Safety Plan & Risk Documentation Templates

Who this page is for. Clinicians who already do this work and have to write it down: therapists, counselors, social workers, psychologists and prescribers. What follows is documentation structure, not clinical guidance. Nothing in this pack assesses risk, works out a risk level, or stands in for a validated instrument, your training or your clinical judgment.

If you are here because you or someone close to you is in crisis, call or text 988 (the US Suicide and Crisis Lifeline) at any time, or call 911.

The /risk template covers the elements clinicians know from SAMHSA's SAFE-T, in this order: the suicide inquiry, risk factors and protective factors, the risk level with its rationale, then the actions taken and the reassessment.

The instruments stay with their owners. A validated suicide risk assessment tool such as the ASQ or the C-SSRS (Columbia Lighthouse Project, cssrs.columbia.edu) structures the screening, and the suicide risk assessment questions you ask come from it and from your own interview. This pack names screeners and records that you used them. It does not reproduce them. The safety plan follows the same six-step outline clinicians know from Stanley-Brown safety planning. It is not the official Stanley-Brown form, which is at suicidesafetyplan.com, and it is no substitute where your agency requires that form.

• Risk factors and protective factors, both. Write down what raises risk and what holds it down. A later reader who needs to follow your decision then sees both sides of what you weighed.
• The reasoning, in your words. Inside /risk the rationale is an empty field whose prompt asks you to tie the level to the findings: what is present, what is absent, and what changed since the last assessment.
• The client's words, never filled in ahead of time. In /safetyplan every client field opens as an empty prompt in brackets and is typed in the client's own words during the session, so no two plans read alike.
• Defaults you can see, and have to check. Several menus open already set to the snippet's shipped defaults, which is the state the pictures on this page show.
  • /risk opens on routine reassessment, denies ideation, no prior attempts or self-harm, and homicidal ideation denied. Two risk factors are selected (hopelessness, a recent loss or major stressor) and two protective factors are selected (engagement in treatment, supportive relationships). The level is set to low. Two actions taken today are selected (safety plan reviewed and updated with the client, crisis resources provided), and reassessment is set to the next session.
  • /sinote opens on the denial, self-harm denied, and "the safety plan remains in place". For a client who has no safety plan, switch that menu to its other option, "no safety plan is indicated at this time".
  • /recordsummary opens on no current ideation, no history of attempts, low, a safety plan in place, and two items that arrive already worded: the "based on" field, and a means-safety option stating there are no firearms in the home.
  Every one of these is yours to set to what you asked, heard and did. A default you did not check is a false record, and so is a pre-selected action you did not take. Read every line before you insert.
• A denial worded as something you asked about. "Client was asked about suicidal and homicidal ideation this session and denies" says what happened in the room. A bare "No SI" does not.
• Screeners named, never reproduced. The screener note records which instrument you gave, how, the result and what you did with it. A PHQ-9 or GAD-7 total is placed in the instrument's published severity range, which you check against the scoring sheet. C-SSRS and ASQ results are recorded in the screener's own categories.
• Editable output, with unfinished lines flagged. Everything lands as text in whatever field you are in, so you can change any word before you sign. No field is required. A prompt you leave unfilled lands in its brackets, which marks the line as unfinished. Look for brackets before you sign or print.
• Where you type, not on paper. A suicide safety plan PDF gives you the steps on a page. A template you insert with a shortcut puts the same structure in your EHR or a document, and what you insert can be edited, printed for the client and saved to the chart.

If your EHR has a suicide risk assessment form or a safety plan that works for you, keep it. This pack is for the places it does not reach: a second system at a group practice, a document you print for the client, a records request, a portal message, and the risk line inside any progress note template, including your EHR's.

If your EHR's form has one field per section, use the shorter snippets field by field: /protective in the protective factors field, /means under means safety, /safetyreviewed where the plan review is recorded, and /nosi or /sinote in the progress note's risk line. Use /risk whole where you have one free-text field.

If you use an AI scribe, the two work together. The scribe drafts the session narrative. The risk line, the rationale and the safety plan are parts most clinicians word themselves, and the template gives those words the same place in every note.

Inside the /safetyplan pop-up

• Header: client, date, your name and your phone number, with your details saved in the template.
• Steps 1 and 2: warning signs, and things the client can do on their own. Three lines each.
• Steps 3 and 4: people and places that help the client feel better (two names with numbers, two places), and people they can ask for help in a crisis (three names with numbers).
• Step 5: professionals and services. Your name and number fill in from the header, followed by your after-hours instruction, a second professional, the nearest emergency department, and a fixed line with 988, the Crisis Text Line and 911.
• Step 6: making the environment safer, with one line for medications and one for firearms or other means. A last field holds the one thing the client names as most important to them.
• For the chart (a toggle, on when the pop-up opens): a paragraph recording that the plan was developed together, the date, which copy the client left with, that means safety was discussed, and when the plan will be reviewed. Apart from two menus, its sentences are fixed text: edit "all six steps completed" when a step was left open, and the means-safety sentence when step 6 did not apply.

Every client field opens as an empty prompt in brackets, such as "[a second coping strategy]". When step 6 has nothing to secure, type that in the step, for example "No firearms in the home, confirmed with client", and change the chart sentence to say that means safety was discussed and no means of concern were identified. A guidance note shows while you write and is not inserted. On a video session, share your screen with the document open and fill the plan in together.

Where your agency requires the official Stanley-Brown form, use that. The chart paragraph and the shorter snippets below still apply.

