Construction and validation of a clinical prediction model for diabetic ketoacidosis

Chen XuXingwen JiangQuanan He^*Peng Xu^*

• Yangtze River Shipping General Hospital, Hubei, China

Background: Diabetes ketoacidosis (DKA) is a common and serious acute complication of diabetes mellitus. Globally, its incidence is on the rise, posing a serious threat to the life, health and quality of life of diabetic patients. In current clinical practice, although a variety of indicators are used to determine DKA, these indicators often have a lag and cannot effectively predict the occurrence of DKA at an early stage, resulting in some patients missing the best time for treatment. Objective: To investigate the risk factors for the development of DKA and establish a prediction model based on the information of type II diabetes mellitus. Methods: A total of 288 patients were collected in this study out of which 74 patients developed DKA. The patients enrolled in this study were randomly divided into a training set and a validation set according to a ratio of 7:3, with 201 patients in the training set and 87 patients in the validation set. The patients' past medical history, dietary habits and relevant information during hospitalisation were collected separately to study the correlation factors affecting the emergence of DKA in patients and to establish a prediction model. Results: Possibly relevant factors were included in a one-way logistic regression, and after analysing the results: history of infection, dietary status, duration of diabetes mellitus longer than 3 years, history of smoking, history of alcohol consumption, abnormalities in liver function, abnormalities in HbA1c, and hypokalaemia were potential risk factors for the development of DKA, P < 0.2; The data obtained were further included in a multifactorial review: history of infection, dietary status (intemperate diet), duration of diabetes mellitus more than 3 years, HbA1c abnormality, and hypokalaemia were predictive factors for DKA (P < 0.05). Conclusion: This model provides a predictive tool for clinicians to identify high-risk patients with DKA at an early stage, which can help to take targeted preventive and intervention measures before the onset of the disease. However, the model was developed and internally validated using hospital-based data and that external validation is required before wider clinical application.