Crampin AC et al. - Reporting of putative source contacts showed little evidence of recall bias and gave estimates of the relative risk of TB associated with identifiable contact. The lower likelihood of HIV-positive cases being named as contacts may reflect reduced infectiousness.