Doctors should not fear that their work will be taken over by neural networks or robots in the near future. There are several factors that explain why this won’t happen.
First, artificial intelligence is a tool, a technology in the hands of a person, that helps doctors today in their daily work. AI-based services free the doctor from repetitive tasks, save him time to communicate with patients, and take care of the most difficult diagnostic cases. But AI algorithms can only find the pathology for which they were trained. In practice, we see that artificial intelligence and doctor’s expertise together have a strong impact and improve the quality of diagnosis. We are moving towards that all medical imaging studies are first interpreted by artificial intelligence algorithms and then verified by a doctor. This approach will reduce the number of diagnostic errors and the burden of radiologists. At the same time, the doctor always has the final say: he ultimately decides what diagnosis to make or what treatment to offer.
Second, AI quickly analyzes and calculates large volumes of data and can perform narrow well-defined tasks. But diagnosing and treating a patient are not linear processes. They require creativity developed through years of clinical thought that algorithms and robots lack. Each case in medical practice has its own characteristics, requires personal attention and involvement of the doctor. If previous doctors have relied on information from simple medical devices when making decisions, the task will remain the same in the future, but a person will have access to data from a large number of “smart” devices, genome research and assessment of the patient’s lifestyle.
Thirdly, advanced technologies will not be able to show empathy and sympathy, understand the patient and help him adopt a new way of life in connection with the diagnosed disease. A positive attitude, often necessary for successful treatment, recovery and rehabilitation, is largely achieved through the support and professional training of medical professionals.
Finally, the development of AI technologies not only replaces doctors, but also opens up new opportunities for medical specialties. At the intersection of informatics and medicine, medtech companies now have professions that are in great demand. For example, developers collaborate and recruit medical professionals with specialized medical training to join their teams. The medical professional interacts with analysts and testers and answers questions that arise during their work: how to understand whether the anatomical region and the pathological region are correctly identified, what are the input data requirements for the model to work correctly, how to interpret the results of the model for the end user.
In order to create high-quality algorithms that analyze visual examinations, it is important to additionally analyze the labels of doctors. The same study can be labeled differently by two different experts, so to increase the accuracy of the data, give the medical professional (already second) opinion, helping to understand why doctors made a different interpretation. At this stage, you can identify some systemic errors or find an important nuance for more accurate training of the algorithm.
The medical professional also participates in the evaluation of the quality metrics of the IT product and comments on the medical part. It helps identify in which images false positive or false negative results occur, reveals the features that unite them (perhaps it is necessary to take into account the age or gender of the patient, equipment settings), and together with the developers formulate a task to improve the algorithm.
Source: Gazeta
Dolores Johnson is a voice of reason at “Social Bites”. As an opinion writer, she provides her readers with insightful commentary on the most pressing issues of the day. With her well-informed perspectives and clear writing style, Dolores helps readers navigate the complex world of news and politics, providing a balanced and thoughtful view on the most important topics of the moment.