Short corrections to recurring mistakes in the way scientific findings, risk figures and vaccine-safety information are described.
More accurate: timing is relevant, but sequence alone does not establish causation. Investigators also consider expected background rates, comparison groups, alternative explanations, biological plausibility and the wider evidence.
More accurate: Yellow Card reports record suspected adverse reactions. They help detect possible signals, but an individual report does not by itself establish that a medicine or vaccine caused the event.
More accurate: a signal is information that warrants investigation. Regulators assess the pattern and other available evidence before deciding whether a causal relationship is supported.
More accurate: a count needs context, including how many people were exposed, reporting patterns, time periods, duplication, background rates and whether the events were causally assessed.
More accurate: an individual study should be judged on its design, relevance, precision and risk of bias, then considered alongside other studies and systematic reviews.
More accurate: expertise and institutional responsibility provide context, but conclusions should still be traceable to methods, data, evidence reviews and stated uncertainty.
More accurate: statistical significance does not show the size, practical importance or freedom from bias of an effect. Read the effect estimate, absolute numbers, confidence interval and study limitations.
More accurate: a non-significant result may reflect no effect, limited precision or insufficient data. The estimate and confidence interval show which effects remain compatible with the data.
More accurate: correlation can arise from causation, reverse causation, confounding, coincidence or analytical choices. Further evidence is needed to distinguish these explanations.
More accurate: different evidence types serve different purposes. Case reports and spontaneous reports can raise questions or signals; stronger comparative designs are usually needed to estimate risk or support causal conclusions.
WHO — adverse-event causality assessment
This page explains general evidence concepts. It does not assess personal symptoms, provide a diagnosis or replace advice from a suitably qualified healthcare professional.