Toward learning from patient safety reporting systems.

PURPOSE: To evaluate the frequency and type of factors involved in incidents reported to a patient safety reporting system and answer specific questions to enhance the value of PSRS data to improve patient safety. MATERIALS AND METHODS: Prospective cohort study of incidents reported from adult and pediatric intensive care units (ICUs) in the United States to the web-based, voluntary, and anonymous Intensive Care Unit Safety Reporting System. Results from July 1, 2002, to June 30, 2004. Main outcome variables were incidents that could or did lead to patient harm. RESULTS: Analysis includes 2075 incidents from 23 ICUs. Median number of reports/ICU/month was 3; 5 hospitals submitted 58% of reports. Harm was reported in 42% of incidents with 18 deaths. Common event types: medication/therapeutics (42%) and incorrect/incomplete care delivery (20%); 48% of line/tube/drain incidents led to physical harm. Deficiencies in training/education contributed to 49% of incidents and teamwork issues 32%; 42% of incidents had 2 or more contributing factors. As the number of contributing factors per incident increased, so did risk of harm. CONCLUSIONS: The Intensive Care Unit Safety Reporting System provides a mechanism for multiple ICUs to identify hazards. Data trends show a correlation between multiple contributing factors and higher rates of harm. Further research is needed to help determine how to use PSRS data to improve patient safety.