Were all the patients accounted for?
Ensure that follow-up is complete and that reasons for patient drop-out
are given. The greater the number of patients "lost to follow-up" the greater
the potential for bias because patients who are lost often have a
different prognosis than those who are retained. If you assume that
all patients lost to follow up had the same outcome, good or bad, would
the study results be significantly different?
Application to Pollock et al's
thyroxine paper
In this study, 3 patients withdrew from the study, 2 from the thyroxine-placebo
group.