Machine Learning in Healthcare
In this era of digitalization almost every industry is affected and healthcare domain is no far behind. The exploitation of healthcare data is much needed by both medical practitioners for automating the process and for patients also to make the process much convenient to use by them..
Machine Learning is a buzzword associated with many domains and it is one of most popular form of Artificial Intelligence in which the processing and finding patterns especially in large datasets is done to automate decision making process.
Expert systems in medical healthcare are not a new thing. MYCIN expert system[1] for detecting cause of severe infections by identifying and studying bacteria causing them, is one such example. It was designed in 1970 by Stanford University and considering the technology and related data availability gave 69 percent accurate prescriptions. Probability based machine learning model Internist1 [2] for diagnosing internal medicine in used in 1980s is another brilliant example but still they were in naive stage. Data had to be entered manually and based on that only doctors could prescribe .It was more of assisting doctors with limited data details but it paved a path for current automated systems with very little intervention by doctors. Popular RX Project [3] by Stanford University was the beginning of automated medical healthcare systems using Machine Learning and Artificial Intelligence. Neural Networks in healthcare took plunge during 1990s.Although it didn’t fit very well in the clinical workflow and could not generalize the clinical process but launched a good platform for research related to artificial intelligence in healthcare in the coming years and today is that era of automated clinical decision support systems handling vast amount of clinical data with a great efficiency .There is so much diversity in digital health data like lab test data, medical imaging, proteomics, genome data, general health record data collected in devices like smart phones and smart watches. Not only in clinical decision support AI and machine learning in healthcare includes many other use cases like automated medical billing, proper framing of clinical care guidelines also. Smart healthcare or smart clinical support using machine learning techniques can be helpful in prevention, diagnosis, treatment, management and can connect and exchange information at any time and location. There is a wide range of potential applications of machine learning in the healthcare out of which the most commonly used are listed below:
1. Artificial Intelligence/Machine Learning assisted radiology and pathology in which machine learning techniques can be used to imaging data like expert radiologists do. Microsoft’s Innereye[4] project is an excellent example of use of machine learning to detect tumors by studying three dimensional radio images.
2. Robotics combined with machine learning for assistance in surgery. Digital surgery robots are great help to surgeons reducing their efforts as well as doing some jobs even more precisely than humans.
3. AI and ML techniques are prominently used by pharmaceutical industries for efficient drug discovery. They are used to find patterns, complex structures in drug discovery and its relationship with various patient’s clinical background and genetic features finding .Deep neural networks, reinforcement learning, association mining are quite useful techniques in this area.
4. Pandemic or epidemic control by analyzing and predicting the outcomes using massive data generated though social media, satellites .Covid 19 study and its control through its predictive analysis using statistics as well as machine learning models is one such example.
An examination of machine learning in healthcare reveals how technology innovation can lead to more effective, holistic care strategies that could improve patient outcomes still data privacy is a challenge for the proper implementation of AI/ML based healthcare systems. Proper enforcement with the consideration of legality and respecting the privacy can standardize the complete system and can be a boon to human race by machines.
References
1. Copeland, B.J.. “MYCIN”. Encyclopedia Britannica, 21 Nov. 2018, https://www.britannica.com/technology/MYCIN. Accessed 19 October 2021.
2. Miller, R A et al. “The INTERNIST-1/QUICK MEDICAL REFERENCE project — status report.” The Western journal of medicine vol. 145,6 (1986): 816–22.
3. Robert Blum, “Discovery, Confirmation and Incorporation of Causal Relationships from a Large Time-Oriented Clinical Data Base: The RX Project”. Dept. of Computer Science, Stanford. 1981
4. https://www.microsoft.com/en-us/research/project/medical-image-analysis
