Game 5 · Study designs, risk and bias
How big is the risk, and how many patients must you treat?
Relative risk, odds ratio, attributable risk, ARR, RRR and NNT.
Read with: Risk and survival90 DNB questions
Play
Read the table, work out the measure
Each round gives a made-up trial or cohort table. Type the measure asked for and see the working from the four cells. Other rounds ask you to find where the odds ratio stops matching the relative risk, and to see why a case-control study gives only an odds ratio. You have three lives.
Loading the first round…
Formulas
One 2 × 2 table, six measures
Put the exposure (or treatment) in the rows and the outcome in the columns. The exposed row holds a (with the outcome) and b (without). The unexposed row holds c and d. Risk in each row is events ÷ everyone in that row.
Relative risk (RR) = [a ÷ (a + b)] ÷ [c ÷ (c + d)]
Odds ratio (OR) = (a × d) ÷ (b × c)
Attributable risk (AR) = risk in exposed − risk in unexposed
Absolute risk reduction (ARR) = control risk − treated risk
Relative risk reduction (RRR) = 1 − RR = ARR ÷ control risk
Number needed to treat (NNT) = 1 ÷ ARR
- a, b
- exposed or treated people with and without the outcome
- c, d
- unexposed or control people with and without the outcome
- AR%
- attributable risk percent = AR ÷ risk in exposed = (RR − 1) ÷ RR
- NNH
- number needed to harm: the same calculation when the treatment adds risk
AR and ARR are the same subtraction. AR is the name used when an exposure causes the outcome, ARR when a treatment prevents it. Round the NNT up to the next whole patient, and state the comparator and the time period. Laupacis, Sackett and Roberts introduced the NNT in 1988.
Relative and absolute
Why the same 40% reduction gives a different NNT
Relative measures stay similar when the baseline risk changes. Absolute measures do not. A drug that cuts risk by 40% saves 10 patients per 100 when the baseline risk is 25% (NNT 10), but only 2 per 100 when it is 5% (NNT 50). Cook and Sackett (1995) gave a real example: with a 40% relative reduction in stroke from treating hypertension, the NNT was 13 in moderate hypertension and 167 in mild hypertension.
| Feature | Relative measures | Absolute measures |
|---|---|---|
| Examples | RR, RRR, OR, HR | ARR (risk difference), NNT |
| Change with baseline risk | Usually similar across baseline risks | Smaller ARR and larger NNT when baseline risk is low |
| Tells a patient | By what proportion the risk falls | How many events are prevented per 100 treated |
| In a trial report | Give both | Give both |
Odds and risk
When the odds ratio matches the relative risk
Risk divides the events by everyone in the group. Odds divide the events by the people who did not have the event. When the outcome is rare, almost everyone is a non-event, so odds and risk are nearly equal and the OR is close to the RR. With risks of 5 and 3 per 1,000, RR is 0.60 and OR 0.599. With 25% and 15%, RR is 0.60 and OR 0.53, further from 1.
Zhang and Yu (1998) pointed out that once the outcome occurs in more than 10% of the study population the OR no longer approximates the RR. For a single table the two are linked exactly:
RR = OR ÷ [(1 − P₀) + (P₀ × OR)]
- P₀
- risk in the unexposed or control group
Risk: events ÷ everyone. Odds: events ÷ those without the event. NNT: 100 ÷ ARR in percentage points.
Case-control studies
Why a case-control study gives only an odds ratio
In a case-control study the investigator decides how many controls to take. The share of cases in the table is then set by the investigator, so a ÷ (a + b) is not a risk and the relative risk cannot be calculated. The odds ratio still works, because the odds of exposure among cases ÷ the odds of exposure among controls equals the odds ratio of disease in exposed versus unexposed people. Cornfield (1951) showed that when the disease is rare this odds ratio approximates the relative risk.
Confidence intervals, matched odds ratios and the Mantel-Haenszel method are worked through in the risk and survival chapter.
Writing the answer
How to write this in a 10-mark DNB answer
- Draw the 2 × 2 table first, exposure in rows and outcome in columns, labelled a, b, c, d.
- Define risk and odds in words, then give each formula from the table.
- Work one example with round numbers: RR, OR, ARR, RRR and NNT from the same table.
- State which design gives which measure: OR from a case-control study, RR from a cohort study or trial, and why.
- Add that the OR approximates the RR only for a rare outcome, with a common-outcome example where they differ.
- For NNT, add that it depends on baseline risk, is rounded up, and needs the comparator and time period. For a question on attributable risk, add AR% and the population attributable fraction.
The one thing to remember
Report a relative measure and an absolute measure together. The relative measure says how strong the effect is. The absolute measure says how many patients it helps.
Past papers
DNB questions this game helps you answer
Questions from NBE theory papers for DNB, DrNB and Diploma courses, most often asked first. The tag after each question is the specialty and the year it was set.
- ×4Differences between odds ratio and relative risk.
- ×4Odds ratio.
- ×3Nested case control study.
- ×2Case control study.
- ×2What is a case control study?
- ×2Advantages and disadvantages of case-control studies.
- ×2Odds Ratio.
