Gavin M Hamilton1, Reva Ramlogan, Anne Lui, Colin J L McCartney, Faraj Abdallah, Jason McVicar, Daniel I McIsaac. 1. From the Department of Anesthesiology and Pain Medicine (G.M.H., R.R., A.L., C.J.L.M., F.A., J.M., D.I.M.) School of Epidemiology and Public Health (D.I.M.), University of Ottawa, Ottawa, Canada The Ottawa Hospital Research Institute, Ottawa, Canada (R.R., A.L., C.J.L.M., F.A., D.I.M.) the Institute for Clinical Evaluative Sciences (IC/ES), Toronto, Canada (D.I.M.).
Abstract
BACKGROUND: Nerve blocks improve early pain after ambulatory shoulder surgery; impact on postdischarge outcomes is poorly described. Our objective was to measure the association between nerve blocks and health system outcomes after ambulatory shoulder surgery. METHODS: We conducted a population-based cohort study using linked administrative data from 118 hospitals in Ontario, Canada. Adults having elective ambulatory shoulder surgery (open or arthroscopic) from April 1, 2009, to December 31, 2016, were included. After validation of physician billing codes to identify nerve blocks, we used multilevel, multivariable regression to estimate the association of nerve blocks with a composite of unplanned admissions, emergency department visits, readmissions or death within 7 days of surgery (primary outcome) and healthcare costs (secondary outcome). Neurology consultations and nerve conduction studies were measured as safety indicators. RESULTS: We included 59,644 patients; blocks were placed in 31,073 (52.1%). Billing codes accurately identified blocks (positive likelihood ratio 16.83, negative likelihood ratio 0.03). The composite outcome was not significantly different in patients with a block compared with those without (2,808 [9.0%] vs. 3,424 [12.0%]; adjusted odds ratio 0.96; 95% CI 0.89 to 1.03; P = 0.243). Healthcare costs were greater with a block (adjusted ratio of means 1.06; 95% CI 1.02 to 1.10; absolute increase $325; 95% CI $316 to $333; P = 0.005). Prespecified sensitivity analyses supported these results. Safety indicators were not different between groups. CONCLUSIONS: In ambulatory shoulder surgery, nerve blocks were not associated with a significant difference in adverse postoperative outcomes. Costs were statistically higher with a block, but this increase is not likely clinically relevant.
BACKGROUND: Nerve blocks improve early pain after ambulatory shoulder surgery; impact on postdischarge outcomes is poorly described. Our objective was to measure the association between nerve blocks and health system outcomes after ambulatory shoulder surgery. METHODS: We conducted a population-based cohort study using linked administrative data from 118 hospitals in Ontario, Canada. Adults having elective ambulatory shoulder surgery (open or arthroscopic) from April 1, 2009, to December 31, 2016, were included. After validation of physician billing codes to identify nerve blocks, we used multilevel, multivariable regression to estimate the association of nerve blocks with a composite of unplanned admissions, emergency department visits, readmissions or death within 7 days of surgery (primary outcome) and healthcare costs (secondary outcome). Neurology consultations and nerve conduction studies were measured as safety indicators. RESULTS: We included 59,644 patients; blocks were placed in 31,073 (52.1%). Billing codes accurately identified blocks (positive likelihood ratio 16.83, negative likelihood ratio 0.03). The composite outcome was not significantly different in patients with a block compared with those without (2,808 [9.0%] vs. 3,424 [12.0%]; adjusted odds ratio 0.96; 95% CI 0.89 to 1.03; P = 0.243). Healthcare costs were greater with a block (adjusted ratio of means 1.06; 95% CI 1.02 to 1.10; absolute increase $325; 95% CI $316 to $333; P = 0.005). Prespecified sensitivity analyses supported these results. Safety indicators were not different between groups. CONCLUSIONS: In ambulatory shoulder surgery, nerve blocks were not associated with a significant difference in adverse postoperative outcomes. Costs were statistically higher with a block, but this increase is not likely clinically relevant.