ClinicalLive · in the productClinical record

TrialOfOne

A randomised A/B trial for one patient, with its uncertainty stated.

Get TrialOfOne Open the appAvailable nowHow pricing works →

What it is

The clinician names two treatments, the outcome, the length of a period and the number of pairs; the order inside each pair is drawn at random when the trial is saved and the database freezes the design. The outcome is either a daily score the patient logs on their portal link (which also tells them which treatment this period is) or the portal’s own “How are you doing today?”, read from the rows it already writes. The result is computed on every read: the mean of each period, the mean on each treatment, the paired B − A difference and an exact randomisation-test p-value over the pairs, printed beside the smallest p-value the design could ever give and a caution in words — one patient’s pattern, not proof.

Who uses it: Built for: Neurology · Psychiatry
Reads: The patient’s daily scores, the portal feeling check
Writes: Trials and their drawn order, diary scores
Never: It does not prescribe, and it does not call a treatment effective: the p-value is shown with its floor (0.25 with three pairs) and the caution that a trend over time, carry-over or chance are not ruled out.
Access: Only people whose role includes it — the clinic owner decides who
Your data: Full PHI. Hosted in Boston, United States. On uninstall: Every record stays; only the workflow leaves.

How it works

Design the trialOn N-of-1 Trials, find the patient, write the question, name treatments A and B, choose where the outcome comes from — a daily score on a scale you set, or the portal’s “How are you doing today?” — and set the days per period, the pairs and the start date.
Let the order be drawnPress Draw the order and start: each pair’s order (A then B, or B then A) is drawn at random and cannot be changed afterwards. The screen tells you the smallest p-value that many pairs can ever give.
Collect the scoresThe patient sees on their link which treatment this period is and taps today’s score; a feeling-check trial reads their daily answer instead. You can also Record score for a day here.
Read the result with its cautionOpen the trial: per-period means, the mean on each treatment, the paired B − A difference and the randomisation p-value, always printed with “one patient’s pattern, not proof”. Mark completed, or Stop early with a reason.

Rules it keeps

What it does not do, by design. A listing with only benefits is an advertisement.

The drawn order and the dates cannot be changed after the trial is created.

Nothing about the result is stored.

AI in TrialOfOne

AskQuestions about what is on screen — “What needs attention in TrialOfOne today?” — answered with the records it read.

DraftSummaries and notes for the team, marked as drafts until a person signs.

Every answer shows its sources. Nothing AI writes is saved until a person accepts it.

See it on screen

TrialOfOne — N-of-1 Trials — Test two treatments for one patient in random order, and read the result honestly — 01N-of-1 Trials — Test two treatments for one patient in random order, and read the result honestlyCaptured from the running demo clinic

01N-of-1 Trials — Test two treatments for one patient in random order, and read the result honestlyCaptured from the running demo clinic

1 of 1

Connected to

Read out of the code, not the brochure: each app below is here because one service queries the other’s tables. Install either side and the hand-over is already wired.

Reads from · 2

Patient Registry RemindRx

See the whole clinic →

Questions clinics ask

What does TrialOfOne do?

A randomised A/B trial for one patient, with its uncertainty stated. The clinician names two treatments, the outcome, the length of a period and the number of pairs; the order inside each pair is drawn at random when the trial is saved and the database freezes the design. The outcome is either a daily score the patient logs on their portal link (which also tells them which treatment this period is) or the portal’s own “How are you doing today?”, read from the rows it already writes. The result is computed on every read: the mean of each period, the mean on each treatment, the paired B − A difference and an exact randomisation-test p-value over the pairs, printed beside the smallest p-value the design could ever give and a caution in words — one patient’s pattern, not proof.

What data does TrialOfOne read and write?

TrialOfOne reads the patient’s daily scores, the portal feeling check and writes trials and their drawn order, diary scores. Its data class is Full PHI. Every app works on the same patient record — nothing is copied into a silo.

What does TrialOfOne deliberately not do?

It does not prescribe, and it does not call a treatment effective: the p-value is shown with its floor (0.25 with three pairs) and the caution that a trend over time, carry-over or chance are not ruled out. The drawn order and the dates cannot be changed after the trial is created. Nothing about the result is stored.

What happens to our data if we uninstall TrialOfOne?

Every record stays; only the workflow leaves.

Is TrialOfOne available today?

Yes — TrialOfOne is live in clinics today. Set up your practice on MedAppz and install it from your dashboard.

Want to see TrialOfOne running on your own patients?

Tell us what you run and we will show you this app inside a practice shaped like yours — or answer the question the page above did not.

More in Clinical

All 437 →

Featured inPsychiatry Neurology