Decision Support

A new report from the Duke-Margolis Center for Health Policy explores some of the policy changes that should be made to enable safer and more effective deployment of artificial intelligence in healthcare. As AI and machine learning become de facto ingredients in many key clinical technologies, a better understanding of how they can best be leveraged for optimal analytics and decision support is the goal of the study, "Current State and Near-Term Priorities for AI-Enabled Diagnostic Support Software in Health Care." WHY IT MATTERS The Duke report takes stock of the existing legal and regulatory landscape for algorithm-based CDS and diagnostic support software, and lays out some essential priorities to work toward in the years ahead to ensure safe deployment of AI in clinical settings. These aren't just theoretical concerns. AI and ML are making inroads all over healthcare, of course, and current legislation and regulatory policy – whether it's the massive 21st Century Cures Act or FDA's new updates to the Software Pre-Cert Pilot Program – are adequate but still not optimal for a future that promises to evolve at a dizzying pace. The Duke-Margolis paper, meant as a "resource for developers, regulators, clinicians, policy makers, and other stakeholders as they strive to effectively, ethically, and safely incorporate AI as a fundamental component in diagnostic error prevention and other types of CDS," looks at some of the major challenges and opportunities facing AI in the years ahead. Stakeholders like those listed about will need to grapple with big questions, more than a dozen researchers and authors write. Such as: Making a case for the value of more widespread adoption of these technologies. Such evidence would include how the software improves patient outcomes, boosts quality and lowers cost of care, gives clinicians relevant information in a manner they find "useful and trustworthy." Assessing the potential risk of using those products in clinical settings. "The degree to which a software product comes with information that explains how it works and the types of populations used to train the software will have significant impact on regulators’ and clinicians’ assessment of the risk to patients when clinicians use this software," said Duke researchers. "Product labeling may need to be reconsidered and the risks and benefits of continuous learning versus locked models must be discussed." Seeing to it that such systems are deployed in a way that's both flexible and ethical. More and more health systems will need to develop best practices that can mitigate any bias that could be introduced by the training data used to develop software, they explained. That's the only way to ensure that "data-driven AI methods do not perpetuate or exacerbate existing clinical biases." Also, these organizations will have to think hard about the data implications as the products scale up into settings that may be different from initial use cases. And, of course, "new paradigms are needed for how to best protect patient privacy," according to the report. THE LARGER TREND As the technological capabilities and clinical applications of AI-enabled decision support continue to expand, the Duke researchers said more regulatory clarity from agencies such as FDA, which has signaled an appetite for much wider approval of machine learning apps, is needed to protect patients from wanton use of the "black box" algorithms that many have warned about. In addition, there are other major areas that need ironing-out. Among them: proper allowances for patient privacy and data access, and the ability for these fast-emerging technologies demonstrate value and ROI for providers. In all of those, hospitals and health systems have an active role to play. Then there are all sorts of other technical questions that exist – but haven't necessarily been answered, certainly not on a consistent or widespread basis. Such as: how new approaches to labeling different software might improve understanding of its inner workings; how to weigh the relative risks and benefits of locked versus continuously learning models of AI; how to evaluate its performance over time most effectively; how to mitigate data bias; how to assess "algorithmic adaptability" and more. ON THE RECORD "AI is now poised to disrupt health care, with the potential to improve patient outcomes, reduce costs, and enhance work-life balance for health care providers, but a policy process is needed," said Greg Daniel, deputy director for policy at Duke-Margolis, in a statement. "Integrating AI into healthcare safely and effectively will need to be a careful process, requiring policymakers and stakeholders to strike a balance between the essential work of safeguarding patients while ensuring that innovators have access to the tools they need to succeed in making products that improve the public health," he said. "AI-enabled clinical decision support software has the potential help clinicians arrive at a correct diagnosis faster, while enhancing public health and improving clinical outcomes," added Christina Silcox, managing associate at Duke-Margolis and co-author of the report. "To realize AI’s potential in health care, the regulatory, legal, data, and adoption challenges that are slowing safe and effective innovation need to be addressed." Twitter: @MikeMiliardHITN Email the writer: mike.miliard@himssmedia.com Healthcare IT News is a publication of HIMSS Media.

Thanks to consumerism in the healthcare industry, providers are looking to improve patient experience and satisfaction in their care. HIMSS19 attendees who are focused on this topic for their organizations should check out these curated sessions and events.

As machine learning rapidly expands into healthcare, the ways it "learns" may be at odds with clinical outcomes unless carefully controlled for, a new study shows.

