Use of an Internet-based community surveillance network to predict seasonal communicable disease morbidity.

Objectives: We designed an Internet-based surveillance network that linked community clinic diagnoses with viral isolation rates and admission patterns at a related children's hospital. We hypothesized that community surveillance would successfully predict subsequent hospital admissions and laboratory viral isolations. Secondarily, we expected the network to monitor trends in disease and that posting this information on a Web site would be useful to physicians in daily practice.
Study Design: Data were collected from December 1999 through August 2000. Information was summarized and posted weekly on a Web site. Active public piloting of the site took place during August 2000, after which the project was evaluated through an electronic mail survey. The predictive ability of the community surveillance data was evaluated by multivariate linear regression.
Results: Increases in the community diagnosis of most syndromes under surveillance, including lower respiratory infections (adjusted R(2) = 0.7086) and gastroenteritis (adjusted R(2) = 0.6532) successfully predicted an increase in subsequent hospital admissions. Community surveillance also successfully predicted laboratory isolation of associated viral organisms. Physicians completing the evaluation (N = 11) indicated that the site provided information useful in daily practice for both physician and parent education.
Conclusions: An Internet-based surveillance network linking a hospital with community physicians is beneficial to the hospital in predicting waves of severe cases requiring admission and reciprocally provides useful information to physicians in daily practice regarding the incidence and cause of seasonal disease in the community.