6
Ryan Wang
Professor Halloran
LAMP-M301
December 5, 2021
Is AI the Answer to the High Cost of Medical Imaging?
In the past couple of years, there has been extensive Artificial Intelligence (AI) application in the health sector. This has been brought about by unprecedented pressure in regards to demand and expectation of information technology use in healthcare delivery. There are mixed feelings about the adoption of Artificial Intelligence in the health industry and, more so, its costs and benefits. Artificial Intelligence (AI) has been incorporated in the healthcare realm, with existing evidence indicating that the technology can effectively reduce cost and enhance efficiency. Medical imaging is one field of health where the use of AI could bring about considerable benefits. There is, however, a question of whether AI is causing an increase in the costs of medical imaging. With Artificial Intelligence and Machine Learning seeing significant progress in generating algorithms to process healthcare records, the technology can be effective in medical imaging as AI can classify patterns in data faster than human beings. Artificial Intelligence, together with image recognition capabilities, can enhance the results of patients at the same time, cut down costs of medical imaging. The emergence of AI in the extensive data period can be beneficial in assisting health specialties to better patient care qualities and equip radiologists with the know-how for enhancing precision and efficacy of analysis and treatment. The incorporation of AI in the health sector, particularly in medical imaging, will assist in keeping costs manageable by improving the delivery of quality services without increasing operational costs.
Artificial Intelligence (AI) is assuming a progressively critical role in clinical decision making. Artificial Intelligence application platforms are presently being established or applied for use in various select healthcare applications and (however not restricted to) medical testing, patient intensive care, and understanding healthcare structures. AI algorithms can help make decisions, “These AI algorithms typically employ computerised predictive analysis algorithms to filter, organise, and search for patterns in big data sets from multiple sources and provide a probability analysis upon which healthcare providers can make fast and informed decisions” (Lysaght et al. 300). Data collection is automated, and statistical analysis is done using the technology. The use of AI algorithms would ultimately result in faster decision making owing to its ability to carry out prognosis of data that filters, organizes, and further identifies patterns allowing the relevant parties to make decisions seamlessly. AI application in medical imaging can provide data sets that radiologists and physicians can use to make informed and fast decisions, which in most cases usually takes a lot of time, especially if it is not an emergency. Error in radiology is also a common predicament. This is especially the case as physicians and radiologists experience persistent burnouts as they have to examine numerous imaging studies daily. Therefore, the chances of not paying attention or being efficient enough are relatively high. AI is software, and therefore no burnout or any burden felt whatsoever. This implies that AI can help ease the burnout experienced by radiologists by complementing their skills.
The efficiency of processes coupled with the accuracy of data presented would save radiologists time of having to repeat tests that may have been erroneous. Reduced time equals reduced operational medical costs. It goes both ways in that patients and radiologists will enjoy faster services and processes. Besides, a critical application of AI has been in analyzing and developing Electronic Health Records (EHRs). If not, industries have already moved to maintain records electronically. The health industry has not been left behind. The use of AI has allowed the establishment of electronic data, which is more secure and accessible. The ability of AI to analyze such big data sets means the provision of clinically appropriate statistics instantaneously to healthcare specialists, health systems administrators, and legislators (Lysaght et al., 301). This way, the unnecessary burdens exerted on radiologists will reduce significantly, translating to the efficient delivery of services. When applied to medical imaging, AI is one measure that can reduce inefficiencies, and consequently, it is a cost-saving measure.
Efficient patient monitoring can help a health institution avoid serious problems. If the health of a patient deteriorates, then the health personnel are required to respond quickly. The digitalization of health services enables easy patient monitoring, owing to the availability of digital data. Artificial Intelligence approaches for monitoring patients have increased. Reddy et al. found that “waveform pattern learning can improve monitoring and analysis of electrocardiographs, electroencephalographs, electromyographs and Doppler ultrasounds in hospitals.” AI basically allows for remote patient monitoring, thus allowing for personalized care. Through the use of AI, there will be a holistic view of the patient. This means that it would be possible to meet patient needs and preferences without requiring radiologists to exert additional manual effort. This is made possible through a machine learning system, which has been used in the treatment of patients with chronic ailments for a long time. For instance, among cardiovascular patients, AI can be used to collect data on vital signs and interpret them. If such machine learning algorithms a system is then adopted by the imaging department, it would mean efficient data collection that can aid in the personalization of patient care. With AI’s personalized care and patient monitoring, radiologists’ time can be saved, and more accurate results can be attained.
