Predicting inpatient hypoglycaemia in hospitalized patients with diabetes: a retrospective analysis of 9584 admissions with diabetes

Kevin Stuart, Nicola Adderley, Tom Marshall, Gerry Rayman, Alice Sitch, Susan E. Manley, Sandip Ghosh, Konstantinos Toulis, Krishnarajah Nirantharakumar

Research output: Contribution to journalArticlepeer-review

14 Citations (Scopus)
228 Downloads (Pure)

Abstract

Aims Inpatient hypoglycaemia in patients with diabetes mellitus is associated with excess mortality, increased length of stay and increased complication rate. The objective of this study was to explore whether a quantitative approach to identify hospitalized patients with diabetes at risk of hypoglycaemia could be feasible by incorporating routine biochemical, haematological and prescription data. Methods A retrospective cross-sectional analysis of all diabetic admissions (n=9,584) from 1st January 2014 to 31st December 2014 was performed. Hypoglycaemia was defined as a blood glucose level of < 4 mmol/L. The prediction model was constructed using multivariable logistic regression, populated by clinically important variables and routine laboratory data. Results Using a pre-specified variable selection strategy, it was shown that the occurrence of inpatient hypoglycaemia could be predicted by a combined model taking into account background medication (type of insulin, use of sulphonylurea), ethnicity (Black and Asian), age (75+ years old), type of admission (emergency) and laboratory measurements (eGFR, CRP, sodium and albumin). ROC curve analysis revealed that the Area Under the Curve (AUC) was 0.733 (95% CI 0.719 to 0.747). The cut-off point chosen to maximize both sensitivity and specificity was 0.15. AUC obtained from internal validation did not differ from the primary model (0.731 (95% CI 0.717 to 0.746)).   Conclusions The inclusion of routine biochemical data, available at the time of admission, can add prognostic value to demographic and medication history. The predictive performance of the constructed model indicates potential clinical utility to identify patients at risk of hypoglycaemia during their inpatient stay.
Original languageEnglish
Pages (from-to)1385-1391
JournalDiabetic Medicine
Volume34
Issue number10
Early online date12 Jul 2017
DOIs
Publication statusPublished - 18 Sep 2017

Keywords

  • Hypoglycaemia
  • Diabetes
  • Inpatient
  • Prediction model

Fingerprint

Dive into the research topics of 'Predicting inpatient hypoglycaemia in hospitalized patients with diabetes: a retrospective analysis of 9584 admissions with diabetes'. Together they form a unique fingerprint.

Cite this