Introduction
It is said that Aristotle (384 – 322 B.C.) was perhaps the last person in the world to have known all that could be known (Qumodo, 2017). As human knowledge continued to increase, The Great Library of Alexandria (285 – 246 B.C.) was said to have housed the sum of all human knowledge (MacLeod, 2000). Today, we are advancing our collective knowledge at an astounding rate (Qumodo, 2017). While today’s Internet virtually brings mountains of data to the fingertips with anyone with computer, it is often either unvetted or too much data for the individual to effectively use without assistance from complex artificial intelligence (AI) algorithms. By leveraging AI technologies, we can effectively mine large amounts of data and make that data available in a useable way, thereby creating value for the individual user (Gregory, 2021). Hygeia, named after the Greek goddess of health, is a proposed AI application that will bring the data needed by clinicians to provide better, faster, and cheaper healthcare.
Business Proposal for Hygeia AI
Situation
The healthcare delivery system is complex and expensive (Rouse and Serban, 2014). The complexity of the system is driven by its interdependencies; research, diagnosis, treatment, and billing, which then effect costs, quality, and access to healthcare (Rouse and Serban, 2014). This inherent complexity is further aggravated by critical shortages in healthcare providers (Johnson, 2022). As a result, healthcare costs around the world are increasing, with the United States spending $4.1 trillion in 2020, or $12,000 per capita (Wager, Ortaliza, and Cox, 2022). The complexity of the healthcare delivery system can, and must be reduced to provide better, faster, cheaper care for everyone (Holley and Becker, 2021).
Background
The evolution and widespread adoption of the electronic medical record (EMR) has successfully created efficiencies in medical care as well as produced an explosion in available data (Shimonski, 2021) (Sidey-Gibbons & Sidey-Gibbons, 2019). Some have estimated that the implementation of EMR systems will save more than $81 billion annually (Hillstead, 2005). Emerging technologies such as artificial intelligence (AI) has also helped deliver high-quality healthcare in an effective and cost-efficient manner (Park & Cha, 2022). AI technologies have already been successfully integrated into a variety of imaging solutions as well as some medical devices (Reddy, 2018), and there is evidence that AI algorithms are performing well at tasks such as analyzing medical images and correlating symptoms and other data from EMRs (Bohr and Memarzadeh, 2020). This success leads many to believe that AI technology “can bring improvements to any process within healthcare operation and delivery” (Bohr and Memarzadeh, 2020).
Assessment
Clinicians have always relied on computers to analyze data to aid in studies and in developing appropriate diagnosis (Davenport and Kalakota, 2019). The advent of artificial intelligence (AI) with its ability to significantly improve the data analysis process has presented healthcare providers with an opportunity to optimize the delivery of care while reducing costs (Reddy, 2018). Because much of the healthcare delivery process is rule-based, AI technologies are well suited to support all aspects of the continuum of care from research to diagnosis, and treatment to medical billing and reimbursements (Davenport and Kalakota, 2019). Insurance companies, the primary purchaser of healthcare through “managed care” have already broken down much of healthcare delivery into rule-based processes in order to standardize care and subsequent billing (Altman, 1999). Cost-effective healthcare can be achieved by reducing the complexities of the delivery system (Holley and Becker, 2021). AI systems are key enablers of reducing that complexity.
Recommendation
Hygeia AI is a proposed suite of application comprising of a variety of AI technologies that are capable of integrating with today’s common healthcare delivery systems that will reduce complexity in today’s healthcare delivery model thereby improving the speed and quality of care and reducing overall costs.
Hygeia AI’s machine learning technology integrates large amounts of external data with the most prevalent EMR systems to provide local clinicians with customized, real-time, diagnosis and treatment plans that have high-probability outcomes particular to their patient’s history and exams. Since these treatment plans are statistically optimal to produce the best outcomes, they are inherently more effective in both quality and cost than today’s model. Additionally, by working directly with regulatory agencies and insurance providers, these treatment plans will be pre-approved with standardized medical billing codes and in line with existing reimbursement guidelines.
This machine learning system also brings global research and clinical trial information to the desktop of subscribed providers. Similar to Lexus-Nexus for attorneys (www.LexisNexus.com), Hygeia AI’s research database brings the collective knowledge of all medical research to the local family practice physician. Hygeia AI uses pattern detection and probability factors within its algorithms to predict medical outcomes to deliver cost-effective, quality care (Reddy, 2018). It works in the background of EMRs to automatically collect variables and share them as data points to be applied to other similar cases (Reddy, 2021).
