Physicians and EHR Documentation strategies

From Clinfowiki
Revision as of 20:11, 28 April 2015 by Jaucoin (Talk | contribs)

Jump to: navigation, search

Introduction and background

Electronic health records (EHR) have a low penetration rate into health care delivery. Input by the physician provider remains one of the largest obstacles. Low utilization is partly due to the difficulty of capturing data in structured format. Also, physicians are trained and used to telling the story of the patient encounter in a narrative format, and then developing the justification for care within or as a conclusion of that narrative. In addition, physicians are pressed for time and will not assume the burden of data entry without significant returns for their efforts. This wiki introduces some of the many methods available to address the barriers that face the physician in completing documentation of patient encounters. Each topic could be a wiki unto itself.

Free text

Free text represents unstructured documentation unless successful Natural Language Processing (NLP, discussed below) can be utilized on the free text. Digital free text entry typically means keyboarding for small passages and dictation for larger passages. Care providers are inefficient at being able to keyboard significant amounts of information and/or will not perform this function reliably. Free text by keyboard input was a documentation necessity in the early renditions of the EHR. Voice recognition is another alternative for documentation of free text and is addressed below. Free text by traditional dictation has the disadvantages of cost, time until the dictation is transcribed and available to other providers, and the lack of providing any discrete data.

Copy and paste

Initially, the word processing tool of copying and pasting, as well as using templates, aided the transition to use of the EHR for electronic documentation. This also was noted to lead to “lengthy, hard-to-read records stuffed with data already available on line.” (1) This author completed a VA study of Computerized Patient Records System (CPRS) charts that showed that nine percent of progress notes they inspected contained copied or duplicated text, and that high-risk author copying occurred once in every 720 notes, and one in ten electronic records contained an instance of high-risk copying. A series of 11 suggestions were made on a ways to decrease the risks associated with copy and paste. Another author has noted that “…the use of NLP may allow a much more precise method to carry text forward from previous notes, which may alleviate some problems caused by uncontrolled cutting and pasting of text.” (2)

Structured documentation (aka structured data entry)

Structured documentation “is a note-writing method that provides the ability to share information between automated systems and to retrieve information easily after it is stored” (3). Most structured documentation format is now in the form of click-select data and facilitates the collection of discrete data related to each patient interaction. The structured data is used to facilitate reporting, decision support applications, quality reporting processes, billing, and patient-oriented clinical research. Most EHR programs will utilize the selected data points to construct textual phrases into an acceptable medical syntax. Systems which allow the addition of free text blocks may be referred to as “semi-structured” documentation. The disadvantage of free text blocks are that they cannot be utilized as discrete data without the use of NLP, and the physician may elect to put what should be discrete data into text block format. Constructing notes from structured or semi-structured EHR programs is typically slower than previously utilized methods (4), and is a significant departure from the usual patterns of dictated or written clinician note entry.

Macros

Within the field of computer science, a “macro” indicates a single, user-defined command that is part of an application and executes a series of commands. Within informatics, it typically means a combination of keys or keystrokes that will generate a user predefined text set, typically a phrase which the clinician is used to using within his written or dictated notes that has a particular meaning. For example, a physician who supervises resident physicians might wish to have a macro that results in “The patient was originally seen by the resident physician, Dr. Xxxxx. I have independently interviewed and examined the patient and have confirmed those findings with the following additions:” The physician using this macro would fill in the “Xxxxx” with the correct name, and then document any additions or changes to what the resident physician had documented. The macro is useful for a shortcut of lengthier text or dictated entry.

Medical scribes

Medical scribes are non-clinical physician documentation assistants who work real time during patient encounters. By obtaining or retrieving parts of the history, transcribing the exam, assessment and plan as it occurs, documenting procedures and checking on progress of ordered tests, the physician is freed from non-productive time in their patient care cycle. Although scribing was created to deal with the need to satisfy CMS’ documentation requirements in the age of paper and pen, they are also a natural answer to adaption to the EHR. Medical scribes, typically pre-med students, are more adaptable to computer use and can be helpful in bridging the physician with their EHR.(5)

Voice recognition

Speech or Voice Recognition software (VRS) has been developed and available for individual use since the 1980’s. Significant advances in the accuracy and usability of VRS have resulted from increasing computer processing power, memory and storage expansion, and the move from a discrete speech recognition to continuous speech recognition. Today’s medical VRS products advertise a 98% accuracy rate and reduced to no training time of the software engine. Besides the obvious value of immediate spoken word to text function, VRS also provides for verbal commands for insertion of templates and macros and navigation through screens of the EHR. Barriers to the use of VRS are license purchase, learning time required to understand and work with the commands of the program, and the need for the clinician to spend edit time after dictating his passage. The product lends itself to documentation by insertion of sentences or short paragraphs that augment an EHR’s structured documentation format, allowing collection of discrete data but with the physicians “story” being relayed within the same document.

