Stroke Severity Versus Dysphagia Screen as Driver for Post-stroke Pneumonia.

Area of Science:

• Neurology
• Infectious Diseases
• Health Services Research

Background:

• Post-stroke pneumonia is a serious complication associated with increased mortality and disability.
• Patients are often kept Nil By Mouth (NBM) post-stroke pending dysphagia screening to ensure safe oral intake.
• Understanding the drivers of post-stroke pneumonia is crucial for effective prevention and management strategies.

Purpose of the Study:

• To quantify the proportional contribution of stroke severity and dysphagia screening to the development of post-stroke pneumonia.
• To apply coalition game theory principles, specifically Shapley value, for a fair distribution of factors influencing pneumonia risk.
• To identify key predictors for post-stroke pneumonia using regression and decision tree analyses.

Main Methods:

• Retrospective analysis of 617 patients admitted to a stroke unit in 2015.
• Data partitioned into training (75%) and validation (25%) sets.
• Shapley value regression and conditional decision tree analysis employed to assess variable contributions to pneumonia.

Main Results:

• Stroke severity (National Institute of Health Stroke Scale/NIHSS) was the largest contributor (72.8%) to post-stroke pneumonia risk.
• Charlson Comorbidity Index (16.2%), age (7.2%), and dysphagia screen (3.8%) had smaller contributions.
• A decision tree identified an NIHSS threshold of 14 for predicting pneumonia, with high accuracy (AUC 0.83).

Conclusions:

• Stroke severity is the primary determinant of post-stroke pneumonia risk, overshadowing the direct contribution of dysphagia screening in this model.
• The study suggests that current dysphagia screening protocols may not fully capture the pneumonia risk associated with dysphagia itself.
• Further research is needed to refine dysphagia assessment and its integration into pneumonia prevention protocols post-stroke.