Aim of the study: The notified incidence of meningococcal disease in European countries varies from <1 case per 100,000 inhabitants to approximately 7 cases per 100,000. Assessing the true burden of disease is important for setting priorities in health services and for estimating the benefit of interventions such as vaccination. Completeness and timeliness of reporting is also essential for the early recognition of outbreaks. The objective of this study was to assess the completeness of surveillance data on invasive meningococcal disease in Austria at the National Reference Center for Meningococci for the year 2002.
Methods: The data stored at the reference center was compared with an independent database containing the main diagnosis documented in the obligatory hospital discharge dataset of all Austrian hospitals (coded in ICD-10 since 2001). All mismatches were reviewed in order to exclude possible errors and identify true cases of meningococcal disease that had not been reported to the reference center. The number of cases not recorded by either data source was estimated using the capture-recapture method.
Results: The first comparison of the two data sources identified 50 cases not registered at the national reference center. Screening of the ICD codes from these 50 patients through the hospitals reduced the number of under-reported cases to 10, of which 6 showed symptoms compatible with meningococcal disease, although microbiological confirmation was missing. Re-evaluation of the case histories of these 6 patients by a clinical expert for meningococcal disease identified them as probable cases. The main reason for correction of the diagnosis in 27 cases was an obvious coding error: these patients had been treated in hospitals for illnesses not related to meningococcal disease. In 72 cases, the two databases were in agreement. Eleven cases of meningococcal disease were notified solely to the national reference center. Addition of the newly recognized cases of invasive meningococcal disease increased the total number of cases from 83 (incidence, 1.03/100,000) to 93 (incidence, 1.16/100,000). Estimation of the "true" number of cases of meningococcal disease, using the capture-recapture method, gave a final total of 95 cases (95% CI, 93-98) and an incidence of 1.18/100,000. The completeness (sensitivity) of the original notification at the national reference center was therefore 87.4% (83 of 95 cases).
Conclusion: All probable cases of meningococcal disease, even those (still) lacking microbiological confirmation, should be reported to the public health authorities as soon as possible, in order to ensure the necessary prompt prophylactic action (e.g., chemoprophylaxis).