• March 22, 2022
• By admin

The Growth of Artificial Intelligence

Advances in Artificial Intelligence (AI) in healthcare

Over the last few decades, artificial intelligence (AI) has played a significant and growing role in the world. Whether you are logging in to social media, e-mail, car ride services, or online shopping platforms, all make use of artificial intelligence algorithms to improve the user experience. Most people are unaware that artificial intelligence manifests itself in a variety of ways that have an impact on daily life. One major area where AI is rapidly expanding is medicine, specifically diagnostics and treatment management. There is a lot of research being done to see how AI can aid in clinical decisions, support human judgement, and improve treatment efficiency. (1)

Increased Presence of Artificial Intelligence in Healthcare

AI-based tools are being explored and integrated in healthcare, just as they are in any other field. Many people, however, are concerned that artificial intelligence will eventually replace or reduce the need for human physicians, especially in the clinical setting. The reality however is completely opposite. AI frequently makes use of a web database, which provides access to thousands of diagnostic resources to doctors and practitioners. Because doctors have received extensive training in their field and are up to date on current research, the use of AI tools greatly enhances the likelihood of a faster outcome that could then be matched with healthcare professional’s clinical knowledge. For example, a patient may exhibit multiple symptoms that can be correlated with various conditions based on both genetic and physical characteristics, causing a delay in diagnosis. Recent research and data, indicate that this tool is more likely to benefit and improve clinical diagnostics and decision making than to reduce clinician need.

AI benefits practitioners not only in terms of efficiency, but it also provides quantitative and qualitative data based on input feedback, improving accuracy in early detection, diagnosis, treatment planning, and outcome prediction. These advantages are augmented by AI’s ability to “learn” from data, allowing for improved accuracy based on feedback responses. This feedback is derived from a variety of back-end database sources, as well as practitioners and research institutions. In healthcare, AI systems operate in real time, which means that data is constantly updated with increasing accuracy and relevance.(1)

AI Machine Learning Allows for More Accurate Diagnosis

AI in healthcare can take many forms. Assembled data is a collection from various sources including medical notes, electronic recordings from medical devices, laboratory images, physical examinations, and demographic data. With this constantly updated collection of information at their disposal, practitioners have nearly limitless resources available to them, in order to improve their treatment capabilities. With a large amount of healthcare data out in the field, AI must sort through it efficiently in order to “learn” and build a network.

Healthcare data, can be roughly classified into two categories: unstructured data and structured data. Structured learning makes use of three techniques: Machine Learning Techniques (ML), a Neural Network system, and Modern Deep Learning. Natural Language Processing is used in the processing of all unstructured data (NLP). Machine Learning techniques use analytical algorithms to extract specific patient traits from all of the data gathered during a patient visit with a practitioner.(2)

Physical exam results, medications, symptoms, basic metrics, disease-specific data, diagnostic imaging, gene expressions, and various laboratory tests all contribute to the structured data gathered. This will be followed by use of machine learning to predict patient outcomes.

Artificial Intelligence Integration in Major Disease Areas

It is critical to use as many resources as possible to help with early detection, diagnosis, and treatment. With cardiovascular disease, neurological disorders, and cancer consistently ranking among the leading causes of death, the use of artificial intelligence improves early detection by highlighting patient’s risk alerts.

Patients at risk for stroke were placed in an early detection stage in one study using AI algorithms based on their presented symptoms and genetic history. During this stage, any abnormal physical movement in the patient was recorded and triggered an alert. Because of this trigger alert, practitioners were able to get patients to an MRI/CT scan for disease evaluation sooner. In the study, the early detection alert provided 87.6 percent accuracy in a diagnosis and prognosis evaluation. Nonetheless, the practitioners were able to begin treatment sooner and predict whether the patient was more likely to have a stroke in the future. Similarly, in 48-hour post-stroke patients, machine learning was used to predict whether the patient would have another stroke or not. In another case, Neural Networking was used in a breast cancer diagnostic process that sorted through 6,567 genes and paired it with texture information from the subjects’ mammograms in a recent study. A more specific tumour indicator outcome was possible due to the combination of logged genetic and physical characteristics.(3)

Smaller-Scale Artificial Intelligence in Telehealth

Although AI is used on a larger scale for high-risk diseases, telehealth tools are being implemented in patients’ homes to help treat and prevent high-risk situations while reducing hospital readmissions. Telehealth tools make it possible to collect, document, and process various metrics in the same way that a larger AI machine would. When a patient reports a high-risk variable, this equipment can alert practitioners immediately. Early detection, faster diagnostics, and an updated treatment plan save time and money for both the patient and the hospital while providing more immediate care. Artificial intelligence enables practitioners to make more efficient and logical decisions, ultimately improving patient care.

Challenges and Future Development in Indian Context

Experts say there are still many challenges ahead. Among them are the high cost and difficulty of obtaining high-quality data, as well as developing and training smart models to detect patterns. Most electronic medical record-keeping systems are incompatible with one another. Because the data is frequently stored on servers at individual clinics or hospitals, it is difficult to create a comprehensive reservoir of medical information.

However, government of India is making efforts to address these challenges. The finance minister proposed launching a ‘Ayushman Bharat Digital Mission,’ which would serve as an open platform for the National Digital Health Ecosystem. It aims to use emerging technologies such as artificial intelligence along with the internet of things, blockchain, and cloud computing to facilitate a holistic digital health ecosystem that can increase equitable access to health services, improve health outcomes, and reduce costs.(4)

References:

1) Yu, KH., Beam, A.L. & Kohane, I.S. Artificial intelligence in healthcare. Nat Biomed Eng 2, 719–731 (2018). https://doi.org/10.1038/s41551-018-0305-z

2) https://blogs.cdc.gov/genomics/2020/09/17/artificial-intelligence/

3) Jiang F, Jiang Y, Zhi H, et al. Artificial intelligence in healthcare: past, present and future. Stroke and Vascular Neurology 2017;2: e000101. doi:10.1136/svn-2017-000101

4) https://abdm.gov.in/publications/ndhm_strategy_overview