For decades, the concept of medical devices has been associated with physical products such as surgical implants, pacemakers, radiology systems, or surgical consumables and instruments. But that paradigm is being radically altered by a new era of healthcare innovation. Software as a Medical Device (SaMD), which is software created to carry out medical tasks autonomously without being a component of physical medical equipment, is one of the most important developments.
SaMD is used for diagnosing, monitoring, and treating diseases within different medical specialties. On the basis of its intended medical function, SaMD falls into five types: Diagnostic software, monitoring software, therapeutic software, disease risk prediction tools, and clinical decision support systems (CDSS). Diagnostic software, including AI-supported radiology solutions, helps diagnose abnormalities such as lung nodules using medical imaging in order to diagnose the condition quickly and accurately. Monitoring software includes programmes that continuously monitor physiological parameters and alert health care providers about clinically relevant changes. Therapeutic software delivers evidence-based treatment directly to patients in a convenient and personalized way: Prescription digital therapies deliver cognitive behavioural therapy (CBT) for insomnia, anxiety, depression. Risk prediction tools use Artificial Intelligence (AI) and machine learning to determine the probability of the development or progression of such diseases as cardiovascular disease, sepsis, or diabetes. Clinical decision support systems (CDSS) combine patients’ clinical data with the help of medical guidelines and predictive algorithms and help healthcare professionals diagnose conditions, choose appropriate treatments and medications, and identify risks.
Aside from diagnosis of diseases, the advancement of AI-SaMD is changing the landscape of the operating room. With commercially available software, surgeons can see the unique anatomical structure of a patient in great detailed 3D models even before an incision is made. This ensures proper preparation for the surgery and leads to improved results. Surgical planning software is also being used for complicated oncology and orthopedic surgeries.
With SaMD making or helping in clinical decisions, how is it regulated? Just like traditional medical devices, SaMD is regulated by authorities like Central Drugs Standard Control Organisation (CDSCO), US Food and Drug Administration (FDA), International Medical Device Regulators Forum (IMDRF), and European Union. IMDRF categorises SaMD to Category I to Category IV based on patient risk. Before SaMD can be used for clinical practice, it must pass through, three-pillar clinical evaluation: 1) Is there a valid clinical association between SaMD output and SaMD’s targeted medical condition? 2) Does SaMD correctly process input data to generate accurate, reliable, and precise output data? 3) Does use of SaMD’s accurate, reliable, and precise output data achieve your intended purpose in target population in the context of clinical care?
As regulatory guidelines and algorithms evolve, the traditional barrier between software and clinical practice will start blurring. SaMD will start moving from its current supportive role to being an indispensable part of clinical practice thereby enabling doctors to make better evidence-based decisions and eventually optimise patient outcomes.
(The views expressed are personal)
This article is authored by Dr Ashok Kumar Moharana, chief medical officer, Healthium Medtech Ltd.
