I would like to hear what approaches have been employed to weight cases when cases and groups have different exposures to a condition. Our Example We have 1,900 patients for whom we have recorded the admissions to hospital, emergency and ambulatory clinics Each record in the file is an admission (some patients can have multiple records) There are two time periods Exposure in this case refers to the individual being in the community (Time 1) or a Continuing Care Centre (Time 2) and potentially entering the Acute care health system.

Time Period 1 - Jan, 01, 1995 to admission to Continuing Care (any time after Jan 01, 1996 to Nov 31, 1997) Period 2 - From Continuing Care Admission date to Discharge Date from Continuing Care or Nov 31, 1997 if not discharged

Thus each exposure period could be different for each case and each case could have different exposure times in the two periods Thus if the question was do individuals enter acute care more often when in the community as opposed to Continuing Care one would have to adjust for the time exposed to both conditions i.e. if you were in the community for 2 months and in continuing care 1 month and if you had 2 admissions while in community and 1 while in continuing care then this would be non-significant My question is would you weight at the individual level or would you total the incidents in each period and total the number of days in each period and then adjust i.e. we know that the sum of ratios is not equal to the ration of sums (unless all denominators are the same)