Hospital laboratory reports of parainfluenza virus (PIV) infections from England and Wales between 1975 and 1997 were analysed with regard to PIV type and seasonality, and in addition, those between 1985 1997 with regard to age, sex and clinical features. Laboratory-based surveillance data highlight striking differences in the seasonality of different PIV types. PIV-3 reports demonstrated a clear annual epidemic cycle, with a peak usually occurring in late spring or summer, whereas peaks of PIV-1 and PIV-2 occurred at one or two year intervals, in the late autumn or early winter. PIV-4 also occurred most frequently in the late autumn or early winter, but a clear epidemic cycle could not be identified. Laboratory surveillance data also provide insight into the age and disease distribution of PIV infection in children and indicate severity of PIV infection in immunosuppressed adults. Of 8221 PIV reports received between 1985-1997, PIV-3 accounted for 70.8%, PIV-1 for 17.2%, PIV-2 for 7.5%, and PIV-4 for 1.1%; 64.1% of reports came from infants under one year, 24.4% from children aged 1-4 years and 7.2% from individuals aged 5 years or older, with an excess of males in all age groups. Bronchiolitis, croup and pneumonia occurred in association with all PIV types. In children under 1 year, PIV-2 infections were more likely to be associated with bronchiolitis than infections with other PIV types. In children under 15 years, croup was more frequently associated with PIV-1 and PIV-2 than with PIV-3 or PIV-4. In 392 (7.2%) of the reported PIV infections between 1989 and 1997 an underlying condition was implicated, which included immunosuppression or chronic cardiac or pulmonary disease. Considerable morbidity is associated with PIV infections in infants and young children and would make the widescale use of a vaccine a valuable public health intervention. Surveillance information is essential to guide the development and use of preventive measures as well as to monitor their effectiveness.