Introducing job agent - get access to jobs faster View jobs

Back to blogs

New Attending Documentation Tips: What Nobody Told You

New attending documentation tips grounded in data: what the 2021 and 2023 E/M rules changed, why pajama time sticks, and habits to build in month one.

Reviewed by O. Daniel Odutola, MD, MBA, MPH14 min read

The most useful new attending documentation tips have almost nothing to do with typing faster. They come down to three changes you probably were not briefed on: you now own every word in the chart, the billing rules were rewritten while you were busy on service, and the charting habits that carried you through residency were built for a system that checked your work behind you. Nobody hands you a memo about any of this on your first day.

You just get a login, a patient list, and a note template somebody built in 2016.

Key takeaways

  • History and physical exam bullet counts no longer set your E/M level for office visits (since 2021) or for hospital, observation, ED, and nursing facility visits (since 2023). Level is driven by medical decision making or total time. [5][6]

  • Since January 1, 2021, Medicare lets the billing clinician review and verify documentation entered by residents, nurses, and students instead of re-documenting it. [7]

  • Residents get measurably faster per patient across a year of clinic, but the share of their EHR time that lands after hours does not budge. Speed alone will not get you out of the chart at night. [3]

  • Average outpatient note length rose 8.1% in the three years after the 2021 coding changes that were meant to shrink it. The rules relaxed; the templates did not. [9]

  • In a typical hospital progress note, 18% of the text was manually entered, 46% was copied, and 36% was imported. Whatever you copy forward, you are now signing. [10]

Why charting as a new physician feels harder than residency

Here is the part that surprises people. The notes get shorter and the work gets heavier.

As a resident, your note was a draft. Somebody read it, co-signed it, and occasionally rewrote your assessment in a way that stung a little and taught you something. There was a net under you. The first time you sign a note as the attending of record and realize nobody is reading it before it becomes the legal document, the permanent record, and the thing a plaintiff's expert will read in four years, the keyboard feels different.

Then the volume arrives. Ambulatory physicians average 16 minutes and 14 seconds of EHR time per encounter, with documentation accounting for 24% of it and chart review another 33%. [1] Family physicians in one longitudinal audit-log study logged 5.9 hours of EHR time inside an 11.4 hour workday, 1.4 hours of it after clinic closed. [2] Those are seasoned attendings. You are starting from behind them.

And the burden is not evenly distributed. American physicians write clinical notes roughly four times as long as their counterparts abroad using the same EHR software. [4] That gap is not clinical. It is billing residue, liability anxiety, and template inheritance.

If your first job is still a few months out, our resident to attending transition guide (/market/resident-to-attending-guide) covers the contract and onboarding side of the same jump.

See how your offer compares

Compare real pay data before you sign anything

Compare pay

New attending documentation tips: six habits for your first 90 days

Treat this as a sequence, not a menu. Most of it is one-time setup work that pays you back every shift.

  1. Audit your own template before you write a single note in it. Open the default note your group handed you and delete every auto-populated section you will not actually read on the next visit. Full labs, full med lists, full problem lists pulled in wholesale. You are the one who has to find your own assessment in there tomorrow.

  2. Write the assessment and plan first. Then let the rest of the note support it. The A/P is the only section anyone reads, including you at 2 a.m. when the patient comes back. Building the note around it forces the thinking to happen once instead of twice.

  3. Ask your billing or compliance team for your specialty's MDM grid, on paper, in week one. Not the coding lecture. The actual table. Then ask which of your top five diagnoses habitually get downcoded and why. Twenty minutes of that conversation is worth more than a year of guessing.

  4. Close notes in the room or immediately after, for at least half your patients. Pick the easy half. Sore throats, medication refills, stable follow-ups. The goal is not perfection. It is shrinking the pile that follows you into the evening. We go deeper on this in our piece on cutting pajama time (/clinical/reduce-pajama-time).

  5. Learn three keyboard shortcuts and one dictation workflow. Stop there. A study of 115 primary care physicians across 48 clinics found no significant association between the standard EHR efficiency tools (message triage, custom quick actions, message pools) and lower inbox time. [15] The tooling is not the lever most people think it is. Find the two things that genuinely fit your hands and ignore the rest of the menu.

