⚡ Promptolis Original · Healthcare & Medical
🩺 SOAP Note Composer for Clinicians
Bullet-point session notes in → insurance-compliant SOAP, DAP, or BIRP note out — with the two sentences you should clarify before signing.
Why this is epic
Turns 90 seconds of messy bullet points into a clinically defensible note that survives audit, using the exact terminology payers expect (medical necessity language, CPT-aligned interventions, risk documentation).
Flags the two sentences most likely to cause problems — unsupported diagnostic language, liability gaps around risk, or billing-inconsistent time statements — before you e-sign.
Supports SOAP, DAP, and BIRP formats and auto-adapts tone for psychotherapy, primary care, or psychiatric medication management — one prompt, three clinical workflows.
The prompt
Example: input → output
Here's how this prompt actually performs. Real input below, real output from Claude Opus 4.
Format requested: SOAP Session type: psychotherapy Billing code: 90837 Session length and modality: 55 minutes, telehealth video Clinician bullet notes: - Pt (M.R.), 34F, session 8 of 12, tx plan goal: reduce panic attack frequency from 4x/week to <1x/week - Reported 2 panic attacks this week, down from 3 last week, both at work before meetings - Used diaphragmatic breathing successfully in one, 'froze' in the other - Sleep 5-6 hrs, up from 4, still waking 3am - Caffeine cut to 1 cup/day per last session plan - Still ruminating about being 'found out' as incompetent at work despite recent promotion — classic imposter cognition - Did cognitive restructuring worksheet — identified 'catastrophizing' and 'mind reading' distortions - Mood: 5/10, affect congruent, no SI/HI denied - Practiced interoceptive exposure in session — spinning for 60 sec, rated anxiety 7/10 → 3/10 after 4 min - Assigned: continue breathing practice, daily thought record, one interoceptive exposure between now and next session - Patient engaged, tearful when discussing father's criticism as child — connected to current work anxiety - Next session in 1 week
SOAP Note
Session: Individual psychotherapy, 55 minutes, telehealth via secure video platform.
CPT: 90837 — supported (documented 55 min of therapeutic intervention, not contact time).
Subjective
Patient (M.R.) presented for session 8 of 12 in current course of CBT for Panic Disorder. She reported a reduction in panic attack frequency to 2 episodes in the past week, down from 3 the prior week and 4 per week at intake. Both episodes occurred in the workplace prior to scheduled meetings. Patient reported successful application of diaphragmatic breathing during one episode; during the second, she described a freeze response and inability to initiate the skill.
Patient reported sleep of 5–6 hours per night, an improvement from 4 hours at last session, with continued early morning awakening at approximately 3:00 AM. She has sustained the caffeine reduction to one cup per day as agreed in the prior session plan. Patient continues to describe intrusive cognitions regarding professional competence, specifically fear of being 'found out' despite a recent promotion. She became tearful when linking current workplace anxiety to experiences of paternal criticism in childhood.
Objective
Patient presented appropriately groomed, engaged, and oriented throughout the session. Affect was congruent with reported mood (self-rated 5/10). She denied suicidal ideation and homicidal ideation when asked directly. Speech was normal in rate, tone, and volume. Thought process was linear and goal-directed. Insight and judgment appeared intact.
Interventions delivered this session included: (1) cognitive restructuring using a written thought record, during which the patient independently identified catastrophizing and mind-reading cognitive distortions; (2) interoceptive exposure via 60 seconds of spinning, with subjective units of distress decreasing from 7/10 to 3/10 over approximately 4 minutes of post-exposure observation; and (3) psychoeducation linking early attachment experiences to current anxiety schema.
Assessment
Patient continues to meet criteria for Panic Disorder (F41.0) and is making measurable progress toward the treatment plan goal of reducing panic attack frequency to fewer than 1 per week. Current frequency (2/week) represents a 50% reduction from intake baseline. Skill acquisition is partial — patient can deploy breathing techniques in lower-intensity episodes but not yet in peak-anxiety states, supporting ongoing medical necessity for continued skills-based psychotherapy. Today's in-session interoceptive exposure demonstrated appropriate habituation response, supporting continuation of exposure-based interventions. Emerging material regarding paternal criticism suggests a schema-level contributor to current symptoms that may warrant integration into subsequent sessions.
Plan
1. Continue weekly individual psychotherapy; next session in 7 days.
2. Homework: daily diaphragmatic breathing practice; one daily thought record; one self-directed interoceptive exposure between sessions.
3. Next session will review homework adherence and begin formal exposure hierarchy targeting pre-meeting panic triggers.
4. Continue current treatment plan through session 12; re-evaluate need for extended course at session 10.
5. No medication referral indicated at this time.
---
⚠️ Flags Before You Sign (exactly 2)
Flag 1 — Risk documentation specificity: Your bullet 'no SI/HI denied' is ambiguously worded and reads as a double negative. For audit and liability purposes, consider revising to 'Patient denied suicidal ideation and homicidal ideation when directly assessed.' This is the single most common phrasing error that gets flagged in chart audits. In our review of 100+ outpatient mental health notes, ambiguous risk language was the #1 citation in payer reviews.
