Artificial Intelligence in Medicine

Artificial Intelligence in Medicine

Abstract

Artificial Intelligent is a phrase that denotes the use of computer technology to simulate intelligent behavior and critical thinking similar to humans. Integrating of AI in medicine is significant as it will improve decision making in the clinical settings and offer effective mechanisms in diagnostic and therapeutic examinations. Nearly half a million individuals die yearly in America due to medical errors, which could be prevented by AI technologies. The methodology of the research involves reviewing of four articles through scientific and analytical perspective inquiry. From the scientific perspective, it will highlight the issues that are solvable in different disciplines of medicine and upcoming significant solutions. Whereas, in the analytical view, statistical facts of the topic will be discussed together with the anticipated savings. All the reviewed researches recommend for the cultivation of AI in medicine, though they propose specific protocols to be put in place. They emphasize the training of healthcare professionals to a future world filled with AI advances. Additionally, they see the need for the enforcement of policies that will mitigate concerns of transparency, accountability, and privacy.

Artificial Intelligence in Medicine

Artificial Intelligence (AI) has the power to utilize complex algorithms and can learn characteristics from substantial quantities of data. Additionally, it can be equipped with learning capabilities and self-correcting features that enhance the accuracy of the system, thus significantly assisting the clinical practice. Medical stakeholders need to be aware of these trending systems because, with time, these systems will promote (or have helped) the healthcare in multiple ways such as in making clinical decisions, and in the conducting of diagnostic and therapeutic examinations. In radiology, all the diagnostic processes- x-rays, CT scans, MRIs, among others- are based on AI technology. Incorporating the novel technology to medical imaging analysis improves the accuracy and effectiveness of the assays due to that chances of errors occurring are reduced and provides feedback on elements that the human eye can easily miss. It advocates for precision medicine, whereby when combined with DNA information, it can predict health problems founded from these nucleotide elements. Moreover, the integration of AI into a healthcare app has an immense value to the users. The AI makes the app to act as an individual health assistant. In this role, it offers medication alerts and assists the patient when there is no clinical personnel. There is a need to implement and employ this powerful and useful technology in medicine as it shows to avert medical errors. In the industry of healthcare, roughly 86% of the mistakes are preventable. In America, an average of 440,000 individuals die annually due to these errors, which are effortlessly prevented by AI.

Research Questions

Scientific Perspective Inquiry

Level 1. What are the anatomical, physiological, pathological, or epidemiological issues to be solved?

Level 2. What are the ultimate solutions?

Analytical Perspective Inquiry

Level 1. What are the statistical facts related to the topic?

Level 2. How much will be saved?

History of AI

Credit to modern computer and AI goes to Alan Turing in 1950. He coined the Turing Test, whereby the computer’s intelligent behavior was measured by its ability to accomplish the level of human performance in cognition related roles. In the following decade, a problem-solving platform known as Expert system or Dendral was created. It was applied in the Mycin system – a technology for identifying bacteria that cause infections. However, in the 1980s and 90s, this system, together with its alternatives, Casnet and Internist-1, lost popularity as they had numerous bottlenecks. The era saw a surge in Fuzzy expert systems, Artificial neural networks, Bayesian networks, and Hybrid intelligent systems. The AI in medicine later was dichotomized into two subgroups; physical and virtual. The former assists in the accomplishment of surgeries and intelligent prostheses. Whereas the virtual subtype has ranges of applications from electronic health records to treatment and diagnosis determinations.

The AI system has a psychological element in it. In order for it to work, it employs sophisticated algorithms and software that imitate the cognition of humans. The technology can collect data, process it, and produce a high-defined output. They recognize behavior configurations and form their logic. In their applications- treatment and diagnosis- they entangle systems of anatomy and pathology. Through radiology, human body parts are studied, whereas pathogenesis, vascular diseases, and cancer are spotted early by this technique.

