How to document de-escalation training competency for a survey

Last reviewed 12 September 2026. Citations verified against primary sources on that date.

Two beliefs are very common in behavioral health and disability services, and both are wrong often enough to cause a deficiency: that de-escalation training must happen annually, and that a sign-in sheet documents it. Neither holds up against the actual regulatory text.

This page sets out what each authority requires, in its own words, with the citation. It is written for the person who has to produce the file when a surveyor asks — a director of nursing, a quality or compliance manager, a program director, a staff development coordinator.

It is a reference, not legal advice. Standards get renumbered and states amend their statutes; check anything load-bearing against your own accreditation manual edition and your own state before you rely on it.

The short answer

1. "Annual" is required by accreditors and by some states — but by no CMS Condition of Participation. CMS says "periodic," on an interval your own policy defines. That changes what the surveyor asks you.
2. For psychiatric residential treatment facilities, annual is not enough. Federal law requires competency to be demonstrated semiannually.
3. The regulations demand demonstrated competency, not attendance. The operative verb throughout federal law is "demonstrate." A sign-in sheet cannot satisfy it.
4. A vendor completion certificate is not competency documentation. The record has to live in your personnel files and reflect your population.

Does anything actually require annual de-escalation training?

It depends entirely on who is looking at you — and the two most demanding federal rules do not say annual.

Where annual genuinely is the rule

Where annual is not the rule

Attendance is not competency

This is the distinction that costs providers citations, and the regulatory language is not ambiguous about it. Federal law never says "attended" or "completed." It says demonstrate.

"Staff must be trained and able to demonstrate competency in the application of restraints, implementation of seclusion, monitoring, assessment, and providing care for a patient in restraint or seclusion…"

"The hospital must document in the staff personnel records that the training and demonstration of competency were successfully completed."

"Staff training must include training exercises in which staff members successfully demonstrate in practice the techniques they have learned for managing emergency safety situations."

"The facility must document in the staff personnel records that the training and demonstration of competency were successfully completed. Documentation must include the date training was completed and the name of persons certifying the completion of training."

Read those together and the gap becomes concrete. A sign-in sheet proves exposure to content. It proves nothing about ability, and it carries no evaluator. Filed in a personnel record under 482.13(f)(4), it is on its face half the required document: it evidences training, not demonstration of competency.

The psychiatric residential treatment rule makes the missing half explicit by naming the two extra data elements — the date, and the name of the person certifying. That is a named evaluator attesting to observed performance, which a sign-in sheet structurally cannot supply. CMS surveyor guidance closes it: "Trainer observations of these exercises must be documented."

In ICF/IID settings the emphasis moves further still, away from paper altogether:

"The primary evidence of an effective staff training program is the observed competent interaction between staff and clients."

What a compliant competency record contains

Assembled from 42 CFR 482.13(f), 42 CFR 483.376(d)–(g) and the CMS survey procedures, a defensible record carries these eight elements. Sign-in sheets typically have the first two.

One more point worth making internally: 482.13(f)(2) requires demonstrated knowledge "based on the specific needs of the patient population." A certificate from a national curriculum evidences that a vendor trained the person. It does not evidence that your employer verified competency on your population. Certificates are the beginning of the file, not the file.

What surveyors actually ask for

These are the published survey procedures, quoted from the surveyor manuals rather than paraphrased.

CMS hospitals — the controlling procedure is Tag A-0208:

"Review a sample of staff personnel records, including contract or agency staff, to determine if the training and demonstration of competency have been completed during orientation and on a periodic basis consistent with hospital policy."

Tag A-0202 adds a second route that has nothing to do with your binder: surveyors are told to "review hospital data (i.e., incident reports, patient injury or death reports) to identify any patterns of patient injuries that may indicate that staff are not adequately trained." And repeatedly across A-0200 through A-0205, the instruction is to "interview staff to determine if they are able to demonstrate the abilities addressed in this requirement." They test the people, not the paperwork.

Psychiatric residential treatment — Tag N0222: "Review a sample of staff personnel files to verify that staff has demonstrated their competence on a six month basis." Tag N0224 adds that training documentation "should be easily accessible and must be current."

ICF/IID — Tag W193 describes what failure looks like: "Inadequate training is evident when staff do not correctly implement behavioral programs, use inappropriate management techniques, cannot explain what intervention is to be used and how it is to be implemented."

CARF — the survey preparation workbook asks, in writing and twice: "Where is the training provided to personnel at orientation documented?" and the same question for the annual training.

OSHA — under the workplace violence enforcement directive, inspectors request the written prevention plan, injury and illness records from five years prior, and "training records… and the schedule for the training that is being provided."

What OSHA does and does not require

There is no OSHA standard for workplace violence. Enforcement runs through the General Duty Clause, section 5(a)(1) of the OSH Act. OSHA says so itself:

"This guidance document is advisory in nature and informational in content. It is not a standard or regulation, and it neither creates new legal obligations nor alters existing obligations created by OSHA standards or the Occupational Safety and Health Act of 1970."

Within that advisory guidance, OSHA recommends initial orientation before duties and annual training thereafter, and notes that "in high-risk settings and institutions, refresher training may be needed more frequently, perhaps monthly or quarterly."

That last clause is worth sitting with, because OSHA's own review of the evidence points the same way. Examining an emergency department study, OSHA observed that the reduction in security codes "was correlated with the number of staff who had been recently trained (in the past 90 days), implying a temporary effect… and suggesting that more frequent training is needed."

Six things vendors commonly overstate

1. "OSHA requires annual workplace violence training." It does not. Only the General Duty Clause is enforceable, and it cites hazards, not missing training.
2. "California SB 553 applies to your clinic or behavioral health agency." Labor Code 6401.9 expressly exempts facilities covered by 8 CCR 3342. Healthcare follows 3342 instead.
3. "New York requires workplace violence training for healthcare employers." Labor Law 27-b covers public employers only. Private hospitals and private nonprofit agencies are outside it.
4. "CMS requires annual restraint and de-escalation training." The text says "periodic basis consistent with hospital policy," and CMS guidance expressly grants flexibility.
5. "Annual is enough for our PRTF." 42 CFR 483.376(f) requires semiannual competency demonstration.
6. "An OSHA healthcare workplace violence standard is coming." The rulemaking has not reached a proposed rule.

Why the interval matters more than the binder

Every authority above is asking a version of the same question: can this person do it now? An annual class answers that question once and then stops answering it. Nothing in the file speaks to month seven.

That is not a rhetorical point — it is what the evidence shows. RAND Europe followed staff after training of this kind and found their confidence back at baseline within six months. No one has demonstrated that the skills outlast it either. Between the class and the incident, most programs measure neither.

Which is why we built the thing we sell the way we did: one training day, then a scored competency check every thirty days, each one dated, named, and exportable. Twelve documented checkpoints a year instead of one, without pulling anybody off the floor to get them.

That is the pitch, and you should weigh it as a pitch. The reference above stands on its own regardless of who you train with.