Supervised analysis of drug prescription sequences

Area of Science:

• Pharmacovigilance
• Clinical Informatics
• Health Data Science

Background:

• Hospitals possess extensive databases suitable for secondary analysis.
• Understanding drug effects in real-world clinical settings is crucial for patient safety and treatment optimization.
• Existing methods may not fully capture the nuanced impact of drug prescriptions within their typical medical context.

Purpose of the Study:

• To develop and validate a method for evaluating drug impact on laboratory results using existing hospital data.
• To assess the influence of drug prescription patterns on specific physiological outcomes.
• To demonstrate the utility of this approach through an analysis of potassium chloride supplementation and its effect on kalemia.

Main Methods:

• Constructing temporal prescription patterns incorporating medical context from hospital databases.
• Analyzing changes in laboratory results during short phases of inpatient stays.
• Comparing observed changes against baseline or monotonous trends to isolate drug effects.

Main Results:

• The developed method successfully identified the impact of potassium chloride supplementation on kalemia.
• The analysis highlighted the importance of considering the temporal and contextual patterns of drug prescription.
• Short-term changes in laboratory results correlated with specific drug prescription patterns.

Conclusions:

• This analytical technique provides a valuable tool for assessing drug effects on laboratory results within their common prescription contexts.
• The approach is applicable to post-marketing surveillance and understanding real-world drug performance.
• Further application of this method can enhance drug safety monitoring and clinical practice.