How this record is built

Methodology, corrections, commercial policy and privacy

Selection, ordering, evidence labels, how the dates work, how corrections are made, and the commercial and privacy policies.

How services are selected, ordered and corrected; what the evidence labels mean; how dates work; and the editorial, commercial and privacy policies. These sections share one page for now, with anchors, and will split when they are long enough to need it.

Inclusion

A service is included when the researcher opened its public pages and could document scope, fees and terms with sources and dates. Inclusion cannot be bought. No service has any commercial relationship with this site today; if one is ever entered into, it will be disclosed on this page and beside the link, and it will not change whether a service is included or where it appears. Seven services are documented today: Felix, Hims, Jill, Maple, Medical Weight Management (a BC public program), Phoenix and Pocketpills. Raven is recorded under Phoenix because the research found the same operator and the same prices. A dedicated record needs enough complete evidence; thin placeholders are not published.

Ordering

Documented matches are listed alphabetically. Rows that are conflicting, not found in the sources checked, explicitly not offered or not applicable are shown in a separate "confirm availability" section and never mixed in as matches. Nothing is sorted by price, because unknown components would make any lowest-price list wrong; readers may compare only the known fees, and the cost worksheet refuses to call an incomplete subtotal a total.

Evidence labels

Regulator-confirmed: an official record was opened directly (Health Canada DPD, a product monograph linked from it, a provincial formulary, a public register).
Directly observed: the researcher saw the page element.
Company self-report: the company states it about itself. This is the normal label for every provider fact. It is not a licence check.
Original research finding, measured test result, estimate, inference, assumption: used where stated.
Unknown: used as a value, never as a blank.

Unknown states

Company self-report: offered The company's own page says it offers the service there. This is not a check of licensing or quality.

Conflicting Two or more of the company's own pages disagree. Both values are shown with their dates.

Not found in sources checked The researcher looked and did not find it. This is not the same as not offered.

Explicitly not offered The company's page says it does not serve that place.

Not applicable The service is a public program limited to one province by design.

Not researched Nobody has looked yet.

Dates

Every fact carries up to three dates: the date the source itself carries, the date the researcher opened it, and the verification date recorded in the update log. They are never merged. "Effective" and "checked" are different things. When a source is undated, the record says "undated" rather than inventing a date.

Seven things kept apart

Authorized in Canada Health Canada has issued a Notice of Compliance and a DIN.
Marketed status in the DPD The Drug Product Database says Marketed, Approved, Dormant or Cancelled. Approved is not Marketed.
Actual pharmacy stock Always unknown on this site. Nothing here says a product is on a shelf.
Canadian indication and age group What the product monograph authorizes, for whom. A US indication is not a Canadian one.
Provincial public coverage Whether a public plan pays, under what criteria, from what date.
Provincial interchangeability Whether a pharmacist may substitute a generic for the brand under that province's rules.
Private insurance Not researched. Every plan differs.

Recheck cadence

Weekly: service fees, province lists and pricing pages (a 7-day recheck assumption for the private services; 30 days for the BC public program). Monthly: DPD status, formulary editions, special authority forms. Quarterly: evidence and care-option pages. Trigger: Health Canada safety alerts, new approvals, formulary changes, major price changes, broken critical sources. A decisive record that is overdue shows "needs recheck" and drops out of sorting until it is reviewed.

Corrections

When a live source disagrees with the record, the source wins and a new dated row is added; old rows are never edited. The update log distinguishes a correction of this site's own error from a later change by the provider or the plan. Every change records the previous value, the new value, the reason, the editor and, where one was done, the reviewer.

Rights

Public facts are summarized with citations. Whole monographs, provider copy, logos and third-party compiled databases are not republished. This site claims the original selection, normalization, annotation and change history, not the underlying public facts. No patient stories and no personal data are collected.

Commercial policy

Today: no paid placement, no affiliate links, no referral buttons, no sponsored cards, no email list, no lead forms. The owner is a licensed pharmacist in British Columbia, and the rules that apply to a pharmacist's advertising and to referral incentives are being reviewed before any of that changes.

