Our Wellness Position Prototype v1.0 Email Us

Why ChatGPT/Claude/etc can say “your health risks are…” :

FDA looks at 3 things: *Intended use + Marketing + Workflow*

FDA says: If you build a structured questionnaire for disease X and output risk for disease X, that’s SaMD. If user types random symptoms into ChatGPT, FDA calls that “patient-initiated inquiry”.

KEY POSITION

Unlike any AI or large organisation that sell devices or medical services led by a disease focused Q and A [eg a CPAP seller asking patients OSA specific questions. Or Boots offering patients disease risk analysis based on AI imaging with no doctor intervention]. we are far closer to regulatory philosophy because we construct the entire package to send to doctors and we point users at accredited third party verification tests so the doctor is even better informed. This can only help time and cost constrained doctors. And we go one step further we give the patient curate wellness tools focused on what they tell us. See how that beats the approved GPT methods whilst staying even closer in line with regulators patient safety objectives

For info: Chat GPT Position:
1. *Free text only, no scoring*: User types “I snore + tired”. AI lists “causes of fatigue: poor sleep, anemia, thyroid…”. No STOP-Bang, no 0-28 score.
2. *No disease-specific flow*: Don’t brand as “OSA screener”. Brand as “general wellness chat”.
3. *Strong disclaimers*: “I’m not a doctor. This is general info…”
Problem: Doctor wants structured eg STOP-Bang PDF, not ChatGPT transcript. And the tech CAN do it safely-safer if the doctor gateway is proactive .

The Pharmaceutical / Device Sales Firms:
Example: How CPAP companies stay on the right side of the regulatory line using our 2 step process that FDA accepts:

*Step 1: Patient- “education +disease focused questionnaire. eg
a.Do you snore?
b.Wake up tired?
c.Have you been told you stop breathing?
You may be at risk for sleep apnea. Take this 8-question screener. Talk to your doctor about a sleep study.

STEP1 : Process features the following regulatory comfort factor:
1. “May be at risk” = not diagnosis
2. “Talk to your doctor” = HCP gate
3. “Sleep study” = test, not treatment
An online example http://SleepTest.com,

*STEP 2 : Doctor-facing =
Doctor does the diagnosis and the prescription
Doctor reviews questionnaire -> orders home sleep test -> confirms OSA → CPAP
As opposed to ->
1. You have sleep apnea”* → That’s diagnosis without HCP-> Warning letter.
2. Buy CPAP now, no doctor needed”* → FDA will shut site fast.

No diagnosis. No treatment. Just: “Here are your answers → Here’s where it sits on scale → Here’s what doctors usually do next” AND HERE IS WHAT YOU CAN DO FROM A WELLNESS PERSPECTIVE.

So doctors get the structured data they want. Patients get clarity. Regulators see a bonafida HCP gate. FDA doesn’t regulate “helping patients prepare for doctor visits”. They regulate “software that diagnoses/treats”.

Regulators 100% prefer “help fix it” over “suspend which is their last resort.

How FDA/MHRA actually work with startups:

**Regulator** **”Help first” program** **What it does**
**FDA** Q-Sub / Pre-Sub meeting Free 60-min meeting. You show your app flow + wording. FDA tells you: “This is CDS” or “This needs 510(k)”. Costs $0 for small companies.
**MHRA** Innovation Service / AI Airlock UK’s “regulatory sandbox”. You describe your AI + they give informal guidance on UK MDR rules.
**EU** Notified Body consultation Before CE Mark, you can pay NB for gap analysis. They’ll say “add this disclaimer, remove this claim”

The Regulatory Process

FDA Position:

1. *Digital Health Center of Excellence*: FDA created a whole team in 2020 just to help startups, not police them.
2. *Pre-Cert pilot*: They know software updates fast. Prefer to coach than chase.
3. *Patient access*: They do not want patients to lose new tools. So they draw lines instead.

CPAP companies, Boots, even ChatGPT got “feedback” before any enforcement. Companies who ask early almost never have a serious problem.

MHRA Position

is one of the most startup-friendly regulators right now. They way rather chat than suspend. MHRA’s “help first” programs is designed with that philosophy:

Their vibe: Pro-innovation, but patients must be safe. Must show HCP gate + no diagnosis. What to send them to get clarity fast:
1.Flowchart of patient journey
2. Screenshots of patient result screen with exact wording
3. Screenshots of PDF you send to doctor
4. Your intended use statement: “To help patients prepare structured info for healthcare professional review”

◈ The Three Pathways At A Glance

Every result on the platform — all 11 Wellness Study conditions and all 19 standard biomarkers — is produced by exactly one of three pathways. The pathway is determined entirely by what kind of input exists; it is never a case-by-case judgement call.

