Phase 2 vs Phase 3
Why later-stage trials generally reduce—but do not eliminate—uncertainty.
Read the guide →METHODOLOGY
The RecompSignal Evidence Maturity Framework.
The M0–M5 framework describes how developed a therapy or research program’s overall human evidence base is. Outcome Evidence Strength separately shows how well a specific result is supported.
Evidence reviewed: September 19, 2026
Educational research
Source-labeled
The M0–M5 scale
M5
Established: multiple strong human trials plus substantial clinical or regulatory experience.
M4
Advanced clinical: large controlled human evidence, usually late-stage development.
M3
Clinical: meaningful controlled human efficacy evidence.
M2
Preliminary human: small, pilot, uncontrolled, or pharmacology-focused human studies.
M1
Preclinical: animal, cell, or mechanistic evidence dominates.
M0
Unsupported: adequate evidence for the claim evaluated has not been demonstrated.
How to read the system: Maturity describes the development of the overall evidence base. Outcome Evidence Strength is scored separately for scale weight, fat/composition, lean mass and strength/function. Source Tiers identify whether support is regulatory, peer-reviewed, registered, sponsor-reported or preclinical.
Five questions we ask
- What phase is the trial? Later phases generally reduce uncertainty.
- Who was studied? Population and sample size affect generalizability.
- What was the primary endpoint? Weight is not the same as body composition or function.
- What was the comparator? Placebo and active-comparator studies answer different questions.
- What did the results actually show? Headlines can overstate what a study measured.
Want the full evidence landscape?
Get the free 2026 Body-Recomposition Evidence Map with evidence maturity, body-composition signals and pipeline status in one reference.
Get the free Evidence MapApply the framework
Outcome Evidence Strength: scoring rubric
Scores describe evidence for the named outcome in the sources reviewed on each card. They are editorial judgments, not a validated clinical grading system. Weight, fat and lean scores concern measured change; the function score concerns demonstrated benefit. A high lean score does not mean lean tissue was preserved.
| Score | Evidence standard |
|---|---|
| 0/5 | Not established in reviewed human outcome evidence; may be unreported, awaiting results or unsupported. Not proof of no effect. |
| 1/5 | Indirect or preclinical support only; no persuasive human outcome result. |
| 2/5 | Preliminary human result, conference abstract, sponsor-only result or exploratory/mixed functional finding. |
| 3/5 | At least one published randomized trial or randomized substudy directly measuring the outcome; limitations remain. |
| 4/5 | Consistent direct results across more than one controlled human study in the relevant setting. |
| 5/5 | Broad, replicated randomized evidence with substantial clinical/regulatory experience for the specific outcome. |
Registration alone receives no efficacy points. Approval for weight management does not earn a muscle-preservation score. A newer phase number does not automatically raise outcome ratings. Each card explains its rating; scores are restricted to cited evidence and may change as sources are added.
Three ratings. Three different questions.
Evidence Maturity describes how developed the overall evidence base is. Outcome Evidence Strength describes support for a specific result. Source Tier identifies what kind of source supports the claim.
None of the three is a treatment recommendation or a prediction of individual response.
Apply the framework
Read the Headline Carefully
Ask about the population, comparator, duration and estimand.
Browse Insights →How to read doses studied
Each regimen is attached to its trial, population, formulation, route, frequency and duration. Target doses are not starting instructions. Dose escalation is identified without creating a self-use protocol. Registered regimens are labeled separately from published results. Where a public source omits the numerical dose, we say so.