Influenza and influenza-like illness in general practice: drawing lessons for surveillance from a pilot study in Paris, France

Abstract

Background: There are two types of inflenza surveillance techniques: qualitative laboratory-based surveillance and quantitative medical practice-based surveillance. The former is of great importance in isolating new strains of the virus, which helps in the decision-making process concerning the composition of the vaccine, and the latter provides estimates of morbidity, mortality or economic impact as a result of infection from the influenza virus. Rapid methods such as immunoflourescence (IF) or immunocapture assays (ICA) are now available for diagnosis of influenza infections. However, little is known about the use of these methods for influenza surveillance purposes.

Aims: To evaluate the feasibility of a rapid influenza diagnosis in ambulatory conditions, and to investigate the therapeutical outcomes of patients suffering from influenza-like illness (ILI) in relation to the virological diagnoses.

Method: During the 1994-1995 influenza season, 130 patients presenting with ILI symptoms (< 36 hours) to 33 general practitioners (GPs) were included in a prospective study. Two nasal swabs and one throat swab per patient were collected and sent to the laboratory within 12 hours. Information concerning therapeutical outcomes was recorded during examination. Specimens were analysed using the immunofluorescence (IF) method and antigen immunocapture assay (ICA).

Results: Sixteen influenza A (12%) and 19 influenza B (15%) infections were diagnosed. The overall rate of influenza positive specimens was 17% in the pre-epidemic period and 31% during the epidemic (P = 0.1). The rates of usable specimens for IF assay, nasal ICA and throat ICA were 46%, 100% and 99% respectively. The combination of these three collections ensured a highly sensitive influenza virological diagnosis. There were no differences in therapeutical outcomes between the influenza positive and negative cases. The GPs prescribed antibiotics in 60% of the cases for a mean duration of 7 days (+/- 1.2). The mean duration of sick leave was 3.4 days (+/- 1.6). Twelve patients (four influenza positive, eight influenza negative) had been vaccinated at the beginning of the winter. The practitioner's intuition concerning 'which patient should be tested for influenza virus' did not prove useful in improving the aptness of virological diagnoses in the field of influenza surveillance.

Conclusion: The only way to estimate the true impact of influenza is to carry out a systematic virological sampling based on a sensitive clinical definition and using sensitive laboratory methods.