For each of the following situations, put an "A" if it is likely to result in the overestimation of the cumulative incidence estimate, a "B" if is likely to result in the underestimation of the cumulative incidence estimate, a "C" if the direction of the bias is variable, or a "D" if no bias is likely to result. Study subjects withdrew because they were sicker over time, leading to a higher risk for censored than for noncensored observations