Skip to main content
. 2023 Sep 6;105(21):1695–1702. doi: 10.2106/JBJS.23.00247

TABLE II.

Multivariable Regression Coefficient Estimates for Factors Associated with the Choice of THA as the Surgery Type*

Coefficient Standard Error P Value
Age, in years
 66-69 Ref.
 70-74 −0.23 0.028 <0.001
 75-79 −0.60 0.027 <0.001
 80-84 −0.91 0.027 <0.001
 85-89 −1.15 0.028 <0.001
 ≥90 −1.36 0.033 <0.001
Sex
 Female Ref.
 Male 0.06 0.015 <0.001
Race/ethnicity
 Non-Hispanic White Ref.
 Black/African American −0.15 0.047 0.001
 Hispanic 0.008 0.044 0.84
 Other −0.05 0.047 0.31
No. of comorbidities −0.04 0.003 <0.001
Low-income status
 No Ref.
 Yes −0.33 0.024 <0.001
Census region of surgical facility
 Northeast Ref.
 South −0.05 0.02 0.02
 Midwest −0.02 0.02 0.30
 West −0.08 0.02 0.001
Candidate instrumental variables
 Overall facility volume of THAs 0.0002 0.00007 0.002
 Proportion of facility’s THAs performed for fracture −0.12 0.043 0.006
 Proportion of femoral neck fracture cases treated with HA −4.03 0.054 <0.001
*

The dependent variable is a binary indicator given the value of 1 if the patient received a THA and 0 if the patient received HA. The model was estimated with use of a probabilistic probit specification to account for the nonlinear, discrete nature of the dependent variable. The estimation procedure included a constant term and accounted for clustering (i.e., multiple observations within the same facility). The regression coefficients represent the adjusted marginal effect associated with a unit change in the indicator covariate, controlling for all other factors. Positive or negative coefficients indicate a higher or lower likelihood, respectively, of patients receiving THA relative to HA. Given the nonlinearity of the dependent variable, the probit regression coefficients do not represent the magnitude of the effect. To provide a sense of the magnitude of the effect for key factors, we utilized these regression coefficients to calculate the predicted (adjusted) probability of receiving a THA (relative to HA) associated with a 1-unit change in the indicator variable (e.g., Black/African American race), while holding all other factors constant at their actual values. These predicted probabilities are presented in the text.