Browse the corpus

Using Quantity of Service Data in Medicaid Claims to Estimate Anesthesia Time. BACKGROUND: Surgical procedure duration and anesthesia time are important variables, with applications including measurement of quality of care and quantity of anesthetic exposure. This study evaluates data from Medicaid anesthesia claims as a proxy for anesthesia time. METHODS: Fee-for-service anesthesia claims for appendectomies and cholecystectomies of beneficiaries under age 65 yr from the Medicaid Analytic eXtract (1999 to 2013) were identified. The quantity of service (QoS) variable in anesthesia claims was evaluated on three criteria: (1) reflection of expected anesthesia time, (2) correlation with Medicaid payment, and (3) state interpretation of QoS as anesthesia time by evaluating payments estimated using state-specific billing rules and correlation with actual payments. Spearman correlations were calculated with coefficients above 0.7 considered to be strong. RESULTS: A total of 902,492 fee-for-service anesthesia claims were identified from 48 states and Washington, D.C., with 14 states reporting QoS values in units, 19 reporting in minutes, and 16 reporting in both units and minutes. The median QoS values for appendectomy and cholecystectomy were 6 and 7 U in claims reporting time in units and 76 and 89 min in claims reporting time in minutes, respectively. The majority of unit states (9 of 14; 64%) and minute states (14 of 19; 74%) met all three criteria. For the 16 states reporting QoS values in both units and minutes, 8 of 16 (50%) reported unit claims and 9 of 16 (56%) reported minute claims that met the three criteria. CONCLUSIONS: The QoS variable reported on state Medicaid anesthesia claims had utility in approximating anesthesia time as minutes or 15-min units when evaluating anesthesia claims from two common surgical procedures. However, this variable does not accurately reflect anesthesia time in every state, and even in states where QoS was deemed to be usable, outliers and misclassified data need to be addressed.