CCLabs Research Pipeline — A Guide for Clinicians

Purpose of This Document

CCLabs produces evidence-based research summaries on nutritional compounds and therapeutic interventions relevant to injury recovery, surgical rehabilitation, and musculoskeletal health. This document explains how that research is conducted, how evidence quality is assessed, and what outputs are produced — so that you, as a clinician, can understand the rigour behind any CCLabs document you read and make informed judgements about its applicability to your patients.

---

What We Research

Our research covers two broad categories:

Nutritional compounds — amino acids, vitamins, minerals, plant extracts, metabolites, and supplements with potential roles in tissue repair, muscle preservation, inflammation management, or functional recovery. Examples include HMB (beta-hydroxy beta-methylbutyrate), creatine, arginine, citrulline, and trace minerals.

Therapeutic interventions — medical procedures, physical therapies, devices, and clinical protocols such as hyperbaric oxygen therapy, photobiomodulation, and compression therapy.

Each research topic is investigated through a structured pipeline that moves from literature retrieval through evidence grading to clinical document production.

---

How We Search the Literature

Sources

Our primary database is PubMed / MEDLINE 🔗, the standard biomedical literature index maintained by the U.S. National Library of Medicine. We supplement this with:

• Cochrane Library 🔗 — the leading source for systematic reviews and meta-analyses
• OpenAlex 🔗 — an open scholarly graph covering over 200 million works, used for citation network analysis
• Semantic Scholar 🔗 — for identifying influential papers through citation metrics
• Europe PMC 🔗 — broader European coverage including preprints and grey literature
• ClinicalTrials.gov 🔗 — to check for registered trials that may not yet have published results

We also access high-impact journals directly, including JAMA, NEJM, BMJ, The Lancet, Nature, Clinical Nutrition, American Journal of Clinical Nutrition, Nutrients, Journal of the International Society of Sports Nutrition, and Journal of Cachexia, Sarcopenia and Muscle.

Sources we do not use: supplement company literature, health blogs, commercial websites, press releases, or grey marketing materials.

Search Strategy

For each topic, we run structured searches in a defined order:

1. Systematic reviews and meta-analyses — our highest priority, preferring the last 10 years unless a seminal older work is essential
2. Randomised controlled trials (RCTs) — the primary source of clinical efficacy data
3. Human mechanistic studies — to explain proposed biological pathways
4. Animal and in vitro studies — used solely to illustrate biological plausibility, never cited as evidence of clinical efficacy

For research-grade and publication-grade outputs, we apply a formal systematic review methodology that includes:

• PICO framing — explicitly defining the Population, Intervention, Comparator, and Outcomes before searching
• Pre-specified inclusion and exclusion criteria — set before searching to prevent cherry-picking
• Title and abstract screening — every retrieved paper is screened for relevance before extraction
• Full structured extraction — capturing study design, population, intervention, dose, comparator, duration, endpoints, effect sizes, confidence intervals, limitations, and funding sources

What We Exclude

We exclude studies that cannot meet basic methodological standards:

• Animal and in vitro studies when used to support clinical efficacy claims
• Combination product studies where the effect of a single ingredient cannot be isolated
• Conference abstracts without a published full-text companion
• Studies where the population clearly does not match the research question
• Retracted publications

All exclusions are recorded with reasons in a retrieval log that accompanies each research run.

---

How We Grade the Evidence

Every significant claim in our outputs is graded across three dimensions using a structured framework adapted from GRADE (Grading of Recommendations, Assessment, Development and Evaluations) 🔗 — the internationally recognised standard for rating the certainty of evidence in healthcare research. Our adaptation retains the GRADE emphasis on study design hierarchy, risk of bias, consistency, and directness, while adding a clinical relevance dimension specific to our audience.

