Thirty-four studies were included: four randomised controlled trials, six controlled trials, one cohort study, one case-control study, 13 pre-test/post-test without control studies, seven cross-sectional studies, one ecological study and one time-series study. Only 15 (those rated as 'strong' or 'moderate') were included in the narrative synthesis. Quality assessment of the studies showed that obe study was rated strong, 14 were moderate and 19 were weak.
Inspection/risk assessment: Five studies evaluated the effectiveness of inspection of food service premises. Two studies found no difference in inspection scores between restaurants scheduled to receive different numbers of inspections per year (6, 9 0r 12 in one study and 2, 4 or 6 in the other). One study found that premises inspected 4 times a year produced better inspection scores than those inspected only on public complaint. One study found that time intervals since the last inspection which were greater than one year were significantly associated with poorer inspection scores, however, time intervals fewer than 3, 4-6 or 7-12 months did not differ significantly from each other. Two studies examined risk assessment, a component of a food inspection strategy which categorises each food premises according to its potential for food-borne disease. One of these studies found that restaurants designated as high risk (inspected every 1-2 months) changed to a lower risk rating whereas medium and low risk premises (inspected every four or 12 months respectively) did not. The other study found that about 10% of low risk food premises had changed to medium risk upon reassessment one year later.
Food handler training: Eight studies, all of moderate quality, evaluated the effect of food handler training programmes. Of these, six studies showed a positive change in post-test measures and the remaining two studies did not show any improvements. One study showed improvements in inspection scores in a group that participated in a workshop but not in a group that received only a food safety manual.
Community based education: Two studies, both of moderate quality, evaluated the effect of community based education. One study evaluated the effectiveness of a supermarket food safety educational tour and noted improvements in food safety knowledge and self-reported practices after the intervention. The other study evaluated a food safety curriculum for elementary students and demonstrated and increase in knowledge scores after the curriculum was introduced.