A controlled study of health education in accident prevention: the Rockland County Child Injury Project. 1966.

The Rockland County Child Injury Prevention Project was designed to test the effects of public health education for parents on the incidence of accidental injuries to children under 7 years of age in a controlled situation. The incidence of accidental injuries, defined as any actual or presumed trauma following an accident for which medical or dental attention was obtained, was determined in a study population before, during, and after exposure to the educational program and in a comparable control group during a corresponding period of time. The study group was exposed to an intensive education program involving neighborhood discussion groups conducted by lay and professional leaders, meetings with organized groups, and a monthly newsletter. The study population was organized into small neighborhood units of about 24 homes within the suburban housing developments. To measure the effect of the education program, accident rates were computed for each three month reporting period, projected on an annual basis. No consistent differences were discernible in the trends in the accident rates between the study and control groups during the three years of the project. On the contrary, the curve for the study group crossed that of the control group on no less than five occasions. There was a sharp decline in accidents during the second half of the education phase. Any conclusions regarding a positive effect of the education program proved untenable in the light of the sharp and disproportionate increase in the accident rate in the study group during the six month phase after education had stopped. After what appeared to be a contrary trend earlier, developments during this period demonstrated vividly the need to continue reporting for a sufficient time in a controlled situation to be sure that fortuitous changes over short periods do not lead to unwarrented conclusion. The flow of reports from hospitals was more consistent than from physicians. This suggested that, in selected control studies, adequate rates for comparative purposes might have been obtained entirely from hospital records.