12 AI Prompts for Healthcare Professionals
Critical HIPAA Warning — Read Before Using Any AI Tool
Never paste protected health information (PHI) into any public AI tool, including ChatGPT, Claude, Gemini, or similar services unless your organization has a signed Business Associate Agreement (BAA) with that provider.
PHI includes: patient names, dates of birth, addresses, medical record numbers, diagnosis codes, treatment details, or any combination of data that could identify an individual.
All prompts in this guide use placeholder variables such as [CONDITION], [PATIENT AGE], and [MEDICATION]. Fill in only de-identified or fictional details unless you are operating on a HIPAA-compliant platform. AI does not replace clinical judgment. All outputs must be reviewed and validated by a licensed clinician before use.
Healthcare professionals spend an estimated 35-55% of their working hours on documentation, administrative tasks, and communication that does not directly involve patient care. ChatGPT and similar tools — used correctly — can cut that overhead significantly. The key is knowing which tasks are safe to delegate to AI, how to structure prompts for clinical accuracy, and where the hard compliance boundaries are.
This guide covers 8 high-leverage workflow areas with 12 copy-paste prompts. Each section includes a realistic time-saved estimate based on published research and practitioner reports. All prompts are designed to be used with de-identified or fictional data.
1. Patient Education Materials
Writing clear, plain-language patient education materials is time-intensive when done from scratch. Clinicians must translate complex medical concepts into language accessible to patients with varying health literacy levels. AI drafts the first version in seconds.
Time saved: 2-4 hours per week
Use this prompt to generate patient-facing educational handouts for common conditions or procedures:
Prompt 1 — Patient Education Handout
Write a patient education handout about [CONDITION / PROCEDURE / MEDICATION]. Target audience: [e.g. adult patients with no medical background / elderly patients / parents of pediatric patients] Reading level: 6th grade or lower Length: 300-400 words Format: short paragraphs with clear section headers Include the following sections: 1. What is [CONDITION / PROCEDURE]? (1-2 sentences, plain language) 2. Why does it matter? (why the patient should pay attention) 3. What to expect (symptoms, timeline, or steps — be specific but not alarming) 4. What to do at home (practical, numbered action steps) 5. When to call your doctor (clear red flags using plain language) 6. Questions to ask your care team (3-4 example questions) Do not include specific medication dosages, dosing intervals, or specific brand name recommendations — leave those as [ASK YOUR PROVIDER] placeholders. Tone: reassuring, clear, non-condescending.
For medication-specific education, use a focused variation:
Prompt 2 — Medication Education Summary
Write a plain-language summary of [MEDICATION CLASS / GENERIC NAME] for a patient starting this treatment. Include: - What this medication does (mechanism in simple terms) - Common side effects to expect and how to manage them - Side effects that require immediate medical attention - General lifestyle considerations (food interactions, activity, etc. — use general guidance only, no specific dosing) - How long it typically takes to see effects Do not include specific doses, dosing schedules, or contraindications for specific conditions — leave those for the prescribing clinician to fill in. Flag any section where the patient should refer back to their provider with [DISCUSS WITH YOUR DOCTOR]. Tone: informative, non-alarming, plain English at a 6th-grade reading level.
2. Clinical Documentation Summaries
Converting raw clinical notes, lab results, or referral histories into structured summaries is one of the highest-volume documentation tasks in outpatient and inpatient settings. AI handles the structural work; the clinician provides and verifies the substance.
Time saved: 3-5 hours per week (varies by specialty and practice volume)
Prompt 3 — Clinical Case Summary Template
Organize the following de-identified clinical information into a structured summary suitable for a care team handoff or specialist referral. Use the following structure: - Reason for visit / chief complaint - Relevant history (past medical, surgical, medications, allergies — as provided) - Presenting findings (from the notes provided) - Working diagnosis / differential - Current management plan - Outstanding items / follow-up required Do not infer, add, or extrapolate any clinical information not explicitly provided. Flag any section where the provided information is incomplete with [INFORMATION NOT PROVIDED]. Source material: [PASTE DE-IDENTIFIED NOTES / DICTATION / STRUCTURED DATA HERE] Important: This output is a draft for clinician review only. Do not treat as a finalized medical record without review by a licensed provider.
