Methodology
How the Patient Score works
Every provider gets the same formula: six dimensions, fixed weights, inputs we track and re-verify. No judgment calls per provider, no exceptions, and nothing a provider can buy. This page is the whole formula.
Six dimensions, fixed weights
ValueTreatment + clinical modelUser experienceTrust + safetyAccessibilitySupport
Value 25%
What you actually pay for what you actually get.
- Advertised monthly price vs the category median, as a percentile
- What the price includes: medication, visits, labs, coaching
- Renewal price vs intro price: surprise jumps cost points
- Public pricing: quiz-gated or hidden pricing is penalised
Treatment + clinical model 25%
The breadth and rigor of care on offer, not outcomes we cannot measure.
- Medication options: brand, compounded, oral alternatives
- Titration protocol and scheduled follow-up cadence
- Lab work: required, offered, or absent
- Prescriber type and clinical oversight disclosed
User experience 15%
How the platform behaves when you actually use it.
- Intake length and effort to reach a clinician
- Platform features: messaging, dose tracking, refill requests
- Member ease-of-use sub-ratings, once a provider has 5+
Trust + safety 15%
Verifiable signals from public records, not vibes.
- LegitScript certification status
- FDA warning-letter database, checked monthly
- Pharmacy partners and prescribing clinicians disclosed by name
- Refund and cancellation terms published and clear
Accessibility 10%
Whether you can actually get treated, where you are, soon.
- State coverage: how many states the program is licensed in
- Wait to first appointment
- Insurance billing: visits, medication, or neither
Support 10%
What happens after the first prescription.
- Support channels and stated response commitments
- Refill turnaround reliability
- Member support sub-ratings, once a provider has 5+
A worked example
Illustrative figures, not a real provider. This shows how the six dimension scores combine into one Patient Score once they are weighted. Live scores appear on a provider’s own page as soon as they are computed.
Value8.8× 25%2.20
Treatment + clinical9.4× 25%2.35
User experience9.2× 15%1.38
Trust + safety9.3× 15%1.40
Accessibility9× 10%0.90
Support8.9× 10%0.89
Patient Score9.1
Reading the scale
06.57.58.510
8.5+Exceptional. Strong on price, clinical model and track record at once. Rare.
7.5 - 8.4Recommended. Solid across the board, usually with one soft spot the review names.
6.5 - 7.4Situational. Right for a specific person or budget; read the watch-outs first.
Below 6.5Caution. Listed for completeness, with the reasons spelled out in the review.
What never moves the score
Money. Affiliate relationships, Featured placements and claimed profiles exist, and none of them touch any input above.
Provider pressure. Corrections of factual errors are welcome and fast-tracked; score complaints change nothing without changed facts.
Single reviews. Member sub-ratings only feed dimensions once a provider has five or more, so one angry Tuesday cannot move a score.
Our own opinions. The editorial review interprets the score; it does not set it. The formula runs on tracked inputs.
When scores change
Prices are re-verified against every provider’s live checkout monthly, and Value recomputes when they move. FDA and LegitScript status changes recompute Trust the day we catch them. Member sub-ratings roll into their dimensions continuously once a provider passes five reviews. And two numbers stay deliberately separate: the Patient Score above is the formula; the star rating next to it is the raw average of moderated member reviews, untouched by us.
What this method cannot tell you
We do not measure clinical outcomes: no comparison site can, honestly. Treatment + clinical model scores the visible rigor of a program, which correlates with good care but does not guarantee it. Member sub-ratings are self-reported and skew toward people motivated to write. And newer providers carry wider uncertainty until review volume builds. Where the method is weak, the written review says so.