Clinical laboratories working with AI should be aware of ethical challenges being pointed out by industry experts and legal authorities
Experts are voicing concerns that using artificial
intelligence (AI) in healthcare could present ethical challenges that need
to be addressed. They say databases and algorithms may introduce bias into the
diagnostic process, and that AI may not perform as intended, posing a potential
for patient harm.
If true, the issues raised by these experts would have major
implications for how clinical
laboratories and anatomic
pathology groups might use artificial intelligence. For that reason,
medical laboratory executives and pathologists should be aware of possible
drawbacks to the use of AI and machine-learning
algorithms in the diagnostic process.
Is AI Underperforming?
AI’s ability to improve diagnoses, precisely target
therapies, and leverage healthcare data is predicted to be a boon to precision medicine and personalized
healthcare.
For example, Accenture
(NYSE:ACN) says that hospitals will spend $6.6 billion on AI by 2021. This
represents an annual growth rate of 40%, according
to a report from the Dublin, Ireland-based consulting firm, which states,
“when combined, key clinical health AI applications can potentially create $150
billion in annual savings for the United States healthcare economy by 2026.”
But are healthcare providers too quick to adopt AI?
Accenture defines AI as a “constellation of technologies
from machine learning to natural
language processing that allows machines to sense, comprehend, act, and
learn.” However, some experts say AI is not performing as intended, and that it
introduces biases in healthcare worthy of investigation.
Keith Dreyer, DO, PhD, is Chief Data Science Officer at Partners Healthcare and Vice Chairman of Radiology at Massachusetts General Hospital (MGH). At a World Medical Innovation Forum on Artificial Intelligence covered by HealthITAnalytics, he said, “There are currently no measures to indicate that a result is biased or how much it might be biased. We need to explain the dataset these answers came from, how accurate we can expect them to be, where they work, and where they don’t work. When a number comes back, what does it really mean? What’s the difference between a seven and an eight or a two?” (Photo copyright: Healthcare in Europe.)
What Goes in Limits What Comes Out
Could machine learning lead to machine decision-making that
puts patients at risk? Some legal authorities say yes. Especially when computer
algorithms are based on limited data sources and questionable methods, lawyers
warn.
How can AI provide accurate medical insights for people when
the information going into databases is limited in the first place? Ossorio
pointed to lack of diversity in genomic
data. “There are still large groups of people for whom we have almost no
genomic data. This is another way in which the datasets that you might use to
train your algorithms are going to exclude certain groups of people
altogether,” she told HDM.
She also sounded the alarm about making decisions about
women’s health when data driving them are based on studies where women have
been “under-treated compared with men.”
“This leads to poor treatment, and that’s going to be
reflected in essentially all healthcare data that people are using when they
train their algorithms,” Ossorio said during a Machine Learning for Healthcare (MLHC) conference
covered by HDM.
How Bias Happens
Bias can enter healthcare data in three forms: by humans, by
design, and in its usage. That’s according to David Magnus, PhD, Director
of the Stanford Center for
Biomedical Ethics (SCBE) and Senior Author of a paper published in the New England
Journal of Medicine (NEJM) titled, “Implementing Machine
Learning in Health Care—Addressing Ethical Challenges.”
The paper’s authors wrote, “Physician-researchers are
predicting that familiarity with machine-learning tools for analyzing big data
will be a fundamental requirement for the next generation of physicians and
that algorithms might soon rival or replace physicians in fields that involve
close scrutiny of images, such as radiology and anatomical pathology.”
In a news
release, Magnus said, “You can easily imagine that the algorithms being
built into the healthcare system might be reflective of different, conflicting
interests. What if the algorithm is designed around the goal of making money?
What if different treatment decisions about patients are made depending on
insurance status or their ability to pay?”
In addition to the possibility of algorithm bias, the
authors of the NEJM paper have other concerns about AI affecting
healthcare providers:
“Physicians must adequately understand how
algorithms are created, critically assess the source of the data used to create
the statistical models designed to predict outcomes, understand how the models
function and guard against becoming overly dependent on them.
