The electronic health record: what physicians say, in their own words

Last updated September 20, 2026.

Read together, the 471 posts on this page describe a promise and its cost. KevinMD published physicians on electronic records for five years before the 2009 HITECH Act paid hospitals and practices to adopt them, and the early posts asked whether they would save money and improve care. The posts from 2010 to 2016, the meaningful use years, answer that they were designed for billing and that the physician, not the software, absorbed the work. The posts since 2017 stop arguing about the record and describe what it does to the physician: hours of clicking, after-hours “pajama time,” alert fatigue, and a screen between doctor and patient. Mentions of burnout went from 1 percent of posts in 2009 to 2012 to 41 percent since 2021, and AI or ambient tools from 1 percent to 56. The record ends, for now, with physicians disagreeing about whether AI scribes are the resolution of that story or its next chapter.

This page is a maintained record of what physicians have written about the electronic health record on KevinMD.com, a physician-authored publication founded in 2004 by Kevin Pho, MD, a board-certified internal medicine physician in Nashua, New Hampshire. It draws on 471 posts with EHR, EMR, electronic health record, electronic medical record, or scribe in the title, published between June 2004 and September 2026 by 207 named contributors, 270 of them bylined by physicians. Every claim on this page is attributed to a named author with the date it was published and a link to the original.

The sections below are organized around the questions physicians, administrators, health IT professionals, and journalists ask. Each opens with a direct answer, followed by what named authors have said, in the order they said it. Where authors cite studies, the study is named alongside the author, and the figures are theirs. This is a page about physician opinion and experience, not a product review.

What do physicians say the electronic record did to the visit?

Physicians on KevinMD describe the electronic record as having put a screen and a keyboard between them and the patient, turned the visit into data entry, and replaced clinical narrative with checkboxes and templates. Posts mentioning the screen, eye contact, or looking at the computer rose from 6 percent in 2004 to 2008 to 47 percent since 2021. A minority of authors, some using electronic records since the 1990s, write that the loss is a matter of technique rather than technology.

Edward Pullen, MD, a family physician who had used an EMR in the exam room since 1997, wrote in June 2010 that the common complaint is that physician and patient “are not able to have eye contact,” and that “it is very possible for the time with patients to be at least a good now with the EMR,” in “How to talk to patients with an EMR in the room.” Philip Green, MD, an emergency physician, wrote in March 2015 of nurses and doctors “with their eyes chained to the screens and a scowl on their faces while they click and type,” and of a timer on his screen reporting him four minutes past the average throughput time, in “EMR horror: Please choose one.”

Edwin Leap, MD, wrote in April 2016 that “it takes about one hour of looking at that screen for me to become exhausted,” and described “clicking and looking, scanning screens and logging on and off until 2 a.m.,” in “Working with today’s EMRs is mentally exhausting.” Suneel Dhand, MD, wrote in October 2016 that his goal every day is to “minimize my time staring at the screen and typing and clicking on a keyboard and mouse,” and that two minutes of preparation is worth ten minutes of turning around to type in front of a patient, in “The time doctors and nurses spend with EHRs is a national disgrace.” Hans Duvefelt, MD, a family physician, wrote in November 2017 that the click boxes physicians fill in print as “stilted English produced by the EMR,” and that anyone reading such notes “would probably just as soon see the original click boxes,” in “EMRs are robbing physicians of their writing skills.” Michael Kirsch, MD, a gastroenterologist, wrote in August 2019 that the record “cannot replicate the experience of pen & paper when physicians could use eye contact,” facial expression, and nodding of the head, in “How have EMRs changed the doctor-patient relationship?” Fred N. Pelzman, MD, an internist, wrote in August 2019 of a note in which “the most important information is almost lost” under “so much noise, so much excessive clicking,” and the drive to fill in all the boxes, in “Just because EMRs can document everything doesn’t mean they should.” Arthur Lazarus, MD, MBA, wrote in March 2026 that physicians “now spend hours each evening completing notes, clicking boxes, and reconciling fields,” and that “the keyboard has often replaced eye contact,” in “The hidden risks and rewards of AI scribes in medicine.”

Who do physicians say the electronic record was built for?