Research Findings
MYCIN was the first application of AI in healthcare for diagnosing blood-borne bacterial infections (Bush, 2018). Since it was not integrated into clinical practice this rule-based system failed to be widely adopted (Davenport and Kalakota, 2019). The most widely known AI system used in healthcare is IBM’s Watson Health. IBM partnered with MD Anderson Cancer Center to create an advisory tool for oncologists. This first major step in integrating AI into healthcare showed a lot of promise. In addition to being a tool for oncologists, Watson Health was to help pharmaceutical companies develop drugs, and help match patients to clinical trials. Unfortunately, Watson Health never really worked and was sold off for parts (O’Leary, 2022). More recently, Google has partnered with Mayo Clinic to “transform patient and clinician experiences, improve diagnostics and patient outcomes, as well as enable it to conduct unparalleled clinical research” (Kurian, 2019). This 10-year strategic partnership is likely to produce some workable proof-of-concept applications that will likely be readily adopted because of the partnership with end users. As recently as this week, Google announced a partnership with Epic, the largest medical records company in the United States, to enable cloud storage of customer data to better leverage AI technology (Landi, 2022). Google entrenched in AI health applications as they are also applying natural language processing to help digitize patient’s EMR (Bousquette, 2022), and looking to develop machine learning models for chronic diseases (Kurian, 2019).
Venture capitalists are backing health AI companies as their value propositions predict an annual healthcare savings of $150 billion in the United States by 2026 (Kalis, Collier, and Fu, 2018). The “10 AI applications that could change healthcare” illustrated below in figure 1 (Kalis, Collier, and Fu, 2018).
Figure 1. (Kalis, Collier, and Fu, 2018)
Issues with the Proposal
AI technology is not without its challenges (Challen, et al., 2018), and there are a few major issues to consider when implementing AI systems (Du and Xie, 2021).
Data Bias
Encapsulated in an AI algorithm, whether through raw data or programming, is the potential for built-in biases (Du and Xie, 2021). Some biases are unintentional, while others are deliberate with the goal to achieve a specific outcome. One example is the misdiagnoses of a particular symptom that may lead to others relying on that data for future diagnosis (Challen, et al., 2018). Unverified, that data would corrupt the reliance on any other data making the system not just worthless, but dangerous. Deployed models must be continuously refined and trained with strict governance (Reddy, 2021).
Cyber-Security
Data Security is of prime importance to healthcare providers and failures can be found in today’s headlines. Health systems across the country have been hit with costly ransomware cyber-attacks. Recently CommonSpirit reported a widespread attack on their systems (CommonSpirit, 2022). Successful AI model development will ensure data security is a key component.
Data Privacy
Aside from health systems being able to protect the data they are entrusted, which still is a major challenge, how businesses use and share that data presents major issues and opportunities for the use of AI in businesses in the future. This includes the transparency to the consumer as to what data is being collected and with whom the data is being shared. The sharing and disclosing of data in healthcare is further complicated by the Health Insurance Portability and Accountability Act of 1996 (HIPPA). HIPPA states that patient data cannot be disclosed with out the patient’s consent. Because of this restriction, patient identifiers, or “personal health information” (PHI), must be scrubbed of that data before it can be transmitted or shared (Shimonski, 2021).
Benefit with Proposal
AI technology has already been successfully integrated into various aspects of the healthcare delivery system (Reddy, 2021). Despite IBM’s early failed attempt with Watson Health, Google and others are positioning themselves to take the next steps. Hygeia AI aims to leverage these lessons learned, with the goal to reduce the complexity of the healthcare delivery system. The benefits of this endeavor could create significant value through the reduction of administrative workflow, fraud detection, dosage error reduction, interconnected machines, clinical trial participation, increased accuracy in preliminary diagnosis and treatment, and the improved security of protected health information. Much like the Google/Mayo Clinic partnership, Hygeia AI must ensure industry and regulatory infrastructure collaboration, as well as partnering with clinicians as the subject matter experts and ultimate adopters (Reddy, 2021).
Ultimately, Hygeia AI will improve the quality of healthcare at reduced costs. By leveraging AI technology needed to integrate systems such as the EMR, and to develop predictive tools to aid clinicians in early diagnosis and optimal treatment of chronic medical conditions, healthcare will become more personalized and effective thereby enabling access equality and the democratization of healthcare (Reddy, 2021).
Future Uses of Hygeia AI in Healthcare
The future of Hygeia AI will undoubtably be the integration of patient care with the AI technology already embedded within the application. Virtual nursing assistants, either in the form of robots or other devices, may be employed with interactive AI technology to allow patients to interface with the system for a number of value-added use cases. Some of these applications may include accessing their medical records, providing translations in a wide variety of languages, and communicating discharge instructions. These interactions can be at the patient’s pace with unlimited opportunities for follow-up questions ultimately helping the patient feel more informed. These in person use cases can also be leveraged into virtual visits providing the same types of information.
