Effect of a Decision Aid on Access to Total Knee Replacement for Black Patients With Osteoarthritis of the Knee: A Randomized Clinical Trial

Abstract

Importance: Black patients with advanced osteoarthritis (OA) of the knee are significantly less likely than white patients to undergo surgery. No strategies have been proved to improve access to surgery for black patients with end-stage OA of the knee.

Objective: To assess whether a decision aid improves access to total knee replacement (TKR) surgery for black patients with OA of the knee.

Design, setting, and participants: In a randomized clinical trial, 336 eligible participants who self-identified as black and 50 years or older with chronic and frequent knee pain, a Western Ontario McMaster Universities Osteoarthritis Index score of at least 39, and radiographic evidence of OA of the knee were recruited from December 1, 2010, to May 31, 2014, at 3 medical centers. Exclusion criteria were history of major joint replacement, terminal illness, inflammatory arthritis, prosthetic leg, cognitive impairment, lack of a telephone, or contraindications to elective replacement surgery. Data were analyzed on a per-protocol and intention-to-treat (ITT) basis.

Exposure: Access to a decision aid for OA of the knee, a 40-minute video that describes the risks and benefits of TKR surgery.

Main outcomes and measures: Receipt of TKR surgery within 12 months and/or a recommendation for TKR surgery from an orthopedic surgeon within 6 months after the intervention.

Results: Among 336 patients (101 men [30.1%]; 235 women [69.9%]; mean [SD] age, 59.1 [7.2] years) randomized to the intervention or control group, 13 of 168 controls (7.7%) and 25 of 168 intervention patients (14.9%) underwent TKR within 12 months (P = .04). These changes represent a 70% increase in the TKR rate, which increased by 86% (11 of 154 [7.1%] vs 23 of 150 [15.3%]; P = .02) in the per-protocol sample. Twenty-six controls (15.5%) and 34 intervention patients (20.2%) in the ITT analysis received a recommendation for surgery within 6 months (P = .25). The difference in the surgery recommendation rate between the controls (24 of 154 [15.6%]) and the intervention group (31 of 150 [20.7%]) in the per-protocol analysis also was not statistically significant (P = .25). Adjustment for study site yielded similar results: for receipt of TKR at 12 months, adjusted ORs were 2.10 (95% CI, 1.04-4.27) for the ITT analysis and 2.39 (95% CI, 1.12-5.10) for the per-protocol analysis; for recommendation of TKR at 6 months, 1.39 (95% CI, 0.79-2.44) and 1.41 (95% CI, 0.78-2.55).

Conclusions and relevance: A decision aid increased rates of TKR among black patients. However, rates of recommendation for surgery did not differ significantly. A patient-centered counseling and educational intervention may help to address racial variations in the use of TKR for the management of end-stage OA of the knee.

Trial registration: clinicaltrials.gov Identifer: NCT01851785.

Figure 1. CONSORT Diagram

The control condition consisted of a booklet about osteoarthritis (OA) of the knee; the intervention, a video examining all treatment options for OA of the knee. ITT indicates intention to treat.

Figure 2. Observed Proportions for Study Outcomes

Data depict the patients in each study arm who received total knee replacement (TKR) surgery within 12 months or a recommendation for TKR surgery within 6 months. The control condition consisted of a booklet about osteoarthritis (OA) of the knee; the intervention, a video examining all treatment options for OA of the knee. ITT indicates intention to treat.

Figure 3. Study Outcomes

Site-adjusted odds ratios (ORs) compare the intervention and control arms for study outcomes.The control condition consisted of a booklet about osteoarthritis (OA) of the knee; the intervention, a video examining all treatment options for OA of the knee. TKR indicates total knee replacement.

Figure 4. Subgroup Analysis of Total Knee Replacement at 12 Months

Odds ratios (ORs) are shown for the intervention vs control groups at each level of age, sex, Western Ontario McMaster Universities Osteoarthritis Index (WOMAC) score (range, 0–100, with higher scores indicating increased pain, stiffness, and functional limitations), and baseline willingness, adjusted for site. P values for interaction were calculated from logistic regression models including the study group, the covariates and their interaction, and site. Age and WOMAC score were continuous, and baseline willingness was ordinal. The control condition consisted of a booklet about osteoarthritis (OA) of the knee; the intervention, a video examining all treatment options for OA of the knee. ITT indicates intention to treat.