Home  |   Subscribe  |   Resources  |   Reprints  |   Writers' Guidelines

E-News Exclusive

Speech Recognition Dos and Don’ts

By Deb Palmer, CHDS; Christina Joseph; Sara Berry, CHDS; and Bonnie Jean Rieber, CHDS

Speech recognition (SR) was introduced with the intention of helping to eliminate human oversights and reduce critical patient safety errors. However, for this to happen, the SR editor must ensure an accurately edited document.

Throughout our editing careers, we have discovered a number of dos and don’ts that we follow when editing SR documents.

1. Do ensure you have the necessary tools needed to perform the job correctly. Ultimately, ensure that you have a comfortable noise-canceling headset with built-in volume control. Your workspace should be ergonomically correct. Your monitor should be 16 to 24 inches from your face. Adjust your sitting position to suit your body. If you have a stand-up desk, ensure the keyboard is at the correct height and you are not looking down or up at the screen. Using a foot pedal can save time and keystrokes.

2. Do ensure that you have proper lighting in your work area to alleviate eye strain. As well, ensure your working environment is quiet and free of distractions, allowing you to focus on the task at hand. Turn off cellphones and social media. Distractions will affect your ability to concentrate, which, in turn, could potentially cause you to miss errors in the SR draft.

3. Do take breaks. Set a timer for every 30 minutes to one hour to remind you to take a few minutes to stand up and stretch.

4. Do synchronize your cursor with the audio file. By doing this, the text cursor will follow along, word by word, with what is being dictated.

5. Do become familiar with keyboard shortcuts for the platform you are working on. Use your keyboard shortcuts to quickly maneuver through the report. However, make sure you build your repertoire gradually, starting with those most commonly used. Attempting to learn too many shortcuts at once could cause you to feel overwhelmed.

6. Do implement a three-pass method. Firstly, listen to the report in its entirety, only editing obvious errors. By correcting these before you begin fully editing the document with audio, you will save time, as you won’t need to stop and start the audio as often. Secondly, go through the report editing as much as you can and researching where necessary. Lastly, go through the report one last time, checking for continuity (does the report flow correctly), context (does it make sense), and completeness (is there any missing or added text).

7. Do proofread for accuracy. Actively listen to what the dictator is saying without reinterpreting it in your mind to what you think the dictator intends to say. This can be a form of distraction and can cause you to overlook homonyms and other more critical medical words that sound alike. Proofread to ensure the document is contextually sound. We must not assume that medical words are correct or that abbreviations have been expanded correctly. Critical thinking and researching play a big role in an editor’s job. Proofread your report for proper spelling, capitalization, grammar, punctuation, and formatting. Reading the report aloud, if possible, is very helpful.

8. Do make sure that your eyes are really seeing what your ears are hearing. There can be a tendency for our eyes to see what should be there rather than what is actually dictated.

9. Do correct the SR text whenever possible instead of deleting the entire report and typing from scratch. If you start over, the SR engine will not “learn.”

1. Don’t trust that the SR engine knows more than you or that the dictator has dictated spelling, formatting, or punctuation correctly. Always research unfamiliar or questionable text. Some dictators speak faster than others, and it might be difficult to make out what was said. Ensure you have removed any erroneous text and added any missed text.

2. Don’t assume the document has been formatted according to the client’s profile. Each facility will have its own rules, and we cannot always count on the SR software to follow them.

3. Don’t trust the numbers, especially for medication dosages, lab values, and vital signs. A dose of 16 mg sounds very close to 60 mg and would be an easy mistake for an SR engine to make. Do not assume the medication dosage in the report is correct. Verify the validity with a credible source.

4. Don’t assume the demographic data have been recorded correctly. It is very important to ensure all names, dates, medical record numbers, etc, are accurate.

5. Don’t be afraid to flag something in the report. Prefixes such as hypo-/hyper- and intra-/infra- can be easily confused in the SR draft. Use critical thinking to determine the correct word based on the context of the report. If it is unclear and you are unable to determine the correct word, flag it. Incorrect text could jeopardize patient safety or alter treatment.

6. Don’t expect to be instantly more productive. Editing SR involves a different set of skills, and it may take some time to develop these.

7. Don’t forget to increase or decrease the audio speed and/or volume if you are having trouble hearing what is being dictated. If this does not help, you may find that taking a short break will help clear your mind and allow you to come back and hear words more clearly.

Deb Palmer, CHDS, is instructional department manager at CanScribe Career College and 2019 AHDI Educator of the Year. Christina Joseph; Sara Berry, CHDS; and Bonnie Jean Rieber, CHDS, are instructors at CanScribe.


2021 State of the Healthcare Documentation Industry Survey

The Association for Healthcare Documentation Integrity (AHDI) and Gary David, PhD, CCS, of Bentley University, have teamed up once again to better understand the current and future states of the health care documentation industry.

The 2007 Survey of Medical Transcriptionists aimed to quantify the characteristics and experiences of the medical transcription professional. It looked at medical transcriptionist demographics, work experiences, experience with technology, and the issues most important to transcriptionists at the time. The landscape of medical transcription was then completely revolutionized by widespread implementation of the EHR, speech recognition, and digital technologies. The 2011 EHR Future Roles Survey was performed in response to those changes. It sought perspective into the new and expanding roles and responsibilities of health care professionals involved throughout the EHR process life cycle. Today, we recognize all those who contribute to the quality and integrity of a patient’s medical record—regardless of their specific role or location.

Your thoughtful responses to AHDI’s 2021 State of the Healthcare Documentation Industry survey will be invaluable to us as we endeavor to serve and support health care documentation professionals in the years ahead. There are no personal identifiers associated with your answers, and your participation is anonymous and confidential.

Thank you in advance for your time and contribution to the future of AHDI. You may access the survey until November 1 at bit.ly/38WHUsJ.

Source: Association for Healthcare Documentation Integrity