Developing a Prognostic Model to Predict Mortality in Patients with Acute Bacterial Meningitis

Area of Science:

• Infectious Diseases
• Medical Informatics
• Public Health

Background:

• Bacterial meningitis is a severe, life-threatening infection where timely treatment is critical.
• Delayed treatment of bacterial meningitis significantly increases mortality risk.
• Predictive tools are needed to identify high-risk patients for prompt medical intervention.

Purpose of the Study:

• To develop and validate a prognostic model for predicting mortality risk in probable bacterial meningitis cases.
• To identify key clinical features associated with increased mortality in bacterial meningitis.
• To provide a decision-making tool for early management of bacterial meningitis patients.

Main Methods:

• A decision tree algorithm was employed to build the prognostic model.
• The model utilized data from 3,923 suspected bacterial meningitis cases from Iran's national registry (2018-2019).
• Key prognostic features, including age and cerebrospinal fluid (CSF) protein levels, were identified.

Main Results:

• The developed model achieved 78% accuracy, 84% sensitivity, and 73% specificity in predicting mortality.
• Cerebrospinal fluid (CSF) protein level and patient age were identified as significant prognostic factors.
• High mortality risk (85.8%) was observed in patients over 65 CSF protein level and under 30 years of age.

Conclusions:

• The prognostic model offers valuable insights for early risk stratification of bacterial meningitis patients.
• Identifying high-risk individuals allows for timely admission to intensive care units (ICUs).
• This tool can enhance public health operations during infectious disease outbreaks and reduce overall mortality.