Pathologists and clinical laboratory scientists may find one hospital’s use of a machine-learning platform to help improve utilization of lab tests both an opportunity and a threat
Variation in how individual physicians order, interpret, and act upon clinical laboratory test results is regularly shown by studies in peer-reviewed medical journals to be one reason why some patients get great outcomes and other patients get less-than-desirable outcomes. That is why many healthcare providers are initiating efforts to improve how physicians utilize clinical laboratory tests and other diagnostic procedures.
This effort came about after clinical and administrative leadership at Flagler Hospital realized that only about one-third of its physicians regularly followed certain medical decision-making guidelines or clinical order sets. Armed with these insights, staff members decided to find a solution that reduced or removed variability from their healthcare delivery.
Reducing Variability Improves Care, Lowers Cost
Variability in physician care has been linked to increased healthcare costs and lower quality outcomes, as studies published in JAMA and JAMA Internal Medicine indicate. Such results do not bode well for healthcare providers in today’s value-based reimbursement system, which rewards increased performance and lowered costs.
Clinical order sets are designed to be used as part of clinical decision support systems (CDSS) installed by hospitals for physicians to standardize care and support sound clinical decision making and patient safety.
However, when doctors don’t adhere to those pre-defined standards, the results can be disadvantageous, ranging from unnecessary services and tests being performed to preventable complications for patients, which may increase treatment costs.
Flagler’s AI project involved uploading clinical,
demographic, billing, and surgical information to the AyasdiAI platform, which then
employed machine learning to analyze the data and identify trends. Flagler’s
physicians are now provided with a fuller picture of their patients’ conditions,
which helps identify patients at highest risk, ensuring timely interventions that
produce positive outcomes and lower costs.
How Symphony AyasdiAI Works
The AyasdiAI application utilizes a category of mathematics called topological data analysis (TDA) to cluster similar patients together and locate parallels between those groups. “We then have the AI tools generate a carepath from this group, showing all events which should occur in the emergency department, at admission, and throughout the hospital stay,” Sanders told Healthcare IT News. “These events include all medications, diagnostic tests, vital signs, IVs, procedures and meals, and the ideal timing for the occurrence of each so as to replicate the results of this group.”
Caregivers then examine the data to determine the optimal
plan of care for each patient. Cost savings are figured into the overall
equation when choosing a treatment plan.
Flagler first used the AI program to examine trends among their pneumonia patients. They determined that nebulizer treatments should be started as soon as possible with pneumonia patients who also have chronic obstructive pulmonary disease (COPD).
“Once we have the data loaded, we use [an] unsupervised
learning AI algorithm to generate treatment groups,” Sanders told Healthcare
IT News. “In the case of our pneumonia patient data, Ayasdi produced nine
treatments groups. Each group was treated similarly, and statistics were given
to us to understand that group and how it differed from the other groups.”
Armed with this information, the hospital achieved an 80% greater physician adherence to order sets for pneumonia patients. This resulted in a savings of $1,350 per patient and reduced the readmission rates for pneumonia patients from 2.9% to 0.4%, reported Modern Healthcare.
The development of a machine-learning platform designed to
reduce variation in care (by helping physicians become more consistent at
following accepted clinical care guidelines) can be considered a warning shot
across the bow of the pathology profession.
This is a system that has the potential to become interposed
between the pathologist in the medical laboratory and the physicians who refer
specimens to the lab. Were that to happen, the deep experience and knowledge
that have long made pathologists the “doctor’s doctor” will be bypassed.
Physicians will stop making that first call to their pathologists, clinical
chemists, and laboratory scientists to discuss a patient’s condition and
consult on which test to order, how to interpret the results, and get guidance
on selecting therapies and monitoring the patient’s progress.
Instead, a “smart software solution” will be inserted into
the clinical workflow of physicians. This solution will automatically guide the
physician to follow the established care protocol. In turn, this will give the
medical laboratory the simple role of accepting a lab test order, performing
the analysis, and reporting the results.
If this were true, then it could be argued that a laboratory
test is a commodity and hospitals, physicians, and payers would argue that they
should buy these commodity lab tests at the cheapest price.