Strength of Evidence

Grade	Indicator	What It Means
High	●●●●	Consistent findings across multiple meta-analyses or several high-quality RCTs with low heterogeneity
Moderate	●●●○	Several RCTs with some heterogeneity, or one strong meta-analysis with notable limitations
Low	●●○○	Small trials, inconsistent findings, or indirect evidence from observational data
Very Low	●○○○	Mechanistic rationale only, animal studies, in vitro data, or highly speculative inference

Direction of Evidence

Direction	Arrow	Meaning
Supportive	↑	Majority of high-quality evidence favours the claim
Supportive but not sufficient	↗	Evidence leans positive but is insufficient for a firm conclusion
Mixed	→	Evidence exists on both sides with no clear consensus
Contrary but not conclusive	↘	Some evidence suggests limited benefit or harm, but not definitively
Insufficient	—	Too few or too low-quality studies to determine a direction
Not supportive	↓	Preponderance of evidence contradicts the claim

Clinical Relevance

Level	Description
Clearly established	Effect is clinically meaningful, replicated, and widely accepted
Likely meaningful	Effect size is large enough to plausibly benefit patients in practice
Modest / uncertain	Effect detected but magnitude is small or variable across populations
Statistically positive but clinically unclear	Statistical significance achieved, but practical significance not established
Not established	No reliable evidence of patient-important benefit

How Claims Are Reported

Every graded claim in our documents follows this structure:

Claim: HMB may modestly improve lean body mass in older adults.
Evidence strength: ●●●○ Moderate
Evidence direction: ↗ Supportive but not sufficient
Clinical relevance: Likely modest
Key limitation: Many studies include mixed populations and combination formulas, complicating attribution of effects to HMB alone.

This format ensures that no claim is presented without its accompanying caveats.

---

Interpretation Guardrails

We apply five explicit guardrails before writing any conclusion:

1. Statistical significance does not equal clinical importance. A significant p-value does not automatically translate to a meaningful patient outcome. When it does not, we say so.
2. No population over-generalisation. A benefit demonstrated in healthy older adults is not extended to patients with diagnosed sarcopenia, frail patients, or other groups without direct evidence.
3. Combination formula attribution. A benefit observed with a multi-ingredient formula is not attributed to a single component unless isolation studies confirm it.
4. Mechanism does not equal efficacy. A plausible biological mechanism is not presented as proof of clinical effect. Mechanistic data are presented as rationale only.
5. Function versus surrogate endpoints. If physical function outcomes are inconsistent with body composition findings, the inconsistency is stated directly — not glossed over.

---

How We Handle Bias

Every research run includes a systematic bias assessment covering:

Bias Type	What We Check
Funding bias	Are key trials industry-sponsored? Do multiple trials share a single funder?
Author conflicts of interest	Declared equity, consultancy, patents, or other COI in primary studies
Publication bias	Are negative or null trials absent from the literature?
Population bias	Are heterogeneous populations pooled as if interchangeable?
Intervention bias	Do trials use different doses, forms, timing, or co-interventions?
Selective outcome emphasis	Are endpoints pre-specified, or do results emphasise post-hoc findings?
Surrogate endpoint reliance	Are patient-important outcomes absent or inconsistent with surrogate findings?
Conclusion-evidence mismatch	Do authors' conclusions exceed what the data support?

Bias findings are reported explicitly in every output document. We do not default to "no bias signals identified" — bias must be actively assessed and documented.

---

How We Report Safety

Every clinical document includes a substantive safety section. We never write "completely safe" or use unqualified "well tolerated" language. Instead, we report:

• Adverse events observed in trials: type, severity, and incidence
• Dropout rates and reasons where reported
• Tolerability profile, including gastrointestinal tolerance and dose dependency
• Organ-function markers relevant to the compound class (e.g., liver enzymes, renal markers)
• Duration basis — the longest trial duration on which any safety statement rests
• Special population concerns — older adults, paediatric populations, pregnancy, hepatic or renal impairment, polypharmacy risk

If safety data are absent, we state this explicitly. The absence of safety data is itself a finding, not a reason to omit the section.

All safety statements are qualified:

"Generally well tolerated in healthy adults at doses of 3 g/day for up to 12 weeks, based on 4 included trials. Long-term safety data beyond 12 weeks are not available from the retrieved literature."

---

How We Report Doses

We report studied doses, not recommended doses. All dose information comes from the peer-reviewed trials in our evidence pack, not from manufacturer labels or product guidance.

Key rules:

• Doses are matched to the specific clinical context being described. If the research focus is muscle preservation during bed rest, we report doses from bed-rest trials — not from resistance-training studies, even if the latter are more commonly cited.
• When doses vary by indication or population, we present an indication-specific dose table rather than collapsing them into a single "typical dose."
• We never extrapolate doses across populations — a dose established in older adults with sarcopenia is not presented as applicable to healthy younger adults.
• We use the phrase "studied dose" rather than "recommended dose" or "optimal dose."

