top of page
Search

Sleep Apnea Diagnosis and Care with AI and ML – What is the potential of new technologies to benefit

Sleep apnea is a common and much underdiagnosed sleep-related breathing disorder that has been reported to affect millions of people globally. The gold standard of in-lab polysomnography overnight tests for diagnosing and managing sleep apnea is currently facing a huge wait and backlog for this test due to the impact of the COVID-19 pandemic. This can be slow and resource-intensive, leading to delayed diagnosis and subsequent treatment. Recently published medical literature in this field is increasing particularly with advancements in artificial intelligence (AI) and machine learning (ML) being poised to transform the Sleep apnoea diagnosis and care of sleep apnea in healthcare settings, tackling and reducing waiting times and expediting treatment.


Understanding Sleep Apnea


Sleep apnea is characterised by repeated interruptions to breathing during sleep, resulting in reduced airway patency leading to disrupted sleep patterns and a number of other potentially severe health consequences such as increased cardiovascular risk. Two types of primary sleep apnea: obstructive sleep apnea (OSA), caused by intermittent collapse of the airway, and central sleep apnea, related to problems in the brain's breathing control centre. Both types of sleep apnoea can exist in the same patient and an accurate diagnostic test, and the correct personalised treatment prescription are vital for effective management of this common condition.


Current Diagnostic and Care Challenges


Polysomnography is considered the gold standard test and is an overnight sleep study conducted in a specialised sleep lab or hospital. Currently, long waiting times exist due to limited availability for this test, the increased need for referrals as the prevalence of the condition is rising globally, and there is the inconvenience of the patient having to take time off work, travel often long distances for an overnight stay to have this test conducted when home sleep tests that are more convenient for healthcare services and patients are now available.


The Role of AI in Sleep Apnea Diagnosis: How can AI and Machine Learning help screen for and assist in the diagnosis of Sleep Apnoea?

  • Early Screening: AI and ML algorithms are being employed to develop more efficient and accurate screening devices and tools for sleep apnea. These devices and tools can analyse data transmitted from wearable devices, such as smartwatches or fitness trackers, to detect potential signs of sleep apnea, like irregular breathing patterns or oxygen saturation levels. These technologies are developing but do need further research to validate their accuracy.


  • Home-based Sleep Apnoea Tests (HSAT) - AI-ML powered devices, like the Sunrise device, can enable patients to conduct sleep apnea assessments in the comfort of their home using a small 3gram sensor that is secured to the tip of the chin. It measures mandibular movements which reflect what is happening in the brain's breathing control centre during sleep. The Sunrise devices use AI and ML algorithms to analyse sleep data, including snoring patterns, breathing pauses (AHI), total sleep time and sleep stages, respiratory arousal from sleep and respiratory effort, airway collapse, body sleep position, and oxygen levels, to identify sleep apnea events. The Sunrise device has been validated against in-lab polysomnography in published clinical trials. If you want to find out more about this novel technology visit our website at: https://www.sefam-uk.co.uk/diagnostics Another HSAT that is widely used and also uses similar methods is the WatchpatOne by Itamar. The WatchPat device monitors blood oxygen levels, heart rate, body position, body movements, snoring intensity, and, crucially, Peripheral Arterial Tone – a key physiological signal that can indicate respiratory disturbances during sleep.


  • Data Analysis: Artificial Intelligence and Machine learning algorithms can process and interpret vast amounts of sleep data quickly. By identifying sleep apnea patterns, they can assist sleep medicine clinicians in making more accurate diagnoses and help to personalise appropriate treatments in a timelier manner thus, reducing the need for extensive lab tests where polysomnography can be prioritised for those that need more detailed tests.


  • Telemedicine: AI can help to facilitate telemedicine consultations for patients suspected of having sleep apnea where the patient can be seen at a distance and this model of care has been well documented and evidence based even before the advent of the COVID-19 pandemic. Sleep medicine services providers can remotely monitor patients' sleep data, such as CPAP use time, efficiency of treatment, and mask fit and can help to troubleshoot issues and then make changes to treatment parameters remotely in a timely manner, reducing the need for in-person appointments. Telemedicine review has been proven to provide equivalent person-centred care, is closer to home, is environmentally friendly and cost-effective. In summary, it has real benefits for patients, healthcare providers, healthcare providers and policymakers.


  • Personalised Treatment in Sleep Apnoea: Machine learning can aid in tailoring treatment plans to individual patients. By considering factors like age, gender, severity of sleep apnea, and comorbidities, AI can recommend the most effective treatment options, which may include lifestyle changes, CPAP therapy, or surgery.


Benefits of AI and ML-driven Sleep Apnea Care


Integrating AI and ML technologies into the diagnosis and care of sleep apnea in healthcare settings can offer several attractive advantages:


  • Reduced Waiting Times: AI-ML powered screening HSAT options and home-based monitoring can significantly reduce the time patients must wait for diagnosis and treatment.


  • Enhanced Patient Engagement: Home-based screening, diagnostic tests and remote monitoring and telemedicine options have the potential to engage patients to take an active role in co-managing their condition, which may see better treatment adherence and health outcomes.


  • Cost Effective healthcare delivery: In a time when the NHS is in a severe financial crisis, AI-ML driven solutions can potentially lower the overall cost of diagnosing and treating sleep apnea by reducing the need for more expensive in lab visits to assess patients. The Sunrise device and WatchPaton one test are already being implemented successfully in the NHS for screening and diagnosing people with OSA.



The addition of artificial intelligence and machine learning in healthcare for sleep apnea screening and diagnosis is an exciting field of research that may make real inroads into expediting diagnosis and treatment for this condition. By harnessing advanced technologies, Sleep Medicine services can reduce waiting times, deliver more personalised care solutions and ultimately deliver more people-centred care leading to improved quality of services and quality of life for our patients. Remobilising Sleep medicine services that have been severely impacted by the global pandemic is an absolute priority for the NHS and these new technologies will help in this monumental task. It must be stressed that these new technologies are tools that require the expertise of trained Sleep Medicine clinicians to use them effectively in the right situation to achieve the best outcomes.

bottom of page