- ×2Case control study vs Cohort study.
- ×2What is nested case controlled study? Discuss in brief the methodology and advantage of the study.
- ×1Discuss biases in case control studies. [3]
- ×1Odds ratio and Risk ratio. [3]
- ×1What is the difference between case control study and randomized control study? [4]
- ×1Briefly describe odds ratio and relative risk.
- ×1Case control and cohort study.
- ×1Define case control study and cohort study.
- ×1Describe case control study with a suitable example.
- ×1Difference between the study design of case series and case control study.
- ×1Differentiate between case control and cohort study.
- ×1Draw the schematic diagrams of the design of case-control and cohort studies. Explain the concept of matching with examples. How will you measure the strength of the association between the risk factor and the outcome in a case-control study?
- ×1Odds ratio and relative risk.
- ×1Relative risk and Odds ratio.
- ×1What do you understand by odds ratio? When is it used?
- ×1Number needed to treat.
- ×1Case Control study (Give 2 examples).
- ×1Compare and contrast between Case control and Cohort study.
- ×1a) Classify various research study design. [3] b) Describe the terms “exposure” and “outcome variable” in a study. [3] c) What is a cross sectional study and how it is different from case control study?
- ×1Cohort and case control studies.
- ×1Compare case-control and cohort studies.
- ×1Discuss about case control study. Define odds ratio & the confounders if any in these studies?
- ×1Discuss the strength and weakness of a case control study.
- ×1Relative risk.
- ×1What are relative risk and attributable risk? Give examples.
- ×1Number Needed to Treat.
- ×1What is the therapeutic index? Discuss the clinical significance of number - needed to treat (NNT).
- ×1Difference between case control and cohort study.
- ×1Differentiate odds ratio and relative risk.
- ×1How is case control study different from a self-controlled study?
- ×1Types of controls in a Case-Control study.
- ×1What are methods to minimise weaknesses of a case control study?
- ×1What are various types of research studies? Compare RCT with Case-control study.
- ×1Population attributable risk.
- ×1Case control studies: a) What are the guidelines for writing a case control study? [5] b) What are the best diseases/conditions for which a case control study is used? [3] c) What is the level of evidence for a case control study?
- ×1Case control versus Cohort study.
- ×1Case-control Study design and its utility
- ×1Compare case control study and cohort study.
- ×1Define bias in epidemiology. Describe four important biases observed in a case control study?
- ×1Describe in detail how you will design a case control study with an example.
- ×1Design of Case-control studies.
- ×1Enumerate differences between case control and cohort studies.
- ×1Write about sample size calculation for case control studies.
- ×1c) What is odds ratio, relative risk and hazard ratio?
- ×1Population Attributable Risk.
- ×1Case control and Cohort studies.
- ×1Define sensitivity, specificity, relative risk and odds ratio.
- ×1Differences between case control studies and cohort studies using examples.
- ×1Odds Ratio. 5 + 5
- ×1Strengths and weaknesses of case-control studies.
- ×1a) Write the difference between correlation and regression analysis. b) List the types of observational study design. c) How is sample size calculated in case-control study?
- ×1i. Case-control study. ii. Cohort study. iii. Interventional study. b) List the advantages and disadvantages of each of these study designs
- ×1Bias in Case Control studies, explain with examples.
- ×1Case Control study design.
- ×1Case control studies
- ×1Differentiate between odds ratio and relative risk.
- ×1How do you calculate the sample size for a case control study?
- ×1Levels of evidences of various study (like case report, cases series, case control study, etc.).
- ×1Methods to control bias in Case Control studies.
- ×1Odds Ratio and Hazards Ratio.
- ×1Define population-attributable risk. How is it different from relative risk? What type of study design is required to measure it?
- ×1Number needed to treat (NNT).
- ×1Numbers to treat (NNT).
- ×1What is nested case control study? Describe its steps with examples.
- ×1Case controlled studies.
- ×1Odds ratio
- ×1a) Confidential interval. b) Odds Ratio.
- ×1a) Impact factor of a medical journal. b) Case controlled studies. c) Five parameters to decide efficacy of any investigation. d) Citation Index.
- ×1Number needed to treat and relative risk
- ×1Cohort control studies Vs Case control studies
- ×1Define a case control study, indicating advantage and disadvantages.
- ×1Submit a plan for a case control study with an illustrative example.
- ×1What do the terms - odds ratio, hazard ratio, relative risk, absolute risk and confidence interval mean and where are these used?
- ×1a) Case control study. b) Its advantages and disadvantages.
- ×1Number needed to treat (NNT) 20/1000 in a trial of new lipid lowering agent.
- ×1Disadvantages of conducting a Case control study.
- ×1Explain in brief case control study.
- ×1What is relative risk (RR) and discuss its implications?
- ×1Describe steps involved in conduct of case-control study. What are the advantages of this design?
- ×1Explain the terms case-control and cohort studies. Indicate the advantages of each of these study designs.
- ×1Describe a case control study. What are the advantages and disadvantages of case control studies?
- ×1What is the difference between nominal and ordinal data? Describe student t test and its use. What is a case control study? What does P value signify?
- ×1Attributable risk
Read next