The incidence of adverse events in an Italian acute care hospital: findings of a two-stage method in a retrospective cohort study
1 General Direction, San Filippo Neri Hospital Trust, Via G. Martinotti 20, Rome 00135, Italy
2 Department of Public Health, Catholic University of Sacred Heart, L.go Francesco Vito 1, Rome 00168, Italy
3 Medical Direction, San Filippo Neri Hospital Trust, Via G. Martinotti 20, Rome 00135, Italy
BMC Health Services Research 2014, 14:358 doi:10.1186/1472-6963-14-358Published: 27 August 2014
The promotion of safer healthcare interventions in hospitals is a relevant public health topic. This study is aimed to investigate predictors of Adverse Events (AEs) taking into consideration the Charlson Index in order to control for confounding biases related to comorbidity.
The study was a retrospective cohort study based on a two-stage assessment tool which was used to identify AEs. In stage 1, two physicians reviewed a random sample of patient records from 2008 discharges. In stage 2, reviewers independently assessed each screened record to confirm the presence of AEs. A univariable and multivariable analysis was conducted to identify prognostic factors of AEs; socio-demographic and some main organizational variables were taken into consideration. Charlson comorbidity Index was calculated using the algorithm developed by Quan et al.
A total of 1501 records were reviewed; mean patients age was 60 (SD: 19) and 1415 (94.3%) patients were Italian. Forty-six (3.3%) AEs were registered; they most took place in medical wards (33, 71.7%), followed by surgical ones (9, 19.6%) and intensive care unit (ICU) (4, 8.7%). According to the logistic regression model and controlling for Charlson Index, the following variables were associated to AEs: type of admission (emergency vs elective: OR 3.47, 95% CI: 1.60-7.53), discharge ward (surgical and ICU vs medical wards: OR 2.29, 95% CI: 1.00-5.21 and OR 4.80, 95% CI: 1.47-15.66 respectively) and length of stay (OR 1.03, 95% CI 1.01-1.04). Among patients experiencing AEs a higher frequency of elderly (≥65 years) was shown (58.7% vs 49.3% among patients without AEs) but this difference was not statistically significant. Interestingly, a higher percentage of patients admitted through emergency department was found among patients experiencing AEs (69.7% vs 55.1% among patients without AEs).
The incidence of AEs was associated with length of stay, type of admission and unit of discharge, independently by comorbidity. On the basis of our results, it appears that organizational characteristics, taking into account the adjustment for comorbidity, are the main factors responsible for AEs while patient vulnerability played a minor role.