◈ Our Approach

1. Ask questions based on established, peer-reviewed indices (Pathway 1) or symptom/risk-factor patterns (Pathway 2) — or, for biomarkers (Pathway 3), ask nothing at all and accept the user’s own already-obtained result.
2. Give a score and/or band only where a validated index supports one (Pathway 1). Never invent one where it doesn’t (Pathways 2 and 3).
3. Suggest seeing a doctor (or a private provider) for further testing, where the index indicates risk, where symptoms suggest it, or where the user’s own reported result warrants it.
4. Aggregate all responses into a single, personalised wellness plan — shareable with a doctor at one click.
5. Implement the wellness plan through the three Wellness Improvement Platforms (Body, Sleep, Mind) — full functionality for Release 1 detailed in the Feature List (Tab 1).

◈ Pathway 1 — Validated Index

Used wherever a peer-reviewed or recognised-professional-body instrument exists. The instrument’s own questions, own scoring, and own published bands are reproduced faithfully — BodySleepMind does not modify or invent any part of the underlying science. The result is shown to the user as a Quartile (Q1–Q4), sourced directly from the instrument’s own band.

Covers: BMI (WHO), VFA (Samouda et al. 2013), Diabetes Risk (Diabetes UK / NICE), GAD-7 (Spitzer et al. 2006), OSA (STOP-BANG, Chung et al. 2008), Insomnia (ISI, Bastien et al. 2001), Circadian Rhythm, and IBS-SSS (Francis, Morris & Whorwell 1997). Full citations and precedent for each are in the Condition-by-Condition table below.

◈ Pathway 2 — Symptom Reflection

Used where symptom or risk-factor questions exist but no validated severity instrument has been identified yet. The platform still names the condition directly and still points to a relevant test and a GP visit — what it withholds is a fabricated number. The card’s only visual signal is a binary one: a light amber background fill when the hidden, never-displayed symptom score crosses its own internal Elevated threshold, otherwise neutral.

Covers: SIBO, Celiac, Lactose. Full citations and precedent for each are in the Condition-by-Condition table below, and the full list of companies and bodies running this same symptom → named condition → test/GP pattern is further down this page.

◈ Condition-by-Condition Outcomes (Pathways 1 & 2)

The framework above, applied — not a defence, the result of running all 11 Wellness Study areas through the same pathway test. Each row states which path the condition landed on, and the index or symptom source that determined it. Biomarkers (Pathway 3) are covered separately above, since they carry no symptom source or index to list here.

◈ Companies & Bodies Using Pathway 2

Symptom → Named Condition → Test/GP Pattern (no score — just an implicit binary symptom-derived statement of risk)

The NHS itself runs this exact pattern for both Celiac and Lactose, with no severity score attached — the two pages cited above. The organisations below run the same structural pattern commercially, for SIBO, Celiac, and Lactose specifically, with no validated severity instrument behind any of them.

SIBO

• RefluxUK
• Healthpath
• Therapy Organics
• SmartSIBOtest.com
• Rachel Scheer
• Dr Hagmeyer
• SIBOtest.com
• driomole.com
• Nourish
• Somerset Health
• Honest Goodness
• Fairfield Nutrition
• Kirsten Greene

Celiac

• NHS
• Coeliac UK
• Celiac Disease Foundation
• Beyond Celiac
• Celiac Canada
• Gluten Free Society
• Top Nutrition Coaching
• Schär
• Thermo Fisher
• Rite Aid

Lactose

• NHS
• GetTested
• Private MD Labs
• Ubie
• CDHF
• Switch4Good

◈ Pathway 3 — User-Reported Lab Results (Biomarkers)

Pathway 3 covers blood test results the user already has — from their own records, a GP visit, or a test they purchased independently. There is no questionnaire, no Quartile, and no algorithmic verdict of any kind to grade. The user enters the number from their own report, then tells the platform themselves what the result was: Normal, Low, High, or Severe — a one-click choice, never a computed one. Any condition or risk their own report named can be typed in their own words; the platform never names a condition from the number itself.

Covers (19 standard biomarkers): Testosterone, TSH, FT4, Ferritin, B12, Folate, Vitamin D, HbA1c, Total Cholesterol, LDL Cholesterol, HDL Cholesterol, Triglycerides, CRP, ALT, GGT, Bilirubin, Creatinine, eGFR, Urea — plus any custom biomarker the user names themselves.

Precedent for Pathway 3

Two categories exist in this space, separated by who does the interpreting:

Log-only — same process as Pathway 3, no platform-side interpretation:

• Blood Test Grapher — the user sets their own reference ranges and logs their own results; the app does not classify them.
• 23andMe Lab Results & Vitals — supports manually entering biomarkers from a doctor’s office for tracking alongside other health data, with no verdict attached.

Interpreted — the platform itself classifies or explains the result on the user’s behalf:

• Carrot Care — logs results, but also generates its own “personalised insights” from them.
• Thriva — sells the test and provides a GP-reviewed interpretation (Normal/Borderline/High) with each result.

BodySleepMind sits in the log-only category for every biomarker: the number, the user’s own one-click result, and any condition their own report named are all recorded verbatim. No threshold comparison and no verdict are ever generated by the platform.

Competition Analysis