ISQUA17-2132IDENTIFICATION OF PATIENTS AT HIGH RISK FOR POOR PAIN MANAGEMENT USING CLINICAL PATHWAYS WITHIN EHRS

Abstract

Objectives: Pain management is a major challenge following surgery. Certain pain medications (prodrug opioids) need metabolic conversion to render them effective for pain management. New evidence suggests that selective serotonin reuptake inhibitors (SSRIs),the most commonly prescribed antidepressant medication, could inhibit the conversion of prodrug opioids, leaving an already vulnerable population at greater risk for poor pain management and increased opioid misuse. Leveraging the power of computational resources for processing the vast amount of medical information residing in EHRs, we aimed to characterize the prevalence of the prescribed combination of SSRIs and prodrugs post-operatively and further examine its effect on postoperative pain control. Methods: Using electronic health records (EHRs) from a single academic medical institute, we identified patients receiving 5 painintensive surgeries using ICD-9/10 codes. We define the fields within EHR and the logic applied to it to specify our variables of interest, e.g. depression, pain, or medications. While structured data are mapped to defined terms, free text are further processed. First we identify clinical terms of interest (e.g. clinical terminology or institution-specific synonyms) and then we include rule-based or machine-learning logic required to transform the relevant free-text to a term category (e.g. ensure a term is not negated). Using these data, we developed descriptive statistics to illustrate the prevalence of combined use of these medications. Maximum pain score at discharge and within 30-days post-discharge were dependent variables used in separate multivariate regression analyses to examine the effect of the combined use of antidepressant and prodrug opioids on post-operative pain. Results: This study included 8811 patients undergoing 5 surgical procedures at a large academic US hospital, 2008-2016. We found 9% were on SSRIs during their hospital stay and 4% received both an SSRI and Prodrug post-surgery. Our sample was predominantly white (62%), women (52%) with a mean age at surgery 59.4 years. Our modeled estimates suggest that patients receiving the combination of medications had higher discharge pain scores (Odds Ratio [OR]: 1.31) and follow-up pain scores (OR: 1.37) after controlling for important patient and clinical characteristics. Conclusion: Prescribing a combination of SSRI's and prodrug opioids post-operatively is common and results in poor pain control. Leveraging t clinical data captured in EHRs may improve the efficiency of pain management through the identification of patients in need of personalized regimes. This study adds to evidence of consequential drug interactions, while the high prevalence of concomitant use suggests that more attention to medication prescribing choices is warranted for post-operative pain management.