3. Discharge Instructions
Discharge instructions are among the most critical patient safety documents in clinical care — and among the most templated. Research consistently shows that poorly written discharge instructions contribute to preventable readmissions. AI generates clear, structured first drafts that clinicians can customize in minutes.
Time saved: 20-40 minutes per discharge (significant at scale)
Prompt 4 — Discharge Instruction Draft
Write discharge instructions for a patient being discharged following [PROCEDURE / ADMISSION REASON — general description, no patient identifiers]. Patient profile (de-identified): [AGE RANGE, e.g. "adult in their 50s"] with [RELEVANT GENERAL CONDITIONS — no names or IDs] Discharge setting: [HOME / HOME WITH SUPPORT / REHABILITATION FACILITY] Format as a patient-facing document with the following sections: 1. Your Diagnosis / What Happened (1 short paragraph, plain language) 2. Activity Instructions (what they can and cannot do — include specific restrictions as [CLINICIAN TO SPECIFY]) 3. Diet and Fluid Instructions ([CLINICIAN TO SPECIFY] placeholders where individualized) 4. Wound / Incision Care (if applicable — leave specific instructions for clinician) 5. Medications (table format: Medication Name | Purpose | [DOSE — clinician to complete] | When to Take | Notes) 6. Follow-up Appointments ([CLINICIAN TO SCHEDULE] placeholders) 7. Warning Signs — Return to ED Immediately If: (use plain language red flags relevant to the condition) 8. Questions? Contact Us: [CLINIC CONTACT — clinician to complete] Reading level: 6th grade. Tone: direct, caring, easy to scan.
4. Research Literature Review
Clinicians and researchers need to synthesize published literature for grand rounds, CME preparation, protocol updates, and clinical decision support. AI accelerates the synthesis and structured comparison phases — though it cannot substitute for accessing actual papers or verify citations independently.
Time saved: 2-4 hours per literature review task
Prompt 5 — Literature Summary and Comparison
Summarize and compare the following research abstracts / excerpts on the topic of [CLINICAL TOPIC]. For each study provided, extract: - Study design (RCT, cohort, systematic review, etc.) - Population studied (key characteristics) - Primary outcome measured - Key finding (one sentence) - Limitations noted by authors - Quality flag (was this peer-reviewed? Funded by industry? Any disclosed conflicts?) After summarizing all studies, provide: 1. Points of consensus across the literature 2. Points of disagreement or conflicting findings 3. Gaps identified — what this body of evidence does not address 4. Clinical bottom line: what a practitioner might reasonably take away (with appropriate uncertainty noted) Note: Do not fabricate citations. If you are unsure of a detail not provided in the text, state "not determinable from provided text." Studies to analyze: [PASTE ABSTRACTS / EXCERPTS HERE — use only text you have legally accessed]
5. Staff Training Materials
Clinical educators and department leads regularly produce orientation documents, competency checklists, policy summaries, and scenario-based training scripts. This writing is templated, high-volume, and time-consuming. AI handles first drafts well when given clear structure requirements.
Time saved: 3-6 hours per training module
Prompt 6 — Staff Training Scenario Script
Write a clinical scenario-based training exercise for [STAFF ROLE — e.g. nursing staff, medical assistants, front desk team] on the topic of [TOPIC — e.g. recognizing sepsis early signs, proper hand hygiene protocol, escalation procedures]. Format: 1. Learning objectives (3-4 bullet points, behaviorally stated) 2. Scenario setup (2-3 sentences describing the situation — use a fictional patient name and de-identified details) 3. What the learner observes (list of findings — use clinical detail appropriate for the role) 4. Decision points (3-4 multiple choice questions with 4 options each — mark the correct answer and explain why) 5. Debrief discussion questions (3 open-ended questions for group discussion) 6. Key takeaways (3-4 bullet points) Tone: educational, direct, non-condescending. Calibrate clinical depth to [STAFF ROLE]. Flag any section where department-specific protocol should be inserted with [INSERT LOCAL PROTOCOL].
