From cell dish to clinic: study types
The index labels every one of its 48 entries by evidence strength, and the kind of study behind a claim is a large part of that label. Below are the study types the index cites, from work outside a living organism to studies that combine human trials. Read the evidence labels for what each tier means and the method page for how they are assigned.
- 1
In vitro
Studied in cells or tissue in a dish or test tube, outside a living organism.
3 of 48 compounds mention "in vitro" in their evidence text.
- 2
Ex vivo
Studied in tissue taken from a living organism and kept alive outside it, such as isolated hair follicles.
1 of 48 compounds mention "ex vivo" in their evidence text.
- 3
In vivo / animal study
Studied in a living organism, most often mice or rats. Results frequently fail to carry over to humans.
21 of 48 compounds mention "animal" in their evidence text.
- 4
Preclinical
Research done before human testing, in cells and animals.
1 of 48 compounds mention "preclinical" in their evidence text.
- 5
Case report
A description of what happened to one patient or a few patients. Useful for spotting rare effects, not for judging how often they occur.
2 of 48 compounds mention "case report" in their evidence text.
- 6
Pilot study
A small early study meant to test feasibility, not to prove an effect.
1 of 48 compounds mention "pilot" in their evidence text.
- 7
Observational study / cohort
A study that follows people without assigning a treatment. It can show associations but not cause and effect.
1 of 48 compounds mention "observational" in their evidence text.
- 8
Randomized controlled trial (RCT)
A study that assigns people by chance to the treatment or a comparison (often placebo), so differences in outcome can be attributed to the treatment rather than to who chose it.
16 of 48 compounds mention "rct" in their evidence text.
- 9
Double-blind
Neither participants nor investigators know who received the treatment until the study ends, which reduces bias in reporting and assessment.
0 of 48 compounds mention "double-blind" in their evidence text.
- 10
Meta-analysis
A study that statistically combines the results of several earlier studies on the same question.
0 of 48 compounds mention "meta-analysis" in their evidence text.
Related terms
Retrospective study
A study that looks back at existing records rather than following people forward under a plan.
Placebo-controlled
A study in which the comparison group receives an inactive look-alike, so that expectation effects are present in both groups.
Open-label / unblinded
Everyone knows who received the treatment. Results are more prone to expectation and reporting bias.
Phase 1, 2 and 3 trials
Phase 1 tests safety and dosing in a small group; phase 2 looks for signs of effect in more people; phase 3 tests effect and safety at scale and usually supports an approval application.
Anecdotal
A personal report, for example a forum post. Unverified, not controlled, and not evidence that something works or is safe.
26 of 48 compounds mention "anecdotal" in their evidence text.