SURVEILLANCE OF INFLUENZA IN WALES - INTERPRETING SENTINEL GENERAL-PRACTICE RATES USING CONTEMPORANEOUS LABORATORY DATA - OPPORTUNITIES ANDLIMITATIONS

Objectives-To evaluate the opportunities and limitations of using labo ratory data to enhance sentinel general practice surveillance of influ enza. Design-Descriptive study of active sentinel surveillance of clin ically diagnosed influenza in general practice and passive total popul ation surveillance of laboratory reports of influenza A, influenza B, Mycoplasma pneumoniae, and respiratory syncitial virus infections. Set ting-Wales. Subjects-Total sentinel practices population (currently 22 8 130); population of Wales (2 913 000, 1994 mid-year estimate). Main outcome measures-Simplicity, flexibility, acceptability, sensitivity, predictive value positive, representativeness, and timeliness of a sur veillance system. Rate of influenza and other respiratory infections. Results-Sentinel general practice surveillance of influenza in Wales i s simple, flexible, acceptable, timely, representative, and relatively sensitive. Current laboratory surveillance is complex and less timely than sentinel practice surveillance but is complete and has a relativ ely high positive predictive value. For the period January 1993 to Sep tember 1996, peaks in rates of influenza reported by sentinel practice s during winters 1993/94 and 1995/96 were temporally associated with i ncreased rates of laboratory confirmed influenza A and respiratory syn citial virus, whereas the peak in 1994/95 was associated with increase d rates of laboratory confirmed influenza B, M neumoniae, and respirat ory syncitial virus. Conclusions-Timely laboratory data can add value to influenza data already obtained from sentinel general practice surv eillance. However continuous audit is essential to resolve the possibl e limitations of either surveillance system.