Prompt 7 — Onboarding Competency Checklist
Create a competency checklist for new [ROLE] staff joining a [SETTING — e.g. outpatient clinic, ICU, long-term care facility]. The checklist should cover onboarding milestones across four phases: - Week 1: Orientation and environment (systems access, safety protocols, team introductions) - Weeks 2-4: Supervised clinical or operational tasks - Month 2: Independent task milestones - End of probation (90 days): Full competency sign-off criteria For each item include: - Competency description (action verb + specific task) - Assessment method (observation, written test, simulation, supervisor sign-off) - Target completion date (relative to start date) - Sign-off field: [SUPERVISOR INITIAL / DATE] Add a section at the end for department-specific items with [INSERT LOCAL REQUIREMENTS] placeholders. Format as a table for easy printing.
6. Quality Improvement Reports
QI teams spend significant hours writing narrative sections of reports, summarizing data findings, and drafting recommendations. AI converts raw data descriptions into structured narrative prose quickly — leaving the subject matter interpretation to the clinician.
Time saved: 2-4 hours per QI report cycle
Prompt 8 — QI Report Narrative Draft
Write the narrative section of a quality improvement report for the following initiative. Use only the data provided — do not invent figures. Initiative title: [TITLE] Department / setting: [DEPARTMENT] Reporting period: [PERIOD] QI framework used: [PDSA / Lean / Six Sigma / other — or leave blank] Data summary to incorporate: - Baseline metric: [VALUE, time period] - Target: [VALUE] - Current performance: [VALUE] - Trend: [IMPROVING / STABLE / DECLINING] - Key interventions implemented: [LIST — dates and descriptions] - Any adverse events or unintended consequences: [DESCRIBE OR "NONE NOTED"] Write four sections: 1. Background and Aim (2-3 sentences — why this initiative was launched, what the aim statement is) 2. Methods (what was done — summarize the interventions in plain language) 3. Results (narrative description of the data provided — do not embellish) 4. Conclusions and Next Steps (what the data suggests, what is recommended next — frame recommendations as suggestions for clinician review, not definitive conclusions) Tone: professional, evidence-referenced, concise. Flag any section requiring additional data with [DATA NEEDED].
7. Referral Letters
Referral letters require accurate transmission of clinical context to a specialist while remaining concise and professionally formatted. Writing them from scratch for each patient visit consumes time that scales poorly with practice volume. AI produces structured first drafts in seconds from a structured input.
Time saved: 15-30 minutes per referral letter
Prompt 9 — Specialist Referral Letter Draft
Draft a specialist referral letter using the following de-identified clinical information. This is a first draft for the referring clinician to review and complete with patient-specific identifiers before sending. Referring provider: [DR. NAME / SPECIALTY — clinician to complete] Receiving specialist: [SPECIALTY] Reason for referral: [CLINICAL INDICATION IN GENERAL TERMS] De-identified clinical context: - Patient demographics: [AGE RANGE, SEX — no names or IDs] - Presenting complaint: [DESCRIPTION] - Relevant history: [LIST — conditions, surgeries, medications] - Relevant recent findings: [LAB / IMAGING / EXAM FINDINGS — general description] - Current management: [WHAT HAS BEEN TRIED OR IS ONGOING] - Specific question for specialist: [WHAT YOU ARE ASKING THEM TO EVALUATE OR MANAGE] - Urgency: [ROUTINE / SEMI-URGENT / URGENT — and why] Format: formal medical letter structure. Opening paragraph establishes context. Body covers history and findings. Closing paragraph states the specific referral request and contact preference. Length: 200-300 words. Do not fabricate any clinical details. Leave [CLINICIAN TO COMPLETE] markers wherever patient identifiers or finalized clinical data must be inserted.
8. Patient Communication Templates
Practices send dozens of routine communications weekly: appointment reminders, results notifications, follow-up instructions, and care gap outreach. AI generates consistent, professional templates that reduce the per-message burden for front desk and clinical staff.
