- From the earlier slides we know that the exposure information in the hazard period will be compared to the exposure information in the control data. There are 2 types of control data we can use in a case crossover analysis. The first type of data and also the most used is the exposure information from a “control period” which has the same length as the hazard period but at different time point. For example, if the length of hazard period is 5 hours, the control period is also a 5-hour period but it can be from the day before or two days before.
- The second type of control data is the exposure information from the past, it can be from the past month or past year. The analysis based on this type of control data will be more complicated.