Everybody warns that Step 3 Day 1 is heavy on biostatistics, multi-paragraph pharmaceutical abstracts, and prognosis/ethics questions. During residency I have not touched relative risk reduction, number needed to treat, or confidence intervals in over a year.
How are residents reviewing biostats and tackling the three-question drug ad sets without running out of time?
For the three-question pharmaceutical abstract or drug ad sets on Day 1, the biggest mistake is reading the entire abstract top-to-bottom before looking at the questions. Skip the abstract completely at first: read Question 1, identify the exact table, figure, or endpoint it asks about, and scan only that section.
Also, leave the drug ad set for the very end of the 30-minute block. Mark all three items, finish the standalone clinical vignettes with 6 to 7 minutes to spare, and then come back to the abstract with zero time anxiety.
Content-wise, spend two evenings reviewing four core areas: (1) interpreting 95% confidence intervals (crossing 1.0 for RR/OR/HR means not statistically significant; crossing 0 for absolute risk difference or mean difference means not significant), (2) calculating NNT = 1 / ARR and NNH = 1 / ARI, (3) recognizing study design biases (lead-time bias in screening survival, confounding by indication, Hawthorne effect, intention-to-treat preserving randomization), and (4) sensitivity/specificity shifts when moving a diagnostic cutoff.
Also brush up on positive and negative likelihood ratios: LR+ = sensitivity / (1 - specificity) and LR- = (1 - sensitivity) / specificity. An LR+ above 10 strongly rules in disease.
For sensitivity and specificity, drawing the 2x2 table with Disease+ on the left and Test+ on top takes five seconds and prevents silly arithmetic mistakes when you are post-call.