Skip to main content

Table 3 Discrimination and calibration of risk prediction tools for morbidity outcomes

From: The evaluation of risk prediction models in predicting outcomes after bariatric surgery: a prospective observational cohort pilot study

 

AUROC POMS-defined morbidity on D3:

AUROC (95% CI; standard error)

HL statistic (p value)

AUROC POMS-defined morbidity or failure to return to preoperative mobility on D3

HL statistic (p value)

EBWL < 50% at 1 year

HL statistic (p value)

AUROC: < 28 days alive and out of hospital

HL statistic (p value)

POSSUM

0.60

(0.50–0.69; 0.04)

1.85 (0.87)

0.63

(0.55–0.72)

1.36 (0.93)

0.60 (0.52–0.69; 0.04)

11.64 (0.04)

0.66 (0.58–0.74; 0.04)

19.74 (0.47)

OS-MRS

0.62

(0.53–0.70; 0.05)

NA

0.63

(0.55–0.71)

NA

0.59 (0.50–0.67; 0.04)

NA

0.62 (0.55 0.70; 0.04)

NA

  1. Note: No HL statistics calculated for the OS-MRS as this is an ordinal scale rather than providing a percentage prediction of the outcome
  2. POSSUM Physiology and Operative Severity Score for the enUmeration of Morbidity and Mortality, OS-MRS Obesity Surgery Mortality risk Score, AUROC area under the receiver-operator-characteristic curve, HL statistic Hosmer-Lemeshow chi-squared statistic, POMS Post Operative Morbidity Survey, EBWL excess body weight loss