Instead of investigating the chance of going 18 months without a root canal your practice is only interested in the chance of having a root canal given various exposures, you do a case-control study. Suppose that for every root canal case from the previous 18 months, 3 controls were selected and both cases (43) and controls (129) were asked the about various exposures.

Q.1: If the controls are a random selection of anyone who did not have a filling in a molar vs. those who did, how might that impact the Odds Ratios?

Q.2: If the general population was used for controls (vs. those with a filling in a molar) should researchers adjust for a variable indicating if a patient had a filling in a molar in the previous 18 months? Why or why not? (This is asking whether filling in a molar should be considered a confounder. Remember, for a variable to be a confounder, it needs to be: 1)Associated with both the exposure and outcome of interest 2) Cannot be in the causal pathway

Q.3: If the controls were patients who also had a molar filling is the OR a good estimate of the RR? Why or why not? (Think about rarity - if the condition is rare, then OR can be a good estimate of RR; if not, then no).

Q.4: Discuss both 1 strength and 1 weakness of this study design (assume they used anyone with a root canal in this practice the last 18 month as cases and any patient from this practice who did not as controls and controls are the same ages as cases)

Q.5: Name one thing you would do to improve the study (as described in q4 above) and explain how it would improve the study.