[Correlation between symptoms and diagnosis in pharmacotherapeutic decision process]

Area of Science:

• Psychiatry
• Pharmacology
• Medical Informatics

Context:

• Psychiatric in-patients often require tailored drug therapies.
• The selection of neuroleptics versus antidepressants can be complex.
• Standard diagnostic criteria and symptom assessments are crucial in treatment planning.

Purpose:

• To evaluate the influence of psychopathological symptoms (using the AMDP System) and diagnoses (ICD-9) on the choice of drug therapy (neuroleptics or antidepressants) in psychiatric in-patients.
• To compare decision-making procedures based on symptoms alone, diagnoses alone, or a combination of both.
• To identify the most accurate model for predicting drug therapy selection.

Summary:

• Statistical analysis of 3745 patients using CART-Analysis explored three decision-making procedures for drug therapy selection.
• Neither symptoms nor diagnoses alone adequately explained the choice of neuroleptics or antidepressants, showing significant reclassification errors.
• A hierarchical model integrating both diagnosis and symptoms achieved the lowest error rates (17.1% for neuroleptics, 10.1% for antidepressants), indicating superior predictive accuracy.

Impact:

• This research provides a data-driven approach to optimize psychotropic medication selection in clinical practice.
• The findings can inform the development of clinical decision support systems for psychiatrists.
• Improved prediction of drug therapy choice can lead to more effective and personalized patient treatment outcomes.