What Healthcare Professionals Need to Know About AI in 2020
Rapid advances in artificial intelligence (AI) promise to revolutionize healthcare in the coming years. The use of AI in healthcare will improve patient care, expand access, reduce costs and produce higher levels of efficiency and innovation.
To prepare for the coming revolution, healthcare professionals need to understand the new benefits, risks and challenges of AI. This article is a roundup of essential reading for nurses and other healthcare professionals on the subject of AI.
The Number and Variety of Digital Tools for Healthcare Application Is Growing Exponentially
The rapid adoption of AI will transform the field of medicine through greater access to healthcare data from digital devices.
However, as a PharmaExec editorial says, the exponential growth in the number and variety of available digital devices — wearables, smartphone apps, and medical devices that allow for remote monitoring — can be overwhelming for healthcare providers.
These new digital tools have roots in traditional telemetry. Business Insider explains that digital devices that are connected to the internet contain a variety of sensors that collect data remotely to support disease prevention, diagnosis and treatment by healthcare providers.
FDA Regulation Is Limited
Over the past five years, the U.S. Food and Drug Administration (FDA) has approved 40 AI-enabled devices. Many of these tools improve early detection, analysis, diagnostics, testing and treatment (as illustrated by this useful diagram in The Medical Futurist).
The regulation of many other digital tools for healthcare applications remains limited, leaving healthcare providers without clear guidance on potential risks.
Digital Devices Provide Greater Transparency on Patient Activities
Among several important benefits, digital devices allow patients to play a more active role in their own healthcare.
Mobihealthnews says these features give patients a new level of transparency and more control over their health.
Meanwhile, healthcare providers gain access to data to support diagnosis, treatment and scientific research.
Digital Devices Offer a New Data Source for AI
Since they typically integrate with other technologies, digital devices also offer a source of seamless healthcare data collection that can be fed into other AI systems.
For example, as reported in Nature, a team of researchers on the West Coast developed a machine learning tool that can detect and diagnose a specific cardiovascular disease from data provided by a wearable biosensor wristband.
Healthcare Providers Need to Understand a New Set of Risks and Challenges
Despite many benefits, AI tools introduce new risks and challenges that can be daunting for healthcare providers.
The Workforce Will Need to Upskill
The AMA Journal of Ethics says that to fully leverage the inherent advantages of AI for healthcare, educators must develop a workforce with new skill sets built around knowledge management and data analytics.
And an article from the Journal of Informatics in Health and Biomedicine outlines the qualifications of a data scientist in healthcare and argues that a competitive talent market will require healthcare providers to train their existing workforce in new skills.
Data Infrastructure Needs More Attention
Healthcare providers will also need access to high-quality, aggregated and broadly representative datasets, according to AMA Journal of Ethics research.
Nature suggests the requisite data infrastructure for using AI effectively does not yet exist. Patient data remains siloed since every healthcare provider uses its own software produced by different vendors. This separation of data will continue to limit how much AI can be leveraged by healthcare institutions.
Healthcare Must Confront Black Boxes and Algorithmic Bias
AI raises the spectre of many new risks, but it’s hard to know how significant those risks are due to a lack of transparency.
For example, complex AI tools often suffer from a “black box” effect whereby the process of analyzing data inputs and generating results can’t be observed without AI expertise.
However, one article in the AMA Journal of Ethics argues that the black box shouldn’t be a major concern from the perspective of patients.
Another concern is algorithmic bias, in which human bias in a program’s design or data collection inputs is echoed in the program’s results. Smithsonian Magazine suggests that black boxes and algorithmic bias in AI could amplify healthcare inequalities rather than reduce them.
Healthcare Must Confront Data Privacy, Security and Consent Issues
Healthcare providers must come to grips with what happens to data after it is collected. They should pay particular attention to how data is used and secured against unauthorized access or tampering.
An AMA Journal of Ethics article asks provocative questions about how much patients need to be informed about the role of an AI "assistant" and provide their consent for collection and use.
FierceHealthcare proposes that healthcare providers will need to invest in cybersecurity measures to prevent the unauthorized release of private health information to avoid costly class-action lawsuits, as reported by CPO Magazine.
The New York Times summarizes a Science journal article warning against cyberattacks that manipulate the operations and results of AI in healthcare. The tampering of data to skew results represents another major risk of using AI for healthcare.
The Medical Futurist recommends healthcare providers grapple with challenges such as ensuring adequate data storage, developing a skilled workforce and investing in computing power needed for analyzing enormous volumes of data.
Other Helpful Reading on AI in Healthcare:
To raise awareness of some of the potential risks and unintended consequences of AI, the National Academy of Medicine recently produced a report designed as a comprehensive resource for healthcare professionals.
Brookings Institution’s Artificial Intelligence and Emerging Technology (AIET) Initiative recently released a report highlighting potential solutions to challenges described above, including improved data infrastructure, better oversight by the FDA and changes to medical education to prepare the healthcare workforce.
Given the many challenges discussed above, a Harvard Business Review article argues that the adoption of AI by the healthcare industry will be slower than anticipated.
Explore more content like this article
Mom, Career Goals Motivate SNHU Grad to Finish Her Bachelor's Degree
Just a year after she decided to go back to school, Shereka Johnson-Bean ’20 is preparing to submit the final project for her bachelor’s degree.
Partnership Helps McDonald’s Employee Earn Her Degree – and Keep Going
Shannon Griffin’s first job was as a crew member at a McDonald’s restaurant. She’s still working at McDonald’s, but her role is a whole lot different thanks to her hard work and the degrees from SNHU.
3 Things Service Companies Should Do to Prepare for AI and Automation
These actions will help employees and service-oriented businesses alike not only survive the predicted influx of AI and automation, but thrive as the business world adapts to new and ever-changing technology.