Time saved: 1-3 hours per week across a practice
Prompt 10 — Normal Results Notification Template
Write a patient portal message template to notify a patient that their [TEST TYPE — e.g. routine bloodwork, imaging, screening test] results are normal. Tone: warm, clear, not overly clinical. The patient should feel reassured but understand this is a template message. Include: - 1-sentence summary of the result ("Your [test] results from [DATE — clinician to complete] look normal.") - Brief explanation of what "normal" means for this test (1-2 sentences, plain language) - What to do if the patient has questions - When the next test of this type is typically recommended (leave as [CLINICIAN TO ADVISE] if it varies) - A clear footer: "This message is for informational purposes. It is not a substitute for a clinical consultation. If you have new or worsening symptoms, contact our office." Leave [PATIENT NAME], [DATE], and [PROVIDER NAME] as placeholders for staff to fill in.
Prompt 11 — Care Gap Outreach Message
Write a patient outreach message for a care gap reminder. The message should encourage a patient to schedule an overdue preventive service. Service due: [e.g. annual wellness visit, mammogram, colonoscopy, diabetes eye exam — general type, not patient-specific] Tone: caring, not alarming, not guilt-inducing. Never imply the patient has been neglectful. Channel: [EMAIL / PATIENT PORTAL / SMS — adjust length and formality accordingly] Include: - Why this screening or visit matters (1-2 sentences, health benefit framing) - How to schedule (leave as [INSERT SCHEDULING LINK / PHONE NUMBER]) - Reassurance that the appointment is routine and brief - An opt-out line for SMS: "Reply STOP to stop receiving these messages." Flag any section that requires personalization with [STAFF TO COMPLETE]. Keep reading level at 6th grade or below.
Prompt 12 — Post-Procedure Follow-Up Message
Write a post-procedure follow-up message template for patients who have recently undergone [PROCEDURE TYPE — general description]. Format: patient portal message, 150-200 words. Include: - Acknowledgment that the procedure is complete and a check-in on recovery - 3-5 specific things to monitor in the days following (appropriate to procedure type — use conservative, evidence-based indicators) - Clear guidance on what warrants calling the office vs. going to the ED - Reminder of follow-up appointment: [DATE — staff to complete] - Encouragement and a warm close Tone: warm, direct, reassuring without being dismissive. Plain language, 6th grade reading level. Add footer: "If you have any concerns, do not wait — call us at [CLINIC PHONE]. In an emergency, call 911 or go to your nearest emergency department." Leave all patient-specific identifiers as [STAFF TO COMPLETE] placeholders.
Total Time Saved Across Workflows
Here is how the numbers stack up across a typical week in a moderately busy clinical or administrative healthcare role:
| Workflow | Time Saved / Week |
|---|---|
| Patient education materials | 2-4 hrs |
| Clinical documentation summaries | 3-5 hrs |
| Discharge instructions | 2-4 hrs |
| Research literature review | 2-4 hrs |
| Staff training materials | 1-2 hrs |
| QI reports | 1-2 hrs |
| Referral letters | 1-3 hrs |
| Patient communication templates | 1-3 hrs |
| Total | 13-27 hrs / week |
For a physician spending 2.5 hours per day on EHR and administrative tasks — a figure well-documented in burnout research — even a 50% reduction represents more than 6 hours returned to direct patient care or recovery each week.
The Critical Limits of AI in Clinical Settings
AI does not examine patients. It does not have access to your EHR, your patient's full history, or the clinical intuition developed over years of practice. Every output from the prompts above is a first draft — a structured starting point that must be reviewed, corrected, and signed off by a qualified clinician before it touches a patient or a medical record.
Where AI helps: eliminating the blank-page problem for templated documents, structuring information you already have, drafting communication at consistent quality. Where AI does not help: clinical reasoning, diagnostic judgment, interpreting ambiguous findings, or making treatment decisions. That line must remain clear.
The 155 prompts in the qarko AI Workflow Guide Core include specialized sections on documentation, research synthesis, training materials, and professional communication — each optimized for Claude, GPT-4o, and Gemini. The HIPAA guidance above applies across all of them.
For a complete workflow system covering documentation, patient communication, research synthesis, and more, see our AI Automation for Healthcare guide.
Related Posts
Get the Full Healthcare Workflow System
155 copy-paste prompts plus step-by-step workflow guides for documentation, patient communication, research synthesis, training materials, and more. Optimized for Claude, GPT-4o, and Gemini.
Just need the prompts?
155 copy-paste prompts for Claude, GPT-4o, and Gemini — documentation, analysis, communication, research, and more.