The Evidentiary Value of Challenge Trials for Three Pre-harvest Food Safety Topics: A Systematic Assessment

Reducing zoonotic pathogens in food animals prior to harvest will reduce the pathogen burden that enters the food chain and the environment. Consequently, the burden of enteric illness in humans may be reduced. Evaluating interventions to reduce a pathogen in animals often begins with challenge trials, in which animals are deliberately exposed to the pathogen under controlled conditions. Challenge trials are subsequently followed by field trials, also known as randomized controlled trials, in which the animals are naturally exposed to the pathogen. Challenge trials can most effectively inform field trials only if they precede field trials, are robust, internally valid and transparently reported. Using systematic review and meta-analysis methodology, we examined the pre-harvest food safety literature for three interventionpathogenspecies combinations: probiotics/competitive exclusion products in ruminants to reduce Escherichia coli O157 shedding, vaccines in ruminants to reduce E.coli O157 shedding and vaccines in swine to reduce Salmonella shedding. We examined two outcomes, prevalence of faecal shedding at the end of the trial and prevalence of faecal shedding throughout the trial period, to compare challenge trials and field trials. We found that challenge trials occurred concurrently with field trials, challenge trials suffered from reporting deficiencies of methodological features, challenge trials tended to report a more favourable outcome than field trials, and there was some evidence of publication bias among all three intervention-pathogen-species combinations. Challenge trials would better serve to inform field trials if they precede field trials, are methodologically sound, include transparent reporting and are published regardless of their results. In addition, due to our findings of greater efficacy reported among challenge trials compared with field trials, risk models predicting the public health benefits of pre-harvest interventions to reduce zoonotic pathogens in livestock might be best served by field trial results alone.