Filled in for a fictional client, one step and the chart paragraph read:

Step 2 - Things I can do on my own to take my mind off things, without contacting anyone:
1. Walk the dog around the block with music on
2. Cold shower, then the 4-7-8 breathing from session
3. Work on the model plane in the garage

For the chart: Safety plan developed collaboratively with Jordan on September 15, 2026, all six steps completed in the client's own words. Client left with a paper copy and a photo of the plan on their phone. Lethal means safety was discussed and the arrangements in step 6 were agreed to by the client. Plan to be reviewed at every session until risk is low.

The step headings, your contact lines, the crisis-line block and the chart paragraph were already typed. The three coping strategies are Jordan's words, typed during the session with Jordan.

Inside the /risk pop-up

• Header: date, your name, and a menu for what prompted the assessment (routine reassessment, intake, a positive screener result, a disclosure in session and others).
• The inquiry: an ideation menu. Pick an option that reports ideation and the detail lines open: frequency and intensity, how long, worst ever, plan, access to means, intent and preparatory behavior. The attempts and self-harm menu opens a detail field when there is a history to describe. The homicidal ideation menu does the same when ideation is present.
• Risk factors and protective factors: two menus you can pick several from.
• Level and rationale: Clinical judgment of acute risk is a menu of low, moderate and high that you set. Nothing in the snippet scores it or changes it. The rationale is an empty paragraph field.
• Actions and follow-up: a menu of actions taken today that you can pick several from, and a menu for when risk will be reassessed.

The pop-up opens on the shipped defaults listed above, the actions and the reassessment included. Set every menu before you insert.

Here is most of the inserted text for a fictional client who reports ideation, with the detail lines open:

SAFETY ASSESSMENT - September 15, 2026
Assessed by: Alice Smith, LPC. Prompted by: a positive screener result.

Suicidal ideation: Client reports active ideation without a plan or intent, described below.
Plan: no specific plan. Access to means: no access to the means considered. Intent: client denies any intent to act and names reasons for living. Preparatory behavior or rehearsal: none.
Suicide attempts and self-harm history: prior self-harm without suicidal intent, most recently more than a year ago.
Homicidal ideation: denied.

Risk factors present: hopelessness, a recent loss or major stressor, history of self-harm, alcohol or substance use, insomnia.
Protective factors present: engagement in treatment, supportive relationships, responsibility to children or pets.

Clinical judgment of acute risk: moderate. Rationale: Active ideation daily for about ten days, new since the separation, alongside poor sleep, increased drinking and a history of self-harm. No plan, no intent, no preparatory behavior, no prior attempts and no firearm in the home. Client took an active part in safety planning, named reasons for living and agreed to a phone check-in tomorrow. Outpatient care with closer follow-up is judged appropriate. A higher level of care was discussed in consultation and will be revisited if a plan or intent emerges.
Risk will be reassessed within 24 hours by phone.

Two lines are left out of the text above. The frequency and intensity line recorded daily thoughts of moderate intensity, present for about ten days, with the worst ever comparable to the current episode. The actions taken today line listed four actions: crisis resources provided (988 and local options), safety plan created with the client, session frequency increased, and consultation with a supervisor or colleague.

The labels, the menu wording and the factor names came from the template. The level and the rationale are a fictional clinician's wording for a fictional client. They show where your reasoning goes. They are not a model rationale.

The other four templates

Nothing carries over between templates: the client's name, the level and your own details are typed or set in each one.

• /sinote, the risk line for a routine progress note. Pick what the client endorsed today and the matching follow-up sentence opens. On the denial, a last menu says whether a safety plan remains in place or none is indicated. Anything beyond passive ideation points to a full assessment instead of summarizing one.
• /screenernote, the screener administration note: instrument (PHQ-9, GAD-7, C-SSRS Screener, ASQ), how it was given, the result, whether it was reviewed with the client, how it compares with last time, and the action taken.
• /screenfu, a same-day check-in message after a concerning screener result, worded for a text message, a voicemail or an email. It never names the item or the score, and it always carries 988 and 911. A live disclosure of intent is a phone call, not a message. Use only a channel the client has consented to, and choose the voicemail wording with whoever else may hear it in mind.
• /recordsummary, a risk summary paragraph for a records request or a coordination-of-care call, written after the release is on file: treatment context, the most recent assessment, the level you judged, safety plan and means status, and how to reach you. The "signed release on file" clause is fixed text. Edit or delete it when you are disclosing on another basis, such as an emergency call from an emergency department. Replace the pre-worded "based on" and means-safety wording with this client's facts.

Shorter snippets for sentences you type often

• /nosi, the No SI line. It inserts exactly this: "Client was asked about suicidal and homicidal ideation this session and denies current suicidal ideation, intent, or plan, and denies homicidal ideation. No self-harm reported since the last session. No change in risk level, and no additional safety measures are indicated at this time." It is the right line only when you asked and that was the answer, with no self-harm reported. For anything else, use /sinote or /risk.
• /reassessed, the follow-up risk line for the session after an assessment.
• /protective, a protective factors sentence. It ships with example factors. Replace them with this client's.
• /rationale, a separate, optional starter paragraph for a low-risk judgment. It ships with wording already in it, so rewrite it to this client's findings or do not use it. It is not connected to the empty rationale field inside /risk.
• /means, the means safety line, with two menus for what was arranged.
• /safetyreviewed, the line recording that the safety plan was reviewed in session.
• /outreach, the chart line for an outreach attempt after a positive screener.
• /crisisline, crisis resources for a message to a client: 988, the Crisis Text Line, 911 and your own number.

The plain-text lines describe a typical session. Edit them whenever this session was not typical. A line that does not match what happened is worse than no line. The same structure each time, with this client's findings in every field, is what lets a supervisor, a covering clinician or a later reader follow what you did.