Every monograph in the atlas carries a Key studies table and an Evidence-by-claim table. Together they are meant to be read as a compact argument: a claim, the human evidence supporting it, the limitations of that evidence, and the atlas's confidence grade. The pieces only work when read in order and when the reader understands what each column asks.
This guide walks through the study-level fields defined in the Evidence grading methodology. It is about evidence literacy, not interpretation for personal decisions. The underlying rules are documented in Evidence grading and Scope and selection.
The Key studies table
Each row in the Key studies table represents one report — a clinical trial registration, a published study, or a retrospective analysis — not a summary of the entire evidence base.
Design and population
The first column tells you what kind of study you are reading. A randomized controlled trial (RCT) is structurally different from a single-arm pilot or a retrospective chart review, and that difference matters more than the reported outcome. The tirzepatide monograph shows the standard: SURPASS-2 is described as "Phase 3, RCT, 40 wk; T2D (n=1,879)" — you immediately know it is prospective, randomized, has a comparator arm, and includes nearly 2,000 participants. Compare that with the BPC-157 row describing "Retrospective chart review of 17 treated knee-pain patients; 16 were reached by phone (12 BPC-157 alone, four BPC-157 plus TB4)." The design label alone signals that the second row carries far less evidentiary weight, whatever the numeric result.
Exposure studied
This column records what was actually administered — dose, route, frequency, and duration. It is not a recommendation. The tirzepatide monograph records a three-arm comparison of escalating doses against semaglutide — a standard active-controlled trial design. The BPC-157 monograph records "Single intra-articular administration; dose not reported in the publication abstract" for its knee-pain study. An unreported dose in a published study is a serious limitation.
Endpoints and result
The atlas records the measured endpoint (what was actually counted or measured) and the result, in the original units. Statistical significance is not treated as clinical importance. Tirzepatide's SURMOUNT-1 reports mean weight loss of 16–22% depending on dose versus 2.4% for placebo, with 57% of the top-dose group achieving at least 20% loss. The result is concrete, the comparator is explicit, and the clinical relevance is self-evident. The BPC-157 chart review reports "11/12 in the BPC-157-only subgroup and 14/16 across both followed groups reported significant pain relief." The denominator changed, the assessment was subjective telephone follow-up, and no control group existed — the numbers cannot be compared to a controlled-trial result.
Important limitations
Every study has limitations. The atlas surfaces them in the last column so they are not buried in discussion sections. Tirzepatide's SURPASS-2 limitation states the semaglutide comparator arm used only a mid-range dose, not the highest approved dose. This is a specific, actionable limitation. BPC-157's limitation row reads "No control group, randomization, standardized outcome instrument, or diagnosis-specific analysis; combination subgroup confounds the overall result." That is not a minor caveat — it describes a study designed in a way that cannot support an efficacy conclusion.
The Evidence-by-claim table
Where the Key studies table lists individual reports, the Evidence-by-claim table synthesizes. Each row assigns a claim (a specific indication or outcome), a development stage, a grade (A–E or X), and the best human evidence for that claim.
Claim specificity
The grade attaches to the claim, not the molecule. Tirzepatide has grade A for "Glycemic control in T2D" (supported by the SURPASS program and FDA approval) but grade B for "OSA with obesity" (supported by a single well-designed trial with pending confirmatory data). The molecule did not change. The claim, endpoint, and regulatory context did.
As the Scope and selection policy notes, inclusion reflects visibility, not endorsement — a research-market peptide may have broad online presence but grade D or E evidence across all claims.
Grade boundaries
A requires current authorization plus adequate controlled trials and post-market context. Tirzepatide's T2D claim meets this.
B means moderate human evidence without current approval for that claim. Tirzepatide's OSA claim sits here because the status table still records the indication as separately approved — the grade reflects the development stage.
C is preliminary human evidence, often small or uncontrolled. BPC-157's interstitial cystitis pilot (n=12, single-arm, no control) earns this.
D is preclinical only. Most BPC-157 claims — musculoskeletal healing, IBD, wound healing — sit here because human efficacy data are absent.
E covers anecdotes and marketing claims.
X signals affirmative contradiction, not merely weak evidence.
Reading cross-table
The real value comes from reading both tables together. Tirzepatide's evidence-by-claim rows carry grade A, and the Key studies table shows large RCTs with credible endpoints. BPC-157's evidence-by-claim rows carry grade C or D, and the Key studies table shows retrospective reviews, single-arm pilots, and an n=2 pharmacokinetic report. The tables agree. When they disagree — a grade that seems too high for the cited study, or a study that does not match the claim — that tension signals where the atlas made a judgment call worth examining.
The semaglutide monograph offers another useful comparison: its Key studies table includes the massive SELECT CVOT (n=17,604) producing a MACE HR of 0.80, while the BPC-157 table shows no completed RCT at all. The contrast between approved-drug and research-market evidence is visible in the table structures themselves.
A parting check
When you read a monograph, test whether the Key studies row matches the claim table's grade. Ask: Is the study design adequate for the claim? Is the endpoint a patient benefit or a surrogate? Is the comparator appropriate? Are the limitations minor or decisive?
For more detail, see the full Evidence grading method.