Artificial Intelligence Leveraged to Write Medical Documentation in Seconds Saving Hours

May 5, 2025

Revolutionizing Clinical Workflow: How AI Streamlines Medical Documentation

Artificial intelligence (AI) is often portrayed as a job-stealer, but before it transforms industries, it’s poised to make our work significantly easier—especially in healthcare. Dr. Seheult walks us through how AI can transform medical documentation, reduce burnout, and save hours for healthcare providers. Whether you’re a clinician or a patient curious about how your healthcare visits are evolving, this is happening in real time, and it’s exciting.

The Growing Burden of Medical Documentation

Medical documentation has come a long way in the last 50 years. Once a simple tool to jog a physician’s memory, it’s now a multifaceted document critical for:

Billing: If it’s not documented, it didn’t happen—and you can’t bill for it.
Legal purposes: Notes detail a provider’s thought process and justify clinical decisions.
Communication: Documentation ensures continuity of care across teams.

However, this evolution has a downside. Documentation now consumes a massive portion of a healthcare provider’s day, often leading to burnout. Studies from over a decade ago showed that hiring scribes—people to shadow providers and handle documentation—dramatically reduced burnout and rekindled clinicians’ passion for medicine. But human scribes are expensive. Enter AI.

AI-powered tools, like Doximity GPT, are stepping in as virtual scribes, generating comprehensive medical notes in seconds. These tools are often free, HIPAA-compliant, and capable of transforming how providers manage their workload. Let’s explore how.

What Does a Typical Inpatient Visit Look Like?

To understand AI’s impact, let’s break down a typical inpatient visit for a hospitalist or intensivist admitting a patient. This process applies to outpatient clinics too, but we’ll focus on the hospital setting.

1. New Patient Admission

When admitting a patient, providers:

Research: Review labs, old records, and patient history (10-15 minutes).
Interview: Speak with the patient or family to understand the situation (5-10 minutes).
Examine: Perform a physical exam (5 minutes).
Order: Enter orders for medications, IV fluids, imaging, or labs (5-10 minutes).
Document: Write a history and physical (H&P) note (10-15 minutes).

The H&P is a structured document with sections like:

History of Present Illness (HPI): A narrative of why the patient is here.
Review of Systems (ROS): A checklist of symptoms across 10 systems.
Past Medical/Surgical/Social/Family History: Relevant background.
Medications/Allergies: Current meds and drug sensitivities.
Physical Exam: Detailed findings from head to toe.
Labs/Imaging: Results like bloodwork or X-rays.
Assessment and Plan (A&P): Diagnoses and treatment plans, often organized by system (e.g., neurological, cardiac).
Additional Details: Code status, DVT prophylaxis, estimated length of stay, and billing time.

This process can take 30-45 minutes per patient, with documentation alone eating up 10-15 minutes. Multiply that by a dozen patients, and it’s easy to see why burnout is rampant.

2. Return Visit (Progress Note)

For follow-up visits, providers:

Review: Check overnight labs, imaging, and events (10 minutes).
Interview/Examine: Assess the patient’s progress (5 minutes).
Collaborate: Discuss with nurses, pharmacists, or respiratory therapists (15 minutes in the ICU).
Document: Write a SOAP note (Subjective, Objective, Assessment, Plan) (10-15 minutes).

The SOAP note updates the H&P’s A&P based on new developments, ensuring the plan reflects the patient’s current status. Like the H&P, it’s time-intensive and must be precise for billing and legal purposes.

How AI Supercharges Documentation

AI doesn’t replace clinical judgment—it streamlines the tedious task of documentation. By inputting raw patient data, providers can generate polished, structured notes in seconds. Here’s how it works, using Doximity GPT as an example.

Step 1: Choose a HIPAA-Compliant Tool

Doximity GPT is a free, HIPAA-compliant AI writing assistant designed for providers. It requires a Business Associate Agreement (BAA) to ensure patient data security, a must for any AI tool used in healthcare. Other HIPAA-compliant options exist, but some require payment. Doximity’s mobile app and desktop platform sync seamlessly, making it versatile for busy clinicians.

Step 2: Craft a Robust Prompt

The quality of AI-generated notes depends on the prompt. A good prompt specifies:

The note type (e.g., H&P or SOAP).
The desired structure (e.g., system-based A&P for ICU notes).
Key details (e.g., ICD-10 codes, prophylaxis, code status, estimated stay).