The machine learning algorithms through AI can be used in medical imaging, in that, during medical imaging, natural language can be used to communicate the findings. This is not just limited to radiologists’ use but can also be used to communicate the appropriate way forward after the imaging to the patient. With such applications, increased medical compliance, in addition to consistent follow-up, can be achieved (Reddy et al., 24). This would ultimately translate into improved service delivery and thus reduced costs. Besides, reduced costs owing to machine learning capabilities exerted by application of AI, could help with image recognition thereby improving patient outcomes. This would also mean reduction is costs associated with medical imaging. As new data is obtained through patent monitoring, radiologists can then make further informed decisions. The medical imaging department should consider extensively applying AI in their processes.
The costs that hospitals incur due to human bias and errors are immeasurable. As applied in medical imaging, AI could help save health institutions from such troubles. Man is to error, and radiologists are no different. Radiologists are prone to make errors and further become bias as they carry out and interpret the medical images. Bias may result from allowing their emotions and even experiences to influence their judgments. AI, a machine, is not prone to emotional effects, and whatever data they analyze, the software only deduce it correctly, providing unbiased information. According to research, “in medical imaging practice, AI has shown impressive accuracy and sensitivity in the identification and characterization of abnormalities leading to enhanced service delivery and quality of patient care” (Antwi, Theophilus, and Benard, 2). As noted earlier, the amount of fatigue that health professionals encounter on a day-to-day basis considerably weighs in on their decision-making capabilities. Repetitive work cannot cause AI technologies to lose focus, and neither can it result in fatigue. Therefore, it is safe to argue that improved quality in medical imaging is heavily reliant on the implementation and adoption of AI technology, without which the hospitals will continue incurring costs related to errors and malpractices. In one study that attempted to understand AI’s ability to diagnose intracranial large vessel occlusions (LVO) in stroke, it was found that when compared to traditional care or instead means of detection, AI was more likely to make the diagnosis. The rate of diagnosis ranged from 87.8 to 97.9 percent (Van Leeuwen et al., 7). This is a very high rate, proving that AI can help diagnose some of the ailments that would not have been discovered through standard care. With heightened detection of LVO, both short-term and long-term costs reduce as the health facility gets to save on general consumption. Many would associate AI in medical imaging with just image interpretation; however, there is more to this as it has been used to enhance image acquisition and enable patient care.
In brief, AI incorporation into medical imaging will ultimately reduce operational costs and provide quality services. The medical imaging field stands to benefit a lot from AI-assisted systems. For instance, AI can help reduce errors common among radiologists, enable efficient and quality decision making, and further allow patient monitoring through easy and timely diagnosis, which is a prerequisite to improved delivery of health services. Such systems push for better and accurate health records, improving service and reducing mistakes and backlogs that make services expensive. While the planet cannot entirely control medical costs, Ai at least helps manage some unnecessary expenses incurred in medical imaging. AI can only strive to manage the costs and improve healthcare delivery like medical imaging. However, some issues should be addressed when AI is introduced in healthcare generally. Accountability is essential when the system makes recommendations that conflict with healthcare professionals’ opinions because there should be no ambiguity on where liability rises. Moreover, the generated algorithms should be explainable and sensible.
Works Cited
Antwi, William Kwadwo, Theophilus N. Akudjedu, and Benard Ohene Botwe. “Artificial intelligence in medical imaging practice in Africa: a qualitative content analysis study of radiographers’ perspectives.” Insights into imaging vol.12, no.1, 2021: 1-9., https://www.researchgate.net/publication/352464282_Artificial_intelligence_in_medical_imaging_practice_in_Africa_a_qualitative_content_analysis_study_of_radiographers%27_perspectives.
Lysaght, Tamra, et al. “AI-Assisted Decision-Making in Healthcare.” Asian Bioethics Review, vol. 11, no. 3, 2019, pp. 299–314., https://link.springer.com/content/pdf/10.1007/s41649-019-00096-0.pdf
Reddy, Sandeep, et al. “Artificial Intelligence-Enabled Healthcare Delivery.” Journal of the Royal Society of Medicine, vol. 112, no. 1, 2018, pp. 22–28., https://journals.sagepub.com/doi/pdf/10.1177/0141076818815510
Van Leeuwen, Kicky G., et al. “Cost-effectiveness of artificial intelligence aided vessel occlusion detection in acute stroke: an early health technology assessment.” Insights into imaging vol.12, vol.1, 2021: 1-9., https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8464539/