The most repeated claim on KevinMD about the electronic record is that it was designed for billing and regulatory documentation rather than for clinical care, and that the 2009 federal incentive program, meaningful use, locked that design in. Mentions of meaningful use peaked at 43 percent of posts in 2013 to 2016; mentions of billing or coding have stayed near half of posts since. The claim comes from physicians who wanted the technology and from the site’s longest-running health IT critic, who is not a physician.

The earliest posts are Kevin Pho’s, and they are not indexed: in January 2007 he wrote that the “holy grail of medicine” had been purchased at huge expense from vendors unwilling to work from a common language, and in October 2008 that most offices were still on paper despite the advantages. Rob Lamberts, MD, a primary care physician and early adopter, wrote in August 2012 that the incentive program had felt like validation, “EHR was so good that the government would pay doctors to adopt it,” and that “something funny happened on the way to meaningful use: I changed my mind,” in “My dream of universal acceptance of EHR has turned sour.”

Mike Koriwchak, MD, wrote in September 2010 that CPT codes “brought elaborate documentation requirements that medical records must fulfill in order to receive payment,” and that physician revenue is limited “by the sheer number of words one must write to properly document” the work, in “Physician resistance to EMR and why CPT should be replaced.” Margalit Gur-Arie, a health IT consultant, wrote in October 2010 that the government wanted physicians to buy a record so they could deliver care at lower cost, and that “there are no EHRs on the market today, that can actually deliver on all these goals,” in “Why doctors need an EHR, and why they should buy it now.” Rosemarie Nelson, a practice consultant, wrote in March 2011 that “Only when such processes are changed do you gain the real operational efficiency,” and that installing software alone returns little, in “Why the return on investment in an EHR is poor.” Adam Sharp, MD, wrote in February 2012 that “If EMRs are so great, why does the government have to essentially” bribe physicians to adopt them through incentives, in “Why EMR is a dirty word to many doctors.” Margalit Gur-Arie wrote in October 2013 that “EMRs are designed to minimize production costs and maximize revenues,” and that seven years was more than enough to refocus them on something other than billing, in “Why physicians will never learn to like EMRs.” Fred N. Pelzman, MD, wrote in June 2014 that his practice faced “another set of administrative hurdles, boxes that need to be clicked” to demonstrate meaningful use, in “The EHR should work for us, not us for it.” Hans Duvefelt, MD, wrote in April 2015 that “clinical notes were just that, clinical, not billing instruments or government compliance documents,” and that physicians now document “for the insurance companies, the government, the legal system and for all kinds of lay people,” in “With the EMR, you have to work backwards from the superbill.” David Mann, MD, wrote in October 2014 that in “an alternative universe where EHR systems were patient-centric instead of being designed to maximize patient billing,” a well-designed record “would be a joy to use,” in “Dissociate EMRs from billing. Here’s how to do it.” Danielle Ofri, MD, PhD, wrote in June 2020 that “diagnosis codes that are driven by billing requirements distort the diagnostic process and can lead to diagnostic error,” in “Does the EMR improve or worsen patient safety?”

Does the electronic record improve safety and quality? Physicians disagree.

Physicians on KevinMD split on this from the beginning. The case for is that a legible, searchable, shared record reduces the errors of paper and makes preventive care measurable. The case against, made with growing specificity since 2012, is that copy-and-paste produces notes no one reads, that alert systems fire so often physicians click through them, and that the studies promised to show quality gains did not. The vendors’ contracts drew their own criticism. Posts citing patient safety have held near 30 percent since 2009.

Robert Wachter, MD, wrote in October 2009 that most IT implementation contracts “insist on gag clauses for clinicians who identify errors caused by faulty software” and hold vendors harmless when an error harms a patient, in “Implementing an EMR or health IT system is harder than it looks.” Kevin Pho, MD, wrote in January 2011 that an Archives of Internal Medicine study of visits from 2005 to 2007 found no consistent association between electronic records and better quality on 20 indicators, and that modernizing paper charts “under the guise of improving patient care is somewhat misleading,” in “Electronic medical records and pay for performance don’t improve care.” Davis Liu, MD, answered in February 2011 that in 2009 only 4 percent of physicians had the comprehensive record he took for granted daily, and that the studies were measuring systems most physicians did not have, in “Why this doctor loves electronic medical records.”