“Data gathered about patient health, diagnostics,
and outcomes become part of the ‘collective knowledge’ of published literature
and information collected by healthcare systems and might be used without
regard for clinical experience and the human aspect of patient care.
“Machine-learning-based clinical guidance may
introduce a third-party ‘actor’ into the physician-patient relationship, challenging
the dynamics of responsibility in the relationship and the expectation of
confidentiality.”
“We need to be cautious about caring for people based on what algorithms are showing us. The one thing people can do that machines can’t do is step aside from our ideas and evaluate them critically,” said Danton Char, MD, Lead Author and Assistant Professor of Anesthesiology, Perioperative, and Pain Medicine at Stanford, in the news release. “I think society has become very breathless in looking for quick answers,” he added. (Photo copyright: Stanford Medicine.)
Acknowledge Healthcare’s Differences
Still, the Stanford researchers acknowledge that AI can
benefit patients. And that healthcare leaders can learn from other industries,
such as car companies, which have test driven AI.
“Artificial intelligence will be pervasive in healthcare in a
few years,” said
Nigam Shah, PhD, co-author of the NEJM paper and Associate Professor of Medicine at Stanford, in the news release. He added that healthcare leaders need to be aware of the “pitfalls” that have happened in other industries and be cognizant of data.
“Be careful about knowing the data from which you learn,” he
warned.
AI’s ultimate role in healthcare diagnostics is not yet fully
known. Nevertheless, it behooves clinical laboratory leaders and anatomic
pathologists who are considering using AI to address issues of quality and
accuracy of the lab data they are generating. And to be aware of potential
biases in the data collection process.
According to Damo Consulting’s 2019 Healthcare
IT Demand Survey, when it comes to spending money on information
technology (IT), healthcare executives believe AI and digital healthcare
technologies—though promising—need more development.
Damo’s report notes that 71% of healthcare providers
surveyed expect their IT budgets to grow by 20% in 2019. However, much of that
growth will be allocated to improving EHR functionality, Healthcare Purchasing News reported
in its analysis of Damo survey data.
As healthcare executives plan upgrades to their EHRs,
hospital-based medical laboratories will need to take steps to ensure
interoperability, while avoiding disruption to lab workflow during transition.
The survey also noted that some providers that are considering
investing in AI and digital health technology are struggling to understand the
market, the news release states.
“Digital and AI are emerging as critical areas for technology spend among healthcare enterprises in 2019. However, healthcare executives are realistic about their technology needs versus their need to improve care delivery. They find the currently available digital health solutions in the market are not very mature,” explained Paddy Padmanabhan (above), Chief Executive Officer of Damo Consulting, in a news release. (Photo copyright: The Authors Guild.)
Providers More
Positive Than Vendors on IT Spend
Damo Consulting is a Chicago-area based healthcare and
digital advisory firm. In November 2018, Damo surveyed 64 healthcare executives
(40 technology and service leaders, and 24 healthcare enterprise executives). Interestingly, healthcare providers were more
positive than the technology developers on IT spending plans, reported HITInfrastructure.com, which
detailed the following survey findings:
79% of healthcare executives anticipate high
growth in IT spending in 2019, but only 60% of tech company representatives
believe that is so.
75% of healthcare executives and 80% of vendor
representatives say change in healthcare IT makes buying decisions harder.
71% of healthcare executives and 55% of vendors say
federal government policies help IT spending.
50% of healthcare executives associate
immaturity with digital solution offerings.
42% of healthcare providers say they lack
resources to launch digital.
“While information technology vendors are aggressively
marketing ‘digital’ and ‘AI,’ healthcare executives note that the currently
available solutions in these areas are not very mature. These executives are
confused by the buzz around ‘AI’ and ‘digital,’ the changing landscape of who
is playing what role, and the blurred lines of capabilities and competition,” noted
Padmanabhan in the survey report.
The survey also notes that “Health systems are firmly
committed to their EHR vendors. Despite the many shortcomings, EHR systems
appear to be the primary choice for digital initiatives among health systems at
this stage.”
