Accurate Speech Recognition at the Core of Radiology

28 April 2022 – During the last two COVID-19 pandemic years, 63% of radiographers reported an increase in their workplace-related stress levels due to a combination of demand, low capacity, major changes to routine departmental protocols and redeployment. However, according to Scribetech (UK) Ltd, a clinical voice-AI solution and disruptive technology provider, the strain the discipline is under was already reaching breaking point prior to the pandemic – and now more than ever – solutions are required to bring back confidence in the profession.

The CQC’s 2018 Radiology Review, which measured slow reporting times across the NHS, already highlighted then, a backlog of 33,400 unreported images at one Foundation Trust alone, despite attempts to recruit more radiologists and reverting to the use of outsourced support at great expense.

Both in that same year and in 2020, the Clinical Radiology UK Workforce Census Reports highlighted a severe lack of radiology consultants. Three quarters of radiology clinical directors didn’t feel they had enough radiology consultants to deliver safe patient care in the working environment. In addition, the 2020 report also found that while Trusts moved to remote/home and hybrid ways of working, this came with an additional set of issues and risk for radiologists and radiology reporting, owing to an increase in IT and technological challenges. Even where the most forward thinking Trusts had already embraced voice/speech recognition applications, few were equipped to make these available for remote access and efficient in connectivity. As a result, most radiologists reverted back to typing their reports or sending audio files for manual transcription.

Shiraz Austin, Co-Founder and Managing Director, Scribetech, stated: “Efforts to improve radiology across the NHS are emerging. The UK Treasury’s 2020 Spending Review earmarked £325 million for investment in new diagnostic imaging equipment. However, radiology remains critically understaffed – by as much as 33% and forecast to increase to 44% by 2025.”

With rising demand for radiology year-on-year, adding more equipment and more staff simply isn’t enough. Existing radiologists need support to work efficiently and accurately today – without adding technology challenges and compromising on attention to detail and risking reporting errors or mistakes.

“In part, replacing outdated equipment and improving connectivity to clinical systems will help reduce errors, drive accuracy and improve patient outcomes. Technology beyond imaging equipment also has an important role to play in helping radiologists feed accurate information into the diagnostics process; cloud-based solutions are the key” added Austin.

Radiology Information Systems (RIS) and Picture Archiving and Communication Systems (PACS) are undergoing a transformation. Innovations like speech recognition are helping radiologists interpret and report on images faster. And, crucially, they are designed to preserve accuracy in a way that simply wasn’t possible a decade ago. Meanwhile, NHS Trusts are moving toward a Vendor Neutral Archive (VNA) – a way to store medical images in a standard format and make them accessible to professionals across a range of systems.

“In recent years, speech recognition (SR) solutions have advanced significantly. With SR the radiologist can dictate the case, edit it (if necessary), and complete the report all at once, which makes the final file available almost immediately. In turn, the clinician can review this report sooner than would have been possible in the traditional reporting method. This leads to better patient care, since they [the patient] can move onto the next step in their workup and diagnosis, or begin treatment sooner.”

“It also leads to a more accurate and informed report as it’s completed while the radiologist is reviewing the images, before details and notes of the case might be forgotten. Plus, there’s less chance of the report getting misfiled or lost in the system. One final consideration is that this type of speech-enabled radiology reporting process can lead to a more satisfied referring clinician, since each report is available sooner, as well as a more satisfied radiologist, due to a sense of completion in knowing that each report will not have to be reviewed again.”

“With reports instantly transcribed and fed into clinical systems, AI driven cloud-based SR solutions such as Augnito, can significantly reduce the burden on radiologists even more. Many users report time savings, but what radiologists really need is secure technology they can feel confident about, knowing that its level of accuracy is such that it is “right first time” – a way to guarantee that information is not only fast to capture, but accurate to the highest standards,” concluded Austin.