Barriers to widespread adoption of artificial intelligence in healthcare

Artificial Intelligence (AI) has the potential to significantly improve healthcare delivery. As AI is able to unlock insights and patterns from very large data sets, it lays the foundation for innovative, high-value, enhanced capabilities such as prediction of patient deterioration, recommendations for appropriate intervention for specific conditions, and high-frequency monitoring of many vital signs in parallel. Analysis and Insights. CalmWave founder and CEO Ophir Ronen discusses the barriers to AI adoption and how the healthcare industry can overcome them.

However, the healthcare industry is particularly cautious in adopting artificial intelligence, according to a recent Brookings Institution report opens a new window. While it's natural to treat new technologies with caution, this is especially true in the healthcare world, where providing the best care for patients involves tremendous responsibility. There are many factors that clinicians worry about when adopting artificial intelligence, including fear of being marginalized, fear that errors caused by AI will have a negative impact on patient health (i.e., death), and fear that conclusions based on black-box AI are not well understood.

Before exploring these questions, it is important to understand what healthcare providers will gain from artificial intelligence, especially when it comes to working conditions.

What can artificial intelligence do for health care?

Artificial intelligence has the potential to revolutionize health care by enhancing clinicians’ ability to identify and treat disease. AI systems can analyze large amounts of data from electronic health records, imaging studies and other sources to find patterns that would be difficult for humans to spot. These analyzes can lead to earlier, more accurate diagnoses, better treatment outcomes and more personalized care.

One area where artificial intelligence can have a significant impact is in reducing clinician burnout. Nurses, in particular, are at risk of burnout due to the high demands of their jobs. Artificial intelligence can help alleviate this problem by providing an objective measure of workload based on the frequency of ICU alerts, patient acuity, and the frequency and complexity of interventions. Enabling hospital administrators and managers to understand clinician workload and likelihood of burnout can promote data-driven opportunities to make the workplace healthier, where clinicians want to stay and follow their passion for treatment.

In addition to reducing burnout, AI can help clinicians make more informed decisions by integrating real-time data to provide actionable insights and predictive analytics. For example, AI algorithms can analyze patient data to identify patients at risk for complications and alert clinicians to take preventive measures. This can improve patient outcomes and reduce healthcare costs by avoiding more serious complications.

Overall, artificial intelligence has the potential to transform health care by enhancing clinicians’ ability to analyze large amounts of data and identify patterns that are difficult for humans to detect. By reducing burnout, providing real-time data and predictive analytics, AI can help clinicians make more informed decisions, improve patient outcomes, and reduce healthcare costs.

Common Barriers to Widespread Adoption of Artificial Intelligence

Artificial intelligence seems to be the key to making the lives of health care workers easier. However, there are several risks in introducing complex and unfamiliar technology into such an important industry. In fact, many health care workers worry that AI will do more harm than good to providers and patients.

Here are a few reasons why healthcare providers may be resistant to AI:

1. Explainability

Perhaps the biggest obstacle to the adoption of AI in healthcare is Mysteries surrounding the mechanics of artificial intelligence. How do these algorithms work? How are the above data points generated? “Black box” AI is a thing of the past, and clinicians (and regulators) expect explanations when it comes to AI-based solutions.

"Explainability" refers to the concept that a machine learning model and its output can be explained in a "meaningful" way at a human-acceptable level. In order to confidently implement artificial intelligence into their operations, healthcare practitioners must demonstrate that it will abide by the Hippocratic Oath, which is to "do no harm." Without a thorough understanding of how AI makes decisions, it will be difficult for practitioners to hand over important responsibilities to machines.

2. Bias and Discrimination

Many health care systems are steadily increasing their efforts to address racial disparities and expand access to services for minority and underserved communities. Unfortunately, medicine has a long history of bias. In some cases, artificial intelligence is used to exacerbate the problem.

Practitioners may worry that AI algorithms trained on specific data sets will systematically ignore companywide initiatives to improve health equity, thereby perpetuating discriminatory practices. Any AI-based technology in healthcare today must consider these dynamics when developing more comprehensive and powerful solutions to improve care for everyone.