Some Healthcare
Providers Starting to Use AI
Even as EHRs receive the lion’s share of healthcare IT
spends, some providers are devoting significant resources to AI-related
projects and processes.
For example, clinical
pathologists may be intrigued by work being conducted at Cleveland Clinic’s Center for
Clinical Artificial Intelligence (CCAI), launched in March. The CCAI is using
AI and machine learning in pathology, genetics, and cancer research, with the
ultimate goal of improving patient outcomes, reported Becker’s Hospital Review.
“We’re not in it because AI is cool, but because we believe
it can advance medical research and collaboration between medicine and
industry—with a focus on the patient,” Aziz Nazha, MD, Clinical
Hematology and Oncology Specialist and Director of the CCAI, stated in an
article posted by the American Medical Association (AMA).
AI Predictions Lower
Readmissions and Improve Outcomes
Cleveland Clinic’s CCAI reportedly has gathered data from
1.6 million patients, which it uses to predict length-of-stays and reduce
inappropriate readmissions. “But a prediction itself is insufficient,” Nazha told
the AMA. “If we can intervene, we can change the prognosis and make things
better.”
The CCAI’s ultimate goal is to use predictive models to “develop
a new generation of physician-data scientists and medical researchers.” Toward
that end, Nazha notes how his team used AI to develop genomic biomarkers that identify
whether a certain chemotherapy drug—azacitidine (aka,
azacytidine and marketed as Vidaza)—will work for specific patients. This is a
key goal of precision
medicine.
CCAI also created an AI prediction model that outperforms
existing prognosis scoring systems for patients with Myelodysplastic
syndromes (MDS), a form of cancer in bone marrow.
Meanwhile, at Johns
Hopkins Hospital, AI applications track availability of beds and more. The
Judy Reitz Capacity Command Center, built in collaboration with GE Healthcare Partners, is a
5,200 square feet center outfitted with AI apps and staff to transfer patients
and help smooth coordination of services, according to a news release.
Forbes described the Reitz command
center as a “cognitive hospital” and reports that it has essentially enabled
Johns Hopkins to expand its capacity by 16 beds without undergoing bricks-and-mortar-style
construction.
In short, medical laboratory leaders may want to interact
with IT colleagues to ensure uninterrupted workflows as EHR functionality evolves.
Furthermore, AI developments suggest opportunities for clinical laboratories to
leverage patient data and assist in improving the diagnostic accuracy of providers
in ways that improve patient care.
Increased use of telemedicine may create opportunities for clinical laboratories to deliver increased value to both physicians and nurses
Recent data shows widespread employer adoption of telehealth services may soon become a reality. However, studies also show virtual provider visits and other telemedicine technologies are unlikely to diminish the role of clinical laboratories in providing the data required for diagnosis and treatment decisions. Instead, laboratories and anatomic pathology groups will likely see changes in how samples are collected from patients using telemedicine and how medical laboratory test results are reported, as access to telemedicine grows.
A recent National Business Group on Health (NBGH) survey indicates that in 2018 “virtually all [large] employers (96%) will make telehealth services available in states where it is allowed.” The survey was conducted between May and June 2017, with 148 large employers participating.
Christine Smalley, Managing Director with consulting firm Claremont Hudson, divides telemedicine technology into three distinct segments:
1. Provider-to-provider;
2. Remote patient monitoring; and,
3. Patient-to-provider.
In an article she penned for MedCityNews, Smalley calls provider-to-provider telemedicine the “most evolved to-date” segment of the telehealth trend. She highlights ICU stroke care with remote consults and monitoring as an example of its “success,” and notes a large potential for growth in remote patient monitoring (RPM). Smalley cites a Berg Insight report that estimates 50-million patients will use remote monitored devices by 2021. However, Smalley also notes consumer acceptance of patient-to-provider telemedicine has fallen short of industry expectations.
While virtual office visits—where patients have access to physicians via telephone or videoconferencing—grab headlines, Smalley argues that “several factors” are hindering adoption.
