Billing and coding errors cost practices significant time and revenue every year, often not because clinicians don't know the codes, but because documentation doesn't clearly support the level of service provided. AI can close that gap, helping you document MDM accurately, look up codes instantly, and format notes for your specific EHR.
Doximity Ask is a HIPAA-compliant environment, which means you can include de-identified clinical context in your billing and coding prompts without privacy concerns. Medical Decision Making (MDM)
MDM is the highest-value element of E/M coding. It is also the most frequently under-documented. AI can write MDM language directly from your clinical bullets, covering all three required elements: problems, data, and risk.
The problems element documents the number and complexity of problems addressed. The data element covers tests ordered or reviewed, external records, and clinical discussions. The risk element captures the highest risk of complications, including prescription drug management.
Copy this prompt: MDM documentation, Internal Medicine: Generate a Medical Decision Making paragraph for E/M coding purposes. Problems: T2DM with poor control (A1c 10.2), hypertension, new peripheral neuropathy. Data reviewed: today's labs, prior A1c trend, podiatry consult note. Risk: adjusting insulin regimen, prescription drug management. Format with three labeled sections: Problems, Data, Risk. |
The output gives you a structured MDM paragraph ready to copy into your note, with each element clearly labeled for audit purposes. Review it before signing to confirm it accurately reflects the visit.
ICD-10 and CPT Lookups
Searching the ICD-10 database manually is slow. AI returns the highest-specificity code with a brief justification in seconds. The same applies to CPT codes and billing level determinations.
Some of the most useful prompt types for coding:
- Specific code lookup: "What ICD-10 code covers T2DM with diabetic peripheral neuropathy? Give me the highest-specificity code."
- Billing level justification: "What E/M billing level is appropriate for this visit? Provide a brief EHR-ready justification."
- Quick-reference table: "Give me a quick-reference table of the top 20 ICD-10 codes I use in Family Medicine, with short descriptions."
- CPT for procedures: "What CPT code covers nail clipping for fungal culture (PAS)? What CPT code covers wound irrigation and debridement?"
- Same-day billing: "Document a separately identifiable condition to support same-day E/M and procedure billing."
In Doximity Ask, you can run these lookups as part of a larger documentation session, combining a coding question with a note request in the same conversation thread. Making Any Prompt EHR-Ready
The most underused technique in AI-assisted documentation is simply naming your EHR in the prompt. The same clinical content, formatted differently for each system, becomes immediately usable instead of requiring manual reformatting.
The template phrase is: "Format this note for direct copy-paste into [EHR name]. Use [EHR]'s standard section headers." This works for every major system:
- Epic: SmartPhrases, dot phrases, .NOTETEMPLATE shortcuts, After Visit Summary, in-basket messages
- Oracle Health EHR (formerly Cerner): PowerChart structure, QuickText templates, structured order sets
- Meditech: paragraph-style progress notes, RVU-optimized A&P structure
- Athena: SOAP format with CPT and ICD-10 embedded, referral letters, macros
You can also use AI to draft order sets and admission templates. Give it the diagnosis, the setting (ED, inpatient, outpatient), and your EHR, and ask it to return a structured order set ready to load into your system. This works for admission order sets, standing order templates, referral letter templates, overnight contingency plans, and discharge instruction bundles.
Copy this prompt: Order set, Emergency Medicine: Draft an ED order set for an adult presenting with community-acquired pneumonia, moderate severity. Include: antibiotics per current IDSA guidelines, IV fluids, oxygen parameters, respiratory therapy orders, nursing parameters, and admission criteria. Format for Epic order entry. |
IMPORTANT AI can suggest billing codes and MDM language, but you are responsible for verifying coding accuracy before signing. AI coding suggestions are a starting point, not a final determination. Always review against the visit documentation. |
