USMLE purpose: Build a repeatable strategy for drug-ad and abstract questions: scan questions first, do math first, eliminate unsupported answer choices.

How to use this page

  1. Scan the questions first before reading the ad/abstract.
  2. Do table, graph, and formula questions first because they are most provable.
  3. Use the strategy table to eliminate answers that sound plausible but are not supported by the data.
  4. Open the transcript only if you want Randy’s original walkthrough.

One-liner

For drug ads and abstracts, do not read everything first. Scan the questions, attack table/graph/math items first, and only choose conclusions directly supported by the data.

Randy’s drug-ad strategy

Step Action Why it works
1 Scan the questions first Prevents wasting time reading irrelevant details
2 Do math/table/graph questions first These are usually provable and fast
3 Eliminate bad choices Drug-ad answers often contain unsupported claims
4 Select only what the data prove Sounding plausible is not enough

Core formulas / tools

Tool Use Trigger
NNT 1 / ARR “How many patients must be treated…”
ARR Control event rate − treatment event rate Comparing two interventions
95% CI crossing 1 Usually not significant for ratios OR/RR/hazard ratio confidence intervals
p-value threshold p ≤ 0.05 Statistical significance
Approximate z comparison (A − B) / √(A + B) Quick estimate for whether two values differ meaningfully

Exam pattern recognition

Common traps