Speech Recognition in Healthcare: Dictation and Transcription
The article provides information on various applications of speech recognition technology in healthcare and the benefits it brings.
The development of speech recognition technology allows dictation and transcription programs to be used in various areas of human activity where it is imperative to reduce the time for documenting personnel activities.
One of these areas is healthcare, with its vast documentation flow. The creation and maintenance of medical records are one of the most time-consuming areas in medical practice. Doctors, nurses, etc., must document their decisions and actions daily, sometimes hourly, using different journals where these decisions and actions are recorded. Therefore, medical personnel belong to the category that especially needs detailed and accurate documentation of their medical activities.
Registration and searching of the necessary information about patients can be performed faster with speech recognition than typing on a keyboard. Actions such as creating case histories, writing prescriptions, preparing recommendations, prescribing treatment, and searching information in a database are easily performed by voice commands.
- Documenting in Medicine
- Manual Input and Transcription in Healthcare
- Peculiarities of Medical Dictation and Transcription
- Speech Recognition (Dictation and Transcription) Speeds up the Documentation Process
- Medical Dictation and Transcription Market Size
- Summary
Documenting in Medicine
According to the Patients’ Rights Act of 02.26.2013 law, in Germany, medical personnel are obliged to document all of the essential information regarding the patient’s illness[1].
Medical documentation is intended to ensure data reliability on the patient’s treatment and the accurate reporting of treatment results.
Electronic storage of medical documentation gives the opportunity to provide better medical services, regardless of geographic distance or location.
Quick access (for example, via the Internet) to the patient’s treatment history (for example, notes on previous or current medications, notification of allergies, etc.) in some (especially critical) cases speeds up the decision-making process for helping a patient in a critical condition.
Detailed recordkeeping is necessary, especially when several doctors are treating a patient simultaneously; otherwise, there is a danger that necessary medical actions either will not be taken or will duplicate each other. Each doctor involved in the treatment of the same patient must carry out their own records.
Deficiencies in documentation may result in professional and legal sanctions against a doctor.
Manual Input and Transcription in Healthcare
Using the example of documenting the results of the medical survey, we consider how the documentation takes place.
The medical survey process consists of two stages:
- Examination of the patient, and
- Documenting the results of the examination.
Studies show that 44% of doctors spend their time on documentation. The doctor should focus on treatment, but from this point of view (the 2nd stage), documentation significantly reduces the doctor’s “useful” time.
A familiar doctor, who works in one of the clinics in Cologne, Germany, describes the process of examination as follows:
- Doctor dictates his considerations to dictaphones, which are available in each ward;
- Then his statements, in the form of audio files from dictaphones, are transcribed to text by a transcriber;
- The next day doctor receives the document as a text document;
- The doctor checks his transcribed documents for accuracy and correctness;
- After checking the text, the doctor saves the document in the main database.
Two points should be noted in this process:
- The need for a staff of transcribers, and
- A rather long time is required for transcription; therefore, the document workflow is significantly slowed down.
It turns out that documenting the results of an examination takes much more time than it does to perform the examination itself.
So, documenting “manually,” from the point of view of the speed of documentation workflow, is the “weak link” in medical practice:
- It takes almost half of the time of doctors for non-professional actions;
- Time-consuming and slow. Although using a transcriber saves the doctor’s time on documentation, this process does not speed up the process. With a one day turnaround time to transcribe a doctor’s recorded speech, there are still many issues, including:
- Expense. Documentation cost for hospitals is ~ 21% of the total staff costs;
- Privacy concerns. Transcription is outsourced to third-party countries (for example, India, Sri Lanka, Philippines, etc.) which do not have data protection rules;
- Need for a repository. For the storage of paper documents, there is a need for a physical repository, service personnel, and maintenance of environmental conditions for preservation.
Peculiarities of Medical Dictation and Transcription
Unlike other dictation and transcription systems, medical dictations and transcription abound with various medical terms that are not used in everyday speech.
So, speech recognition and its dictation and transcription functions, designed for medical use, should be able to recognize medical terminology.
If a medical transcriber accidentally printed the wrong medicine or an erroneous diagnosis, and if the doctor did not check the document for accuracy, the patient’s life could be in serious danger.
