The Use of General Anesthesia for Cesarean Delivery: A Population-Based Sample from New York State
Abstract Number: OP2-2
Abstract Type: Original Research
INTRODUCTION : The proportion of general rather than neuraxial anesthesia (GA) for cesarean delivery (CD) has been proposed as a quality measure for obstetric anesthesia,(1) and yet, little empiric data is available about its use at a population level in the United States. We analyzed New York Inpatient administrative data, which includes a categorical variable for anesthesia, to identify temporal trends and predictors of the use of GA for CD.
METHODS : Hospitalizations for delivery were extracted for 1998-2008 using an enhanced algorithm of ICD-9-CM diagnostic and procedural codes. Patients were included in our analysis if ≥90% of the CD records from their delivering hospital reliably specified type of anesthesia (i.e. regional or GA). Logistic regression models including all CDs clustered by hospital, were used to identify independent predictors of the use of GA, and to assess for temporal trends after adjusting for relevant demographic characteristics, obstetric complications, and hospital-level variables.
RESULTS : We identified 2,638,853 admissions for delivery, including 757,256 CDs in 195 unique hospitals. 258,886 women from 95 hospitals met inclusion criteria. Across all years, GA was used in 8.7% of all CDs; this rate decreased from 15.5% in 1998 to 4.7% in 2008, (P<0.001 for trend). This temporal trend persisted after adjusting for patient age and race, obstetric complications, and hospital level variables (adjusted odds ratio [aOR] 0.82 per year increase, 95% confidence interval [CI] 0.77, 0.88, P<0.001). The strongest predictors of whether a woman would receive GA for CD were: 1) the occurrence of additional surgery, specifically an exploratory laparotomy (aOR 7.7, 95% CI 5.0,12.0) or hysterectomy (aOR 5.1, 95% CI 3.4, 7.6), and 2) payer mix (compared with >72% private insurance; 60-72% aOR=1.8, 95% CI 1.2, 2.7; 48-59.9% aOR=6.5, 95% CI 2.3, 18.7; <48% aOR=4.7, 95% CI 2.8, 8.0). In 2008, the general anesthesia rate was below 15% in 95% of hospitals, below 10% in 82%, below 5% in 50%, and below 2.5% in 23% of institutions analyzed.
CONCLUSION : In this sample of hospitals, the use of GA for CD decreased markedly during the 11 year study period; adjustment for changes in maternal and obstetric characteristics did not significantly account for this change. Nevertheless, the use of GA for CD continues to vary by a factor of 5 between institutions. The fact that the delivering hospital’s payer mix is among the most important predictors of whether a woman would receive GA for CD suggests that institutional resources, socioeconomic differences in patients and/or the availability of epidural analgesia exert a powerful influence over anesthetic selection for CD. The GA rate for CD appears to demonstrate many desirable attributes of a useful quality measure,(2) including sensitivity to the equitable distribution of healthcare.