For example, I created an H&P prompt for Doximity GPT that dictates the exact order of sections, from HPI to A&P, and includes ICU-specific systems (neurological, respiratory, etc.). After inputting patient data, the AI formats it into a professional note.

Step 3: Input Patient Data

Providers can type or dictate patient information, even in a disorganized format. For instance, I dictated a fictional case:

A 63-year-old nursing home resident with altered mental status, hypotension (BP 70/50), and tachycardia (HR 120s).
History of urinary tract infections, diabetes, hypertension, and dementia.
Labs showed elevated WBC (15,000), creatinine (2.5), and lactic acid (3.0).
Treatment included IV fluids, vasopressors, and antibiotics for suspected septic shock.

Despite my dictation being a rough stream of sentences, Doximity GPT transformed it into a polished H&P in seconds, complete with:

A concise HPI in clear English.
Organized ROS, past history, and exam findings.
A system-based A&P with ICD-10 codes, prophylaxis details, and critical care time.

Step 4: Update for Progress Notes

For return visits, I used a separate prompt to update the A&P. After copying the previous day’s A&P and inputting overnight developments (e.g., vasopressors discontinued, creatinine improved to 2.0), the AI generated an updated SOAP note, seamlessly integrating new data while preserving unchanged elements.

Step 5: Review and Finalize

AI isn’t perfect—always review the output for accuracy. Doximity GPT flags this, reminding providers they’re the authority. Once verified, notes can be copied into an electronic medical record (EMR), faxed, or texted.

Beyond Documentation: AI as a Clinical Ally

While this video focused on documentation, AI can do more. For example, I asked Doximity GPT:

“What complications is this patient at risk for?” It identified healthcare-acquired infections, acute kidney injury, and thrombotic events.
“What antibiotic is best for ESBL E. coli?” It correctly recommended a carbapenem (e.g., meropenem).

These insights can guide management, though I recommend cross-checking with clinical guidelines.

Why This Matters

AI like Doximity GPT can:

Save Time: Cut documentation from 15 minutes to seconds, freeing providers to focus on patient care.
Reduce Burnout: Less paperwork means more time at the bedside, reigniting clinicians’ passion.
Improve Accuracy: Structured prompts ensure notes meet billing and legal standards.
Enhance Care: By streamlining workflows, providers can see more patients without sacrificing quality.

For patients, this translates to more attentive providers and potentially shorter hospital stays, as documentation bottlenecks are eliminated.

Get Started with AI Documentation

Ready to try AI in your practice? Here’s how:

Choose a Tool: Start with Doximity GPT (free, HIPAA-compliant) if you’re a provider.
Use Our Prompts: Visit MedCram’s blog and scroll to the “Useful Links” section. You’ll find a post titled “Revolutionizing Clinical Workflow: How HIPAA-Compliant AI Like Doximity GPT Streamlines Patient Care.” It includes Dropbox links to my H&P and SOAP prompts, tailored for ICU and adaptable for other settings.
Customize: Modify the prompts for procedure notes, outpatient visits, or specialties like cardiology or surgery.
Explore MedCram: While you’re on our site, check out our videos on ECG interpretation, acid-base disorders, and more for continuing medical education (CME) credit.

The Future of Healthcare

AI isn’t here to replace clinicians—it’s here to empower us. By handling repetitive tasks like documentation, it lets us focus on what matters: treating patients. As AI evolves, expect even more integration into clinical workflows, from decision support to predictive analytics.

Have you tried AI for documentation? Share your experience in the comments below, and subscribe to MedCram on YouTube for more insights on medicine and technology. Let’s embrace this revolution together!

LINKS / REFERENCES: Doximity GPT (Doximity) | https://www.doximity.com/doximity-gpt…

10 top AI tools in healthcare for 2025 (TechTarget) | https://www.techtarget.com/healthtech…

13 AI healthcare companies to watch in 2025 (TechTarget) | https://www.techtarget.com/healthtech…

Is ChatGPT HIPAA Compliant? (The HIPAA Journal) | https://www.hipaajournal.com/is-chatg…

Heidi (Heidi Health) | https://www.heidihealth.com