  6. Put a hard stop on your calendar and let notes be late instead of letting your evening be. A note finished at 8 a.m. the next morning with a clear head beats one finished at 11 p.m. with three errors in it. Check your group's policy on closure windows first, then use the whole window.

Start your first job charting faster with Physicians' Copilot.

The E/M rule changes nobody handed you on graduation day

This is the single largest gap between what residents are taught and what the rules require. Many attendings are still documenting to a 1995 standard because that is what the template, and the senior partner, and the muscle memory all expect.

What you learned in training

What the rules require now

In effect since

Your level depends on counting HPI elements, ROS items, and exam systems

Perform a "medically appropriate history and/or examination"; level is set by medical decision making or total time

Office visits 99202-99215, Jan 1, 2021 [5]

Hospital, observation, ED, and SNF notes need documented history and exam elements to support the level

History and exam documentation requirements removed; code on MDM or total time

Jan 1, 2023 [6]

Observation care has its own code set

Observation codes folded into the initial, subsequent, and discharge visit codes

Jan 1, 2023 [6]

The attending must re-document the resident's history and exam

The billing clinician may review and verify (sign and date) documentation already in the record

Jan 1, 2021, Transmittal 10639 / CR 11862 [7]

Only your face-to-face minutes count toward a time-based level

Total time on the date of service, including non-face-to-face work such as chart review and care coordination

Office visits, Jan 1, 2021 [5]

On a shared visit, whoever spent longer with the patient bills it

Substantive portion means more than half the total time, or a substantive part of the MDM

Jan 1, 2024, MLN MM13592 [8]


Two practical consequences. First, if you are attesting on resident notes by retyping the HPI, you are doing unpaid transcription; a review-and-verify attestation that documents your own participation and your own medical decision making is what the rule actually asks for. [7] Second, for inpatient admissions the AAFP's coding walkthrough notes that time spent past midnight on the calendar day of admission may still be applied to the reported date of service, which is not how office visits work. [6]

Confirm the specifics with your own compliance office and payer mix. Rules vary by payer and by contractor, and this article is educational rather than coding or compliance advice for your practice.

What the data actually says about getting faster

You will get faster. That part is real and it is measurable. In a study of 622 resident physicians tracked over 12 months of ambulatory practice, EHR time per patient fell and patients seen per day rose as experience accumulated. [3]

Now the uncomfortable half of the same finding. The proportion of EHR time spent after hours did not change. [3] Residents got more efficient and still took the same share of the work home.

Read that as a warning about strategy. Raw speed redistributes work inside the clinic day and leaves the evening untouched, because the evening is where the leftover work goes by default. What moves pajama time is structural: closing notes at the point of care, killing template bloat, and negotiating your schedule and support staff. The efficiency gospel oversells itself here.

Epic Research makes the same point from a different angle. Across 1.7 billion notes from 166,318 outpatient providers at 384 organizations, average note length rose from 4,628 characters in May 2020 to 5,002 characters in April 2023, an 8.1% increase, even though the 2021 coding changes were designed to shrink documentation. Time per note did fall, from 5.4 minutes to 4.8 minutes, an 11.1% drop. [9] Providers got faster at producing longer notes that the rules no longer required. About 40% did manage to cut their note length. Be in that 40%.

Physician burnout is now at 41.9% by the AMA's 2025 measure, down for a third consecutive year, with ineffective EHR systems and excessive administrative tasks named among the recurring drivers. [13] The trend is genuinely encouraging. It is also still four in ten.

Copy forward, note bloat, and the liability you inherit

In 23,630 notes from 460 clinicians at UCSF, the typical note was 18% manually entered, 46% copied, and 36% imported. Residents manually entered just 11.8% of their note text. [10] If you trained recently, copy forward is not a bad habit you picked up. It is the water you learned to swim in.

It becomes a different animal when you are the one signing. A copied exam finding you did not perform is a documentation integrity problem and a plaintiff's gift. A copied assessment that no longer matches the patient is worse, because it is the section that demonstrates your reasoning.

