Skip to main content
Journal of Patient-Reported Outcomes

Table 2 Prediction model estimates and bootstrap estimates

From: Development and internal validation of a predictive risk model for anxiety after completion of treatment for early stage breast cancer

Predictor

MI estimate

MI bootstrap estimate

B

SE

95% CI

Bb

SEb

Z

Bias

B 95% CI

HADS-A

0.734

0.034

0.67,0.80

0.734

0.027

27.12a

0.000

0.68, 0.79

HADS-D

0.094

0.043

0.01, 0.18

0.095

0.041

2.28b

0.001

0.02, 0.18

Age

−0.011

0.011

−0.03, 0.01

−0.010

0.008

−1.30

0.001

− 0.03, 0.01

Caring responsibility

0.485

0.274

−0.05, 1.03

0.488

0.200

2.43a

0.002

0.08, 0.87

Homeowner

−0.426

0.326

−1.07, 0.22

−0.432

0.263

−1.62

−0.006

−0.95, 0.08

Constant

2.515

0.793

0.95, 4.08

2.475

0.587

4.28b

−0.040

1.39, 3.67

  1. MI estimate: B = MI observed coefficient, SE B = standard error of B, 95% CI (confidence intervals)
  2. MI Bootstrap estimate: Bb = MI bootstrap estimates of coefficient, SEb = standard error of Bb, z = bootstrap estimate divided by the standard error, bias = bias for the parameter estimate, B 95% CI = bias corrected 95% CI
  3. All estimates are based on MI data (M = 50) and Bootstrap distribution across 1000 results (10,000 random samples with replacement) ap < 0.05 b p < =0.001
  4. Simple 6-month predicted anxiety = 2.5 + (HADS-A score × 0.7) + (HADS-D score × 0.1) + (age x − 0.1) + 0.4(if carer) + − 0.4(if homeowner)
Back to article page