The value of inpatient rehabilitation after uncomplicated knee arthroplasty: a propensity score analysis

Objective: To compare the effectiveness of rehabilitation after total knee arthroplasty (TKA) in models with or without an inpatient rehabilitation component.

Design, setting and participants: A propensity score-matched cohort of privately insured patients with osteoarthritis who underwent primary, unilateral TKA in one of 12 Australian hospitals between August 2013 and January 2015 were included. Those discharged to an inpatient facility because of poor progress or who experienced significant complications within 90 days of surgery were excluded.

Intervention: Discharge after surgery to an inpatient rehabilitation facility or home.

Main outcome measures: Patient-reported knee pain and function (Oxford Knee Score; at 90 and 365 days after surgery) and health rating (EuroQol “today” health scale; at 35, 90 and 365 days). Inpatient and community-based rehabilitation provider charges were also assessed.

Results: 258 patients (129 pairs) from a sample of 332 were matched according to their propensity scores for receiving inpatient rehabilitation; covariates used in the matching included age, sex, body mass index, and markers of health and impairment. The only significant difference in outcomes was that EuroQol health scores were better on Day 35 for patients not undergoing inpatient rehabilitation (median difference, 5; IQR, –10 to 19; P = 0.01). Median rehabilitation provider charges were significantly higher for those discharged to inpatient therapy (total costs: median difference, $9500; IQR, $7000–11 497; P < 0.001; community therapy costs: median difference, $749; IQR, $0–1980; P < 0.001).

Conclusions: Rehabilitation pathways incorporating inpatient rehabilitation did not achieve better joint-specific outcomes or health scores than alternatives not including inpatient rehabilitation. Given the substantial cost differences, better value alternatives should be considered for patients after uncomplicated TKA.