30 This sample seems to provide some support for this view, with the veterans reporting PTSD symptoms, Depressive symptoms, and Somatic symptoms with some frequency. There is a clear need here for more work on the classification of disorders. As regards PTSD, the mean score reported by these war veterans was within the range of other studies (58). The distribution of the CAPS scores suggests that a cut-off of 19 on the Frequency subscale and 21 on the Intensity subscale would be appropriate for Zimbabwe. This preliminary finding needs to be extended by further work, probably with a two-stage procedure using the CAPS as a screen, validated by subsequent independent psychiatric assessment. The screening and assessment procedure seems to work well, and there were strong relationships between all three instruments and their derived measures. As can be seen from Table 7 above, there was a strong correlation between the SRQ-20 and the CAPS, and good correlations within the instruments too. The SRQ-20 total score correlated moderately well with its two subscales, 0.48(p=0.001) and 0.42(p=0.01) for Anxiety and Depression respectively, whilst the two scales of the CAPS showed a very strong correlation (0.63; p=0.001). These results together suggest both good internal and external validity for the two instruments, and also provide some small support for the instruments' reliability. Additionally, it may be that the the symptom checklist, the Present State of Health, and its two subscales may provide a useful short-form screen for assessing the consequences of organised violence. This too deserves further work, again using a two-stage validation procedure.