How AI is shaping the current state of the healthcare industry.
The topic of Artificial intelligence (AI) is a complex one, as the world continues to learn both about how AI works and how we can take advantage of its computing abilities while remaining ethical. AI softwares have now been introduced across the U.S. within hospitals and health systems, often making physician workflows more efficient and diagnosis results more accurate.
AI technology has begun to encompass just about every aspect of the healthcare field – from combatting physician burnout with technology, battling health misinformation, AI-specific health research, and health systems coming together to regulate and control the use of AI as a healthcare tool.
Read on as Repertoire Magazine breaks down the latest in AI healthcare news.
AI and Health Information
Cedars-Siani study shows that AI can evaluate cardiovascular risk during CT scan
AI has largely taken off in healthcare because the technology can assist researchers and physicians in more accurately identifying both diseases and their treatments.
A recent study developed by researchers at Cedars-Siani found that AI can accurately evaluate cardiovascular risk during patient CT scans without contrast. This accurate, technology-based imaging method measures coronary calcium and sizes of patient’s heart chambers and muscle, which makes identifying cardiovascular risk in patients much less expensive.
Research conducted in the Cedars-Siani study incorporated two different artificial intelligence models to evaluate data on coronary calcium and heart muscle chamber sizes from nearly 30,000 patient imaging records. Researchers were able to determine that these new measures using AI are a better indicator of cardiac risk than an individual radiologist’s identification of abnormalities.
Generative AI to fight health misinformation
Physicians and healthcare leaders across the industry are searching for ways to keep patients and communities safe from costly and deadly medical misinformation. The COVID-19 pandemic and social media exacerbated the negative impacts of medical misinformation, which threatens the well-being of patients and leads to burnout for public health officials.
To combat misinformation, researchers at the University of California Davis launched the Health Cloud Innovation Center (UCDH CIC) in November of 2021, with the goal of solving issues within digital health innovation.
The UCDH CIC, in collaboration with The University of Pittsburg and The University of Illinois Urbana-Champaign (UIUC), started the groundwork for Project Heal, an open-source AI/ML-based toolkit concept that assists public health officials in protecting communities from misleading and incorrect health information. Project Heal provides public health officials with the tools to manage workloads more efficiently. It also allows them to shift from reactive to proactive behaviors, allowing community education to empower individuals to make informed health decisions.
Project Heal will give public health professionals the ability to generate, edit, tweak, and adapt counter messaging of false health claims, with messaging specific to certain populations. To achieve this, Project Heal will use generative AI foundation models (FMs) through Amazon Bedrock, such as Amazon Titan. Through this technological advancement, Amazon Bedrock can generate more personalized messaging by combining trusted information and user preferences.
Study shows potential for using AI tools to detect HAIs
Healthcare-associated infections (HAIs) are a serious, costly disease in both hospitals and healthcare environments. According to the CDC, there were approximately 687,000 cases of HAIs in acute care hospitals in the U.S., and 72,000 HAI-related deaths among hospital patients in 2015. About 3% of all hospital patients have at least one HAI at any given time.
A recent study published in the American Journal of Infection Control (AJIC) shows that AI technology can be of great assistance in accurately identifying cases of HAIs in clinical scenarios. Researchers at Saint Louis University and the University of Louisville School of Medicine evaluated the performance of two AI-powered tools for accurate identification of HAIs. One tool was built using OpenAI’s ChatGPT Plus and the other was developed using an open-source large language model known as Mixtral 8x7B.
For all six cases tested, both AI tools identified the specific HAI accurately when given clear prompts. The researchers found that missing or ambiguous information in the descriptions could prevent the AI tools from producing accurate results. This research exemplifies the potential for incorporating AI technology as a component of routine infection surveillance programs and highlights the need for clear and consistent language when programming AI tools within healthcare.
Many hospitals have a variety of systems in place to monitor and control infection risk, however they often require extensive training and resources. AI technology could potentially serve as a cost-effective alternative to current surveillance programs for cost savings and better protection for high-risk patients.
AI and Physician Burnout
Using AI to improve EHR
According to American Medical Association’s (AMA) Dr. Jesse Ehrenfeld, complicated data management systems can cause stress for physicians as they navigate extra administrative responsibilities in the workplace.
To combat this issue, healthcare practices have begun to use AI tools that integrate data automation to make the process of triaging and responding to patient messages much more efficient, noted Dr. Ehrenfeld. Administrative AI tools leverage technology to more efficiently accomplish time-consuming tasks, alleviating work-related burdens for physicians.
According to Dr. Ehrenfeld, the healthcare field is also starting to see practices experiment with AI scribes to cut down on the time physicians spend on documentation. This process uses generative AI to address long-standing problems within healthcare and physician burnout by addressing the most time-consuming tasks.
