Publication Date


Document Type

Doctoral Dissertation

Academic Program

Clinical and Population Health Research


Quantitative Health Sciences

First Thesis Advisor

John E. Ware, Jr., PhD


Knee Replacement, Knee Arthroplasty, Patients, Pain, Quality of Life, Reproducibility of Results, Patient Outcome Assessment, Psychometrics


Dissertations, UMMS; Arthroplasty, Replacement, Knee; Patients; Pain; Quality of Life; Reproducibility of Results; Patient Outcome Assessment; Psychometrics


Background: Patient reports of pain and function are used to inform the need for and timing of total knee replacement (TKR) and evaluate TKR outcomes. This dissertation compared measurement properties of commonly-used patient surveys in TKR and explored ways to develop more efficient knee-specific function measures.

Methods: 1,179 FORCE-TJR patients (mean age=66.1, 61% female) completed questionnaires before and 6 months after TKR. Patient surveys included the knee-specific Knee injury and Osteoarthritis Outcome Score (KOOS) and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and generic SF-36 Health Survey. Tests of KOOS and WOMAC measurement properties included evaluations of scaling assumptions and reliability. Item response theory methods were used to calibrate 22 KOOS function items in one item bank; simulated computerized adaptive tests (CAT) then were used to evaluate shorter function scores customized for each patient. Validity and responsiveness of measures varying in attributes (knee-specific versus generic, longer versus shorter, CAT versus fixed-length) were compared.

Results: KOOS and WOMAC scales generally met tests of scaling assumptions, although many pain items were equally strong measures of pain and physical function. Internal consistency reliability of KOOS and WOMAC scales exceeded minimum levels of 0.70 recommended for group-level comparisons across sociodemographic and clinical subgroups. Function items could be calibrated in one item bank. CAT simulations indicated that reliable knee-specific function scores could be estimated for most patients with a 55-86% reduction in respondent burden, but one-third could not achieve a reliable (≥ 0.95) CAT score post-TKR because the item bank did not include enough items vi measuring high function levels. KOOS and WOMAC scales were valid and responsive. Short function scales and CATs were as valid and responsive as longer KOOS and WOMAC function scales. The KOOS Quality of Life (QOL) scale and SF-36 Physical Component Summary discriminated best among groups evaluating themselves as improved, same or worse at 6 months.

Conclusions: Results support use of the KOOS and WOMAC in TKR. Improved knee-specific function measures require new items that measure higher function levels. TKR outcomes should be evaluated with a knee-specific quality of life scale such as KOOS QOL, as well as knee-specific measures of pain and function and generic health measures.



Rights and Permissions

Copyright is held by the author, with all rights reserved.