It addresses the dangers of a common feature of EHR's used recklessly: copy-and-paste.
EHRs: “Sloppy and paste” endures despite patient safety risk
Copying and pasting information is common within EHRs, but the practice sometimes can lead to confusion and endanger patient care.
By Kevin B. O'Reilly, amednews staff. Posted Feb. 4, 2013.During the winter holidays, a patient at Yale-New Haven Hospital in Connecticut had a large pressure ulcer with an abscess. A surgical intern made a note in the patient’s electronic health record that said, “Patient needs drainage, may need OR.”The problem? The same note appeared for several consecutive days, even after a surgical team successfully drained the abscess. The intern had copied and pasted the previous day’s note, but failed to appropriately update it to reflect the fact that the drainage was done. The note confused the consulting infectious disease team and nearly led to an unneeded change in the patient’s antibiotic regimen.
The practice of carelessly copying and pasting previous information, often dubbed “sloppy and paste,” is on the decline at Yale-New Haven Hospital but is widespread across medicine and can lead to mix-ups that sometimes harm patients, research shows.
“It’s especially problematic when you have multiple teams taking care of the patient and we’re communicating through the chart, which happens very often nowadays because physicians don’t see each other as often as we used to,” said Dr. Horwitz [General internist Leora Horwitz, MD], assistant professor of medicine at Yale University School of Medicine. “We do rely on the chart in many cases, and it can lead to genuine confusion.”
When you rely on an information system and the information system contains incorrect information (for whatever reason), patients are put at risk. That the systems are implemented without simple controls on copy-and-paste (such as permanently embedding substantive metadata in the output) is a significant flaw.
From Sec. II of Aguilar v. Immigration and Customs Enforcement Div. of U.S. Dept. of Homeland Sec. 2008 WL 5062700 (S.D.N.Y. Nov. 21, 2008), available at this link:
... Substantive metadata, also known as application metadata, is “created as a function of the application software used to create the document or file” and reflects substantive changes made by the user. Sedona Principles 2d Cmt. 12a; Md. Protocol 26. This category of metadata reflects modifications to a document, such as prior edits or editorial comments, and includes data that instructs the computer how to display the fonts and spacing in a document. Sedona Principles 2d Cmt. 12a. Substantive metadata is embedded in the document it describes and remains with the document when it is moved or copied. Id
Microsoft Word's "Track Changes" feature is an example of substantive metadata being displayed.
The available stats on the phenomenon are of great concern:
... A study in February’s Critical Care Medicine found that copying and pasting is the rule in EHRs rather than the exception.
Using a software program that can detect identical matching word sequences, researchers examined the assessment-and-plan portions of more than 2,000 progress notes for 135 patients created by 62 residents and 11 attending physicians working in a Cleveland medical intensive care unit. For the residents, 82% of the notes contained 20% or more copied text, while 74% of attending doctors’ notes also exceeded that rate of copying and pasting.
A similar study in the January-February 2010 issue of Journal of the American Medical Informatics Assn. found a copy-and-paste rate of 78% in sign-out notes generated by internal medicine residents. The rate of copied text in progress notes was 54%, the study said.
... Other times, patients are harmed. In a July/August 2007 case study in AHRQ WebM&M, an online patient safety journal, William Hersh, MD, described the case of a 77-year-old woman hospitalized for diarrhea and dehydration after chemotherapy.
An intern noted that the patient would receive heparin to prevent venous thromboembolism. The note was copied and pasted for four days in a row and signed by a resident and an attending physician, who appeared to believe the heparin had been ordered and administered. Ultimately, the patient was discharged without ever receiving the preventive medicine and two days later was rehospitalized and diagnosed with a pulmonary embolism. Only then did physicians realize the patient never got the correct prophylaxis.
“The problem is getting worse now with the rise of EHRs,” said Dr. Hersh, professor and chair of the Dept. of Medical Informatics and Clinical Epidemiology at Oregon Health & Science University in Portland.
HHS OIG (Office of the Inspector General) announced that it plans to review multiple EHR notes for the same patient by the same physician to see whether doctors are copying and pasting the identical note from visit to visit. The practice is sometimes called cloning and could be implicated in fraudulent coding and billing practices.
John Halamka, MD, calls for a more radical fix.
“The way we document in medicine has grown up over decades for medical reasons, for billing, for medical-legal justification,” said Dr. Halamka, chief information officer at Beth Israel Deaconess Medical Center in Boston. “You wind up with 17 pages of replicated and duplicated and challenging-to-read documentation. I propose we blow up the way we do documentation altogether and replace it with a Wikipedia-like structure.”
Such an approach would allow physicians to edit the progress note collaboratively, just as the popular open-source encyclopedia is updated. Dr. Halamka hopes to pilot-test the idea within the next year. “With that concept, you wouldn’t ever really need to copy and paste,” he said.
An interesting concept and experiment. My questions:
- Do we first rigorously investigate and understand the causes of the copying, i.e., cryptic and difficult-to-use data entry methods that significantly slow clinicians down?
- Do we get informed consent from patients for the experiment?
- If so, what do we tell them? We are conducting an experiment in charting, risk unknown, to solve cheating and risks due to poorly-designed EMRs?
- Would not simpler solutions (such as the embedded-metadata identifiers indicating text has been copied as I described above) be important to implement first, before experimenting with medical documentation?