The following is the summary of “Score-based prediction model for severe vitamin D deficiency in patients with critical illness: development and validation” published in the December 2022 issue of Critical care by Kuo, et al.

The risks of death, infections, and many other disorders are greatly amplified in those with severe vitamin D deficiency (SVDD). According to research, vitamin D deficiency is more common among critically ill patients than in the general population. To better predict SVDD in critically sick patients, a score-based model is created and tested in this multicenter retrospective cohort study.

There were 662 critically sick patients enrolled between October 2017 and July 2020. A serum 25(OH) D concentration of less than 12 ng/mL (or 30 nmol/L) was used as the diagnostic threshold for SVDD. Based on enrollment dates, the data were split into 2 groups: a derivation cohort and a validation cohort. The derivation cohort was subjected to multivariate logistic regression (MLR) to create an SVDD prediction model. On top of the MLR model, a score-based calculator (the SVDD score) was developed. Finally, the validation cohort was used to assess the accuracy and precision of the model. 

About 16.3% of the derivation cohort and 21.7% of the validation cohort had SVDD. 8 predictors were used to create the MLR model, which was then incorporated into the SVDD score. In the validation cohort, the SVDD score had an AUC (area under the curve) of 0.848 [95% CI 0.781-0.914] and an AUC (area under the curve) for precision-recall of 0.619 [95% CI 0.577-0.669]. This work aimed to create a straightforward score-based model for SVDD prediction in critically ill patients.