Skip to main content

Table 2 Regression coefficients and odds ratios with 95% CI from the original and internally validated models

From: Development and internal validation of a prediction model to identify older adults at risk of low physical activity levels during hospitalisation: a prospective cohort study

 

Original Model

 

Model after internal validation

Variable

Regression coefficient

Odds Ratio (95% CI)

p-value

Regression coefficienta

Model 1—Cut-off value for low or high PA levels at 50.0% of the cohort (64.4 min standing/walking)

 Intercept

2.042

-

.000

1.942

 SPPB

-.251

.778 (.677 – .894)

.000

-.239

 AM-PAC (independent)

-.894

.409 (.159 – 1.054)

.064

-.850

 Sex (female)

-.519

.595 (.269 – 1.313)

.199

-.493

Model 2—Cut-off value for low or high PA levels at 33.3% of the cohort (47.2 min standing/walking)

 Intercept

7.008

-

.022

6.255

 SPPB

-.305

.737 (.608 – .894)

.002

-.275

 AM-PAC (independent)

-1.124

.325 (.115 – .921)

.034

-1.012

 Age

-.078

.925 (.861 – .994)

.034

-.070

 Walking aid

    

  Crutch/Cane

-.006

.994 (.248 – 3.977)

.993

-.006

  Walker

1.281

3.601 (1.317 – 9.843)

.013

1.153

To estimate the individual probability of low PA levels during hospitalisation:

  Model 1: P(Low PA) = 1 / (1 + e(−(1.942 − .239*SPPB − .850*AM−PAC − .493*Female))) *100%

  Model 2: P(Low PA) = 1 / (1 + e(−(6.255 − .275*SPPB − 1.012*AM−PAC − .070*Age − .006*Crutch/Cane + 1.153*Walker))) *100%

  1. CI confidence interval, PA Physical Activity, SPPB Short Physical Performance Battery, AM-PAC Activity Measure for Post-Acute Care Inpatient Basic Mobility short form
  2. aRegression coefficients after adjustment for overfitting by shrinkage (shrinkage factor model 1 = 0.95 and model 2 = 0.90); the intercept was re-estimated