Artificial Intelligence: Current and Future Landscape in Health

Artificial Intelligence (AI), since its introduction in the 1950s, has become pervasive and an indispensable part of our lives. For the uninitiated, AI refers to the complex algorithms that are designed to perform tasks automatically without explicit instructions and direct involvement of humans. With the announcement of Tesla’s ambitious project on automated and driverless cars, our complete dependence on AI looks futuristic. It has already emerged as a game changer in industries like banking, finance, transportation, media, education, transportation, e-commerce, cybersecurity, logistics, manufacturing, and AI is now also cementing its place in the healthcare industry.

AI in Health — Current Landscape

AI found its way into healthcare in the 1960s, starting with domain specific research in medicine, and it progressed with the advent of various applications in the following decades. AI has immense potential in further evolving innovations using machine learning and deep learning.

According to a report from PwC, AI is estimated to contribute over $15 trillion to the world economy by the year 2030, and the greatest impact of artificial intelligence will be in the field of healthcare.

Optimizing literature research and healthcare delivery

NLP (Natural Language Processing) is a field that uses AI applications like speech recognition, text interpretation, and language translation to understand and communicate with humans. NLP systems in healthcare are used to extract relevant information from academic documents like published literature for the purpose of biomedical research; and clinical documents like physicians notes or diagnostic reports to be converted into EHRs (Electronic Health Records). NLP can also transcribe the spoken word into text to create reports. Nuance, a pioneer in AI speech recognition technology recently launched a healthcare AI solution called Dragon Ambient Experience (DAX) that deploys NLP to automatically document conversations in virtual consultations. This reduces both time and administrative workload for doctors and enables satisfactory experience for patients.

Medical diagnostics

CNN (Convolutional Neural Network), an algorithm that comes under the umbrella of deep learning technology of AI, can be utilized in medical diagnostics. The functioning pattern is analogous to that of visual processing in the human brain. CNNs can be trained to recognize and analyze medical images of the organs taken for diagnosis and prognosis of diseases like diabetic retinopathy, tumors, cancers, cardiac diseases, and fractures that are otherwise difficult to detect. Some medical anomalies are detected and understood better by ascertaining their location, size, structures by AI-powered medical imaging tools that identify subtle variations with precision.

According to a survey research finding, CNN had a major breakthrough when GoogLeNet used it to detect cancer at an accuracy of 89% while human pathologists could achieve the accuracy of only 70%.

Supporting Clinical Decision-Making

AI algorithms, when coupled with predictive analytics, help to provide clinical decision support. This works to autonomously process and segregate huge and unstructured historical healthcare data to make predictions about future health outcomes of patients and detect early signs of deterioration. It works by autonomously studying similar patients’ histories and treatment procedures, enabling healthcare providers to accordingly provide proactive care by implementing the most appropriate treatment protocols. Predictive algorithms are being improvised further to be applied in community health setups to forecast the likelihood of disease outbreak in a population.

Revolutionizing Drug Discovery

AI technologies have also started to make progress in drug discovery and biomedical research. To understand the interaction dynamics of potential drug candidates with different biomolecules, molecular modelling is required. It uses computational methods to mimic the behavior of real molecules using virtual simulation. AI algorithms help in assessing large databases of molecular libraries (billions of molecules with elucidated structures) and drug experiment results, a feat that is impossible to achieve by humans with the same speed. The potential use cases of AI in drug design can spare years of research and human effort, with a promise to save lives in cost-effective ways.

Some leading biopharmaceutical companies like Pfizer, use AI solution IBM Watson for cancer drugs research. IBM Watson uses cloud technology and machine learning to build predictive models of drug design by drawing data from medical journals, databases, and patent research. A life-threatening disease such as breast cancer is sometimes attributed to the heritable genetic mutations that can be detected earlier with technologies of genome sequencing. Deep Genomics, a leading biotechnology company, uses techniques of deep learning and mines human genome to find disease-causing genes. It can also help discover potential therapeutic approaches to treat those diseases.

Future of AI in Health — An Exciting Road Ahead

AI innovation is unfolding with applications in fields like precision medicine, which involves a personalized approach to disease prevention and management that takes into account unique characteristics of individuals like genetics, age, lifestyle, and environmental factors to accordingly customize treatment plans. AI solutions can empower this novel approach of medicine by processing an ever-increasing plethora of healthcare data points. Mobile app-based POCT (point-of-care-testing) involves rapid diagnostic procedures without the need of visiting labs. AI algorithms can be coupled with diagnostic devices to identify abnormalities and pathological conditions, and also guide clinical decisions.

Widespread implementation of AI has some potential risks like lack of patient’s data privacy, threat of unemployment to healthcare workers, possibility of errors leading to patient injuries, and the inability to detect bias and discrimination from data.

Despite some challenges and risks, demand for AI solutions in the healthcare sector is projected to surge. Moreover, traditional healthcare is inclined to undergo an imminent transformation because of the accelerated digitalization of the world. It is catering to the patients who are willing to take control of their health and make informed decisions for prevention and treatment. To prevent and manage non-communicable diseases, AI has enabled remote patient monitoring by physicians through wearable devices, promoting behavior change in the form of lifestyle modification, and medication adherence.

In the future, AI promises to augment the capabilities of healthcare providers as they can leverage its various applications for automated diagnosis with a high accuracy, efficient primary care delivery, prediction of the most effective treatment plans, and robot-assisted surgeries. AI-driven solutions can also make up for the shortage of healthcare workers and reduce their workload.

Some studies have also shown a change in the trends in patients’ compliance and physicians’ acceptability of telehealth usage due to the COVID-19 pandemic. As a result, implementation of AI is also going to see an expansion in optimizing the delivery of telehealth services.

To conclude, AI has started to bring about a positive transformation in the healthcare industry that benefits both healthcare providers and consumers, and its use cases in this industry are expected to expand in the coming years.

Anjali Dhingra is a Scientific Content Writing Intern at Saathealth, an AI powered, chronic care digital health platform.