The Spravato® audit question:
could you produce it?

The four things you would be asked for

Strip away the format and the paperwork, and a review of a Spravato® session is four questions about one afternoon. They are worth writing on a wall, because every process decision in a clinic is ultimately a bet about how quickly they can be answered.

• The monitoring record for a given session. What happened in the room, with times: dose, devices, vitals at their checkpoints, observations as they occurred, the discharge assessment.
• The REMS form for it, and proof it arrived. Not the form you have a copy of — evidence that the program received it, and when.
• The drug record for the carton that was used. Which lot, which serial, where it came from, and what happened to the rest of it.
• A log of who did what. Which staff member administered, which observed, who amended a record afterwards and when.

Notice that only the first is about clinical care. The other three are about bookkeeping, and they are the three that clinics routinely cannot answer inside a working day.

Capture and retrieval are different problems

Almost every clinic that struggles here captured the information correctly. The vitals were taken. The lot was written on the carton log. The form was faxed. The failure is not that the facts were never recorded; it is that they were recorded in four places that do not know about each other, so answering one question about one patient means assembling four sources and hoping they agree.

This is why “we keep good records” is not an answer to the audit question. A clipboard in a treatment room, a shared spreadsheet, a fax confirmation page in a drawer, and a paper drug logbook are four good records. Producing one coherent account of a session from them is a research project, and it gets harder every month as the volume grows and the people who remember the details move on.

Where clinics actually come up short

The gaps are remarkably consistent, and none of them is about diligence. Each is a place where the record depended on somebody remembering to make it a second time.

• The exact administration time. Everyone knows roughly when the session started. The record often says “1:00” because that was the appointment, not because that was the first device. A monitoring window can only be proven to have run its length if its start is a recorded fact rather than a scheduled intention.
• Proof of filing, as opposed to filing. Clinics that file diligently frequently cannot show it. A sent fax and a delivered fax are different claims, and only one of them is evidence. Keep the transmission confirmation — destination, time, page count — as a permanent record, not a printout that lives until somebody tidies the desk.
• The forty-minute check. The one with no immediate consequence and therefore the one that slips. Its absence is invisible on the day and obvious in a record months later.
• The lot number chain. The lot on the monitoring form and the lot in the drug ledger are meant to be the same fact. Transcribed by hand into two places, they are two facts that agree by luck, and a single transposed character breaks a chain you would need intact for a recall as much as for a review.
• Waste with nobody behind it. An 84 mg dose reduced to 56 mg leaves a device that is wasted rather than returned. Recorded at the time with a reason and a witness it costs fifteen seconds; reconstructed later it cannot honestly be reconstructed at all.

Why filling a gap afterwards is worse than the gap

There is a strong and understandable temptation, on finding a hole, to complete the record from memory and notes. It is worth being blunt about why that is the wrong instinct. A gap is a documentation failure with a date on it. A gap filled in later, without saying so, is a record that misrepresents when it was made — and if that is ever established, every other record you hold becomes a question rather than an answer.

The defensible move is the boring one: record what is known, record when you recorded it, and record why it is late. An honest late entry is a finding about one session. A record that cannot be trusted is a finding about the clinic. This is also the practical argument for a system where entries are timestamped by something other than the person making them, and where amendments are visible as amendments rather than silently overwriting what was there before.

It is not only the REMS program asking

Esketamine is a Schedule III controlled substance, so the drug record answers to controlled-substance recordkeeping and to your state board as well — often with requirements of their own on top of the federal baseline. Those reviews ask a different question from the REMS program's: not “was this session documented” but “does what you received reconcile with what you used, wasted and still hold?”

A clinic that treats the two as separate filing systems does the work twice and still has the seam between them to defend. A clinic where the dose given and the carton decremented are one action has no seam. That is the whole argument for keeping the drug ledger and the session record in one place, and it is worth settling before the first delivery rather than after the first count that does not balance.

The self-audit, and how to run it

You do not need permission or a consultant to find out where you stand. Take ten completed sessions at random from the last few months — genuinely at random, because choosing them defeats the exercise — and answer these for each one, with documents rather than recollection:

• Was the patient enrolled before this session, provably, rather than around the same time?
• What time was the first device administered, and who administered it?
• Are the pre-dose, forty-minute and pre-discharge vitals all present, with times?
• Did the monitoring window run its full length, and if it ended early, is the reason recorded?
• Which lot and serial was used, and does the drug ledger agree with the monitoring form?
• Was the adverse-event review completed, including an explicit finding of none?
• Was the form filed within seven days, and can you show the delivery confirmation?
• Can you show who recorded each part, and whether anything was amended afterwards?

Score it honestly: eight of eight, from documents, in under five minutes per session. Most clinics running on paper score six or seven and take considerably longer, and the two that fail are almost always the administration time and the delivery confirmation. That result is not a verdict on the staff. It is a verdict on where the record lives, which is a decision somebody made — often by default — before the clinic opened.

What “provable” looks like when the system is built for it

None of the above requires software; a rigorous paper clinic can pass it. What software changes is whether passing depends on rigor being sustained by busy people for years. For the sake of being concrete rather than promotional, here is how Lucido answers those eight questions, because a vendor that cannot be specific about this is not worth evaluating:

• The window is timed from the first device, recorded when it happens, and discharge below two hours is refused by the database unless a reason is entered — so an early end exists as a documented decision rather than a silent one.
• A completed session is immutable once written. The narrow amendments an owner can make are confined to unfiled records and leave a trail; nothing is quietly overwritten.
• Filing is one action, and every transmission keeps a permanent delivery receipt with destination, time and page count. Receipts are deliberately exempt from the retention purge: they are evidence.
• Days remaining are tracked per session, with a reminder before anything slips and a Monday digest of what is still outstanding.
• The lot is captured once at dosing and carries to both the form and the ledger, so the number on the form and the number in the ledger are one number, not two. If the clinic switches on carton scanning, that capture is a scan and pins the exact serial rather than the lot alone.
• Waste is recorded with a reason and, if the clinic switches it on, a second witness. Physical counts that disagree become durable discrepancy records an owner resolves with a note.
• Every PHI access is written to a seven-year append-only audit log that the application itself cannot rewrite.

All of that is included at one flat price per location — there is no compliance tier, because a clinic that needed to buy the evidence separately would be exactly the clinic that could not produce it.

The short version

You cannot control whether you are asked. You can control the answer, and the answer is decided long before the question — by where the record is made, whether it is made once or three times, and whether proof of filing is a document or a memory. Run the ten-session drill. Whatever it tells you is true today whether or not anybody is asking.

Related guides

• Spravato® REMS requirements, explained
• The seven-day filing deadline, and how clinics miss it
• Spravato® inventory tracking, done properly
• The two-hour monitoring window: what it actually requires