Targeting symptoms for self-care health education: a multivariate analysis of physician contacts.

The purpose of this study was to identify common acute symptoms that were significantly related to the decision to seek medical care. Data were from the Seattle Virus Watch, a 4-year study of 149 young families designed to monitor viral infection, regardless of medical care contact. Illness diaries were maintained by the families and checked for accuracy every 2 weeks during home visits by study personnel. Age sex, family unit, medical contact, duration and presence or absence of 80 symptoms in 17 categories were recorded for each illness. In 11,583 person-months of observation, 3,929 illness records were compiled. Using physician contact as the dependent variable and the illness information as independent variables, a series of discriminant analyses were performed on a 50% random sample of illnesses. Success of each analysis was tested on the remainder of the illnesses. Fever and ear symptoms were significant predictors in all age groups and cough was significant in four of five age groups. These symptoms are suggested as logical choices for health education programs in self-care of acute illnesses.