1. You are conducting a case-control study to measure whether drinking alcohol during pregnancy is associated with prete
Posted: Tue Jul 05, 2022 8:48 am
1. You are conducting a case-control study to measure whetherdrinking alcohol during pregnancy is associated with pretermdelivery. Cases are women with preterm deliveries and controls arewomen with term deliveries; they are all recruited from the samehospital. You are concerned that women may under report alcoholintake during pregnancy as they are told not to drink duringpregnancy. You believe that cases (women with premature infants)will be accurate in reporting alcohol use due to their desire tohelp their infants but that controls will be subject to the socialdesirability bias and will under report drinking.
Is this differential or non-differential misclassification (i.e.information bias) and why?
How will this bias affect the estimated OR from this study?
1b. You are conducting an RCT to examine the effect of anintervention to convince young adults to get vaccinated forCOVID-19. You recruit unvaccinated 20-35 years olds and assign halfto get the intervention (E+) and the other half get only a pamphlet(E-). You follow the participants for six months to see how manyget vaccinated (D+) or are unvaccinated (D-). Over the course ofstudy follow-up, more participants in the untreated group drop outof the study – they withdraw consent so you cannot find out if theyare vaccinated or not. You believe that most of the people who werelost to follow-up are unvaccinated at the end of the sixmonths.
What form of bias might this create and why?
Imagine the true RR for the intervention is 2.0 (i.e. the truerisk of vaccination in people who get the intervention is 2 timesthat of people who didn’t get the intervention). Based on theinformation provided, describe how this bias could move theobserved RR in the study.
Is this differential or non-differential misclassification (i.e.information bias) and why?
How will this bias affect the estimated OR from this study?
1b. You are conducting an RCT to examine the effect of anintervention to convince young adults to get vaccinated forCOVID-19. You recruit unvaccinated 20-35 years olds and assign halfto get the intervention (E+) and the other half get only a pamphlet(E-). You follow the participants for six months to see how manyget vaccinated (D+) or are unvaccinated (D-). Over the course ofstudy follow-up, more participants in the untreated group drop outof the study – they withdraw consent so you cannot find out if theyare vaccinated or not. You believe that most of the people who werelost to follow-up are unvaccinated at the end of the sixmonths.
What form of bias might this create and why?
Imagine the true RR for the intervention is 2.0 (i.e. the truerisk of vaccination in people who get the intervention is 2 timesthat of people who didn’t get the intervention). Based on theinformation provided, describe how this bias could move theobserved RR in the study.