Embedded dictation

To address the barriers to physician adaption of VRS, Cerner Corporation (developer of PowerChart, an EHR), has utilized back end VRS embedded on voice clips that are captured as part of the structured document. During the creation of the “point and click” discrete data entry, the physician can elect to insert a dictated sentence(s) or paragraph(s). The voice clip will exist within the document as a voice file which is converted to text by a medical transcriptionist utilizing server based VRS. Until the voice clip is converted to text, an icon remains in the document which can be listened to by readers of the document. This eliminates barriers of physician adaption to front end VRS as well as the disadvantage of delay to information availability due to turn around time of traditional dictation transcription.

Natural Language Processing (NLP)

NLP (also has been referred to as MLP for Medical natural Language Processing) is the newest model of potential input of structured data. The goals of NLP are "to receive spoken sentences, understand them, parse them into concepts, and then store them in the appropriate place in a database." (3) The attractiveness of this method is the ability to convert a physician dictated note, which is the current physician preferred method of unstructured documentation and convert it into structured documentation with all the benefits of same. This also results in the ability for a clinician to search the patients EHR documents for specific topics much in the way that Google searches the web.

References

1. Hammond KW, Helbig ST, Benson CC, Brathwaite-Sketoe BM. Are electronic medical records trustworthy? Observations on copying, pasting and duplication. AMIA Annu Symp Proc. 2003:269-73.

2. Johnson SB, Bakken S, Dine D, Hyun S, Mendonça E, Morrison F, Bright T, Van Vleck T, Wrenn J, Stetson P. An electronic health record based on structured narrative. J Am Med Inform Assoc. 2008 Jan-Feb;15(1):54-64. Epub 2007 Oct 18

3. Zinder DJ. Structured documentation. Otolaryngol Clin North Am. 2002 Dec;35(6):1211-21.

4. Payne TH, Perkins M, Kalus R, Reilly D. The transition to electronic documentation on a teaching hospital medical service. AMIA Annu Symp Proc. 2006:629-33.

5. Scheck, A. The next big thing: Medical Scribes: Scribes push emergency medicine closer to adoption of electronic medical records. http://www.em-news.com/pt/re/emmednews/abstract.00132981-200902000-00015.htm;jsessionid=KhQCJvPYFJhnYGqW85FztvyGx5T2xph9ZXV4hxy7PvyJ20Vl1F9k!-1775402713!181195628!8091!-1

6. Devine EG, Gaehde SA, Curtis AC. Comparative evaluation of three continuous speech recognition software packages in the generation of medical reports. J Am Med Inform Assoc. 2000 Sep-Oct;7(5):462-8

7. Kashyap V, Turchin A, Morin L, Chang F, Li Q, Hongsermeier T. Creation of structured documentation templates using Natural Language Processing techniques. AMIA Annu Symp Proc. 2006:977.

8. Morrison FP, Li L, Lai AM, Hripcsak G. Repurposing the clinical record: can an existing natural language processing system de-identify clinical notes? J Am Med Inform Assoc. 2009 Jan-Feb;16(1):37-9. Epub 2008 Oct 24.

9. Guglielmo WJ. What a scribe can do for you. http://medicaleconomics.modernmedicine.com/memag/article/articleDetail.jsp?id=278018

10. Beats, J. et al. (2003). Speech Recognition in the Electronic Health Record (AHIMA Practice Brief). Web site: http://library.ahima.org/xpedio/groups/public/documents/ahima/bok1_022192.hcsp?dDocName=bok1_022192

11. Kliner, DJ. Improve your ICD-9 coding with voice-recognition macros. http://www.aafp.org/fpm/20060600/39impr.html

12. Multi-modal entry for electronic clinical documentation. http://www.freshpatents.com/Multi-modal-entry-for-electronic-clinical-documentation-dt20080724ptan20080177537.php

Submitted by C. Blake Schug, M.D.