AI and Medicine
‘ while the results were promising, they were not yet accurate enough to be introduced into the clinical setting ’ (Day,2022).
Explainability and interpretability: Ensure that the AI models provide transparent explanations for their predictions or decisions so that healthcare providers can understand and trust them. This is paramount as AI becomes a day-to-day tool for healthcare professionals and is used more and more. If AI is not easy to explain, it becomes very hard to correct when it makes mistakes, particularly due to the complex nature of DL algorithms. Regulatory compliance: Ensure that the AI models comply with all relevant regulations and standards, such as HIPAA (Health Insurance Portability and Accountability Act) in the US or MHRA (Medicines and Healthcare Products Regulatory Agency) in the UK. The legal regulations put in place regarding this technology is a subject that will be discussed in more detail when considering the ethics of its use. Continuous monitoring: Continuously monitoring the AI models in real-world settings to detect any issues or biases that may arise and address them promptly. This can be done in two different ways. You can adjust the data set that the neural network is drawing from or you can change how the algorithm is set up (though this requires the algorithm to be explainable). Human oversight: Ensure that healthcare providers are involved in the decision-making process and that the AI models are used as decision support tools rather than replacing human judgement. This is universally accepted as the correct approach to AI application in medicine. AI, should it ever reach large-scale clinical use, will be a tool that assists clinicians, and does smaller and more repetitive tasks that are an in efficient use of doctors’ time. This function necessitates human oversight. It is important that when AI is introduced there are quality checks at every stage of its development to ensure that it is capable of carrying out its specific function, with the most benefit to the patient, medical staff and the healthcare provider as a whole. AI can therefore be relied upon, provided that these checks and procedures are implemented.
Personalized medicine and improved patient engagement
AI in healthcare has the potential to transform the field of medicine and significantly improve patient outcomes (Bizopoulos,2019). One way medical AI can do this is by enabling personalized medicine and improving patient engagement. Personalized medicine or precision medicine involves tailoring medical treatment to an individual's specific needs, based on their unique genetic, environmental, and lifestyle factors. Medical AI can facilitate this by analysing large amounts of patient data, including medical histories, lab results, and genetic information, to identify patterns and make predictions about which treatments are most likely to be effective for each patient (Bhinder, 2021). This has been explored within the field of oncology. In this same paper, we see how predictive models based on large datasets are being used to improve early diagnosis, staging and grading of lung cancer. This has massive implications for patients as it means that they have far more information accessible at an earlier stage with which to make informed decisions – while there are still options available to them. Along similar lines, an article from 2019 documented research surrounding an AI which had been designed to make a prognosis for breast cancer patients based on certain genetic markers (Shimizu, 2019).
38
Made with FlippingBook - PDF hosting