[Reporting of incidents. Experiences with medical registration systems]

Current incident reporting systems in Denmark are primarily focused on litigation. Epidemiological studies in the United States have shown that only 2.7% of adverse events that qualify for litigation are identified in such systems. Existing Danish reporting on adverse effects and complications is not exhaustive and often focused on new medicine/technologies and only a few specialties. Published findings on the quality of reporting systems are mostly based on local systems, certain specialties, procedures, or products. Observational biases are found in both mandatory and voluntary reporting systems. Current documentation does not support theories of higher coverage in mandatory reporting systems than in voluntary systems. Reporting systems run by authorities have shown significantly lower coverage than those run by medical professions. Anonymous and confidential systems have higher coverage than open systems. Fast, relevant, and constructive feedback to the informants increases the quantity and quality of reports. Risk managers and locally based systems could increase reporting through better possibilities for direct feedback, although local systems have difficulty in initiating major prophylactic actions. Foreign epidemiological studies have shown an adverse events incidence of 3-13%. However, more than half of the adverse events occur at such a low frequency that it is unlikely that isolated hospital systems can produce sufficient information for prophylactic action.