Last week, Deqing county hospital in Guangdong Province launched free consultations featuring artificial-intelligence (AI) cameras to detect ocular fundus diseases, which are major causes of blindness, according to a report by Xinhua News. About 300 residents from Zhaoqing City in Dequing County attended the free consultation sessions. The hospital became the first to use the device, co-developed by China's search engine Baidu and Sun Yat-sen University, to serve the general public. The instrument is capable of diagnosing three types of fundus disorders -- diabetic retinopathy, glaucoma and macular degeneration. It scans the eyes and generates a report in 10 seconds, all done without the need for an ophthalmologist to be present. Baidu’s AI-powered camera was first unveiled in China in November 2018, according to a MobiHealth News article. Fundus diseases are a major cause of blindness in the developing world, where the short supply of eye doctors and instruments has stymied timely diagnosis and treatment. China, with a population of 1.39 billion, has only thousands of ophthalmologists capable of analysing fundus photos screening. “As a doctor working at the grassroots level, I believe AI can greatly help in all aspects of screening. For instance, there is so much imaging data during medical checkups and it takes up a lot of time and energy for doctors to physically look at this data, which is simply not efficient. In ophthalmology, the use of AI to verify test results from fluorescein angiography and OCT examinations can help doctors expedite their analysis, which saves time and improves their efficiency,” said Dr Honghu Xia, Director of Ophthalmology at Deqing county hospital. Xu Yanwu, a Baidu engineer developing the instrument, said the AI cameras were specifically designed to address the lack of medical instruments and ophthalmologists at grassroots health facilities. "It is easy to use and can be operated by a non-professional. Its 94-percent sensitivity and specificity at analysing photos can match a senior doctor at a tertiary hospital," Xu said. As of 10 January 2019, Baidu already has four of said AI cameras operating in four hospitals in Deqing County to assist ophthalmologists in fundus screening. It is estimated that by end of March, Guangdong Province will have 14 hospitals using the AI camera instrument.

Big Blue's collaboration the nation's largest Parkinson's foundation aims to apply AI to a vast dataset with the aim of learning more about how the disease grows.

“I think the biggest trend (in healthcare) is towards greater integration. Traditionally, healthcare has been very fragmented, where many different groups serve specific clinical needs without necessarily coordinating with each other. But going forward, the trend is towards integration – not just of things like databases and systems, but integration of the way we process the data and how this influences the clinical workflow,” said Prof Ngiam Kee Yuan, Group Chief Technology Officer of National University Health System (NUHS) in Singapore, in response to what he thought would be the key trends that will impact healthcare systems in future. It was with the same motivation and mission to best use the healthcare data for research at NUHS that led to the building and development of the DISCOVERY AI platform, which started about four years ago and the platform was officially announced in July 2018. The platform is what Prof Ngiam describes as a ‘sandbox’ that allows the staff at NUHS to develop AI tools in a safe and equitable way – the platform is scalable and can be applied to more than one system within the organisation. “We saw the opportunity because we had datasets which were large enough to support the development of these AI tools. And one of the advantages we have at NUHS is that we have clinicians and allied health professionals who understand, and are willing to develop these tools. I cannot emphasise enough the importance of having the clinicians onboard throughout the development process.” Currently, a randomised control trial of a system as part of the platform is a free-text diagnosis machine at the Accident & Emergency (A & E) department. When doctors input a certain set of findings as part of clinical documentation, the machine would automatically provide a suggestion for a diagnosis. The team is exploring diagnosing appendicitis for a start. The trial is slightly under halfway through and Prof Ngiam hopes by the later part of 2019, they would be able to have results, which would be the basis for them to operationalise the AI tool. One of the early milestones for the NUHS DISCOVERY AI platform is its ability to sustain multiple proof-of-concept projects. With the platform, individual projects are no longer fragmented and there is the ability to aggregate, link and share large data sets. “It took us four years to get to this point and the next milestone for us is to finish our trials and to actually launch them as “software as medical devices”. Again there are some hurdles to get through before we get there but seeing where we are right now, it is very likely we should be able to get through them.” Prof Ngiam also pointed out that as the platform is unlike anything they had before and behaves like an advanced form of clinical decision support system (CDSS), which are not based on a set of rules but based on a set of complex trained weights and multiple factors that affect a certain outcome. Despite its complexity, the AI tools need to be thoroughly trialed before it can be used in routine clinical practice. “This is why we are running it as a trial now in the hospital. In essence, the platform is run under the ambit of a trial to mature the workflow and test the system under real world conditions. Operationally, what the doctors are doing during the trial is no different from what they would have done, except that we collect the data on the basis of the trial,” he added. Looking to the future, Prof Ngiam and his team is working towards completing the trials in 2019 and hopefully heading towards registration in terms of the platform being used as a medical advisory device.

Artificial intelligence can help support and guide care providers' decision-making processes in an era of increasing volume and complexity of data in clinical situations, says Dimitri Fane, director of product management of TrakCare at InterSystems.  

Its homegrown Canopy system is an example of the type of platform that value-based models of the future will need to deliver evidence-based care in a complex care ecosystem.

Partners HealthCare researchers show how they mine override comments to detect areas where clinical decision support systems could be improved.