Aims: To determine the incidence of avoidable cardiac arrest among patients who had received resuscitation in a district general hospital. To establish how location and individual or system factors influence avoidable cardiac arrest in order to develop an evidence-based preventive strategy.
Methods: Expert panel review of case-notes from 139 consecutive adult in-hospital cardiac arrests over 1 year.
Results: There were 32,348 adult admissions in 1999 with 1,023 deaths. The cardiac arrest team was activated 139 times: 118 were for primary in-hospital cardiac arrest. The cardiac arrest rate excluding 'do not attempt resuscitation' (DNAR) cases was 3.8/1000 admissions. In 88.5% of deaths there was a DNAR policy. Survival to hospital discharge following resuscitation was 14%. Among the 118 cases, the panel unanimously agreed that 61.9% of arrests were potentially avoidable, rising to 68% when emergency department arrests were excluded (66 and 73% for majority opinion). Cardiac arrests were more likely at the weekend than during the week (P = 0.02). The odds of potentially avoidable cardiac arrest was 5.1 times greater for patients in general wards than critical care areas (P < 0.001); patients in critical care areas were more likely to survive (P < 0.001). The odds of potentially avoidable cardiac arrest was 12.6 times greater for patients nursed in a clinical area judged 'inappropriate' for their main complaint (P < 0.002, Fisher's exact test) compared to those nursed in 'appropriate' areas. The panel agreed that 100% of potentially avoidable arrests were judged to have received inadequate prior treatment. Clinical signs of deterioration in the preceding 24 h were not acted upon in 48%, and review was confined to a house officer in 45%.
Conclusion: The majority of treated in-hospital cardiac arrests are potentially avoidable. Multiple system failures include delays and errors in diagnosis, inadequate interpretation of investigations, incomplete treatment, inexperienced doctors and management in inappropriate clinical areas.