Part 1: Machine Learning Making Waves in Healthcare
The disruptive influence of artificial intelligence (AI) and machine learning (ML) is perhaps best felt in the healthcare domain, especially in the exciting shift from curing to ongoing caring.
Learning computers can, after all, be expected to compare symptoms and genetic details against millions of possible diagnoses in a fraction of the time it would take human doctors, not to mention process massive volumes of information no human could possibly manage.
In this and the next blog post, we’ll examine a few representative cases of how this technology is being put to beneficial use in the healthcare sector.
Improving Clinical Trial Safety and Efficacy
Intel has its sights set on helping to improve clinical trials, which often face challenges that drive up R&D costs and slow the delivery of promising treatments. To achieve this goal, Intel introduced its Pharma Analytics Platform in 2018, an edge-to-cloud artificial intelligence (AI) solution that employs remote monitoring to capture continuous clinical data from wearable device sensors.
The platform critically applies machine learning to develop objective measures for assessing symptoms and quantifying therapy impact. Intel’s solution can reportedly enable pharma companies to gather rich sensory data (such as skin temperature and activity levels), subject this data to sophisticated analytics to assess patient symptoms and identify subtle patterns, and gain real-time awareness of protocol adherence. Intel ultimately claims to supplement manually collected subjective data with device-aggregated objective data to help reduce clinical trial operational costs, increase patient adherence, and improve evidence quality — all contributing to accelerated drug and treatment development.
Empowering Heart Patients with “The Ultimate Wearable”
People with heart conditions can also benefit from the use of AI and machine learning in healthcare. Consider for example a pacemaker — perhaps “the ultimate wearable” — developed by Boston Scientific to leverage real-time patient-generated data to improve treatments and support healthier lives.
The Wi-Fi-enabled device tracks every single heartbeat, with its AI checking for any abnormal rhythm, and even monitors its own voltage settings and battery life. The pacemaker helps patients feel safe and in control of their lives by constantly reporting whether their heartbeat is in the safe zone, and it even enables all the data to be transmitted to their doctors over Wi-Fi too.
This represents a truly transformational time for healthcare, the future of which will be marked by such disruptive trends as patient-generated data, patient-empowering IoMT (the Internet of Medical Things) sensors and wearables, smart home applications, and mobile access to doctor care.
Mitigating Neurological Conditions
Another area of healthcare where machine learning is proving great value is epilepsy care. Epilepsy is among the most severe and erratic neurological disorders.
A team of doctors and neurologists aiming to help cope with the condition has developed MyCareCentric Epilepsy, which is an engineering solution combining the Microsoft Band wristband, a mobile app with machine learning, and care records and data analysis tools.
The solution tracks and stores integrated patient data — either logged manually by patients themselves or detected by the Microsoft Band in the process of monitoring motion and exercise habits, temperature, heart rate, galvanic skin response, and sleep patterns. It delivers this information to medical staff, keeps doctors constantly informed on patient health, and notifies them to have epilepsy specialists on standby well in advance when hospitalization may be needed.
A wearable speech recognition device developed by a team of MIT researchers similarly helps keep an eye on people suffering from speech disorders. Rather than settle for analyzing a patient’s voice for half an hour or less during a therapy session, this device enables doctors to improve treatment efficacy with access to 24 hours of real-time data per day. Machine learning-driven speech recognition technology is also tapping into other healthcare applications, such as early warning and diagnostics devices for diseases affecting cognitive function as well as products designed to monitor speech for early signs of a range of neurological conditions, including Parkinson’s, Dementia, and Alzheimer’s.
Remote Patient Monitoring
Over the past two years, remote patient monitoring (RPM) has seen significant gains in the United States as a result of the Covid-19 pandemic. As the pandemic undermined our nation’s health and severely challenged our hospitals and health systems, healthcare systems across the U.S. increasingly implemented RPM to deal with the strain.
RPM utilizes the latest advances in artificial intelligence and, more specifically, machine learning by passively collecting, analyzing, and extracting valuable non-explicit information from massive volumes of patient data. RPM value is in its power to give providers the ability to analyze and track real-time changes in a patient’s health.
With hospitals’ resources stretched thin and half of America’s physicians reporting burnout, patented AI-powered wearables like Somatix’s SafeBeingᵀᴹ smartband provide a way for physicians to provide high-value care outside of the traditional healthcare setting. For example, Somatix’s RPM wearable tracks over fourteen different metrics, including:
- Walking
- Sleeping
- Falling
- Hydration
- Medication intake
- Smoking
- Activity levels
- Wandering
- Heart rate
- SpO2
- UTI risk
- Pressure sore risk
- Readmission risk
Caregivers and providers then can review alerts, insights, and reports on an easy-to-use cloud-based dashboard. Their technology utilizes wearable-enabled gesture detection to yield powerful clinical insights and predictive analytics for providers. Furthermore, Somatix can be used without any hardware installation. Many location monitoring solutions require sensors to be installed in patient rooms or within the facilities. Other companies that don’t require hardware installations include numerous devices like blood pressure monitors and weights for patients to measure for clinicians. Somatix, however, allows all information to be tracked using their wearable, significantly enhancing patient experience.
So far, RPM has mostly been leveraged to facilitate the discharge of Covid patients, help prevent hospital readmissions, and care for patients suspected of Covid infection without putting healthcare workers at even greater risk. As more and more patients and providers hop on the RPM bandwagon, machine learning will only further improve the technology’s predictive capabilities and allow for more timely intervention and management of concerns. It’s only the beginning.
