Crosswalks Between Patient-Reported Outcome Measures After Hip Arthroscopy

Abstract

Background:

Hip injuries such as femoroacetabular impingement syndrome (FAIS) and labral tears have become increasingly diagnosed and treated. However, different surgeons have utilized different hip-specific patient-reported outcome (PRO) scores, which makes between-database comparisons difficult. The aim of this study was to develop statistical crosswalks between commonly recorded hip-specific PRO indices for patients undergoing hip arthroscopy for FAIS.

Hypothesis:

High-fidelity statistical crosswalks would be generated between common hip-specific PROs.

Study Design:

Cohort study; Level of evidence, 3.

Methods:

A retrospective study was performed using a large single-institution database from 2014 to 2025. All patients undergoing hip arthroscopy for FAIS with both Hip disability and Osteoarthritis Outcome Score (HOOS) and modified Harris Hip Score (mHHS) at the preoperative, 3-month, 6-month, 1-year, 2-year, and 5-year visits, as well as sex, age, and race data, were included. Equipercentile equating was performed to create statistical crosswalks between the HOOS and mHHS PROs.

Results:

In this study of 677 patients, equipercentile equating was statistically warranted due to the following assumptions being verified: (1) the scores measured the same underlying construct, with correlation values >0.75; (2) the scores were unidimensional, as demonstrated by an eigenvalue ratio >3; and (3) there was invariance across subpopulations, as demonstrated by a root expected mean squared difference <0.11. Statistical crosswalks were successfully created that converted from HOOS to mHHS and vice versa. A chart demonstrating one-to-one conversion between mHHS and HOOS was generated. For example, an HOOS of 50 converts to an mHHS of 45, while an HOOS of 90 converts to an mHHS of 81. Similarly, an mHHS of 60 converts to an HOOS of 66. Sufficient accuracy of the crosswalks was demonstrated by Bland-Altman plots demonstrating minimal differences between actual and crosswalked scores.

Conclusion:

Statistical crosswalks with sufficiently high fidelity may be generated to convert between various hip-specific PROs. These conversion factors will allow pooling and meta-analysis of PRO data from patients undergoing FAI in various hip arthroscopy databases.