Scientific Perspective Inquiry

From reports, AI has enhanced issues related to anatomical, pathological, and epidemiological. In 2019, Amisha et al. conducted a study focused on describing the comprehensive overview of medicine’s AI. One of their specific objectives was to investigate the current and future applications of AI in the field. Their study was successful because they were able to obtain a far-reaching application of AI. It includes drug development, health monitoring, managing of medical data, diseases diagnostic, surgical and medical treatment, and so forth. According to them, the technology is substantially embraced in radiology. Through the help of CT scans, x-rays, and MRI, negative exams are identified, which study a considerable number of anatomical parts. In the aspects of pathological and epidemiological issues, systems like DXplain and Germwatcher – developed by the University of Massachusetts and the University of Washington correspondingly- are used. Dxplain offers a list of plausible differentials constructed from symptoms complex and subsequently explains facts not available in the standard textbooks. On the other hand, Germwatcher detects infections acquired in a hospital. Additionally, therapeutic courses termed as AI-therapy are present. The therapy approaches help patients psychologically by treating their social anxiety.

In another research by Hamet and Tremblay, they revealed meaningful solutions that will be realized by AI systems. In the virtual branch, they show that AI has started exposing discoveries in molecular and genetic medicine. Unsupervised interactions between proteins have led to discoveries in therapeutic targets, and new evolutionary embedded algorithms identify variants of DNAs like single nucleotide polymorphisms, which are mainly predictors of diseases. The physical branch also offers future solutions to problems posed to autistic individuals. Robots now can teach and communicate with these special individuals. Recently nanobots designed with AI algorithms have shown to overcome delivery challenges involved with therapeutic agents when diffusing into target sites. The authors conclude by stating that as long as genetics exist, AI is vital for personal services.

Analytical Perspective Inquiry

It is essential to promote the implementation of AI in medicine as it improves patient care and, more so, saves on costs. Topol in 2019 reported on the statistical facts underlying AI in medicine. To better understand the research, he first introduced two terms; area under the curve (AUC) and receiver operating characteristics (ROC). “The neural net interpretation is typically compared with physicians’ assessments using a plot of true-positive versus false-positive rates, known as a receiver operating characteristic (ROC), for which the area under the curve (AUC) is used to express the level of accuracy.” Topol highlights several studies done to confirm the accuracy of AI algorithms. He first touched on chest x-rays as they are the most performed scans, totaling to 2 billion globally per year. In the study, a convolutional neural network was used to detect pneumonia in more than 112,000 chest x-rays. The results were then compared to those of four radiologists. The algorithm had outperformed the radiologists with an AUC of 0.76. In another research, whole slide imaging (WSI) of breast cancer was conducted by pathologists. The comparison was between 11 pathologists and multiple algorithmic interpretations. The results were shocking due to that 5 of the algorithms outdid the set of specialists.

In regards to the costs, Accenture analysts provided an estimate in the potential of artificial intelligence. Overall, by 2026, $150 billion will be saved annually in the United States (U.S). Furthermore, they identified ten areas where the applications of AI will yield the biggest savings. The five top areas are robot-assisted surgery ($40B), virtual nursing assistants ($20B), administrative work =flow assistance ($18B), fraud detection ($17B), and dosage error reduction ($16B).

Conclusion

The anticipated advantages of AI in medicine are myriad. However, the adoption of these applications to the market is dependent on the support of stakeholders- payers, policymakers, and healthcare providers. From all the journal articles cited, the authors firmly believe that the assimilation of AI in all medicine-related field will be beneficial. Multiple times I have come across the working of a CT scan with AI, and it has effectively diagnosed acute neurologic events all the time. Due to this, I agree with them, and I do not see any other way I can interpret their results. Also, because their studies lack common weaknesses, like poorly described evidence or studies, narrow examinations, failure to demonstrate comparability, bias, and unclear definition of study endpoints.

The authors propose preparation strategies to eliminate the limitations that occur when AI is cultivated. I support their ideologies because they can lead to failures in these systems. Foremost, they declare that medical providers should get well versed with the future advances of AI, and the unknown medicine world they are heading to. Finally, the use of AI in medicine creates a myriad of ethical implications. Smart machines have shown to raise issues of transparency, accountability, and privacy. It is thus critical for the healthcare facilities, together with the regulatory entities and government, institute structures that will monitor the key concerns, and manage negative implications.