Aretrospective cohort study was conducted using the Truven Health AnalyticsMarketScan Commercial Claims and Encounters (CCAE) and Red Book databases. TheCCAE database is an administrative healthcare database comprised of longitudinalcommercial claims from over 150 million health plan enrollees, mostly whom arecovered under large employers in the United States. Contents include detailed patient-levelinformation on prescription drug claims, as well as inpatient and outpatientmedical claims.
The Red Book is a pharmaceutical reference that containsinformation on prescription drug package size, strength, generic/proprietarynames, and quantity. Collectively, the CCAE and Red Book files enabled the identificationof subjects that underwent SRM, and the subsequent characterization ofparameters pertinent to the present study. All observations were de-identifiedand validated by Truven Health Analytics to ensure completeness, accuracy, andcompliance with the Health Insurance Portability and Accountability Act of 1996.
Consequently, this research was deemed exempt from review by an InstitutionalReview Board.The study population was initially selected on the basis ofhaving at least one outpatient services claim indicating SRM in any position betweenJanuary 1, 2010 and December 31, 2015. SRM was identified using the CommonProcedural Terminology (CPT) Code “19318.” The earliest claim indicating SRMfor each patient will be referred to as the “index date” henceforth. A study period of sixteen months (twelve months prior and fourmonths following the index date) was established a priori. A one-year pre-surgery period was selected to facilitate athorough and accurate capture of each study subject’s characteristics atbaseline.
To ensure completeness of our data, subjects were examined forcontinuous enrollment in a health plan with medical and prescription drugbenefits during the study period. Subjects with evidence of opioid exposure duringthe baseline period (i.e., between15 days and 1 year prior to the index date) were excluded. Lastly, only subjectswith at least one outpatient drug claim for an opioid medication during theperioperative period – defined as between 14 days prior and 14 days followingthe index date – were included in our analysis. To determine each patient’s opioid exposure throughout the studyperiod, the Red Bookwas cross-referenced to prescription drug claims in the CCAE database using NationalDrug Codes (NDC). Subsequently, all prescription claims for opioid drugs wereidentified by generic name using the following ingredient names: “oxycodone,” “hydrocodone,””codeine,” “hydromorphone,” “oxymorphone,” “fentanyl,” “morphine,” “meperidine,””buprenorphine,” “propoxyphene,” “methadone,” and “tramadol.
” Once identified,opioid prescriptions were evaluated at three intervals of the study period, asportrayed in Figure 1. The presenceof at least one opioid prescription during the postoperative period (between 15days and 120 days following the index date) was noted, and served as an indicatorof a subject’s refilling of their perioperative prescription. Opiatequantity, days supply, generic product name, and milligram strength werecollected for each study subject. Using these variables, the daily morphinemilligram equivalent (MME) of perioperative opioid prescriptions wasestablished by dividing the product of drug strength, drug quantity, and therelevant conversion factor by days supply indicated in each claim. Conversionfactors were derived from data published by the Centers for Disease Control andPrevention and Centers for Medicare and Medicaid Services.16