Glossary

Likelihood ratios

Likelihood ratios (LRs) help you understand how much a test result changes the chance of a disease. In simple terms, a positive likelihood ratio (LR+) tells you how much more likely it is for a patient with a positive test to have the disease compared to someone without it. Conversely, a negative likelihood ratio (LR–) shows how much less likely a patient with a negative test is to have the disease. If the LR+ is high (for example, above 10) or the LR– is very low (below 0.1), the test result significantly alters your initial estimate (pre-test probability) of disease. When LRs are close to 1, the test adds little value because it barely changes your estimate. This way of thinking makes it easier to decide whether a test result is truly meaningful in guiding treatment decisions. 

Meta-analysis

You take several separate studies on the same topic and combine their results into one big study. By pooling all this data, you get a clearer overall picture of what’s really happening. This combined approach, called a meta-analysis, often provides stronger, more reliable evidence than you’d get from looking at each study on its own. 

Number Needed to Treat (NNT)

• First you find the absolute risk reduction (ARR), which is the difference in the event rate between the control group (often given a placebo) and the treatment group.
• You then calculate the NNT by taking the inverse of the ARR. Mathematically, this is: NNT= 1/ARR
• Worked example for (NNT):
  • In the control group, 10% of patients experience a heart attack.
  • In the treatment group, only 5% of patients experience a heart attack.
  • The ARR is 10% – 5% = 5%, or 0.05 in decimal form.
  • The NNT is then: NNT=1/0.05 = 20 
  • Interpretation: You need to treat 20 patients with this drug to prevent one heart attack.

Number Needed to Harm (NNH)

• Similarly, if a treatment increases the risk of a negative event (an adverse effect), you calculate the absolute risk increase (ARI).
• The NNH is the inverse of the ARI: NNH= 1/ARI

Randomised Controlled Trial 

You randomly split participants into different groups – one receives the new treatment, and another might receive a “dummy” pill (also known as a placebo) or standard care. Because you assign them randomly, any differences in outcomes are more likely to come from the treatment itself rather than other factors. This method is called a Randomised Controlled Trial (RCT).