I thank Dr Vetter for his cordial letter. The question at issue is not whether or not it is possible to create a statistical model that can predict psychiatric breakdown in military recruits - that is certainly possible, as the experiences of the Second World War psychiatrists showed. The question is with what accuracy can one make such a prediction and what are the consequences for those both correctly identified and, even more importantly, those who have been incorrectly identified (the false positives). Dr Vetter does not provide sufficient information for us to make that judgement. What is needed is the sensitivity, specificity and most importantly the positive predictive value of whatever collection of variables he and his colleagues are using to determine the risk of future illness. It is this statistic that enables us to assess the utility of the proposed model.

Furthermore, we do not know what were the consequences of being labelled as at risk of psychiatric breakdown. Were these people denied military service? Switzerland is one of the increasingly few countries that still has compulsory military service. Serving in the Armed Forces is a fundamental part of the life of every Swiss citizen and enables a person to form social networks that operate for many years. Are people disadvantaged from being denied that opportunity? Given that the Swiss are also famed for their neutrality, the fall in psychiatric morbidity as a result of screening is not likely to be because those denied military service are not exposed to the risks of the battlefield. Instead it may be that their subsequent breakdown merely happens in another sector of Swiss life. Without data from a randomised controlled trial, it is impossible to decide whether any public health benefit has resulted from introducing psychiatric screening. Given the weakness of the individual predictor variables, the timing of screening (at the end of adolescence) and the fact that to date no programme of psychiatric screening for events that have yet to happen (i.e. future breakdown) has been shown to be effective in a randomised controlled trial, I think that I am entitled to stay with my conclusions that psychiatric screening to detect vulnerability to future breakdown remains unproven and continues to have the potential to do more harm than good. Until such evidence is forthcoming, it may be more useful to devote resources to increasing resilience through support and training, and providing better and more acceptable services to help those who do succumb to the rigours of military life.