If I'm not mistaken, there seems to be a discrepancy between total sleep time variables in SHHS data (Visit1). The variable slp_time appears to be accurately calculated, however the variable slptime appears incorrect. As an example time in bed for subject 205804 is 397.5 minutes, however, total sleep time as per slptime is 557 mins, which is longer than time in bed. On the other hand, slp_time=357 appears accurate. Cross-checking the times based on the hypnograms, slp_time appears to be the accurate variable.
EDIT: It appears there is another variable SlpPrdP, which is similar to slp_time and accurate.
Ankit Parekh, PhD.
Hey Ankit - thanks for using the site and for bringing this to our attention. Your finding is correct about slptime. The documentation in the original SHHS datasets about this variable was not clear, so I have identified it for removal from future versions of the dataset (https://github.com/nsrr/shhs-data-dictionary/issues/1107). slpprdp is typically our preferred variable for total sleep time.