Of the 4164 hospital admissions sampled from the participating hospitals, 3745 patient charts (89.9%) were eligible for a full screening by the stage 1 reviewers (). Of these, 1527 (40.8%) were assessed as positive for 1 or more screening criteria () and were sent for detailed review by the physician reviewers.
Fig. 1: Review process for the Canadian Adverse Events (AEs) Study. *Reasons for ineligibility were hospital stay less than 24 hours (n = 261), obstetrics patient (n = 56), patient transferred from other hospital (n = 48), cardiac arrest on arrival and (more ...)
In stage 2, the physician reviewers identified a total of 1133 injuries or complications in 858 charts. In 401 (46.7%) of these charts the injuries resulted in death, disability at the time of discharge or prolonged hospital stay. In 255 of the charts one or more of the AEs were rated 4 or higher on the 6-point causation scale (). The total number of AEs in these charts was 289, and 27 (10.6%) of the charts indicated more than 1 AE. After weighting for the sample frame, the overall AE rate was 7.5% (95% CI 5.7%–9.3%).
The proportion of AEs by the timing of occurrence and detection relative to the index hospital admission is displayed in .
Fig. 2: Timing and occurrence of AEs relative to index hospital admission. Two of the 289 AEs were excluded because of incomplete timing data. O = occurrence, D = detection of AE.
There was a trend for AEs to occur more frequently in the teaching hospitals than in the large community or small hospitals (). The trend was significant for AEs across the 3 hospital types (p < 0.001) but not for preventable AEs (p = 0.8). When we adjusted for comorbidities, age and sex, the adjusted rate for teaching hospitals was significantly higher than the adjusted rate for the nonteaching hospitals ().
Of the 255 patients who experienced AEs, 106 (41.6%) were judged to have 1 or more AEs with a high preventability rating (4 or more on the 6-point preventability scale ). In 39 (15.3%) of the 255 patients, preventability was judged to be “virtually certain.” A brief description of the clinical details of AEs occurring in the 255 patients, grouped according to the maximum preventability score, is provided in Appendix 3 (www.cmaj.ca/cgi/content/full/170/11/1678/DC3
). When these results were adjusted for the sampling strategy, we calculated that highly preventable AEs occurred in 36.9% (95% CI 32.0%– 41.8%) of the patients with AEs. Similarly, death was estimated to have occurred in 20.8% (95% CI 7.8%– 33.8%) of those with AEs, and 9% of these AEs were judged to have been highly preventable. The weighted rate of preventable AEs was similar across all 3 hospital types, ranging from 2.5% in the large community hospitals to 3.3% in the small and teaching hospitals.
Most (64.4%) of the AEs resulted in no physical impairment or disability, or in minimal to moderate impairment with recovery within 6 months. However, 15 (5.2%) of the AEs resulted in permanent disability, and 46 (15.9%), occurring in 40 patients, resulted in death (). When these results were adjusted for the sampling strategy, we estimated that death would be associated with an AE in 1.6% of patients with similar hospitalizations in Canada. The rate of preventable AEs across all hospitals was 2.8% (95% CI 2.0%– 3.6%), and the rate of deaths from preventable AEs was 0.66% (95% CI 0.37%– 0.95%).
We found that patients who experienced AEs had longer stays in hospital than did those without AEs (). The physician reviewers, using their professional judgement, estimated that the 255 patients with AEs required an additional 1521 days in hospital directly related to their AEs.
For 51.4% of the AEs, the service most responsible for the delivery of care was surgery, for 45.0% it was medicine and for 3.6% it was another service (e.g., dentistry, physical therapy, podiatry). The most common types of AEs were related to surgical procedures, and the next most common were associated with drug- or fluid-related events (). In the medicine service, AEs resulting from errors of omission (the failure to carry out necessary diagnosis or treatment) were more common than those resulting from errors of commission (57.1% v. 42.9%). In the surgery service, the frequency of these errors was assessed as being roughly equal (50.8% v. 49.2%).
The mean age (and standard deviation) of patients was significantly higher among those experiencing an AE than among those who did not have an AE (64.9 [16.7] v. 62.0 [18.4] years; p = 0.016). There was no difference between female and male patients in their risk of AE.