Challenges Of Delivering Clinical Impact With Artificial Intelligence
Artificial intelligence is not magic. AI tries to understand the patterns from the data in much more efficient way, a fast process which is a limitation to humans. Then it creates models for future predictions and inferences.
Artificial Intelligence in Healthcare
History stands strong on how human brains and expertise are the life and soul of all the possible innovations and revolutions. Whether it is the 7 wonders of the world, or the invention of steam engine and airplanes, the ideas generated from the wires of human brains have always eased the way of living. Today, technology has been massively sped into the healthcare industry. With the current pandemic, the usage is quick, and the impact is massive. Starting from a medical chatbot to detecting critical diseases like breast cancer, the industry is booming with the benefits of artificial intelligence. A person staying at a rural area in India is now aware of the video consultation and assessing the covid risk by answering some survey questions in vernacular language. The day humans realized that data is the secret sauce to create possible innovations, there is no looking back. According to the Amazon CEO, Jeff Bezos – “I think healthcare is going to be one of those industries that is elevated and made better by machine learning and artificial intelligence.” The intensity of the COVID19 pandemic would have been much more if there would have been no contact tracing process happening. AI/ML applications have proved wonders in assessing the covid risk of an individual and to plan out the vaccination drive strategically.
Let’s take a moment and realize, how our daily lives have been completely dependent on artificial intelligence. Right from ordering food from our favorite food chains to getting daily recommendations on OTT channels, AI has successfully disrupted almost every industry. But when it comes to healthcare industry, the impact is yet to be realized fully. There are few challenges in delivering significant clinical impact with the use of AI.
The Healthcare industry is one such arena where human interaction and empathy is mandatory. Artificial intelligence is not magic. AI tries to understand the patterns from the data in much more efficient way, a fast process which is a limitation to humans. Then it creates models for future predictions and inferences. The power of reasoning, learning new patterns apart from the past data and showcasing of empathy is yet to be realized in this technology. As mentioned by Nick Bilton, an eminent tech columnist in New York Times, “The upheavals of artificial intelligence can escalate quickly and become scarier and even cataclysmic. Imagine how a medical robot, originally programmed to rid cancer, could conclude that the best way to obliterate cancer is to exterminate humans who are genetically prone to the disease.” With reasoning features, these problems can be eliminated. The potential of BIG DATA is huge. Data varies from one individual to another and from one population to another. If an AI model learns patterns of a particular sample and becomes attuned to it, then there will be problems of underfitting, overfitting and curve fitting. The model will fail to work for some different set of samples. Again, the process of robustification of the model is extremely time consuming. There are models which can detect heart rate, Blood pressure, oxygen saturation and stress level through some features present on the face. But owing to the varied facial features in individuals like colour, puffiness of face and placement of organs on the face, the model will fail to deliver accurate clinical impact if robustification is not done. When you are dealing with models related to human body, the accuracy of the results is expected to be very high. Hence, the issues of performance of the models and diversification of models over varied population affects the accuracy. The issues of data privacy, security and breach has also risen many ethical concerns in successfully conducting research activities.
Barriers in Adoption
Covid-19 pandemic era has seen the maximum adoption of technology in healthcare sector. It is indeed booming. But what will happen in the post covid era? Will there be a relapse and the patients will reason with visiting a physician in person rather than going for a video consultation? There has always been a challenge in adoption of newer technology in the healthcare industry. Owing to the Hippocratic Oath, the physicians are abiding by the words “First do no harm”. Whether it is a small error in the treatment module, coming to terms with any ethical standards or adopting newer technologies like AI, healthcare leaders are not ready to compromise with the health of their patients. On the other hand, the patients are keener to take medical advice from a specialist rather than a machine. Human interaction is the bread and soul of a medical consultation. The patients look for some soothing words, empathy, and a healing touch after a diagnosis rather than talking to a chatbot. Yes, it is indeed true that telemedicine and virtual medical consultation is on a rise currently. But it won’t generate much value for the diagnosis and treatment of life-threatening chronic diseases which requires a complete health investigation.
Artificial Intelligence has indeed showed a new picture of healthcare industry and it is going to stay. So, the big question is, with so many challenges, how can we leverage AI in generating maximum clinical impact in healthcare industry? Artificial Intelligence should be used in augmenting the 3 pillars of delivering a healthcare service – Quality, Cost and Access rather than replacing human expertise. A physician will easily adopt a model which can reduce his burn out and make the diagnosis faster with maximum accuracy. Further, adoption can be increased by adding courses on usefulness of AI in healthcare in academic curriculum in medical institutions. At the same time, a strong regulatory policy specifying the requirements for data collection and AI based models is in need to avoid any dangerous research activities and unethical products/services. Finally, the best way to maximize clinical impact by creating a harmonious ecosystem between the IT industry and medical professionals. A strong involvement of doctors in creating AI models will revolutionize the industry faster with some disruptive ideas and innovations.
Disclaimer: The views expressed in the article above are those of the authors' and do not necessarily represent or reflect the views of this publishing house
Around The World