Key takeaways
- A statistically significant result may still be clinically unimportant.
- Paediatric applicability depends on age, setting, baseline risk, and outcome definition.
- Trial limitations belong in the synthesis when they could change interpretation.
Clarify the question and the population
Before reading the result, I identify the trial's population, intervention, comparator, and outcome. In paediatrics, age bands, developmental stage, disease severity, and care setting can materially change whether the findings transfer to the intended reader.
Separate effect size from statistical significance
I record the absolute effect, confidence interval, and number needed to treat when available. A relative reduction can sound impressive while producing little absolute benefit in a low-risk population. Conversely, a modest-looking effect may matter when the outcome is serious or the intervention is low burden.
Check how the outcome was measured
Surrogate outcomes can be useful, but they are not interchangeable with outcomes families or clinicians value directly. I check who measured the outcome, whether assessors were blinded, how missing data were handled, and whether the analysis followed the prespecified intention-to-treat population.