If a paid model is adopted after that review, these rules will apply and will be published here first:

Placement is labelled as sponsored, sits outside the results, and cannot change the order of anything.
Any material connection is disclosed next to the link, not on a distant page.
Outbound links to a provider carry a plain tag naming this site and nothing else; no health information is ever placed in a link.
No referral payment or incentive is accepted where a professional rule prohibits it.
No service can buy inclusion, exclusion or a better row.

Privacy

No cookies, no advertising or social media pixels, no accounts, no uploads, and nothing that follows a reader from one visit to the next. Three parts of this site send something to a server when you use them, and each is described below: the question page, the contact forms and the anonymous visit counters. Typefaces, icons and every other file are served from this site. Two scripts come from Cloudflare, the company that hosts the site: its cookieless Web Analytics on every page, and Turnstile on the contact page, a spam check that does not set advertising cookies or build a profile, to keep automated messages out of the inbox. Outbound links go to the provider's own site, which has its own privacy practices, and carry a plain tag naming this site and nothing else.

Every page is served by Cloudflare, which hosts the site. Like any web host, Cloudflare processes the technical details of each request (the network address, the browser's identifying string and the page requested) to deliver the page and protect the site from attack, under its own privacy policy. This site does not export or keep those request logs and does not store network addresses. The owner can see traffic totals, which automated crawlers visited, and the security events Cloudflare records when it blocks or challenges a request, in the Cloudflare dashboard, where Cloudflare keeps them for a limited period under its own terms.

The comparison filter and the cost worksheet run in your browser. Nothing you type into the cost worksheet is sent anywhere. The only thing recorded from either tool is an anonymous count: which province, service type and language were picked in the filter, and which service was chosen in the worksheet, as described under counting visits below.

When you ask a question there, the text of your question goes to this site's own endpoint. That endpoint matches it against an index built from the published records, and then sends the question together with the matching records to Anthropic, the provider of the language model that words the answer, in the United States. The model is given the records and is instructed to use nothing else. This site keeps the question, the answer it showed and the records it cited for ninety days, with no address, cookie or identifier attached, so the maintainers can see which records people ask for and where the answers fall short. A question the page refuses, such as one about your own health, is counted by category only; its text is never kept. Under Anthropic's published terms for business use of its API, the question and the answer are deleted within thirty days, may be kept for up to two years if its automated safety systems flag them, and are not used to train its models without permission. To stop one machine flooding the service there is a counter that allows twelve questions a minute; it is keyed to a one way hash of your network address and expires after sixty seconds, so no address is stored.

The contact forms are the other part that sends something. What you type goes to this site's own endpoint, which does two things with it. It stores a copy in a database hosted by Cloudflare so a correction can be followed through to a published change; a daily job deletes anything older than ninety days. It also emails a copy to the people who maintain the record, through Resend, an email delivery service in the United States that keeps sent messages and delivery logs for thirty days. That email arrives in the maintainers' inbox, which is hosted by Google, and the inbox copy is not deleted automatically; you can ask for it to be deleted, as described below. No network address, no cookie and no identifier is stored with a message, only what you chose to type and the country your request came from. Nothing is used for advertising, sold, or passed to any service recorded here.

The forms are built not to want health information, and the endpoint removes it anyway. Before a message is emailed or stored, anything that reads like a dose, a measurement or a description of your own treatment is replaced, the removal is noted, and the original is never written down. The page says the same thing in plain words next to the box, because the most useful correction contains a page address and a source, not a medical history.

Because the question text leaves this site, the page says so in plain words above the box and asks readers not to type personal health details. The page refuses questions about a reader's own situation before anything is sent, so a question about your own dose or eligibility is answered by a refusal from this site rather than by a model.

Counting visits

Since 23 September 2026 the site counts its own visits, so the maintainers can see which records are read, where readers come from and what is missing. Neither way of counting uses a cookie, stores a network address or gives a visitor an identifier, so no visit can be linked to another visit or to a person.

The server counts each page it serves: by page and day, the kind of site that sent the reader (a search engine, an AI answer engine, a social site, another website, or none), that site's name, the country the request came from, and whether the device is a phone, a tablet or a computer. It also counts crawlers by name, downloads of the dataset files, addresses that do not exist, and visits that arrive on links carrying campaign tags. Each of these is a daily total, not a record of a visit.

