MW Wellness Labs
Making Traditional Chinese Medicine data computable.
Most of what patients say about Traditional Chinese Medicine is written in free text and never counted. We turn real-world experiences of TCM care into structured, searchable research data — conditions, body regions, modalities and patient-reported outcomes, each one traceable back to the words it came from.
First dataset: 776 public Google reviews from 3 multidisciplinary clinics in British Columbia, de-identified and structured.
The dataset
One selection, applied everywhere.
Choose a cohort and a discipline. Every tile, chart, answer and list below is recomputed from that one selection — nothing on this page is a stored figure.
Cohort
Discipline
A discipline filter selects reviews that name a modality from that discipline. These are multidisciplinary clinics, so a review can appear under more than one.
Reviews analyzed
776
In the current selection
Reviews with text
692
89% of this selection
Reviews naming a modality
361
47% of this selection — the rest describe care without naming it
Clinics
3
Multidisciplinary, British Columbia
Conditions identified
22
Distinct lexicon terms found in this selection
Modalities identified
10
Across Acupuncture & TCM, physiotherapy, massage therapy and chiropractic
Ratings
759 of 776 are 5★
Ratings are not an outcome measure. People who leave a public review are not a random sample of patients, so this distribution is a selection effect — we do not chart it.
Reviews mentioning Monica
53
The practitioner cohort inside this dataset
Source: public Google reviews from the three clinics, published here de-identified — no reviewer names, no contact details. Dates are approximate, derived from Google’s relative timestamps, so they are shown as months.
Modalities named, by discipline
How often each modality is named in the reviews in this selection.
- Acupuncture & TCM
- Physiotherapy
- Massage therapy
- Chiropractic
- Hatched = named alongside another modality
Acupuncture & TCM
Acupuncture
Cupping
Chinese herbal medicine
Physiotherapy
Physiotherapy
Exercise & rehab
IMS / dry needling
Manual therapy
Shockwave therapy
Massage therapy
Massage therapy
Chiropractic
Chiropractic
A mention means the review named that modality — not that it was the only care the person received. These are multidisciplinary clinics; acupuncture and TCM, physiotherapy, massage therapy and chiropractic are counted separately and never pooled.
Ask the dataset
Ask about a condition, a body region or a modality. Answers are counts of what reviewers wrote in the current selection.
Rule-based lookup over this dataset — the question is matched against a fixed lexicon and answered with counts. It is not a chatbot, it does not write prose about your health, and it cannot tell you whether a treatment works.
Conditions mentioned
Select a condition to open it
How many reviews in this selection name each condition.
Back pain
54 reviews · Back
Co-mentioned conditions
- Shoulder pain10
- Neck pain9
- Injury & strain4
- Disc problems3
- Hip pain3
- Knee pain3
- Muscle tension & stiffness3
- Sleep problems3
Body regions named
- Back54
- Shoulder10
- Neck9
- Leg, ankle & foot6
- Hip3
- Knee3
Modalities named, per discipline
Acupuncture & TCM
- Acupuncture21
- Cupping4
- Chinese herbal medicine1
Physiotherapy
- Physiotherapy10
- Exercise & rehab2
Massage therapy
- Massage therapy13
Chiropractic
- Chiropractic1
Counts are mention frequency, not effectiveness. Disciplines are listed separately and never added together — a review may describe physiotherapy, massage therapy or chiropractic rather than TCM.
How it works
Free text in. Structured records out.
A review is written as prose. To count anything, it has to become fields — and the fields have to stay attached to the words that produced them.
From a review to a record
Real reviews from the dataset. Every extracted field carries the phrase that triggered it, so the record can be checked against the text.
Review text
Good Vibes TCM® Healing Centre · ≈ Sep 2024 · 5★ · de-identified
I had a wonderful experience at the clinic with Dr. Arica. She was very professional, friendly, and attentive to my needs. She explained the procedure clearly and made me feel comfortable throughout. The acupuncture session was very relaxing and effective. I felt a noticeable improvement in my stress levels after the treatment. I highly recommend Dr. Arica and the acupuncture clinic to anyone looking for a natural and holistic way to heal their body and mind.
Extracted record
Conditions
Modalities
Patient-reported outcome
Attribution
Extraction is deterministic and rule-based today: a fixed lexicon with negation handling, run over the review text. Nothing here is a diagnosis, and nothing here says a treatment worked — only what the reviewer wrote.
Architecture
Three sources, one vocabulary.
The long-term work is connecting traditional knowledge, published research and real-world data through the same terms. One of the three is live today; the other two are planned, and are labelled as such.
Traditional TCM knowledge
PlannedAcupuncture points, TCM patterns, herbal formulas and classical indications, as structured references.
Published research
PlannedClinical trials, systematic reviews and biomedical studies, linked to the same terms.
Real-world data
Live todayDe-identified patient-reported experience from multidisciplinary clinics — the dataset on this page.
776 reviews structured · 3 clinics
AI research layer
Extraction, normalisation to a shared vocabulary, and retrieval across the three sources. Its job is to make the underlying records findable and countable — every statement it produces stays attached to the records it came from. It does not diagnose, prescribe or decide whether a treatment works.
Structured records
Conditions, body regions, modalities and patient-reported outcomes, per review.
Searchable counts
Frequency of what people describe, always split by discipline.
Traceable evidence
Every number opens into the de-identified reviews behind it.
Roadmap
What exists, and what comes next.
Written plainly, so the difference between what is built and what is intended stays visible.
Today
- 776 public reviews from 3 multidisciplinary clinics, de-identified and structured
- A fixed lexicon for conditions, body regions, modalities and patient-reported outcomes
- An explorer where every count opens into the reviews behind it
Next
- A published methodology page: lexicon, extraction rules and known failure modes
- Consented, de-identified intake and follow-up records alongside the review data
- Inter-rater checks on a sample of extractions, with the disagreement rate published
Later
- Linking the same vocabulary to published trials and systematic reviews
- A structured reference layer for TCM patterns, points and formulas
- An open dataset release for other researchers, subject to review
What this is, and what it is not
- This is an exploratory analysis of patient-reported experience — what people chose to write in a public review, counted.
- It is not efficacy evidence. Nothing here shows that a treatment caused a change, and no count should be read as a success rate.
- It is not diagnosis or medical advice, and it is not a substitute for assessment by a regulated health professional.
- These are multidisciplinary clinics. A review may describe physiotherapy, massage therapy or chiropractic care rather than TCM, so disciplines are counted separately and never pooled.
- People who leave a public review are not a random sample of patients, and the ratings in this dataset are overwhelmingly positive. Treat every proportion as a selection effect first.
- Reviews are published de-identified: no names, no contact details. Dates are approximate, so they are shown as months.