BACKGROUND: The estimation of incidence rates of infectious diseases based on the sentinel surveillance data is rather rare. We attempted to estimate these in 2000 in Japan by the surveillance data, and to evaluate their biases. METHODS: We used the incidences of influenza-like illness and 12 pediatric diseases in each of the sentinel medical institutions in Japan based on surveillance data in 2000. The incidence in all medical institutions was estimated under the assumption that the sentinel medical institutions were randomly selected. The possible bias of this estimate was evaluated in comparison with the hypothetical true incidence obtained as the total incidence in all medical institutions estimated by a regression model using the numbers of all disease outpatients per day from the National Survey of Medical Care Institutions of Japan. RESULTS: The estimated annual incidence rate was 75.6 (95% confidence interval: 72.3-78.7) per 1,000 population in influenza-like illness, and ranged from 1.1 (95% confidence interval: 1.0-1.2) to 285.2 (95% confidence interval: 270.2-300.3) per 1,000 population aged 0-19 years among 12 pediatric diseases. The ratio of the estimated incidence to the hypothetical true one was 1.06-1.26 among influenza-like illness and the 12 pediatric diseases. CONCLUSIONS: The incidence rates of influenza-like illness and pediatric diseases in 2000 in Japan were estimated from sentinel surveillance data. The rates obtained provide some useful but not always accurate information. Thus, further research is necessary.
BACKGROUND: The estimation of incidence rates of infectious diseases based on the sentinel surveillance data is rather rare. We attempted to estimate these in 2000 in Japan by the surveillance data, and to evaluate their biases. METHODS: We used the incidences of influenza-like illness and 12 pediatric diseases in each of the sentinel medical institutions in Japan based on surveillance data in 2000. The incidence in all medical institutions was estimated under the assumption that the sentinel medical institutions were randomly selected. The possible bias of this estimate was evaluated in comparison with the hypothetical true incidence obtained as the total incidence in all medical institutions estimated by a regression model using the numbers of all disease outpatients per day from the National Survey of Medical Care Institutions of Japan. RESULTS: The estimated annual incidence rate was 75.6 (95% confidence interval: 72.3-78.7) per 1,000 population in influenza-like illness, and ranged from 1.1 (95% confidence interval: 1.0-1.2) to 285.2 (95% confidence interval: 270.2-300.3) per 1,000 population aged 0-19 years among 12 pediatric diseases. The ratio of the estimated incidence to the hypothetical true one was 1.06-1.26 among influenza-like illness and the 12 pediatric diseases. CONCLUSIONS: The incidence rates of influenza-like illness and pediatric diseases in 2000 in Japan were estimated from sentinel surveillance data. The rates obtained provide some useful but not always accurate information. Thus, further research is necessary.