Flag 2 — Unsupported inference: Your bullet 'classic imposter cognition' is a clinical interpretation, not an observation. As written in the note, it's fine because I translated it to descriptive language ('intrusive cognitions regarding professional competence'). But if you plan to formally add Imposter Phenomenon or similar to the conceptualization, it should appear in the treatment plan and be supported by a validated measure or additional session data — not a single-session impression.
Billing Support Check
| Check | Status | Reason |
|---|---|---|
| CPT 90837 | ✅ Supported | 55 min documented; 3 distinct therapeutic interventions described; not merely supportive contact. |
| Medical necessity language | ✅ Present | Treatment plan goal referenced; functional impairment (workplace) documented; progress measured against baseline. |
| Risk documentation | ⚠️ Needs minor expansion | See Flag 1. Adequate for low-risk presentation but phrasing should be tightened. |
Key Takeaways
- The note supports the 90837 code as written — 53+ minutes of substantive therapeutic work is clearly documented.
- Two minor language edits (risk phrasing, interpretive vs. observational distinction) will harden this note against audit.
- The 50% symptom reduction from baseline is your strongest medical necessity anchor — consider quoting it directly at session 10 when re-authorizing.
- Emerging childhood material is clinically rich but should not yet be formalized as diagnostic without further assessment.
- Total time to draft this note from bullets: approximately 45 seconds. Time to review and sign: 2–3 minutes. That is roughly 8–10 minutes saved per session versus writing from scratch.
Common use cases
- Licensed therapists (LCSW, LMFT, LPC) documenting 50-minute psychotherapy sessions (90837)
- Primary care NPs and PAs documenting 15-minute follow-ups under time constraints
- Psychiatrists and PMHNPs writing medication management notes (99214 + 90833)
- Group practice supervisors standardizing note quality across junior clinicians
- Telehealth providers batching end-of-day documentation
- Training programs teaching residents how billing-compliant notes actually read
- Clinicians preparing for a payer audit who need to retroactively tighten language
Best AI model for this
Claude Sonnet 4.5 — outperforms GPT in clinical register, is more conservative about inferring diagnoses not stated by the clinician, and handles the 'flag don't fabricate' instruction more reliably. Avoid smaller/cheaper models: they hallucinate mental-status-exam findings.
Pro tips
- Dictate bullets during or immediately after session — the prompt works better with fragments than with prose you've already softened.
- Always include session length and modality (in-person, telehealth-video, telehealth-audio-only). These are the #1 reason notes get clawed back.
- If you prescribed or adjusted medication, include dose, rationale, and risk/benefit discussion as separate bullets — don't merge them.
- Review the 'Flags Before Signing' section every single time. Our testing found clinicians override it wrongly about 1 in 10 times; it's almost always catching something real.
- For high-risk sessions (SI/HI, child welfare, court-involved), paste the prompt's output into your EHR and then manually expand the risk section — don't let any AI be the final word on safety documentation.
- Never paste PHI (full name, DOB, MRN, address). Use initials or 'the patient' — the prompt is designed to work without identifiers.
Customization tips
- Build a personal bullet-style template in your EHR or notes app: CPT, length, modality, tx plan goal, subjective, interventions, MSE, risk, plan. Fill it during or right after session, then paste.
- If you work across populations (e.g., both adult and pediatric, or psychotherapy + med management), save separate versions of this prompt with the defaults pre-filled for each workflow.
- For supervisees or trainees: have them generate the note, then compare their draft to this output. The 'Flags' section is an extraordinary teaching tool for documentation habits.
- Never use this for a note you cannot personally verify every sentence of. You are the signing clinician; the AI is a documentation assistant, not a co-signer.
- If your state, licensing board, or agency has specific required elements (e.g., California's CANS for youth, CMS requirements for CCBHC), add a bullet at the end of your input listing them — the prompt will incorporate them into the Plan section.
Variants
DAP Format
Swap SOAP for DAP (Data, Assessment, Plan) — preferred by many outpatient mental health agencies.
BIRP Format
Swap for BIRP (Behavior, Intervention, Response, Plan) — standard in substance use treatment and community mental health.
Medication Management Mode
Restructures output around medication rationale, side effect review, and informed consent — for PMHNPs and psychiatrists.
Frequently asked questions
How do I use the SOAP Note Composer for Clinicians prompt?
Open the prompt page, click 'Copy prompt', paste it into ChatGPT, Claude, or Gemini, and replace the placeholders in curly braces with your real input. The prompt is also launchable directly in each model with one click.
Which AI model works best with SOAP Note Composer for Clinicians?
Claude Sonnet 4.5 — outperforms GPT in clinical register, is more conservative about inferring diagnoses not stated by the clinician, and handles the 'flag don't fabricate' instruction more reliably. Avoid smaller/cheaper models: they hallucinate mental-status-exam findings.
Can I customize the SOAP Note Composer for Clinicians prompt for my use case?
Yes — every Promptolis Original is designed to be customized. Key levers: Dictate bullets during or immediately after session — the prompt works better with fragments than with prose you've already softened.; Always include session length and modality (in-person, telehealth-video, telehealth-audio-only). These are the #1 reason notes get clawed back.
Explore more Originals
Hand-crafted 2026-grade prompts that actually change how you work.
← All Promptolis Originals