Background: During the 20-month period September 1993 to April 1995, a health education injury prevention programme focusing on home injuries among the young (< or = 18 years old) and elderly (> or = 65 years old) on the Greek island of Naxos was undertaken, its effectiveness was evaluated by comparing the subsequent injury experience in sentinel population groups in Naxos as well as in Spetses, another island of similar sociodemographic profile, where no such intervention programme had been formally implemented.
Methods: On the island-of Naxos an injury prevention campaign was initially undertaken involving virtually all opinion leaders and implemented through lectures, workshops and publicity in the local media. The main intervention focused on 172 households on the island of Naxos and was done by trained local collaborators who visited each household weekly to provide injury prevention advice and assess home safety. Similar visits were done by untrained collaborators in 177 households on the island of Spetses in order to assure collaboration of household members in the comparative evaluation stage of the programme. The process evaluation was based on ascertained changes of safety features and attitudes in the participating households, whereas the outcome evaluation was based on the incidence of injuries among members of the participating households in the two islands over a period of 8.5 months (255 days).
Results: On the intervention island of Naxos there were statistically significant improvements with respect to 11 of the 28 examined variables, whereas on the island of Spetses, such improvement was only noted for one variable. The age-adjusted incidence rate ratio of injuries overall among the target groups, contrasting the intervention and the control households was 0.85 with 90% confidence interval (CI): 0.69-1.05. With respect to home accidents the corresponding ratio was 0.79 with 90% CI: 0.60-1.04.
Conclusions: An intensive and focused injury prevention intervention had only modest success when injuries themselves were the outcome variable.