A small script on each page adds, in broad bands, how long the page stayed open and how far it was scrolled, the outside site and page a link led to, and the choices picked in the comparison filter and the cost worksheet. It never sends an amount, a message or anything typed into a box. If your browser sends a Global Privacy Control or Do Not Track signal, the script sends nothing at all.

Cloudflare also measures visits and page speed on these pages with its Web Analytics, which uses no cookies and no fingerprinting, under Cloudflare's own privacy policy.

Daily totals are kept for 400 days, and the question log and contact messages for ninety days. None of it is sold, shared with any service recorded here or used for advertising, and only the site's owner can see it, on a private page protected by Cloudflare sign-in.

Where it is processed

Cloudflare, Anthropic, Resend and Google operate mainly from outside Canada, chiefly the United States. What they handle for this site may be stored or processed there, and while it is, it may be accessible to the courts, law enforcement and national security authorities of that country under its laws.

Who is responsible, and how to ask

The site's owner is accountable for the personal information this site handles and acts as its privacy contact. Privacy questions and requests go through the contact page, choosing "A privacy request" in the reader form, or by email to [email protected]. You can ask what is held from a message you sent, ask for it to be corrected, or ask for it to be deleted early, including the copy in the inbox. Because nothing ties a message to an account, include enough of what you sent (the date, the page it named, a phrase from it) for it to be found. A request is answered within thirty days.

This site sends no marketing email and keeps no mailing list. A reply to a message you sent is only a reply. The site, the question page and the contact forms opened to the public on 22 September 2026, and on 23 September 2026 the real data flows described above were mapped against British Columbia's Personal Information Protection Act, the federal PIPEDA and Canada's anti-spam law. The mapping was brought up to date on the same day, when visit counting started. Anything that would collect more about readers, such as a newsletter or an advertising pixel, will be described here before it starts.

