Jean Guglielminotti1, Catherine Deneux-Tharaux, Cynthia A Wong, Guohua Li. 1. From the *Department of Anesthesiology, Columbia University College of Physicians and Surgeons, New York, New York; †Département d'Anesthésie-Réanimation, APHP, Hôpital Bichat-Claude Bernard, Paris, France; ‡INSERM, UMR 1137, IAME, Paris, France; §INSERM, U1153, CRESS, EPOPé, Paris, France; ‖Department of Anesthesia, University of Iowa Carver College of Medicine, Iowa City, Iowa; and ¶Department of Epidemiology, Columbia University Mailman School of Public Health, New York, New York.
Abstract
BACKGROUND: Marked variation across hospitals in adverse maternal outcomes in cesarean deliveries is reported, including anesthesia-related adverse events (ARAEs). Identification of hospital-level characteristics accounting for this variation may help guide interventions to improve anesthesia care quality. In this study, we examined the association between hospital-level characteristics and ARAEs in cesarean deliveries and assessed individual hospital performance. METHODS: Discharge records for cesarean deliveries, ARAEs, and patient characteristics in the State Inpatient Database for New York State 2009 to 2011 were identified with International Classification of Diseases, Ninth Revision, Clinical Modification codes. The hospital reporting index was calculated as the sum of International Classification of Diseases, Ninth Revision, Clinical Modification codes divided by the number of discharges. Data on hospital characteristics were obtained from the American Hospital Association and the Area Health Resources files. Multilevel modeling was used to examine the association of hospital-level characteristics with ARAEs and to assess individual hospital performance. RESULTS: The study included 236,960 discharges indicating cesarean deliveries in 141 hospitals; 1557 discharges recorded at least 1 ARAE (6.6 per 1000; 95% confidence interval [CI], 6.2-6.9). The following factors were associated with a significantly increased risk of ARAEs: Charlson comorbidity index ≥ 1 (adjusted odds ratio [aOR], 1.2), multiple gestation (aOR, 1.3), postpartum hemorrhage (aOR, 1.5), general anesthesia (aOR, 1.3), hospital annual cesarean delivery volume <200 (aOR, 2.3), and reporting index (aOR, 1.1 per 1 increase per discharge). Fifteen percent of the between-hospital variation in ARAEs was explained by the hospital annual cesarean delivery volume and 6% by the reporting index. Eight hospitals (6%) were classified as good-performing, 104 (74%) as average-performing, and 29 (21%) as bad-performing hospitals. Compared with good-performing hospitals, a 2.3-fold (95% CI, 1.7-3.0) and 5.9-fold (95% CI, 4.5-7.8) increase in the rate of ARAEs was observed in average- and bad-performing hospitals, respectively. Bringing up bad-performing hospitals to the level of average-performing hospitals would prevent 466 ARAEs (30%). CONCLUSIONS: Low cesarean delivery volume is the strongest hospital-level predictor of ARAEs in cesarean deliveries and the main determinant of between-hospital variation. Future study to identify other factors and interventions to improve performance in bad-performing hospitals is warranted.
BACKGROUND: Marked variation across hospitals in adverse maternal outcomes in cesarean deliveries is reported, including anesthesia-related adverse events (ARAEs). Identification of hospital-level characteristics accounting for this variation may help guide interventions to improve anesthesia care quality. In this study, we examined the association between hospital-level characteristics and ARAEs in cesarean deliveries and assessed individual hospital performance. METHODS: Discharge records for cesarean deliveries, ARAEs, and patient characteristics in the State Inpatient Database for New York State 2009 to 2011 were identified with International Classification of Diseases, Ninth Revision, Clinical Modification codes. The hospital reporting index was calculated as the sum of International Classification of Diseases, Ninth Revision, Clinical Modification codes divided by the number of discharges. Data on hospital characteristics were obtained from the American Hospital Association and the Area Health Resources files. Multilevel modeling was used to examine the association of hospital-level characteristics with ARAEs and to assess individual hospital performance. RESULTS: The study included 236,960 discharges indicating cesarean deliveries in 141 hospitals; 1557 discharges recorded at least 1 ARAE (6.6 per 1000; 95% confidence interval [CI], 6.2-6.9). The following factors were associated with a significantly increased risk of ARAEs: Charlson comorbidity index ≥ 1 (adjusted odds ratio [aOR], 1.2), multiple gestation (aOR, 1.3), postpartum hemorrhage (aOR, 1.5), general anesthesia (aOR, 1.3), hospital annual cesarean delivery volume <200 (aOR, 2.3), and reporting index (aOR, 1.1 per 1 increase per discharge). Fifteen percent of the between-hospital variation in ARAEs was explained by the hospital annual cesarean delivery volume and 6% by the reporting index. Eight hospitals (6%) were classified as good-performing, 104 (74%) as average-performing, and 29 (21%) as bad-performing hospitals. Compared with good-performing hospitals, a 2.3-fold (95% CI, 1.7-3.0) and 5.9-fold (95% CI, 4.5-7.8) increase in the rate of ARAEs was observed in average- and bad-performing hospitals, respectively. Bringing up bad-performing hospitals to the level of average-performing hospitals would prevent 466 ARAEs (30%). CONCLUSIONS: Low cesarean delivery volume is the strongest hospital-level predictor of ARAEs in cesarean deliveries and the main determinant of between-hospital variation. Future study to identify other factors and interventions to improve performance in bad-performing hospitals is warranted.
Authors: David Metcalfe; Cheryl K Zogg; Elliott R Haut; Timothy M Pawlik; Adil H Haider; Daniel C Perry Journal: Int J Epidemiol Date: 2019-12-01 Impact factor: 7.196