Hans Duvefelt, MD, wrote in March 2017 that vendors “inserted these mandated items in sometimes very illogical places in the medical record,” where they serve billing “but not to document clinical thinking,” in “EMRs are dangerous. Let’s change that.” Margalit Gur-Arie wrote in November 2014 that the record’s dashboards treat reviewing a patient’s history as an administrative task, when “actively taking notes helps you synthesize information, internalize, memorize and understand the narrative,” in “EMRs remove the soul of the medical record. So, what’s next?” Elizabeth Hipp wrote in July 2012 that copying “pages of labs from an unrelated issue” into a note was the first way copy-forward harms care, in “How copy and paste in electronic medical records affects patient care.” Yul Ejnes, MD, wrote in October 2014 that customizing his system eliminated many clicks and that he was “better caught up with reviewing labs and communicating results,” in “The EHR report card 2014: Has it gotten better?” Robert Rowley, MD, a family physician who helped build an electronic record, wrote in September 2015 that he despised his EHR and was still using it, and that “while adoption is up, satisfaction has plummeted,” in “I despise my EHR. But I’m still using it.” Fred N. Pelzman, MD, wrote in September 2019 that the record helps when it lets a team care for patients “more efficiently, more collaboratively, and without a lot of clicking of boxes and excess busywork,” created by someone who does not understand the work, in “How EMRs can actually help physicians.” Scott Keeney, DO, wrote in January 2019 that alerts should be reserved for serious interactions and safety, and asked how to avoid “a computer questioning our clinical judgment and giving us extra clicks,” in “How EMR alert fatigue overwhelms physicians.” Danielle Ofri, MD, PhD, wrote in June 2020 that copy-and-paste yields notes “that resemble those online terms of service agreements,” and that “by far, the biggest landmine for medical error and patient harm is the automated alert system,” in “Does the EMR improve or worsen patient safety?”

Do scribes fix it?

Physicians on KevinMD began writing about hiring a person to type for them in 2010, within a year of the incentive program, and the posts since divide into three positions: scribes restore the visit, scribes are an expensive workaround for a record that should not need them, and scribes raise questions of privacy and of who is really writing the note. Posts mentioning scribes rose from 6 percent in 2009 to 2012 to 34 percent since 2021, by which point the scribe was usually software.

Douglas Perednia, MD, wrote in August 2010 that “If you’re a doctor who sees an average of 30 patients per day,” a scribe would take almost 10 percent of gross practice revenue, in “Doctors hiring scribes because of electronic medical records,” and in November 2010 that a scribe company estimated physicians could see eight additional patients over a ten-hour shift, “hiking Medicare revenues alone by $91 an hour,” in “Scribes lead to unintended consequences from electronic medical records.” Shadowfax, MD, wrote in December 2012 of his first months with a scribe that “Almost never has the presence of the scribe occasioned any further comment or discussion,” in “My experience with a scribe in the emergency department.” An anonymous scribe wrote in March 2014 that scribes “are purported to decrease physician burnout considerably,” and described the billing pressures visible from the scribe’s chair, in “The disturbing confessions of a medical scribe.”

Joel Sherman, MD, wrote in April 2014 that “Scribes are a possible solution to the imposition of EMRs whose benefit to the patient and practice are frequently unclear,” or negative to begin with, in “Do medical scribes threaten patient privacy?.” Jim Pagano, MD, wrote in August 2014 that “scribes put some humanity back into the practice of medicine,” in “Scribes put humanity back into the practice of medicine.” Suneel Dhand, MD, wrote in November 2015 that physicians spend “as little as 10 percent of their day” in direct patient care and that “5 minutes with a patient and then 30 minutes documenting it on a computer” is the problem scribes solve, in “Why medical scribes are the ultimate answer to our health IT woes.” Torie S. Sepah, MD, wrote in September 2019 about a remote human scribe speaking through a desk speaker who was “quickly becoming the most important person” in his life, virtual or otherwise, in “Virtual scribes are game-changers for physicians.” Janet Tamaren, MD, wrote in May 2022 that with 20 or more patients a day documentation “could easily take another two hours,” and that a scribe is worth the expense for a heavy load, in “To scribe or not to scribe? That is the question.”

What do physicians say about AI scribes?