---

Overall Conclusion Level

Every research run concludes with an overall evidence conclusion level:

Level	Criteria
Strong	Multiple independent, consistently replicated RCTs or high-quality meta-analyses; low bias risk; patient-important outcomes demonstrated; adequate safety data
Moderate	Several RCTs with meaningful but qualified evidence; some heterogeneity or bias flags; effects detectable but gaps remain
Cautious	Limited or inconsistent RCTs; notable bias risks; surrogate endpoints predominant; safety data short-term or narrow
Weak	Mechanistic rationale or preclinical data constitute the primary evidence; no adequate RCTs; major safety unknowns

A Cautious or Weak conclusion level does not mean the topic is unworthy of attention — it calibrates the language we use and the confidence with which findings are presented. Language is proportional to the evidence:

• Strong — "evidence demonstrates"
• Moderate — "evidence suggests"
• Cautious — "limited evidence indicates, though findings require confirmation"
• Weak — "preliminary evidence from preclinical studies proposes"

---

What We Produce

Structured Conclusion

Every document ends with a four-part conclusion:

1. What the evidence supports — with appropriate hedging
2. What remains uncertain — specific gaps, not boilerplate
3. Where the intervention may be most useful — the clinical or demographic context in which benefit is most plausible
4. Studies still needed — specific designs, populations, or endpoints that would resolve key uncertainties

Output Documents

Depending on the depth of research commissioned, we produce some or all of the following:

Document	Audience	Purpose
Clinical Evidence One-Pager	Clinicians, investors	Rapid-scan evidence summary: evidence strength indicators, direction arrows, key findings, safety, clinical takeaways — designed to be read in 30 seconds
Research Paper	Researchers, clinicians	Publication-quality literature synthesis (2,500–4,000 words): structured abstract, methods, mechanisms, clinical trial evidence, meta-analysis evidence, safety, limitations, clinical implications
References	All audiences	Verified Vancouver-style reference list — every citation confirmed against its source
Informative One-Pager	Patients, general public	Consumer-friendly evidence summary, written in accessible language but grounded in the same evidence base
Educated Blog	Informed consumers	Longer-form evidence-led article for website publication
Short Blog	General consumers	Concise summary for broad readership

Research Depth Modes

Mode	Depth	Papers Reviewed	Documents Produced
Scan	Quick signal check	Up to 6	Clinical one-pager only
Briefing	Citable evidence summary	10–12	Clinical one-pager + references
Research	White paper	15–20	Clinical one-pager + research paper + references
Publication	Full pipeline	20–25+	All documents including blogs

---

Language Standards

All clinical documents use neutral scientific language. We maintain a strict vocabulary:

We Use	We Never Use
associated with	breakthrough
may improve	miracle
modest benefit	revolutionary
preliminary evidence	guaranteed
evidence suggests	proven
results are mixed	dramatic effect
some evidence indicates	completely safe
findings suggest	definitively shows

Conditional language ("may," "appears to," "evidence suggests") is used throughout. Statistical significance is never presented as clinical certainty.

---

Citation Verification

Every citation in our documents is verified before publication. We use the Crossref 🔗 REST API to confirm author names, journal, year, title, and DOI for each reference. Our reference lists use Vancouver-style numbered citations, the standard in biomedical publishing.

We never fabricate citations, author names, trial results, statistical figures, or publication years. If an effect size cannot be verified from the original source, we describe the direction of effect qualitatively rather than reporting an unverifiable number.

---

How to Read Our Documents

When reviewing a CCLabs clinical one-pager or research paper, look for:

1. The Quick-Look Verdict (one-pagers) — a boxed summary showing overall evidence strength and a one-sentence bottom line
2. Evidence strength indicators — visual markers (e.g., filled/unfilled circles) showing the quality of evidence behind each finding
3. Direction arrows — indicating whether the evidence supports, contradicts, or is mixed on each claim
4. The Safety section — always present, always qualified with duration and population
5. The Key Limitation block — an honest, prominent appraisal of the most important caveat
6. The Conclusion — structured as what is supported, what is uncertain, who may benefit most, and what research is still needed

If any of these elements are absent from a document, the review process is incomplete.

---

Contact and Further Information

CCLabs research outputs are produced to support evidence-informed clinical decision-making. They are not intended as treatment guidelines, prescribing advice, or substitutes for clinical judgement.

For questions about the evidence behind any CCLabs product or document, contact: research@cclabs.uk

---

This document describes the CCLabs research methodology as of March 2026.