3.Risk and Comfort

Technology will never be perfect. Healthcare providers strive for perfection because anything less than perfect can mean lives are compromised. The stakes in health care are high, and so are the expectations for any new medical technology. AI-based products are very accurate, but not perfect. Therefore, new technologies based on artificial intelligence may still lead to some errors or failures that may lead to misdiagnosis or mistreatment of critically ill patients. This expectation is not unique to AI, but it does create a high and sometimes unrealistic bar that slows adoption. Additionally, legacy systems face ongoing challenges.

Different organizations have their own systems and methods of patient care. Suppliers often view familiarity and consistency as more important than sophistication and accuracy. A technology is not necessarily good or accurate enough, but it is equally important to consider the clinician's comfort level using and understanding the technology.

4. Lack of Regulation

Although the FDA has approved hundreds of artificial intelligence medical devices, there are no relevant regulations for non-commercial artificial intelligence algorithms in healthcare. The challenge in crafting these regulations stems largely from the speed at which artificial intelligence is developing. This seeming lack of oversight and accountability is understandable for healthcare workers, who would prefer to know that the new technology has been approved by regulators and adhered to certain standards, particularly around privacy and anonymity. aspect.

How to Introduce Artificial Intelligence into Healthcare

Despite the reservations of clinicians, artificial intelligence can and will change the face of healthcare. However, to successfully implement AI-based tools, clinicians must be at the forefront of designing, testing, and training new medical technologies.

Design

To give trust to an AI system, healthcare practitioners must be directly involved in its design and implementation. You can’t blame clinicians, who expect AI developers to share their goals and be fully aware of their concerns.

Hospitals are complex ecosystems with critical workflows. Successfully integrating AI into healthcare systems requires a comprehensive look at existing workflows to improve them rather than add more work. Including health workers in the design phase is critical to ensuring that AI prioritizes usability and integrates seamlessly into daily workflows.

Transparency

Developers of AI systems must provide practitioners with full visibility and transparency into the AI ​​decision-making process. Provide users with not only the end result of the process, but also data to support decision-making. Without this basic requirement, the future of AI in critical care functions seems remote. Clinicians must feel they agree with the design of the algorithm and the data processed by the AI ​​to deliver the desired results.

User Testing

To this end, healthcare workers should have adequate opportunities to test artificial intelligence in clinical settings. These real-world interactions will ultimately reveal which use cases support care delivery for practitioners and patients, and which use cases create unnecessary complications.

Simply rolling AI technology into hospital wards without providing user testing to clinicians will exacerbate clinician concerns about unfamiliarity, bias and risk of failure. Getting clinicians accustomed to using the technology from the outset will alleviate their concerns and improve integration. Additionally, feedback from healthcare professionals will ultimately help AI companies continually improve their technological capabilities to streamline daily tasks and address practitioners’ most pressing needs.

Clinical Evidence

There is one thing that guarantees approval from a healthcare provider: certification. Much healthcare follows a clinical evidence-based approach. Clinical evidence-based medicine (EBM) is a medical practice approach that emphasizes the use of the best available research evidence to guide clinical decisions. The goal of EBM is to improve the quality of patient care by ensuring that treatments and interventions are based on the latest and most reliable scientific evidence. The key word here is evidence.

While this takes more time and may seem like a huge inconvenience and adoption barrier, it is often a necessary step to ensure a safe and sustainable solution. To be clear, there are varying degrees of evidence and the healthcare industry (including regulators) must adapt to conditions, scenarios and exceptions to provide appropriate flexibility to accelerate the use of the technology. Putting evidence behind technology not only improves patient care but also instills the confidence clinicians need to drive adoption.

Health care providers: The purpose of artificial intelligence is to enhance your capabilities

The healthcare industry’s reservations about artificial intelligence are undoubtedly reasonable and deserve to be taken seriously. Start by acknowledging the changes that AI will bring and dispelling the notion that the introduction of AI will immediately modernize the industry.

It is important for healthcare professionals to know that without their input, AI will not be adopted and any AI initiative will have clearly defined goals, values ​​and evidence. Clinicians can, should, and must have a say in the design, testing, and implementation of AI technologies. There is no health care without clinicians. As more healthcare workers have the opportunity to become an integral part of healthcare AI-enhanced technologies, making them more aware of their new capabilities: reducing stress, improving working conditions, and improving patient outcomes, barriers to widespread AI adoption will gradually disappear.

The above is the detailed content of Barriers to widespread adoption of artificial intelligence in healthcare. For more information, please follow other related articles on the PHP Chinese website!