Gaps that still limit what this site can say

Gaps that still limit what this record can say
Gap	Topic	State	What it blocks	Resolution step	Priority	Status after round 2
G01	Factual verification gap: current DPD detail/status for indexed products	source inaccessible	Which product presentations may be described as authorized or marketed	Reopen exact DIN records and compare current Health Canada monograph and status; retain indexed-only status until successful	P0 before publishing affected product	CLOSED by R2-1: all 32 DINs opened as DPD detail records; Sevmia Approved not Marketed (2026-06-29). See R21 rows.
G02	Factual extraction gap: complete current Rybelsus, Victoza, Contrave, Imcivree and additional-DIN monographs	not researched	Exact indication, population, device and safety statements	Finish current Canadian monograph extraction for each proposed launch product; omit unaudited fields or defer those products	P0 before publishing affected product	CLOSED by R2-1: Rybelsus PM 2026-07-13 (CV indication in PM), Victoza PM 2025-08-21, Contrave and Imcivree Marketed with current PMs extracted.
G03	Factual extraction gap: province × product/DIN × indication × plan benefits and interchangeability	not researched	Whether a national coverage checker can return an accurate benefit result	Complete only pilot jurisdictions first; join current criteria, plan eligibility, DIN, substitution policy, effective date and payment basis; retain all others as unverified	P0 before coverage claims	PARTLY CLOSED by R2-1 for BC and Ontario (benefit status, SA/LU criteria, interchangeability groups). Other provinces: generics listed; QC and PEI interchangeability still unknown. Territories not researched.
G04	Factual access gap: BC current decision/SA documents and New Brunswick current criteria	source inaccessible	BC pilot coverage facts and any NB benefit statements	Reopen official documents and check for superseding decisions; do not substitute historical bulletins or indexed snippets for completed current criteria audit	P0 before affected coverage claims	CLOSED by R2-1: BC Wegovy decision PDF opened (Non-benefit, 2025-12-18); NB SA criteria PDF dated 2026-09-10 recovered.
G05	Factual extraction gap: territorial/NIHB eligibility and product benefits	not researched	Territorial expansion and avoiding blanket coverage assumptions	Extract each applicable plan separately, including DIN and indication criteria; verify alternative service and dispensing routes before territorial expansion	P1 expansion gate	OPEN: territories and NIHB not researched in Round 2 (by design).
G06	Factual evidence gap: unavailable primary-study full texts and methods	source inaccessible	Whether complete effect, withdrawal and missing-data comparisons can be published	Obtain original articles through lawful access; extract estimands, denominators, attrition, uncertainty and funding; omit unchecked numerical comparisons	P0 before affected clinical comparison	OPEN: primary-study full texts not part of Round 2.
G07	Factual evidence gap: broader mental-health, eating-disorder and supplement evidence	not researched	Scope of clinical education beyond the narrow audited launch subset	Conduct a separately scoped current guideline and primary-evidence review if those topics enter publication; keep them outside initial claims meanwhile	P1 expansion gate	OPEN: not in scope of Round 2.
G08	Factual contradiction: Victoza future discontinuation dates	conflicting	Any definitive national discontinuation or switching statement	Reconcile current official supply and manufacturer notices by product and date; do not infer stock or a universal stop date	P0 before affected claim	PARTLY CLOSED by R2-1: Drug Shortages Canada report 240246 (updated 2025-10-06) shows anticipated discontinuation 2027-01-01, actual date blank; no newer notice found. Do not publish a stop date.
G09	Factual contradiction: Maple service-specific province availability	conflicting	Whether a province filter can show Maple as a confirmed match	Resolve which service and plan each official page describes; retain conflict outside overlapping Ontario claim until current public terms or later authorized confirmation settles it	P0 before unconditional match	STILL CONFLICTING per R2-2 (Codex and Claude, 2026-09-17): Maple help page (all except NS, NU, NT) vs obesity landing page (Ontario only) vs assessment page (BC and ON). Publish as dated conflict.
G10	Factual contradiction: Phoenix Wegovy card and comparison-table prices	conflicting	Any lowest-price claim or drug-cost calculation using Phoenix	Recheck exact product, quantity, strength, conditions and both page elements; preserve conflict and omit resolved-price ranking until reconciled	P0 before price claim	STILL CONFLICTING per R2-2 (2026-09-17): Phoenix card $399.99 vs table $339.99 on same page; Raven $399.99. Publish both figures with date; no lowest-price claim.
G11	Factual discovery gap: Jill/Pocketpills weight-service province lists and Hims territory reach	not found in sources checked	Provider matching by province or territory	Check service-specific public terms first; do not infer prescriber reach from pharmacy delivery or all-provinces language; later contact only if separately authorized	P0 before affected match	PARTLY CLOSED by R2-2: Hims says all provinces (territories unnamed); Jill and Pocketpills weight-service province lists still not found in sources checked.
G12	Factual discovery gap: complete care, laboratory, dispensing, supply and cancellation charges	not found in sources checked	Any reconciled treatment total or provider savings comparison	Obtain product/quantity/period-specific fee schedules and distinguish requisition from actual testing, included fees from extras, and promotions from recurring charges; keep subtotals incomplete	P0 before total-cost claim	PARTLY CLOSED by R2-2: care fees, follow-up fees, shipping and fill fees captured per service; lab testing fees and full cancellation terms still incomplete for several. No totals.
