Skip to main content
. 2023 May 29;481(12):2368–2376. doi: 10.1097/CORR.0000000000002703

Table 4.

Multilevel linear regression of factors associated with PROMIS Physical Function including PCS, accounting for nesting by clinic location

Regression coefficient
(95% CI)
p value Δ-AIC
Pain Catastrophizing Scale -0.34 (-0.68 to 0.00) 0.053 1.7
Employment status
 Unemployed, student, or retired Reference value
 Employed 3.08 (0.23 to 5.94) 0.03 2.4
Assistive devices
 Crutches Reference value 12.5
 Knee scooter 1.44 (-4.24 to 7.11) 0.62
 Walker -2.64 (-6.75 to 1.46) 0.21
 Other 1.82 (-2.41 to 6.05) 0.40
 None 7.52 (3.66 to 11.38) < 0.001
Weightbearing status 6.5
 Full weightbearing Reference value
 Nonweightbearing -6.30 (-10.03 to -2.58) 0.001
 Partial weightbearing -2.64 (-6.31 to 1.03) 0.16

SHAI and PROMIS Depression have been dropped because of multicollinearity with PCS. Week to injury and insurance status have been dropped due to multicollinearity with weightbearing status and employment status, respectively. Δ AIC = Akaike Information Criterion; AIC of the full model was compared to the model without each variable. Higher values indicate better fit.