Predicting inadequate postoperative pain management in depressed patients: A machine learning approach

Area of Science:

• Pharmacology
• Clinical Medicine
• Data Science

Background:

• Prodrug opioids require CYP-2D6 liver enzyme activation for analgesia.
• Selective serotonin reuptake inhibitors (SSRIs) inhibit CYP-2D6, potentially reducing opioid efficacy.
• The interaction between SSRIs and prodrug opioids in postoperative pain management is not well understood.

Purpose of the Study:

• To identify patients using both SSRIs and prodrug opioids postoperatively.
• To examine the impact of this combination on postoperative pain control.
• To develop predictive models for pain management outcomes.

Main Methods:

• Utilized electronic health record (EHR) data from an academic medical center over 9 years.
• Developed and validated natural language processing (NLP) algorithms for extracting depression-related information.
• Employed machine learning models with 10-fold cross-validation to predict pain scores.

Main Results:

• 14.1% of 4,306 surgical patients with depression were on both SSRIs and prodrug opioids.
• Patients on SSRIs and prodrug opioids reported significantly worse pain control post-discharge (p < .01).
• Machine learning models accurately predicted pain score changes (AUCs: 0.87, 0.81, 0.69).

Conclusions:

• SSRIs significantly impair prodrug opioid effectiveness, leading to poorer pain control in depressed surgical patients.
• Current prescribing may overlook this critical drug interaction.
• Direct-acting opioids may be a better choice for depressed patients on SSRIs.