Objective - To develop an epidemiological database of morbidity in the
community as presented to general practitioners and to support epidem
iological research in general practice. Design - The project is a sent
inel network of 23 general practices in Northern Ireland that report t
he incidence of a list of selected diseases to a central unit for anal
ysis. Results - Results are presented for depression, diabetes, and my
ocardial infarction. The age - sex distributions of diabetes and myoca
rdial infarction are comparable with other published data but the inci
dence of depression is under estimated. Conclusions - The importance o
f complete accuracy of data reported within sentinel networks depends
on the purposes for which information is to be used. For some diagnose
s, such as diabetes and acute myocardial infarction, the accuracy of t
he reported incidence is high while for other diseases, such as depres
sion, where diagnostic behaviour varies more between doctors, the figu
res are much less reliable.