Built on fifteen years
of clinical evidence.
GutIQ is not a wellness app dressed in clinical language. It is a structured methodology synthesised from fifteen-plus years of gastroenterology practice, peer-reviewed literature across thirty medical knowledge sources, biomedical AI models, and prospective validation across a 25,000-person cohort.
"Patient-reported symptoms, when systematically structured, can recover the clinical pattern recognition that defines the best gastroenterology practice — and make it accessible at scale."
The four pillars of the methodology
Each insight delivered by GutIQ rests on four reinforcing inputs. No single source — clinical, literary, computational, or empirical — is allowed to determine a recommendation alone.
Clinical Practice Data
Fifteen-plus years of de-identified gastroenterology consult records, symptom-cluster annotations, and longitudinal outcome data drawn from clinical practice in functional GI medicine.
Peer-Reviewed Literature
Systematic synthesis across thirty medical knowledge sources, including PubMed, the Cochrane Library, society guidelines (AGA, ACG, BSG), the Rome Foundation criteria, and the Monash FODMAP framework.
Biomedical AI Models
Open biomedical language models and embedding systems — including BioBERT, PubMedBERT, BioGPT, and Med-PaLM-class architectures — assist with literature synthesis, pattern extraction, and cross-citation reconciliation under clinician oversight.
Prospective Validation
Pattern classifications and pattern-specific protocols are refined through prospective validation across a 25,000-participant cohort, with paired symptom tracking, baseline-to-12-week outcomes, and inter-rater agreement studies.
How the methodology was built
The framework is not a snapshot. It is the cumulative output of more than a decade of clinical observation, literature review, and validation.
Clinical Foundation
Pattern observations across thousands of functional GI consults begin to cluster into recurring presentations. Early symptom-cluster annotations form the seed of what becomes the GutIQ pattern framework.
First Pattern Articulation
The twelve-pattern framework is formalised, mapped against Rome IV functional-GI criteria, and cross-referenced with the emerging microbiome literature. The first three archetypes are described.
Literature Synthesis at Scale
Open biomedical language models are integrated into the literature workflow, enabling structured synthesis across more than seven hundred peer-reviewed studies per pattern. All AI-extracted evidence is reviewed by clinical staff before incorporation.
First Validation Cohort
A 5,000-participant validation cohort produces the first quantitative test of pattern reproducibility, inter-rater agreement, and pattern-specific protocol response rates. Findings drive the addition of the seven-overlay framework.
25,000-Person Refinement
Expansion to a 25,000-person multi-site cohort enables refinement of pattern-specific scoring thresholds, overlay sensitivity, and protocol-response prediction. The current public methodology is the direct output of this cohort.
Continuous Re-Validation
Quarterly re-validation cycles incorporate new literature, refined cohort outcomes, and pattern-specific tracking data. Every recommendation surface in the product carries an internal evidence stamp visible to the clinical team.
Where the evidence comes from
We synthesise across the same authoritative sources used in academic gastroenterology — and apply the same standards of citation, recency, and methodological quality.
Medical knowledge bases
Society guidelines
Specialty frameworks
Open biomedical AI models
AI models support literature triage and citation extraction. They do not generate clinical recommendations independently. Every recommendation that reaches a user has been reviewed by a clinician on the working group.
How a recommendation enters the product
Every claim, threshold, dose, and pattern definition in GutIQ passes through a four-stage validation pipeline before it reaches a user.
Literature Triage
Biomedical AI surfaces all relevant literature for a candidate claim. Clinical reviewers grade studies on the GRADE framework and discard low-quality evidence.
Clinical Working Group
Practising gastroenterologists, registered dietitians, and behavioural-health clinicians review and adjudicate. A two-thirds majority is required for inclusion.
Cohort Validation
Approved claims are tested against the 25,000-person cohort. Recommendations that fail to predict reported outcomes are rejected or refined.
Quarterly Re-Review
Every claim is re-evaluated against new literature each quarter. Stale or superseded evidence is replaced; depreciated recommendations are explicitly withdrawn.
The pattern system at a glance
Symptoms are scored against a structured framework of primary patterns, archetypes, and overlays. Every individual receives a distinct profile.
A focused 48-question instrument covering symptoms, lifestyle, hormonal context, and stress dynamics. Validated against clinical interview agreement at 0.81 kappa.
Slow Transit, Fast Transit, Fermentation Sensitive, Stress-Reactive, Visceral Sensitivity, Fat/Bile Sensitive, Upper GI/Reflux, Meal-Timing Sensitive, Inflammatory/Leaky-Prone, Low Diversity, Protein-Heavy/Fiber-Poor, Balanced/Resilient.
Higher-order clinical groupings that synthesise multiple primary patterns into a single phenotype with shared therapeutic implications.
Modulating features (Gas-Dominant, Immune-Reactive, Motility-Impaired, Gut-Brain Dominant, Environmental/Seasonal, Hormonal/Cycle-Reactive, Variable Pattern) that can layer on top of any primary pattern.
What we hold ourselves to
Evidence transparency
Every claim that appears in a recommendation, report, or article cites its underlying evidence. Where evidence is mechanistic or extrapolated rather than direct, we say so explicitly.
Clinical adjudication for every recommendation
No supplement, dose, or protocol enters the product without review by a practising clinician on the working group. AI accelerates literature work; it does not author recommendations.
We are not a substitute for medical care
GutIQ supports informed self-management. It does not diagnose disease, replace your gastroenterologist, or override the judgement of a licensed clinician who knows your individual history.
No supplement-brand kickbacks
When a specific brand is named in a recommendation, the citation is illustrative — to indicate quality marks, standardisation, or formulation. GutIQ does not earn affiliate commissions on supplement purchases.
User data is private by default
Assessment data is encrypted at rest and never sold. Aggregate, de-identified cohort data is used internally to refine the methodology. Users may delete their data at any time.
Methodology is publicly inspectable
The pattern definitions, scoring rules, and recommendation logic are described in our research library and accompanying guides. We invite clinician scrutiny and correspond with any researcher who requests detail.
We use biomedical AI extensively in the literature workflow because it would be irresponsible not to — the rate of publication in gastroenterology, microbiome, and gut-brain research exceeds what any individual team can keep up with manually. But AI in our pipeline is a research accelerant under clinician supervision, not a recommendation generator. The line between those two is the difference between a credible methodology and a wellness app pretending to be one.
See the methodology
applied to your data.
A ten-minute assessment maps your unique pattern profile across 48 clinical parameters and produces a personalised report grounded in the same evidence base described on this page.