Extalvo Navigator
There is no standard taper, because there is no standard patient. Extalvo Cloud designs a protocol for the person in front of you, predicts where that person is likely to struggle, and adapts the schedule as the taper runs — with the prescribing clinician setting, reviewing, and approving every step.
 

Tapering is a data problem in disguise
A prescriber starting a patient on a benzodiazepine has decades of dosing guidance behind them. A prescriber stopping one has almost nothing: no approved product for the indication, no dosing table for the last milligram, and no way to know whether this particular patient is about to have a hard week.
So the taper gets run on instinct, in ten-minute appointments six weeks apart, using whatever the patient remembers of the six weeks in between. Symptoms get attributed to the wrong cause. A patient who was one hold away from finishing is put back on the drug instead.
The irony is that tapering is one of the most information-rich problems in medicine and one of the least instrumented. Every taper produces a dose-by-dose record of exposure and a day-by-day record of how one nervous system responded to it. Almost none of it has ever been captured.
Extalvo Navigator exists to replace guesswork with observation — and then to turn that observation into a protocol built for one patient.
 

Personalized medicine, applied to stopping
Personalized medicine has usually meant matching the right drug to the right patient. In tapering the drug is already chosen. The variable that decides whether a patient succeeds or fails is not which molecule — it is the trajectory: how fast, in what size steps, with which pauses, and through which weeks.
That trajectory is where the variation lives. Two patients on the same 20 mg of diazepam, for the same number of years, can need schedules that differ by months. Half-life and receptor pharmacology, length of exposure, what happened the last time they tried to stop, age, concomitant medications, and their own sensitivity to each reduction all move the answer. A schedule built for the average of those two patients fits neither of them.
So the unit of personalization here is the taper schedule itself. Extalvo Navigator treats every schedule as a hypothesis about one patient — generated from what is known about them at the outset, tested against what they actually report as they go, and revised when the two diverge.

 

There is no average patient at 0.4 mg. The taper that works is the one built for the person taking it.

 

What the protocol is personalized on
The engine assembles a schedule from what is known about the drug, the patient, and — once the taper is running — the patient’s own response to it.



 

The taper engine

The engine is built on retrospective outcome data from real tapers — which patients came off which drugs, at what rate, with what symptoms, and where they failed. From that it produces an individualized schedule for the patient in front of you and, more usefully, a prediction of where that schedule is going to get hard.

It is worth being precise about what this is not. It is not therapeutic drug monitoring and it is not biomarker titration. The signal is the symptom course, not the plasma level. Withdrawal severity is what patients feel, and what makes them stop; that is what the engine is trained to anticipate.

 

What the clinician sees:

• A predicted severity course for this patient across the whole taper, plotted against the observed course of a matched cohort of patients who came off the same drug from a comparable starting dose.

• Symptom hotspots — the weeks in which patients like this one have historically run into trouble.

• Failure windows — the periods in which tapers of this shape have historically broken down, flagged before the patient reaches them, so a step can be slowed or held in advance rather than reversed afterwards.

• The patient’s actual course against the prediction, updated as symptom scores and dispensing data come in.

• The reasoning behind each recommendation — which factors drove it and which comparison population it is drawn from — so the clinician can weigh the advice rather than take it on trust.

 

The clinical value is not the schedule. It is knowing which week is going to be the bad one, before it arrives.

 

A protocol that adapts

A taper designed on day one and never revisited is still a guess — a better-informed one, but a guess. The point of running a taper on an instrumented platform is that the plan can answer to the patient.

Every dispense from the device and every symptom check-in flows back in. While the patient tracks the prediction, the schedule holds. When they run harder than predicted, the engine says so — typically weeks before it would surface at the next appointment — and proposes the specific adjustment: hold at the current rung, halve the next reduction, lengthen the interval, or step back up for a fortnight.

Those are precisely the maneuvers that rescue a taper, and precisely the ones that rarely happen today, because each of them means a phone call, a new prescription, and a trip to the pharmacy. Here each is a change to a schedule, approved by the clinician and on the patient’s device the same day.

The loop is deliberately short. The cost of noticing late is a patient who stops, restarts, and stays on the medication for another two years.
 

Protocol builder

Hyperbolic or linear. Percentage per step, interval between steps, hold rules, and the number of steps down to zero — all set by the prescriber rather than dictated by what the packaging allows.

The engine’s proposal is a starting point, not an answer. Every parameter is the clinician’s to override, and the basis for every suggestion is on the screen next to it. Change one parameter and the whole schedule — along with the prediction that goes with it — is recomputed.

The result is a complete protocol, every step from the current dose to zero on a single page, that goes straight to the patient’s device. A 20 mg diazepam taper might render as twenty-four steps over roughly forty-six weeks.
 

Withdrawal, or relapse

This is the question the field most often gets wrong, and it is the question the platform is best positioned to answer.

Withdrawal and relapse are genuinely hard to tell apart in a ten-minute appointment. They are much easier to tell apart when you can see the shape of the symptom course and its timing relative to each dose reduction. A symptom that appears a few days after a step and settles within a fortnight behaves like withdrawal. One that emerges independently of the schedule and does not settle behaves like something else.

Personalization sharpens the distinction. The comparison is not only against the general pattern for the drug but against this patient’s own predicted course — so the question becomes whether what they are reporting is what their taper was expected to produce, or something that arrived on its own schedule.

Getting this right is not an academic matter. It determines whether a patient is held at the current dose for two weeks or put back on the medication for another two years.
 

The clinician prescribes. The software informs.

Extalvo Navigator is designed as clinical decision support, and every design decision follows from that.

• The clinician sets, reviews, and approves every protocol and every change to one. Nothing reaches the patient’s device without that approval.

• Every recommendation shows its basis — the factors that produced it and the comparison population behind it — so the prescriber can evaluate it independently rather than defer to it.

• The software does not diagnose, does not initiate therapy, and never changes a dose on its own authority.

• Recommendations stay inside the boundaries the prescriber authored and the concentrations the formulation actually supports.

That is the difference between a tool that makes an expert faster and one that asks a clinician to trust a black box. The prescribers who take on tapering patients are, as a rule, exactly the people who will not accept the second.

The name is the job description. A navigator computes the course; the captain decides whether to take it, and the captain is the one who signs.

Everyone on the same record

A taper runs for months and involves more people than the prescriber. Extalvo Navigator gives each of them the view they need and nothing they do not.

Every completed taper improves the next one

Each taper that runs on the platform produces something the field has never had at scale: a dose-by-dose record of what was taken, when it was taken, and what the patient felt. De-identified, that record goes back into the engine, and the system’s estimate of where the hard weeks fall improves with every patient who finishes.

This is what makes the personalization compound. The first generation of protocols is built on retrospective data from tapers that were never designed to be measured. Every taper run on the platform is measured by design — which means the answer to “what should this patient’s schedule be” is a better answer each year than it was the year before.

Over time this is the asset. A formulation can be copied. A dispensing and outcome record covering hundreds of thousands of real tapers cannot.

Extalvo’s products are investigational and have not been approved by the FDA. Nothing on this page is medical advice. Tapering should be undertaken only under the supervision of a prescribing clinician, and benzodiazepines and antidepressants should never be stopped abruptly.