Both the doctor and the medical transcriber should ensure that the transcribed or dictated text is correct and accurate. The doctor should speak slowly and concisely while dictating medications or details of diseases and conditions.
A medical transcriber must have medical knowledge and good reading comprehension.
Even though some speech recognition systems initially include medical vocabulary dictionaries, medical speech recognition systems should allow the user to add new words to the dictionary, thus adapting the system to work in a specific medical department.
Speech Recognition (Dictation and Transcription) Speeds up the Documentation Process
The speech recognition system, and its dictation and transcription applications, greatly simplify and make the creation and maintenance of electronic medical records more efficient.
It is the rapid development of speech Recognition technology that contributes to its practicality and feasibility of use.
So, how does speech recognition speed up the documenting process? The answer to this question is simple: most people speak faster than they type on the keyboard. An experienced operator can print a 300-word message in about six minutes. By using speech recognition technology, time is reduced three times - to two minutes.
1. The doctor dictates to speech recognition system and sees the text immediately;
2. The doctor proofreads and, if necessary, edits the result instantly.
Product 1: Dictation App for Entering Text via Speaking to both Phone and PC
Installed on the smartphones of doctors and nurses, as well as all computers at the hospital;
The mobile app pairs with any PC at the hospital and allows the text to be entered on any program on PC (for example, MS Word) via speaking into the phone;
Allows notes to be taken with Speech-to-Text on the phone while on the go (because doctors are not at the office all of the time);
Allows transferring notes from phone to PC after returning to the office.
Product 2: Automatic Transcription for Converting Prerecorded Audio Files to Text
Useful for transcribing audio recordings of meetings with patients, conference speeches, etc.
Upload audio or video files to the website;
Wait for the automatic transcription produced by the speech recognition technology;
Edit the transcription quickly with a text editor, which automatically plays audio fragments for a selected sentence.
A working demo (not for healthcare, but general domain: newspaper, education, etc.) can be found at: https://voicedocs.com/en/transcriber
Medical Dictation and Transcription Market Size
Over the years, various approaches have been used to simplify and speed up the process of creating medical documents: from engaging operators for registration and documentation to using special online forms, where a specialist could tick necessary data.
It’s apparent that the need for medical speech recognition software is increasing every day. This new technology continues to capture more and more new application areas.
In Germany, around €33.4 billion was paid to doctors and nursing services in 2013. Documentation costs amounted to 21% of this amount[2]. For example, in a clinic with 450 hospital beds and an annual staff cost of ~ € 26 million, documentation cost is about €5.5 million.
By using speech recognition technology, a 20% increase in documentation efficiency could be achieved; in this case, that would equate to 1.1 million euros per year, the study’s authors claim (please see the link above). Another study[3] claims a 26% increase in productivity.
Summary
Currently, doctors, nurses, medical assistants, pharmacists, administrators of medical organizations, and operators successfully use speech recognition technology.
It is possible to use speech recognition software in two ways:
Some users find it more convenient to dictate text directly to a computer and correct any recognition errors immediately that occur during documentation. Such health workers can use dictation software.
Others prefer to write text to a voice recorder and then transfer the recording to a computer. In this case, they can use the Transcription software.
Speech recognition technology helps professionals register and search for the necessary information about a patient much faster than using a keyboard.
Activities such as creating case histories, writing prescriptions, preparing recommendations for rehabilitating patients, prescribing treatment, and searching information in a database can be easily performed with speech recognition software.
Today, many medical organizations are gradually switching to electronic document management to ensure the safe storage of information in archives and provide their employees with remote access to electronic documents. Speech Recognition technology can significantly simplify the transition to a “paperless office” by giving any employee the ability to create, edit, and search for electronic documents using only their voice.
Choosing Voicedocs dictation and transcription services allows health workers to save time on creating medical documents and, consequently, increase the volume of medical services provided to patients.
References:
[1] https://www.patientundanwalt.de/ratgeber/patientenrechtegesetz?gclid=Cjw...
[2] https://www.aerztezeitung.de/Wirtschaft/Zeitfresser-Dokumentation-249186...
[3] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4642384/