient outcomes and clinical improvement one year after THA. Moreover, there remains an opportunity to improve capture of PROM data from the American population—at present, only about 4 percent of patients who undergo a procedure reported to AJRR have linked PROMs, and only 20 percent of those have linked one-year postoperative PROMs.
Despite this relatively limited group, the data set was able to identify certain strong predictors of clinical improvement, the authors concluded. As longitudinal capture rate of the data increases, the data points available could be leveraged for more robust predictive analytics for outcomes following THA.
The study will be presented today as Paper 727 at 2:20 p.m. in Ballroom 6B.
Dr. Gray’s coauthors of “Predictors of One Year Outcomes after Total Hip Arthroplasty via the AJRR” are Prakash Jayakumar, MD, PhD; Paul A. Rizk, MD; James I. Huddleston III, MD; Kevin J. Bozic, MD, MBA; and Hari K. Parvataneni, MD.
Mita De, PhD, is AAOS director of data science. She can be reached at firstname.lastname@example.org. Terry Stanton is senior medical writer for AAOS Now. He can be reached at email@example.com.