Conclusion
The complexity of healthcare stems from its complicated delivery model with disparate systems and endless rules to navigate which ultimately increase the cost of healthcare. Hygeia AI is well suited to address these complexities. By aiding clinicians to predict and diagnose patients with the most up to date research, to aid in the development of customized treatment plans that will produce the best outcomes with the highest degree of certainty and integrate the EMR in a manner that protects patient privacy while sharing valuable medical data with the rest of the world, is just what the doctor ordered!
References
AI Multiple. (2020, September 26). Future of AI according to top AI experts of 2020: In-depth guide. https://research.aimultiple.com/future-of-ai/
Altman, R. B. (1999). AI in medicine: The spectrum of challenges from managed care to molecular medicine. AI Magazine, 20(3), 67. https://doi.org/10.1609/aimag.v20i3.1467
Bohr, A., and Kaveh Memarzadeh, K. (2020). Chapter 2 – The rise of artificial intelligence in healthcare applications. Artificial Intelligence in Healthcare 25-60. https://doi.org/10.1016/B978-0-12-818438-7.00002-2
Bousquette, I. (2022, March 24). Mayo Clinic assessing Google AI tool as it seeks more insights from patient records. The Wall Street Journal. https://www.wsj.com/articles/mayo-clinic-assessing-google-ai-tool-as-it-seeks-more-insights-from-patient-records-11648159790
Bush J. (2018). How AI is taking the scut work out of health care. Harvard Business Review. https://hbr.org/2018/03/how-ai-is-taking-the-scut-work-out-of-health-care
Challen, R., Denny, J., Pitt, M., Gompels, L., Edwards, T., and Tsaneva-Atanasova, K. (2018). Artificial intelligence, bias and clinical safety. BMJ Quality & Safety Journal. http://dx.doi.org/10.1136/bmjqs-2018-008370
Cheng, M., Li, X., & Xu, J. (2022). Promoting healthcare workers’ adoption intention of artificial-intelligence-assisted diagnosis and treatment: The chain mediation of social influence and human-computer trust. International Journal of Environmental Research and Public Health, 19(20). https://doi.org/10.3390/ijerph192013311 .
CommonSpirit. (2022, October 17). CommonSpirit Update. https://www.commonspirit.org/update
Davenport, T., & Kalakota, R. (2019). The potential for artificial intelligence in healthcare. National Library of Medicine. Future Healthcare Journal 6(2): 94-98. https://doi.org/10.7861/futurehosp.6-2-94
Dilmegani, C. (2020, October 21). Future of AI according to top AI experts: In-depth guide for 2022. AI Multiple. https://research.aimultiple.com/future-of-ai/
Du, S. and Xie, C. (2021). Paradoxes of artificial intelligence in consumer markets: Ethical challenges and opportunities. Journal of Business Research, V129, 961-974. https://doi.org/10.1016/j.jbusres.2020.08.024
Gregory, R.W., Henfridsson, O., Kaganer, E., and Kyriakou, H. (2021). The role of artificial intelligence and data network effects for creating user value. Academy of Management Review. https://doi.org/10.5465/amr.2019.0178
Hedgepeth, W. O. (2018, December 6). The growth of artificial intelligence in our society. In Cyberdefense. https://incyberdefense.com/editors-picks/growth-artificial-intelligence/
Heillila, M. (2022, October 20). A bias for AI will help catch unfair algorithms faster. MIT Technology Review. https://www.technologyreview.com/2022/10/20/1061977/ai-bias-bounty-help-catch-unfair-algorithms-faster/
Hillstead, R., Bigelow, J., Bower, A., Girosi, F., Meili, R., Scoville, R. and Taylor, R. (2005). Can electronic medical record systems transform health care? Potential health benefits, savings, and costs. Health Affairs. https://doi.org/10.1377/hlthaff.24.5.1103
Holley, K., & Becker, S. (2021). AI-First Healthcare (1st edition). O’Reilly Media, Inc.
Johnson, S.R. (2022, July 28). Staff shortages chocking U.S. health care system. U.S. News & World Report. https://www.usnews.com/news/health-news/articles/2022-07-28/staff-shortages-choking-u-s-health-care-system
Kalis, B., Collier, M., and Fu, R. (2018, May 10). 10 promising AI applications in health care. Harvard Business Review. https://hbr.org/2018/05/10-promising-ai-applications-in-health-care
Khan, M. A., Ashraf, I., Alhaisoni, M., Damaševičius, R., Scherer, R., Rehman, A., & Bukhari, S. A. C. (2020). Multimodal brain tumor classification using deep learning and Robust feature selection: A machine learning application for radiologists. Diagnostics, 10, 565.
Khan, M. A., Sharif, M. Akram, T., Raza, M., Saba, T., & Rehman, A. (2020). Hand-crafted and deep convolutional neural network features fusion and selection strategy: An application to intelligent human action recognition. Applied Soft Computing, 87, 105986.