“Reimbursement is not yet universal,” she notes. “But consumers are growing used to paying more out-of-pocket with high-deductible plans. Physicians have long resisted change in how they practice, and many remain lukewarm at best about telemedicine. It’s no coincidence that many of the innovations and pioneering models have come from outside of healthcare delivery … The barriers that loom the largest may likely be consumer awareness and trial.”
The Center for Connected Health Policy (CCHP) reports that 35 states have laws governing private payer reimbursement of telehealth, a number that has not changed since 2016. According to a CCHP press release, some state laws require reimbursement be equal to in-person visits, though not all laws mandate reimbursement.
Adopting Existing Retail Models to Promote Telemedicine to Patients
Smalley contends “smart marketing” will be needed to get consumers to leverage the telemedicine options that are becoming available to them. She says simply offering video or telephone visits is not enough. She encourages integrated delivery systems to take a page out of retailers’ playbooks.
“Look at how retailers, like Walmart, integrate online shopping and the store experience by offering side-by-side options supporting product delivery and in-store pickup. Telemedicine options ultimately need to be offered in a way that feels integrated and seamless to the health consumer,” she suggested, in her MedCityNews article. One example, she notes, would be providing an easy-to-navigate link to a virtual visit on a healthcare network’s urgent care webpage.
Telemedicine isn’t just about the office visit. Pathologists such as J.B. Askew, MD, PA (above), have embraced telepathology technology to bring pathology interpretation services to remote and resource strapped areas worldwide. (Click on image above to watch a video of Askew demonstrating the use of a telepathology imaging system.) (Image/video copyright: J.B. Askew, MD, PA, North Houston Pathology Associates/Meyer Instruments.)
Healthcare Spending Could Increase Due to Telehealth
While health plans have zeroed in on telehealth as a way to drive down healthcare costs, a 2017 RAND Corp. study published in Health Affairs found virtual visits to physicians might not decrease spending, though access to care is improved.
“Instead of saving money by substitution [replacing more expensive visits to physician offices or EDs], direct-to-consumer telehealth may increase spending by new utilization [increasing the total number of patient visits],” a MedCityNews article suggests.
“Given that direct-to-consumer telehealth is even more convenient than traveling to retail clinics, it may not be surprising that an even greater share of telehealth services represents new medical use,” noted Lori Uscher-Pines, PhD, a RAND Policy Researcher. “There may be a dose response with respect to convenience and use: the more convenient the location, the lower the threshold for seeking care and the greater the use of medical services.”
Telehealth in Clinical Laboratories
Will telehealth services offered by hospital networks and healthcare providers impact clinical laboratories? While a physical visit is still required for drawing blood, collecting urine, or performing pathology testing, interpretive digital pathology, such as Whole Slide Imaging (AKA, Virtual Slide), does enable pathologists to provided distance interpretation services of blood tests to remote and/or resource deficient areas of the world, as Dark Daily reported in past e-briefings. This could become a substantial revenue stream in the future if telepathology’s global popularity continues to rise.
Weaker finances at the nation’s hospitals causes administrators to further shrink the budgets for clinical laboratory and anatomic pathology services
Hospital admissions across the country continue to be flat or in decline over recent years. The result is less revenue for many hospitals. As a result, administrators continue to shrink the budgets of hospital service lines—including clinical laboratory services. For pathologists and clinical laboratory leaders, this poses the challenge of setting innovative strategies that take into account the changes in payment and delivery models.
Hospital Inpatient Admissions Have Been Declining over Recent Years
Pathologists should take note that an increasing number of patients who want genetic tests are complaining when they learn their insurance plan will not pay for such tests
Concerned about the increased cost of genetic tests, health insurers are becoming reluctant to pay for many types of molecular diagnostics and gene tests. When refusing to pay for these tests, however, they face a buzz saw of angry patients—many of whom see a genetic test as their last resort for a diagnosis and selection of a therapy that might just work for them.
Reuters recently reported that health insurance companies are reluctant to pay providers for genetic-sequencing tests until more research becomes available. This is a sign for pathologists and clinical laboratory managers that enough patients have been affected by this situation to justify news coverage by a major news source. (more…)