The Partnership for Health IT Patient Safety published four safe-practice recommendations for copy and paste in 2016 that still hold up: make copied material easily identifiable, keep its provenance available, train people on appropriate use, and monitor the practice. [11] Most of that is organizational. Here is the personal version, as a checklist you can run in 30 seconds per note:

☐  Every exam finding in this note is one I personally elicited today.

☐  The assessment reflects today's thinking, not last week's, and says what changed.

☐  No lab or imaging result is pasted without my interpretation attached.

☐  Nothing in this note contradicts anything else in this note.

☐  A colleague covering me tomorrow could read only the A/P and know the plan.

☐  I would be comfortable if the patient read this today, because they can.

That last one is not rhetorical. Under the ONC information blocking rule, patients have had near-immediate electronic access to clinical notes since April 5, 2021. [14] Write the note to the patient and the covering physician. Both are reading it.

For the full framework behind these habits, see our clinical documentation guide (/clinical/clinical-documentation-guide).

Should a new attending use an AI scribe?

Probably yes, if your organization offers one and you can afford the learning curve during a season when everything else is also new.

The Permanente Medical Group reported results from more than 2.5 million patient encounters across 7,260 physicians using an ambient AI scribe, with statistically significant reductions in note-taking time outside standard hours, time per appointment, and pajama time, and an estimated time savings equivalent to 1,794 working days in a single year. [12] Notably, the top third of users accounted for 89% of activations, and heavy users saved roughly two and a half times more time per note than occasional users. [12]

That last detail is the real finding. Ambient scribes reward commitment. Dabbling produces very little. If you turn it on for the complicated visits only, you will conclude it does not work.

Two cautions for year one. You still have to read every line before you sign it, and the draft will be confidently wrong sometimes in ways that are easy to skim past when you are tired. And a scribe that produces a beautifully organized note does not fix a template that pulls in 900 characters of stale problem list. Fix the template first; let the scribe write into something clean.

No AI tool replaces clinical judgment. Verify all outputs against primary sources and your own assessment.


Where to start tomorrow

Open your default template. Delete what you do not read. Write the assessment first. Ask your coder which of your codes get downgraded. Then pick one hard stop in your week and defend it.

The documentation load in American medicine is a system problem, and no habit list fixes a system. But the gap between a new attending who sets up deliberately in month one and one who inherits whatever the EHR handed them is measured in hours per week, every week, for years.

Start your first job charting faster with Physicians' Copilot.


REFERENCES

  1. Overhage JM, McCallie D. "Physician Time Spent Using the Electronic Health Record During Outpatient Encounters: A Descriptive Study." Annals of Internal Medicine, January 14, 2020; 172(3):169-174. doi:10.7326/M18-3684. Data period: 2018 (approximately 100 million encounters, 155,000 physicians, 417 health systems). Retrieved via PubMed. https://doi.org/10.7326/M18-3684 (accessed September 30, 2026)

  2. Arndt BG, Beasley JW, Watkinson MD, et al. "Tethered to the EHR: Primary Care Physician Workload Assessment Using EHR Event Log Data and Time-Motion Observations." Annals of Family Medicine, September/October 2017; 15(5):419-426. doi:10.1370/afm.2121. Data period: July 1, 2013 to June 30, 2016; 142 family medicine physicians. https://doi.org/10.1370/afm.2121 (accessed September 30, 2026)

  3. Holmgren AJ, Lindeman B, Ford EW. "Resident Physician Experience and Duration of Electronic Health Record Use." Applied Clinical Informatics, August 4, 2021; 12(4):721-728. doi:10.1055/s-0041-1732403. Data period: July 2017 to June 2018; 622 resident physicians. Retrieved via PubMed. https://doi.org/10.1055/s-0041-1732403 (accessed September 30, 2026)

  4. Downing NL, Bates DW, Longhurst CA. "Physician Burnout in the Electronic Health Record Era: Are We Ignoring the Real Cause?" Annals of Internal Medicine, May 8, 2018; 169(1):50-51. doi:10.7326/M18-0139. Retrieved via PubMed; figure on note length also summarized by Stanford Medicine, May 7, 2018. https://doi.org/10.7326/M18-0139 (accessed September 30, 2026)

  5. American Medical Association. "E/M office-visit changes on track for 2021: What doctors must know." AMA, page updated January 26, 2026. https://www.ama-assn.org/practice-management/cpt/em-office-visit-changes-track-2021-what-doctors-must-know (accessed September 30, 2026)

