A prospective cohort of 260 people was followed for 1 year. Half of the study participants were exposed (N = 130 exposed), and half were unexposed (N = 130 unexposed). Based on this information, please answer a - d.

a) In the cohort study described above, 15 of the exposed and 10 of the unexposed participants were lost to follow-up. Consequently, only 235 participants and their outcomes were observed based on their questionnaire responses. During 1-year of follow-up, there were 18 participants diagnosed with the disease among the exposed group and 5 diagnosed with the disease among the unexposed group. You decide to conduct the analysis only using participants with complete data. Calculate the risk ratio of the outcome comparing the exposed and unexposed groups among the 235 participants who completed the study.

b) Suppose we were able to identify the disease outcome status through medical records for those who were lost to follow-up. We found that among those lost to follow up, there were 3 with the outcome in the exposed group and 7 with the outcome in the unexposed group during the 1-year follow up. What would have been the true risk ratio of the disease, comparing those with exposure to those without exposure in the cohort of 260 participants, if there were no loss to follow up?

c) Compared to the true risk ratio (from 11b), was the observed risk ratio (from 11a) biased toward the null or away from the null? Explain your answer in one sentence.

d) Loss to follow-up is likely to introduce what type of bias?