The Evidence
Field Guide

Start with the exact claim

Write down the ingredient, formulation, route, population and outcome. A study of an approved product does not establish what is inside a differently sourced vial. “Recovery” is too broad to compare with a specific measured outcome such as a functional test.

Try it: rewrite a headline as “In which people, with which product, compared with what, and over how long?”

Separate humans from models

A cell experiment can help explain a pathway. An animal study can suggest a research question. Neither establishes a useful or safe effect in people. Human studies also differ: an uncontrolled series cannot answer the same causal question as a randomized comparison.

Look for: study design, enrollment, comparator and whether participants resemble the population named in the claim.

Read endpoints, not headlines

An endpoint is the thing a trial measures. A biomarker, body-weight change and improvement in daily function are different outcomes. Read the stated time point, missing-data handling, discontinuations and adverse events alongside the reported benefit. A relative change without the starting value can be misleading.

Look for: the registered primary outcome and time frame, followed by the full results. Trial registration alone does not mean the treatment worked.

Keep approval separate from evidence

Approval belongs to an exact product and labeled use in a particular jurisdiction. A positive study, an investigational trial or a compounding review is not the same as approval. Accelerated approval may rely on an endpoint expected to predict clinical benefit and require confirmation.

Example: the Forzinity profile separates its narrow Barth syndrome approval from broader mitochondrial research and RUO SS-31 products.

Audit safety and uncertainty

The number of participants and the observation period limit what a trial can show about uncommon or delayed harms. “No reported problem” does not prove no risk. Product identity, amount, contamination and preparation conditions are separate questions from a molecule’s mechanism.

Look for: the product label or study safety results, who was excluded, follow-up duration and the source of the material. A supplier documentation score is not a safety certificate.

Treat every blend as a new intervention

Evidence for individual ingredients does not establish the behavior of a mixture. Ratio, total contents, formulation, compatibility and route all matter. A familiar blend name may describe different formulas. The app calculates from the actual component amounts you enter; it does not establish that combining them is appropriate.

Use the library to understand the research question, and the tracker to preserve accurate records from your own instructions. Neither replaces clinical guidance.

Primary-source reading tools

ClinicalTrials.gov provides registered designs, outcomes and study status. DailyMed provides product labeling. Check the source dates and exact product.

For a worked approval example, see the Elamipretide / Forzinity profile and its FDA source.

Educational edition updated September 23, 2026. This is a reading framework, not medical advice or a dosing guide.