Proposal Implementation of Artificial Intelligence for Improving Patient CareJason P BronzoPurdue University

Table of Contents

Proposal: Implementation of Artificial Intelligence for Improving Patient CareJason P. BronzoPurdue University NorthwestProposal: Implementation of Artificial Intelligence for Improving Patient CareBackground of Problem Healthcare, a 3.5 trillion-dollar industry in the U.S., is a field that is rapidly progressing to meet the complex needs of an increasing quantity of patients that are living longer and longer, and subsequently growing more difficult to treat. Due to this, healthcare is met with many difficulties such as understaffing with high-census, complex diagnoses and treatments, difficult surgeries, increased cost of healthcare, and simply human error due to the complex critical thinking skills needed in delivering care. According to the article Criminal Prosecution of Health Care Providers for Unintentional Human Error, human error is inevitable and solutions are needed to be developed to better defend against these errors before effecting patients (The American Association of Nurse Attorneys, 2012). According to the US Census Bureau, in 2010, 12.7% of the national population was 65 or older, while in 2018, 16% were 65 or older with even higher numbers in certain locations (US Census Bureau, 2019). This is alarming because the patient population is increasing faster than the workforce cannot keep up. Additionally, in a study between nine Central Florida hospitals, it was found that 526 patients were turned away due to capacity issues, and nearly half of all ED patients waited an hour or longer to transfer once admitted (Roth, 2019). These statistics show that interventions need to be done to reduce hospital admissions, length of stays, improve patient outcomes, improve overall efficiencies, and more accurately diagnose and treat patients. There are a multitude of ways these issues are being dealt with, such as mandated nurse-to-patient ratios being made law in some states to decrease nurse burnout and improve patient care and outcomes, however, hospitals often do not have the staff to support this system. One of those ways, and likely the best way, is with artificial intelligence, or AI. Technology Alan Turing, who is responsible for the founding of the term and science of AI, defined it as the science of creating intelligent machines that can act as (or more) intelligently as humans (Meetoo & Rylance, 2018). According to Davenport & Kalakota, AI in healthcare has already been proven to perform as well or better than humans in certain tasks such as diagnosing, reading radiology scans, and guiding researchers in organizing and constructing groups for clinical trials (Davenport & Kalakota, 2019). AI is still in its early forms, but even still, it is rapidly growing and causing large changes in healthcare, not seen since the electronic health record, or EHR. There are different forms of AI that are used in healthcare, with some technologies being used since at least the 1970’s. These include machine learning which assists in finding treatment protocols that are most likely to succeed for a patient, Natural Language processing (NLP) which can examine and extract text from unstructured documents and recognize and understand speech, and finally, physical robots which can be used in surgery, or in patient care with moving, repositioning, or communicating (Davenport & Kalakota, 2019). According to Sumi Menon, healthcare AI is expected to grow 40% by 2021, showing how fast the technology is being implemented into healthcare. Menon also mentions that AI can potentially improve outcomes by 30-40% and even cut some costs of healthcare in half, estimating 150 billion dollars in annual savings by the year 2026 (Menon, 2018). An example of a technology that has already been implemented in almost all hospitals that can benefit from AI is EHRs. Nurses spend on average 25% of their work time performing routine administrative activities, primarily being documentation (Davenport, 2019). In comes AI, and now many of these routine processes can be automated and automatically upload to the EHR, saving nurses valuable time that could be spent delivering patient care and possibly even reducing the risk of nurse burnout. This same concept can apply to physicians when charting notes on patients. Many computers now include a tool for physicians to do voice to text that employs the use of AI to understand and translate the dictations into text. This is fascinating because of how well it can work even if a physician has an accent that is difficult to understand.Quality Medical professionals spend much of their time using electronic health records. This is where all patient information is documented and can be found. As stated previously, nurses spend on average 25% of their work time on administrative activities such as