Kurian, T. (2019). How Google and Mayo Clinic will transform the future of healthcare. Google Cloud Blog. https://cloud.google.com/blog/topics/customers/how-google-and-mayo-clinic-will-transform-the-future-of-healthcare
Lancet. (2017). Artificial intelligence in health care: within touching distance. The Lancet, 390(10114):2739. https://doi.org/10.1016/S0140-6736(17)31540-4.
Landi, H. (2022, November 16). HLTH22: Google, Epic ink deal to migrate hospital EHRs to the cloud to ramp up use of AI, analytics. Fierce Healthcare. https://www.fiercehealthcare.com/health-tech/google-epic-ink-deal-migrate-hospital-ehrs-cloud-ramp-use-ai-analytics
Larsson, S., and Heintz, F. (2020, May 5). Transparency in artificial intelligence. Internet Policy Review 9-2. DOI: 10.14763/2020.2.1469. https://policyreview.info/concepts/transparency-artificial-intelligence
Lee, D., & Yoon, S. N. (2021). Application of artificial intelligence-based technologies in the healthcare industry: Opportunities and challenges. International Journal of Environmental Research and Public Health, 18(1), 271–. https://doi.org/10.3390/ijerph18010271
MacLeod, R. (2000). The Library of Alexandria: Centre of learning in the ancient world. I.B. Tauris. Pp. 1-18.
Merkusheva, D. (2020, March 25). 10 humanoid robots of 2020. https://www.asme.org/topics-resources/content/10-humanoid-robots-of-2020
Mudgal, S. K., Agarwal, R., Chaturvedi, J., Gaur, R., & Ranjan, N. (2022). Real-world application, challenges and implication of artificial intelligence in healthcare: an essay. The Pan African Medical Journal, 43, 3–3. https://doi.org/10.11604/pamj.2022.43.3.33384
O’Leary, L. (2022). How IBM’s Watson went from the future of health care to sold off for parts. Slate. https://slate.com/technology/2022/01/ibm-watson-health-failure-artificial-intelligence.html
Park, S. H., & Cha, W. C. (2022). Application Strategies for Artificial Intelligence- based Clinical Decision Support System: From the Simulation to the Real-World. Healthcare Informatics Research, 28(3), 185–187. https://doi.org/10.4258/hir.2022.28.3.185
Qumodo. (2017, April 30). The last person to know everything. Artificial Intelligence. https://qumo.do/blog/3
Reddy, S. (2021). Artificial intelligence: applications in healthcare delivery. Productivity Press. https://doi.org/10.4324/9780429317415
Rehman, A., Sadad, T. Saba, T., Hussain, A., & Tariq, U. (2021). Real-time diagnosis system of COVID-19 using X-Ray images and deep learning. IEEE IT Professional. doi:10.1109/MITP.2020.3042379
Rouse, W.B., and Serban, N. (2014). Understanding and managing the complexity of healthcare. MIT Press. https://books.google.com/books?hl=en&lr=&id=1vPvAwAAQBAJ&oi=fnd&pg=PP8&dq=health+care+complexity+and+costs&ots=G2VRRkIepB&sig=TGnOalx4eIczYFW13V8DTUK1zrI#v=onepage&q=health%20care%20complexity%20and%20costs&f=false
Saba, T., & Khan, A. (2021). Intelligent computing applications for COVID-19: predictions, diagnosis, and prevention (Amjad R. Khan, Eds.; first edition.). CRC Press.
Shimonski, R. (2021). AI in healthcare: How artificial intelligence is changing IT operations and infrastructure services. Wiley.
Sidey-Gibbons, J., and Sidey-Gibbons, C. (2019). Machine learning in medicine: a practical introduction. BMC Medical Research Methodology 19, 64. https://doi.org/10.1186/s12874-019-0681-4
Wager, E., Ortaliza, J, and Cox, C. (2022). How does health spending in the U.S. compare to other countries? Health System Tracker, Peterson-KFF. https://www.healthsystemtracker.org/chart-collection/health-spending-u-s-compare-countries-2/#GDP%20per%20capita%20and%20health%20consumption%20spending%20per%20capita,%202020%20(U.S.%20dollars,%20PPP%20adjusted)
WGU. (2021, September 21). How AI is affecting information privacy and data. Western Governors University. https://www.wgu.edu/blog/how-ai-affecting-information-privacy-data2109.html#close
White House. (2016). Preparing for the future of artificial intelligence. https://obamawhitehouse.archives.gov/sites/default/files/whitehouse_files/microsites/ostp/NSTC/preparing_for_the_future_of_ai.pdf
Yuba, M., & Iwasaki, K. (2022). Systematic analysis of the test design and performance of AI/ML-based medical devices approved for triage/detection/diagnosis in the USA and Japan. Scientific Reports, 12(1), 16874–16874. https://doi.org/10.1038/s41598-022-21426-7
