What Counts as Healthcare Evidence?

#Start with the question

Healthcare evidence is information used to understand health, illness, tests and treatments. It can come from laboratory experiments, studies involving people, health records and research into people's experiences. No single type of evidence answers every question, so the starting point is what someone wants to know.

Laboratory research can explore how a biological process works. Interviews can reveal what living with an illness feels like. Studies of health records can identify patterns across large groups. These findings can all be useful, but they support different kinds of conclusions.

#Match the design to the task

Randomised trials are especially useful for comparing the effects of interventions because random allocation helps make groups comparable. Observational studies follow what happens without assigning an intervention. They can help investigate uncommon harms, long-term patterns and questions that would be difficult or unethical to study through randomisation.

Studies of diagnostic tests ask how accurately a test identifies a condition in a particular setting. Qualitative research examines experiences, meanings and barriers to care. Systematic reviews bring together relevant studies using explicit methods. Their usefulness depends on both the review methods and the evidence available.

#Judge quality, not just labels

A study's label is not a guarantee of reliability. Important questions include how participants were selected, what was measured, whether comparisons were fair and how missing information was handled. Funding and competing interests also deserve attention, although they do not automatically make a finding wrong.

Confidence usually grows when well-conducted studies produce compatible findings across relevant settings. A single promising result is rarely the whole story. Evidence also needs interpretation alongside people's priorities and circumstances; research can estimate likely benefits and harms without deciding which trade-offs matter most to everyone.

#Common misunderstandings

Healthcare evidence is not simply “strong” or “weak” because of the study label. A randomised trial can have important flaws, while a carefully conducted observational study can provide valuable information about uncommon harms or everyday care.

Peer review is a check, not a guarantee that findings are correct. Similarly, a published result does not become reliable just because it attracts headlines or is repeated widely online. Several reports may all draw on the same underlying study.

“Statistically significant” does not necessarily mean a difference is large enough to matter in daily life. And finding no statistically significant difference does not prove that two treatments are equivalent: the study may have been too small to detect a meaningful difference.

Personal experiences can highlight symptoms, priorities and possible problems. They usually cannot establish whether a treatment caused an outcome. Evidence is most useful when its strengths, limitations and relevance are considered together.

#Questions worth asking a clinician

• Does this study’s design fit the question we’re asking about benefits, harms, patient experiences or how the treatment works?
• How were participants selected, outcomes measured and other possible explanations addressed, and could those choices have biased the findings?
• Were harms tracked for long enough and in enough people to detect uncommon or delayed problems?
• Does the evidence show better health outcomes for patients, or only a biological change that might not translate into benefit?
• How does this conclusion account for all relevant studies, including conflicting findings and evidence from people whose circumstances resemble mine?