Since 2024 the scribe on KevinMD is ambient AI software, and physicians describe it as the first tool that has given time back rather than taken it. They also describe the new risks in unusual detail: notes that capture what a physician would have filtered out, technical failures that leave no note at all, a consent checkbox that is not consent, and audit rules that treat a note signed seconds after the visit as cloned documentation. AI or ambient tools appear in 56 percent of posts since 2021, against 16 percent in 2017 to 2020. The measured benefit is smaller than the pitch, and by September 2026 the question had become whether the minutes saved turn into anything a practice can bill.

David Canes, MD, a urologist, wrote in February 2025 that AI notes “work beautifully 99 percent of the time,” that the failures leave the note that never was, and that the software “captures what you would have filtered out,” in “Why AI scribes are changing medicine and the hidden risks you must know.” Ashten Duncan, MD, a family physician, wrote in December 2025 that before the tool, giving patients full attention meant “many hours of charting that I would have to do without pay after work,” and described the first visit in which he “nervously clutched onto my scrap paper and clipboard” and launched the app, in “How an AI medical scribe saved my practice.” Arthur Lazarus, MD, MBA, wrote in March 2026 that one implementation across more than 2.5 million encounters reported “saving physicians an estimated 15,700 hours of note-writing in a single year,” in “The hidden risks and rewards of AI scribes in medicine.”

Junaid Niazi, MD, MPH, wrote in August 2026 that roughly a third of clinicians now have access to an ambient scribe, and that a study reported by STAT found adopters saved about 16 minutes of documentation time and 13 minutes of total EHR time per eight hours of patient care, against a pitch that runs “Talk to your patient, walk out with a finished note, get your evenings back,” in “AI scribes save 16 minutes, not your evenings.” Erin J. Silvertooth, MD, a psychiatrist, wrote in July 2026 that the consent conversation “can derail the opening of the appointment, the minutes where the patient decides how much of themselves to bring,” into the room, in “Your ambient AI scribe consent checkbox is not consent.” Kaif Ruman wrote in August 2026 that thin ambient notes are being flagged in the 2026 CMS audit cycle as cloned documentation, with recoupments averaging $2,400 a month per provider, and that a note signed within seconds of the visit is used by auditors “to argue that no meaningful clinical review occurred,” in “Why ambient AI scribe notes fail the 2026 CMS audit.” Lynn McComas, DNP, ANP-C, wrote in March 2026 that “the note has become so overloaded that it often competes with the patient for the clinician’s attention,” and asked whether the time saved could go to teaching, in “Can AI scribes give clinicians time to teach again?” Karan Kanwar wrote in September 2026 that in a May 2026 MGMA poll of practice leaders “fewer than half, 46 percent, will say it’s made providers more productive,” that an NEJM AI editorial concluded “AI scribes are not productivity tools yet” with savings of “roughly 7 to 22 minutes per provider a day,” and that in a 150-day primary care study billed wRVUs barely moved, in “Why haven’t ambient AI scribes boosted productivity?”

How has the electronic record conversation changed since 2004?

The corpus shows three periods. From 2004 to 2008 the question was adoption: whether to buy, what it would cost, whether it would save money. From 2009 to 2016, the HITECH and meaningful use years, KevinMD published 37 to 47 posts a year on the record, and the subject was the incentive program, the vendors, and the discovery that the software served billing. Since 2017 the volume has fallen to a dozen posts a year and the subject has changed: the record itself is settled, and physicians write about its cost to them, “pajama time,” burnout, and now the AI tools sold as the way out. Mentions of burnout went from 1 percent of EHR posts in 2009 to 2012 to 41 percent since 2021.