DAX Copilot to automate the creation of clinical documentation and reduce physician burnout at Stanford Health Care
Nuance Communications and Stanford Healthcare have deployed an AI system that reduces heavy administrative workloads that lead to physician burnout, called the Nuance Dragon Ambient eXperience (DAX) Copilot.
The DAX Copilot software expands access across the healthcare industry to personalized and high-quality care, primarily by automating the creation of clinical documentation during patient exams. The technology automatically and securely drafts clinical summaries of exam room and telehealth conversations quickly for review and entry into the Electronic Health Record (EHR). It gives physicians more time to see more patients each day by reducing the time needed to create clinical documentation.
According to Standford Health researchers, in a preliminary survey of Stanford Health Care clinicians using DAX Copilot, 96% of physicians stated that it was easy to use, and 78% reported that it expedited clinical notetaking. About two-thirds reported that DAX Copilot saved time.
DAX Copilot optimizes administrative efficiency, enhances care quality and value, increases access to care among communities, and trains the next generation of clinicians to advance precision care.
Mayo Clinic and Google Cloud collaborate to improve clinical workflows with generative AI
To determine treatment options and help physicians with clinical definitions, conditions, and diagnoses, healthcare professionals often use information from various sources including medical records, research papers, and health guidelines. Healthcare data is often stored in clinical offices in numerous different locations and in a variety of formats, making it difficult at times for physicians to quickly find needed health information.
To make clinical workflows easier and more efficient for patients, Mayo Clinic and Google Cloud have collaborated for the benefit of the healthcare industry to create the Enterprise Search in Generative AI App Builder (Gen App Builder), which improves the efficiency of clinical workflows, makes it easier for clinicians and researchers to find the information they need quickly and easily, and ultimately helps to improve patient outcomes.
The Google Could Gen App Builder technology works by unifying data across dispersed documents and databases, making it more efficient to quickly search through and analyze needed health information. Mayo Clinic is an early adopter of the technology and is currently exploring how Google-quality search and generative AI can bring much needed information to doctors quickly and easily.
AI Partnerships
AI can help cancer patient receive personalized and precise treatment faster
With the industry-wide adoption of AI, some health systems have begun to partner with technology leaders to implement more effective healthcare solutions and improve patient care. Providence, a 51-hospital healthcare organization across the Western U.S., has developed a research prototype of AI tools, with the help of Microsoft, that sort through patient databases. The technology aims to find the best therapies available to patients and advance their cancer treatment.
Patient information exists in a variety of formats – electronic medical records (EMRs), imaging scans, genomics, and many varieties of lab tests. The same information might be noted using different formats, and the core of the patient information requires synthesizing a large amount of unstructured data. Fortunately, AI is very good at summarizing unstructured data in text form.
According to Providence and Microsoft, AI can assist doctors in treating cancer patients with more refined lab tests, scans, and genetic analyses that can help promote a better understanding of an individual patient’s case.
Summarizing these types of data and results into personalized therapies can adapt treatments and medications to each patient’s genetic biomarkers. Assessing all this information into usable data, for one doctor alone, is an unfathomably huge task and AI cuts the processing time for these tasks in half.
Providence is working with Microsoft on more AI prototype tools that improve care for cancer patients and accelerate progress in understanding cancer. Providence and Microsoft are also working together on new technologies for machine learning using diverse data generated and managed by Providence.
New consortium of healthcare leaders form Trustworthy and Responsible AI Network (TRAIN)
To increase the regulation and accessibility of AI technology, numerous healthcare organizations have come together to ensure the quality, safety, and trustworthiness of AI softwares within the healthcare field.
Organizations including AdventHealth, Advocate Health, Boston Children’s Hospital, Cleveland Clinic, Mercy, Mount Sinai Health System, Northwestern Medicine, Providence, UT Southwestern Medical Center, and many more, in partnership with and Microsoft (as the technology enabling partner), launched the Trustworthy & Responsible AI Network (TRAIN), which aims to operationalize responsible AI principles across the healthcare industry.
According to UT Southwestern Medical Center, AI holds the potential to transform the healthcare industry, can be used to help screen patients, develop new treatments and drugs, and enhance overall public health.
Across the UT Southwestern health system specifically, researchers are utilizing artificial intelligence to break down complex biological processes in human health to improve patient care. Examples include the development of a deep learning system to predict which melanoma cell lines will metastasize, and the deployment of an AI-powered location system to track equipment and improve efficiency in the Radiation Oncology facilities, according to UT Southwestern.
All of TRAIN’s members will help improve the quality and accessibility of AI by sharing best practices among the organization – including information on the reliability of AI algorithms and the skillsets required to manage the technology, providing tools and best practices for studying the efficacy of AI, and facilitating the development of a federated national AI outcomes registry for organizations to share and learn from.