  6. Millette KW. "The 2023 Hospital and Nursing Home E/M Visit Coding Changes." Family Practice Management (AAFP), January 15, 2023; 30(1). https://www.aafp.org/pubs/fpm/issues/2023/0100/hospital-em-coding.html (accessed September 30, 2026)

  7. Centers for Medicare & Medicaid Services. CMS Manual System, Transmittal 10639, Change Request 11862, Medicare Benefit Policy Manual Pub. 100-02 Chapter 30. Effective January 1, 2021; implemented April 12, 2021. https://www.cms.gov/files/document/r10639bp.pdf (accessed September 30, 2026)

  8. Centers for Medicare & Medicaid Services. "Updates for Split or Shared Evaluation & Management Visits." MLN Matters MM13592, released May 3, 2024; policy effective January 1, 2024. https://www.cms.gov/files/document/mm13592-updates-split-or-shared-evaluation-management-visits.pdf (accessed September 30, 2026)

  9. Epic Research. "Two Years After Coding Changes Sought to Decrease Documentation, Notes Remain Bloated." July 6, 2023. Data period: May 2020 to April 2023; 1.7 billion notes, 166,318 outpatient providers, 384 organizations. https://www.epicresearch.org/articles/two-years-after-coding-changes-sought-to-decrease-documentation-notes-remain-bloated/ (accessed September 30, 2026)

  10. Wang MD, Khanna R, Najafi N. "Characterizing the Source of Text in Electronic Health Record Progress Notes." JAMA Internal Medicine, August 2017; 177(8):1212-1213. doi:10.1001/jamainternmed.2017.1548. Data period: January 10 to August 31, 2016; 23,630 notes, 460 clinicians, UCSF Medical Center. https://doi.org/10.1001/jamainternmed.2017.1548 (accessed September 30, 2026)

  11. Partnership for Health IT Patient Safety (ECRI). "Health IT Safe Practices: Toolkit for the Safe Use of Copy and Paste." February 2016. https://www.ecri.org/resources/hit/cp_toolkit/toolkit_copypaste_final.pdf (accessed September 30, 2026)

  12. "Ambient Artificial Intelligence Scribes: Learnings after 1 Year and Over 2.5 Million Uses." NEJM Catalyst Innovations in Care Delivery, April 7, 2025; doi:10.1056/CAT.25.0040. Figures as summarized by Permanente Medicine, "Analysis: AI scribes save physicians time, improve patient interactions and work satisfaction," April 2025. https://permanente.org/analysis-ai-scribes-save-physicians-time-improve-patient-interactions-and-work-satisfaction/ (accessed September 30, 2026)

  13. American Medical Association. "Physician burnout rate continues to decline, falling to nearly 42%." April 16, 2026. AMA Organizational Biopsy / National Physician Comparison Report; approximately 19,000 physician responses, 2025 data. https://www.ama-assn.org/practice-management/physician-health/physician-burnout-rate-continues-decline-falling-nearly-42 (accessed September 30, 2026)

  14. OpenNotes. "U.S. Federal Rule Mandates Open Notes" (ONC 21st Century Cures Act information blocking rule). Compliance date April 5, 2021; electronic health information definition expanded October 6, 2022. https://www.opennotes.org/onc-federal-rule/ (accessed September 30, 2026)

  15. Bundy R, et al. "Assessment of EHR Efficiency Tools and Resources Associated with Electronic Health Record Inbox Time." Journal of General Internal Medicine, May 8, 2024. doi:10.1007/s11606-024-08761-3. Data period: July 1, 2021 to June 30, 2022; 115 primary care physicians, 48 clinics. https://doi.org/10.1007/s11606-024-08761-3 (accessed September 30, 2026)

FAQs

There is no benchmark you should hold yourself to in month one, and chasing one will make you worse. For context, the average outpatient note took 4.8 minutes of documentation time in 2023 across a very large national dataset, and total EHR time per ambulatory encounter averages about 16 minutes when chart review and ordering are included. [9][1] Expect to be well above that at first and to improve over roughly a year. [3]

Related articles