Artificial intelligence (AI) is already making an appearance in diagnostic sleep labs throughout the world. A team of sleep medicine experts from Nox Medical came together to share their thoughts about how AI can help streamline workflows and ease the potential burden of labor shortages. All views expressed in this article are those of Nox Medical employees and not reflective of AAST.
Sleep technologists make a living by monitoring people’s sleep, but that is far from the only task they do. The busy life of a sleep technologist involves hooking patients up to diagnostic equipment, carefully placing electrodes onto patients’ skin, understanding physiological sleep measurements and combing through diagnostic data. The job can be tedious and time-consuming, and coupled with the labor shortages that some sleep labs are experiencing, the day-to-day work of a sleep technologist can become hectic.
“This is not a straightforward job,” says Helgi Helgason, MS, RPSGT, RST, a Nox Medical product specialist in the United States. “There are a lot of moving parts that go into it.”
According to a survey of 8,074 sleep technologists from the Board of Registered Polysomnographic Technologists (BRPT), at least half of all those surveyed believe that there is a sleep technologist shortage in the United States. Anecdotally, the pandemic caused this shortage to grow worse — more sleep labs seemed to cut down on the ratio of sleep technologists to patient beds, which put greater strain on technologists already working. Many sleep labs also closed temporarily while others never opened again, says Nox Health’s Director of Program Strategy and Innovation Nigel Ball, PhD.
Fortunately, there are many ways to cope with staffing shortages and hiring woes in the sleep lab, with artificial intelligence being a top way to increase efficiency. In this context, AI is referring to algorithms that data scientists have developed to score sleep study data. These algorithms can cut the time it takes to score an in-lab sleep test in half compared to a human scorer, who may typically take up to an hour to manually score a test according to Nox Medical Product Training Specialist Jay Hemnani, RPSGT.
In some sleep labs, AI-driven sleep software has changed and challenged the way that they operate, optimizing workflows and saving valuable time and resources. One example of an AI sleep software is Noxturnal, which has both manual and automatic sleep-scoring.
Of course, AI can’t speed up the process of hooking patients up to the various electrodes, but it can help sleep technologists work faster when scoring tests.*
“AI can alleviate the sleep technologist shortage that is out there,” says Nox Medical Product Specialist Byron Jamerson, RPSGT. At Fusion Sleep in Georgia, Noxturnal helped to fill a void when the sleep lab struggled to hire a new sleep technologist.
“We had a sleep technologist job posted everywhere and it still took over a year to fill the opening,” says Hemnani, who was Fusion Sleep’s sleep lab manager before joining Nox. “The sleep lab even hired an outside recruiter to work on filling the position with help from the company’s internal human resources department, but still, there was a struggle to fill the position.”
According to Hemnani, when no one came forward to fill the position, AI via Noxturnal made things easier with scoring sleep studies.
“AI has definitely reduced the amount of time taken to score a sleep study,” says Hemnani. “The output for the average sleep scoring technology could be doubled.”
While output could be doubled with Noxturnal, accuracy remains high. Recent research has found that Noxturnal’s respiratory analysis is reliable when compared to AHI scored manually by a human sleep technologist.
Of course, the results of autoscoring must be reviewed by health care professionals, but overall, AI is increasing efficiency while maintaining the high degree of accuracy that sleep medicine demands.
*The automatic analysis results should always be reviewed by a technologist or a physician prior to diagnosis.