Limitations of Voice Recognition Technology in EMRs

There are a few studies which have recently shown that voice recognition software which generates EMR clinical documentation is error- prone. One of these studies is the JAMA study. This was the study which was focused mainly on identifying and analyzing the mistakes at every stage of the voice recognition process. The main aim of this study was to provide the right recommendations for ensuring that EMR documentation is maintaining a high level of accuracy.,

Today there is a high rate of adoption when it comes to speech recognition technology in most of the healthcare institutions. The main reason behind this adoption is to reap the benefits of this software. These technological adoptions are becoming popular mainly because documentation in the healthcare system is the most time-consuming job even today. Voice recognition systems aim to improve the efficiency of literature and save a lot of time for clinicians.

After review by a transcriptionist, there was a 0.4 % decrease in error rate and 0.3 % was the decreased error rate in the final version of EMR documentation, which is reviewed, signed by clinicians. Even though error rates were comparatively reduced after the clinical review, it demanded an investigation by clinicians. So, this was the administrative burden which rose after the implementation and adoption of voice recognition in EMRs.

The aim was to reduce the pressure related to documentation time on the clinicians and make them review the notes superficially. When the responsibility got shifted from transcriptionists to clinicians, this led to enhanced documentation errors since, in many cases, clinicians were not able to review the notes correctly.

It is possible that healthcare organizations can implement speech recognition both at the front end and even at the back end. Back end systems can automatically convert the dictations of clinicians into text. Later medical transcriptionists can quickly review and if needed, edit the clinical documentation for ensuring the accuracy.

When the responsibility of reviewing the note is transferred to medical transcriptionists, this can free up physicians. This is the reason they can start focusing on many other aspects like caring for patients. This will result in boosting clinical productivity, along with reducing the administrative burden. There are many healthcare organizations which have already implemented voice recognition systems at their front end. This required clinicians to edit their notes after reviewing themselves. This will enhance the administrative burden.

Even though the main focus of the voice recognition system in EMR was to ease most of the burden that was related to documentation, there is no clarity on time available for physicians to review dictated notes. They sometimes even do not find time for reviewing these notes superficially. Inaccuracies will happen when there is no sufficient review in the final version of notes by physicians. When we think of clinically relevant information getting affected by documentation errors, that may seriously pose some threat in patient safety.

When there is the presence of dictation errors mainly in the EMR documentation and this when combined with a clinical review which is insufficient, now it’s time to focus on the integration of voice recognition review along with its use into existing clinical workflows. This should be done by developers of voice recognition software.

Final thoughts

One among the biggest headaches which healthcare organizations are facing today is documentation. When it comes to patient care, timely and accurate documentation is essential. This helps in accurate billing and all other insurance and legal purposes. Implementation of voice recorders in clinic notes, physical reports, procedures, consultation reports, and in ER reports, health reports should get converted into clear transcripts.

Speech recognition software is an application, which is developed to make the medical transcription highly efficient. This helps through recognizing the human voice and then converting it to digital records in just a few minutes. There is a rise in the adoption rate of Speech recognition software, and it has seen an increase from 21 percent to 47%. But it is true that there will be pros and cons for any application, and it is true with Speech recognition software as well.

Majority of the clinicians have the opinion that this voice recognition software helps in improving the workflow. When it gets completely integrated into some office environment, it will allow automatic queuing of the dictations from many authors. This helps in speech recognition, predefined assistance, and even selection routing of the dictation files. There is one more advantage of this; it also helps in reducing the turnaround time. This results in timely decision making. Accurate and prompt information acquired will help in saving lives and reducing the patient care days. Along with all these, the best part is there are healthcare organizations which are experiencing financial savings through the implementation of automated transcription software. Speech recognition software is intelligent enough to convert clinical dictations into the formatted draft documents.

But we have to look at the other side as well. Almost all the tools can make mistakes, and the worst part is when we used them in noisy places, they can be prone to errors. In some cases of implementation of speech recognition software, physicians are facing issues like overused punctuation, missing punctuation, poor grammar, and even disorganized dictations. These can lead to a lot of questions, and also if the review is arranged, again that will be a burden on physicians. If the voice is with heavy accents, then the software may not recognize.

So, it is observed that human intervention is necessary for training and making the software help in identifying voice pattern. There is help needed even for identifying the dictation style of the end user. Hence there is a great challenge with this voice recognition software in the EMR system, and yet the need for transcriptionists still prevail to a huge extent.