Examqo › Examqo library › Family medicine › Evidence-based medicine, family and law · IMLE
Study design, bias and risk measures
The study design decides what can be concluded: a randomised controlled trial shows causation, a cohort study measures relative risk, a case-control study yields an odds ratio — and every design has its typical bias.
Why it matters in the exam
Every family medicine paper has 6–10 research questions: identify the design from a short description, choose the right design for a question (rare disease — case-control; rare exposure — cohort), calculate NNT, and recognise a bias (recall, selection, lead time) or a confounder.
Other regular questions: what an intention-to-treat analysis means, the difference between incidence and prevalence, and what a meta-analysis adds and what its limits are.
The skill being tested is critical reading: ask who was studied, how the groups were chosen, what was measured and what could explain the result apart from the treatment itself. A candidate who identifies the design in the first line of the question solves most of these questions without any calculation.
A standard calculation: an event rate of 10% in the control group and 6% in the treatment group. The absolute reduction is 4%, the relative reduction is 40%, and the NNT is 25 — meaning 25 patients must be treated to prevent one event. Many questions build their distractors on confusing these three numbers. Other questions describe a study of the link between an exposure and a disease and ask what it must include for the conclusion to be valid: a suitable comparison group, uniform measurement of the exposure, adjustment for confounders and long enough follow-up.
Key points
- Randomised trial — causation; cohort — relative risk; case-control — odds ratio.
- Rare disease — case-control; rare exposure — cohort.
- NNT = 1 divided by the absolute risk reduction.
- Lead time and length time imitate a benefit of screening; mortality is the true measure.
- Intention-to-treat analysis preserves randomisation.
Summary table
| Design | Suits | Typical measure | Typical bias |
|---|---|---|---|
| Randomised controlled trial | Comparing treatments, causation | Relative risk, absolute risk reduction, NNT | Drop-out; without intention-to-treat analysis randomisation is broken |
| Cohort study | Rare exposure, several outcomes, incidence | Relative risk | Loss to follow-up; confounding |
| Case-control study | Rare disease or one with a long latency | Odds ratio | Recall bias; selection bias of controls |
| Cross-sectional study | Prevalence at one point in time | Prevalence | Cannot tell what came first |
| Systematic review and meta-analysis | Summarising all the evidence | Pooled measure | Publication bias; heterogeneity between studies |
Free exam questions
מחקר השווה בין 8,276 חולי סוכרת עם דמנציה לעומת 8,276 חולי סוכרת ללא דמנציה התואמים להם מבחינת גיל ומין. במחקר זה נבדקו התרופות אותן נטלו המטופלים לסוכרת, ונבדק הקשר בין נטילת תרופות שונות לסוכרת לבין התפתחות דמנציה. מהו המדד המתאים לתיאור גודל האפקט בסוג המחקר המתואר? (the question as examined, in Hebrew)
- א. Odds Ratio
- ב. Relative Risk
- ג. Confidence interval
- ד. Kaplan-Meier analysis
במחקר קליני, מה הנוהל לגבי)SAE( Event Adverse Serious ? (the question as examined, in Hebrew)
- א. חומרתו של SAE היא תמיד Severe
- ב. בד"כ SAE .הוא אירוע תמותה
- ג. על החוקר לדווח עליו בתוך 24 שעות מרגע שהוא יודע שהתרחש, לוועדת האתיקה במשרד הבריאות
- ד. חומרתו של SAE נקבעת על ידי החוקר
מה ההבדל בין מחקר קליני מיוחד למחקר קליני שאינו מיוחד? (the question as examined, in Hebrew)
- א. מחקרי פאזה שניה הם תמיד מחקרים מיוחדים
- ב. מחקר קליני שאינו מיוחד יכול להכיל אוכלוסיות בוגרים כשירים לחתום על הסכמה מדעת אך לא אוכלוסיות מיוחדות כגון קטינים או חסרי ישע
- ג. המחקר הקליני המיוחד מתבצע באוכלוסיות מיוחדות כגון קטינים או חסרי ישע
- ד. המחקר המיוחד מאושר על ידי מנהל בית החולים ולא על ידי ועדת האתיקה של משרד הבריאות
In the full version for subscribers
- A simple algorithm
- Memory aid
- Exam traps
- 5 review cards
- Enough for IMLE
The full version is open to subscribers of: Family medicine, Medical licensing exam (IMLE), Internal medicine, Psychiatry, Public health.
Subscriber? Read the full topic Plans and prices 25 free questions
Sources
- Gordis Epidemiology (Celentano DD, Szklo M), 6th edition (2019) — Study designs, measures of association, bias and confounding
- Straus SE, Glasziou P, Richardson WS, Haynes RB. Evidence-Based Medicine: How to Practice and Teach EBM, 5th edition (2018) — Therapy and harm
- CONSORT and STROBE statements, 2010 / 2007 — Reporting of randomised trials and observational studies
Sources are given for further reading; the text is written afresh and not copied from them.
Reviewed by the Examqo AI Professor · 2026-11-06. The text may contain errors; when in doubt, check the source.