Understanding Society User Support

Could you point me to anywhere in the documentation that describes how the death of a BHPS/US respondent is handled and recorded in the data. I have found only one very brief mention in each of the technical reports and user guides. Lui and Weale (NIESR DP no 384) were able to estimate mortality functions for BHPS. I'd like to know if this is possible with US. Not only would this be an excellent way of assessing the representativeness of the survey as Lui and Weale did for BHPS, the much larger N's of US could turn it into an excellent resource for looking at the correlates of mortality.

Thanks
JM