Every day it seems we are alerted to more incidents of medical record errors; in 2014, it was reported that preventable medical errors are the number 3 killer in the U.S. and that 400,000 people are lost each year to these. That is certainly not an insignificant number. What are these errors? The Journal of Patient Safety suggests that Preventable Adverse Events (PAEs) “may be separated into these categories:
• Errors of commission
• Errors of omission
• Errors of communication
• Errors of context
• Diagnostic errors¹

On September 22, 2015, the National Academies of Sciences, Engineering, and Medicine published a report through the Institute of Medicine detailing how diagnostic errors, the largest category of medical errors, can be prevented, and that “urgent change is warranted to address this challenge.² So what is being done and/or proposed to prevent errors? The NASEM recommends 8 goals for improving diagnosis in healthcare. One of these goals is to “Develop and deploy approaches to identify, learn from, and reduce diagnostic errors and near misses in clinical practice.”
Some say the EHR is to blame for these errors. In an article published by The Atlantic, Dr. R. Gunderman states that “A recent study at Johns Hopkins University indicated that hospital interns… spend only about 12 percent of their time interacting with patients. By contrast, they spend 40 percent of their time…interacting with hospital information symptoms. The flesh-and-blood patient is getting buried under gigabytes of data.³

In the same article, Dr. Gunderman relates a story of a newly admitted patient whose intern reported that the patient was “status post BKA (below-knee amputation).” When questioned about the patient, it was discovered on exam that the patient’s extremities were completely intact. Apparently, this patient’s chart had been produced by a speech recognition system and had changed DKA (diabetic ketoacidosis) to BKA—four hospital admissions earlier! How could this happen? Simple: The “new technology” of SR is inherently flawed and can never be replaced by human eyes. The report from NASEM mentions nothing about SR errors.

Speech recognition is not a new technology. Developed in the 1950s, it has, in fact, been in widespread use since the 1990s with the advent of Dragon Dictate. SR was then believed to be an answer to the cost of medical transcription. It was thought that costs could be cut because medical transcriptionists would be editing the documents for errors in SR, instead of transcribing, and that this would take a lot less time. Pay for MTs then was cut in half for editing, despite the fact that some reports were/are so poor that it would take less time to transcribe them from scratch. A provider using SR can edit their own notes; however, this takes precious time away from patient care. If a speech-generated document is allowed to be part of the permanent medical record without editing, chances are high that it will contain errors. This is not the solution to documentation errors.

What are some solutions to these errors? Traditional medical transcription is one. MTs can be the answer to front-line error capture, whether in SR-generated documentation editing or full-document transcription. It is well known among MTs that providers struggle occasionally with dictation, and MTs are professionals who can spot discrepancies and recognize “guesses” (yes, it does happen, too often) at medication and other pronunciations. MTs are trained and experienced experts in creating healthcare documentation. Cons: Traditional transcription is dictated by the provider, and the voice file is uploaded; the documentation may take up to 72 hours to be created and finalized.

What about medical scribes? Scribes can also be a great solution to both error prevention and capture. A recent error found in a medical record could have been caught by a scribe: A patient with simple hypothyroidism was given an incorrect diagnosis code for “Congenital Hypothyroidism with Diffuse Goiter.” The patient noted this error upon reviewing her electronic summary post visit. However, a scribe would have seen and questioned this diagnosis error at the point of care before it became a part of the patient’s permanent electronic record.

Traditionally, scribes are pre-med students who have had some training in producing healthcare documentation. With both classroom and on-the-job training, it can take 6-12 months for a traditional scribe to be up to speed and able to solo scribe without additional training. After the scribe has completed his/her onsite training, it is then up to the providers to continue teaching the scribe. Since a scribes on a pre-med track may be taking advantage of this opportunity during his or her “gap year,” at the end of 12 months the scribe may be leaving for school.

However, at AHDPGTM, in addition to a traditional scribe program, we have an exciting new opportunity for Allied Health Professional scribes. These candidates may be experienced medical assistants, medical transcriptionists, etc., who undergo our expedited program and then are ready to scribe without requiring any additional training time from providers. Thus, you have a professional set of expert eyes to create and edit documentation free from errors. These unique scribes are career employees who are not on track to enter medical school, and you can count on them to be there consistently for your healthcare documentation needs.

Medical errors are preventable with good stewardship of electronic health records; however, an experienced set of human eyes is key to critical thinking and judgment for better prevention of errors.

1. James, PhD., John. "A New, Evidence-Based Estimate of Patient Harms Associated with Hospital Care." Journal of Patient Safety 9.3 (2013): 122-28. Journal of Patient Safety. Wolters Kluwer. Web. 12 Oct. 2015. <http://journals.lww.com/journalpatientsafety/fulltext/2013/09000/a_new,_evidence_based_estimate_of_patient_harms.2.aspx>.
2. "Improving Diagnosis in Healthcare--Quality Chasm Series." Institute of Medicine (National Academies of Science, Engineering, and Medicine), 22 Sept. 2015. Web. 12 Oct. 2015.
3. Gunderman, MD, PhD, Richard. "The Drawbacks of Data-Driven Medicine." Atlantic 5 June 2013. Print.