Term 2004 to 2008 (80 posts) 2009 to 2012 (143 posts) 2013 to 2016 (128 posts) 2017 to 2020 (61 posts) 2021 to 2026 (59 posts)
Meaningful use or HITECH 5 percent 38 percent 43 percent 18 percent 17 percent
Billing or coding 5 percent 28 percent 54 percent 33 percent 49 percent
Vendor 5 percent 32 percent 32 percent 20 percent 24 percent
Alerts or clicks 1 percent 31 percent 47 percent 51 percent 41 percent
Screen, eye contact, or looking at the computer 6 percent 23 percent 38 percent 44 percent 47 percent
After hours, inbox, or pajama time 0 percent 19 percent 27 percent 49 percent 44 percent
Patient safety or error 5 percent 30 percent 34 percent 30 percent 31 percent
Copy-and-paste or note bloat 2 percent 5 percent 9 percent 5 percent 12 percent
Burnout 0 percent 1 percent 5 percent 26 percent 41 percent
Scribe 0 percent 6 percent 12 percent 21 percent 34 percent
AI or ambient 0 percent 1 percent 2 percent 16 percent 56 percent

The earliest posts are Kevin Pho’s, and they are not indexed: skepticism of vendors in January 2007, a poll on whether records would save money, and in October 2008 the observation that adoption remained distressingly low despite the advantages. In 2009 Robert Wachter, MD, warned that implementation was harder than it looked and that vendor contracts gagged clinicians who found errors. In 2010 Douglas Perednia, MD, noticed physicians hiring people to type. From 2012 to 2014 the meaningful use posts arrived from both directions: Rob Lamberts, MD, changing his mind, Adam Sharp, MD, on bribes, Margalit Gur-Arie on billing design, David Mann, MD, on dissociating the record from billing, and Yul Ejnes, MD, grading it. Philip Green’s 2015 “please choose one” and Edwin Leap’s 2016 exhaustion are the turn toward the physician. Danielle Ofri, MD, PhD, summarized the safety case in 2020. From 2024 the record is AI scribes, and the last four posts in the corpus, from July to September 2026, are about consent, minutes saved, audits, and productivity.

The KevinMD electronic health record corpus by the numbers

The figures below describe the set of KevinMD posts this page draws on, as of September 17, 2026. They are counts of what KevinMD has published, not survey data about physicians.

Measure Value
Posts in the corpus 471
Date range June 23, 2004 to September 17, 2026
Named contributors 207
Posts bylined by an MD or DO 270
Posts under Kevin Pho’s byline, mostly short commentaries from 2004 to 2011 120
Posts in the corpus that are not indexed, counted but not cited 108, of which 103 are Kevin Pho’s short commentaries or pre-2009 posts and 5 are short contributor posts or podcast episodes without transcripts
Posts with scribe in the title 37
Podcast episodes with full transcripts 12
Total words About 356,000
Peak year 2014, 47 posts
Posts a year since 2020 5 to 14
Most frequent contributors Margalit Gur-Arie (16); Fred N. Pelzman, MD (14); Hans Duvefelt, MD (11); Rosemarie Nelson (10); David Mann, MD (8); Rob Lamberts, MD (8)

How this page was built and how it is updated

The corpus was assembled by searching KevinMD post titles for EHR, EMR, electronic health record, electronic medical record, and scribe, and removing the false matches for the last term. Term frequencies were computed against the full text of each post on the corpus assembled for the September 17, 2026 build; this page was converted to the current standard on September 20, 2026 without rebuilding that corpus, so the counts and figures are as of September 17, 2026. Posts are counted whether or not they are indexed; citations on this page are limited to posts that are indexed and can be verified at the link. One hundred eight posts in the corpus are not indexed, Kevin Pho’s short commentaries and every post published before 2009, plus a few short contributor posts and untranscribed podcast episodes, and they are counted but not cited. Quotations are taken verbatim from the original posts. Author credentials are as they appeared in the byline at publication. Every source is linked in the sentence that cites it, and the full list appears at the end of the page. The complete feed is at the Health IT and AI in Medicine archive. Related records: Physician burnout: what physicians say, in their own words, Artificial intelligence: what physicians say, in their own words, Prior authorization: what physicians say, in their own words, Medical errors: what physicians say, in their own words, and Medical malpractice: what physicians say, in their own words.

This page is updated as new essays on the electronic health record and AI scribes are published on KevinMD. When it is updated, the date at the top changes, the counts in the tables are recomputed, and new named claims are added to the relevant section. Nothing is removed unless the original post is removed. An author who believes a quotation on this page misrepresents them can write to Kevin Pho and the page will be corrected.

To cite this page: Pho K. The electronic health record: what physicians say, in their own words. KevinMD.com. Updated September 20, 2026. https://kevinmd.com/electronic-health-record

The 44 KevinMD posts cited on this page, in order of publication