Using triggers in primary care patient records to flag increased adverse event risk and measure patient safety at clinic level.

AIM: Using triggers to identify adverse events is proposed as an efficient means of consistently measuring, and tracking events that result in harm to patients. We aimed to test whether using triggers in our large provincial general practice could provide meaningful directions for improving safety.
METHOD: A literature review identified potential triggers and established the number of patients whose records we should review. Two teams independently reviewed 170 randomly selected patients' records for trigger presence and for evidence of harm relating to that trigger. All triggers were tested for sensitivity and specificity: triggers with low specificity were removed. Logistic regression was used on both initial and refined trigger sets to measure the odds ratio (OR) of harm occurring if a trigger was present.
RESULTS: Initially 36 triggers were identified. Applying these to 109.6 patient-years of records for 170 patients, we identified harm in the records of 46 (27.1%) patients. There were 7 occurrences of harm per 100 consultations (harm rate per consultation=0.07 (95% confidence interval [CI] 0.05-0.09) and 41 per 100 consulting patient years (95%CI 29-55). All harms related to medication use. The initial triggers were sensitive (0.98) but non-specific (0.08): removing triggers with low specificity left only 8. The OR for harm occurring using the initial triggers was 4.0 (95% 0.5-30) and using the refined trigger set OR=6.3 (95%CI 2.7-14.8).
CONCLUSION: 8 selected triggers are a useful way of measuring progress towards safer care for patients in primary care practice.