G13	Factual verification gap: practitioner licences, pharmacy registrations and some legal operators	not researched	Use of independently verified professional or operator labels	Match named practitioners and pharmacies to current jurisdictional registers and exact entities; current CPSO check remains narrow and does not certify clinic quality	P0 before verified-credential label	PARTLY CLOSED by R2-2: CPSO matches for Broussenko 109143, Dehghani 91114, Ko 112759. OCP pharmacy register not opened (maintenance window); CPSBC not opened; four CPSO candidates unmatched. No verified-credential labels yet.
G14	Factual discovery gap: provider language, continuity and plan-specific cancellation details	not found in sources checked	Language/access filters and accurate service commitments	Complete public service-specific documentation for launch rows; distinguish a French webpage from French clinical availability and a generic refund policy from accepted-plan terms	P1 before affected filter or promise	PARTLY CLOSED by R2-2: cancellation and refund terms captured for Felix, Maple, Phoenix, Jill; clinical-service language still mostly not stated by providers.
G15	Factual contract gap: private affiliate action, payout, attribution, reversals, channels and assignment	not found in sources checked	Whether referral revenue is measurable, permitted and transferable	After an authorized diligence step, obtain the actual proposed MPA/EIO and written answers; calculate no partner-specific revenue until qualifying event and net terms are known	P0 commercialization gate	OPEN: affiliate terms require authorized outreach; not a research item.
G16	Future diligence: owner eligibility and programme acceptance	not researched	Whether this pharmacist-owned publication can participate in a Canadian programme	Assess professional fit first, then apply or seek written eligibility only if separately authorized; keep public application availability distinct from approval	P0 before affiliate implementation	OPEN: owner eligibility requires application; not a research item.
G17	Future interpretation: actual pharmacist ownership, paid layout and payment-event compatibility	not researched	Which publishing, placement, price-display and compensation model can operate	Review concrete prototype, owner licence/employment context, wording, linked destinations and proposed contracts under current HPOA and advertising framework; obtain a written actionable assessment	P0 commercialization gate	PREPARED by R2-5: HPOA bylaws v2026.1 located (D.2 Std 2(d) p202; D.4 s2(2) p207), six layouts and seven compensation types mapped, 12 specialist questions. Needs operative wording and Schedule A.1 check, then lawyer or CPBC practice line.
G18	Future interpretation: actual privacy flows and Quebec/French applicability	not researched	Permissible analytics, vendors, outreach and targeted Quebec expansion	Map real collection and transfers, configure minimal data, assess notices/consent and applicable jurisdictions; assess French obligations before targeted commercial expansion	P0 before affected feature or expansion	PREPARED by R2-5: PIPA, PIPEDA, CASL and Quebec P-39.1 sections mapped to four data flows; Quebec view: English-only BC site less likely caught but not exempt. Needs implementation map before launch.
G19	Future test: reader demand and improvement over existing alternatives	not researched	Whether the proposed asset warrants construction and continued upkeep	Run the planned consented task tests against an incumbent, record failures and time/confidence changes, and interview target readers; no interview or task result exists yet	P0 investment gate	OPEN: reader task test not yet run; prototype build is the prerequisite.
G20	Future test: willingness to pay, pilot purchases and renewals	not researched	Commercial feasibility and choice of buyer/product	Use the concrete proposed brief or placement mockup in later authorized buyer tests; distinguish interest, signed agreement, accepted delivery, cash received and renewal	P0 investment gate	OPEN: buyer test not yet run.
G21	Future measurement: clean consumer-AI baseline and actual audience/conversion evidence	not researched	Whether discovery changes lead to verified citations, visits, actions or revenue	Run fresh unprimed sessions under recorded settings and preserve errors/zero mentions; measure actual site and partner events only after a launch, using distinct denominators	P1 measurement gate	PARTLY STARTED: 12 Gemini consumer captures and 12 work-account Copilot captures on 2026-09-17, repetitions minutes apart. ChatGPT, Claude, Google AI Mode, Perplexity, Grok not run. Not yet a baseline.
G22	Factual access gap: sampled robots responses and community rule pages	source inaccessible	Crawler-control observations and any community distribution proposal	Fetch actual public robots responses through a working permitted method and inspect current community rules; failure is not evidence of crawler exclusion or permission to post	P1 technical/distribution gate	OPEN: robots and community rules not in Round 2 scope.
G23	Future technical verification: remaining rendering, accessibility, indexability and reporting checks	not researched	Readiness of a future public site and reliability of measurement	Test actual launch pages: initial versus rendered HTML, keyboard/screen-reader operation, canonical responses, schema, mobile tables, sitemap, firewall and available Search Console reports; one sampled mobile check is not site-wide assurance	P0 public-launch gate	OPEN: applies to the built site.
G24	Future rights diligence: source reuse, exports, logos and compiled data	not researched	Whether a maintained brief, downloadable dataset or transferable asset can lawfully reuse sources	Record source-specific terms and permitted quotation/reuse; obtain permission where needed, prefer linked factual summaries, and review rights before bulk exports or resale	P0 paid-data/export gate	OPEN: rights register not started; R2-5 notes reuse limits.
G25	Future test: maintenance capacity, real economics and transferable contracts	not researched	Whether the business can sustain updates and survive founder absence	Time a narrow update cycle; reconcile actual review/vendor costs and cash receipts against hours; test handoff and assignment requirements; do not convert assumed budgets or acquisition examples into forecast returns	P0 continuation/scale gate	OPEN: applies to operations after build.