documentation, so naturally, reducing this number will allow for more time performing patient care. Routine processes such as checking vital signs, intake and output, etc. can be automated with implementation of AI driven software and technology, subsequently reducing time spent on these routine activities. NLP can be used for extraction of data from notes and different areas of the EHR to gain a more accurate perspective on the patient’s condition. Additionally, time spent on documenting can be reduced with the implementation of voice recognition and dictation using natural language processing, a form of AI. To help decrease the quantity of people being seen in the emergency department, implementation of intelligent virtual assistants can be used to for access to the public to answer questions and behave similar to a doctor (not replace) if the need arises. This has the potential to increase patient engagement, improve self-management skills, and decrease the quantity of patients seen in the ED. In medicine, AI should be implemented into radiology, diagnosis, and treatments. While AI does not replace the physician, it aids them in more accurately and efficiently performing tasks. Implementation of the program Watson by IBM into these areas could vastly improve patient outcomes. Watson is able to look through millions of pertinent medical documents, patient’s EHRs, and images to come to a conclusion on the best course of action, or even detection of bleeds or malignant tumors in scans. Stakeholders Initially, to begin implementation of these changes, I would need to hold a presentation to hospital/company executives in which I would present them with the data research that shows the vast benefits of AI in healthcare, potential for benefits in our hospital specifically, cost of implementing, possible areas of difficulty and limitations, and time needed to transition the changes. With recognition of the potential immediate and long-term benefits, I would hope to secure their buy-in at this point. Upon approval of changes, a given company would be selected by executives to provide resources (technology, software) that best fits the budget allowed. Roles would need to be assigned to those responsible for planning and managing this process, making sure as smooth a transition as possible. With these gross changes in technology, primarily in workflow, all staff will be involved in the change. Those primarily involved however will be IT, physicians, nurses, project managers, nurse educators, and the software companies (spokesperson, IT). The first role, that of the project managers will be to manage this whole process and to create a schedule with goals to have certain aspects completed by given dates. Initially, spokespersons and IT from the company selling their product will come on site and have classes at this point just for management and educators. IT from the company will be there to demonstrate their AI, its best uses, and to install required devices and software in preparation of the launch. Once this is all setup, classes can then be held by educators for the rest of the staff such as nurses and physicians. Feedback will be the roles of nurses and physicians and it is as this point that staff will take a survey on personal opinions toward these changes. Informatics nurses will work with staff and IT to make appropriate changes and adjustments on things that aren’t functioning properly, are bothersome, or anything that simply could be improved to best satisfy the staff using the products.Evaluation of Outcomes (1-3 pages)ConclusionReferencesDavenport, T., & Kalakota, R. (2019). The potential for artificial intelligence in healthcare. Future healthcare journal, 6(2), 94–98. doi:10.7861/futurehosp.6-2-94. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6616181/Meetoo, D., & Rylance, R. (2018). AI: revolution or apocalypse? British Journal of Nursing, 27(19), 1092. http://pnw.idm.oclc.org/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=132531110&site=ehost-liveRoth, M. (2019). AdventHealth uses AI to balance capacity issues among 9 hospitals. Healthcare Leadership Review, 38(11), 8–11. http://pnw.idm.oclc.org/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=139696406&site=ehost-liveMenon, S. (2018, September 24). How Artificial Intelligence is Changing the Healthcare Industry. Retrieved February 7, 2020, from https://www.cabotsolutions.com/how-artificial-intelligence-is-changing-the-healthcare-industryThe American Association of Nurse Attorneys (2012). Criminal Prosecution of Health Care Providers for Unintentional Human Error. Journal of Nursing Law, 15(1), 33–35. http://pnw.idm.oclc.org/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=104473104&site=ehost-liveUS Census Bureau. (2019). Retrieved February 7, 2020, from https://data.census.gov/cedsci/table?q=age&hidePreview=false&tid=ACSST1Y2018.S0101&t=Age and Sex