Centers for Medicare and Medicaid Innovation is considering adding clinical laboratory services to bundled payments in its proposed Oncology Care First model
CMMI, an organization within the Centers for Medicare and Medicaid Services (CMS), is charged with developing and testing new healthcare delivery and payment models as alternatives to the traditional fee-for-service (FFS) model. On November 1, 2019, CMMI released an informal Request for Information (RFI) seeking comments for the proposed Oncology Care First (OCF) model, which would be the successor to the Oncology Care Model (OCM) launched in 2016.
“The inefficiency and variation in oncology care in the
United States is well documented, with avoidable hospitalizations and emergency
department visits occurring frequently, high service utilization at the end of
life, and use of high-cost drugs and biologicals when lower-cost, clinically
equivalent options exist,” the CMMI RFI states.
With the proposed new model, “the Innovation Center aims to build on the lessons learned to date in OCM and incorporate feedback from stakeholders,” the RFI notes.
How the Oncology Care First Model Works
The OCF program, which is voluntary, will be open to
physician groups and hospital outpatient departments. CMMI anticipates that
testing of the model will run from January 2021 through December 2025.
It will offer two payment mechanisms for providers that
choose to participate:
A Monthly Population Payment (MPP) would apply
to a provider’s Medicare beneficiaries with “cancer or a cancer-related
diagnosis,” the RFI states. It would cover Evaluation and Management (EM)
services as well as drug administration services and a set of “Enhanced
Services,” including 24/7 access to medical records.
Of particular interest to medical laboratories, the RFI also
states that “we are considering the inclusion of additional services in the monthly
population payment, such as imaging or medical laboratory services, and seek
feedback on adding these or other services.”
In addition, providers could receive a
Performance-Based Payment (PBP) if they reduce expenditures for patients
receiving chemotherapy below a “target amount” determined by past Medicare
payments. If providers don’t meet the threshold, they could be required to
repay CMS.
Practices that wish to participate in the OCF model must go through an application process. It is also open to participation by private payers. CMS reports that 175 practices and 10 payers are currently participating in the 2016 Oncology Care Model (OCM).
Medical Lab Leaders Concerned about the CMMI OCF Model
One group raising concerns about the inclusion of medical laboratory service reimbursements in the Monthly Population Payment scheme is the Personalized Medicine Coalition. “Laboratory services are crucial to the diagnosis and management of many cancers and are an essential component of personalized medicine,” wrote Cynthia A. Bens, the organization’s senior VP for public policy, in an open letter to CMMI Acting Director Amy Bassano. “We are concerned that adding laboratory service fees to the MPP may cause providers to view them as expenses that are part of the total cost of delivering care, rather than an integral part of the solution to attain high-value care,” Bens wrote.
She advised CMMI to “seek further input from the laboratory
and provider communities on how best to contain costs within the OCF model,
while ensuring the proper deployment of diagnostics and other laboratory
services.”
Members of the coalition include biopharma companies, diagnostic companies, patient advocacy groups, and clinical laboratory testing services. Lab testing heavyweights Quest Diagnostics (NYSE:DGX) and Laboratory Corporation of America (NYSE:LH) are both members.
CMS ‘Doubles Down’ on OCM
The proposal received criticism from other quarters as well. “While private- and public-sector payers would be well served to adopt and support a VBP [value-based payment] program for cancer care, we need to better understand some of the shortcomings of the original OCM design and adopt lessons learned from other successful VBP models to ensure uptake by providers and ultimately better oncology care for patients,” wrote members of the Oncology Care Model Work Group in a Health Affairs blog post. They added that with the new model, “CMS seems to double down on the same design as the OCM.”
Separately, CMMI has proposed a controversial Radiation
Oncology (RO) alternative payment model (APM) that would be mandatory for
practices in randomly-selected metro areas. The agency estimates that it would
apply to approximately 40% of the radiotherapy practices in the US.
These recent actions should serve to remind pathologists and
clinical laboratories that CMS continues to move away from fee-for-service and
toward value-based care payment models, and that it is critical to plan for
changing reimbursement strategies.
At present, medical laboratories are collecting blood specimens for testing by authorized public health labs. However, clinical laboratories should prepare for the likelihood they will be called on to perform the testing using the CDC test or other tests under development.
“We need to be vigilant and understand everything related to the testing and the virus,” said Bodhraj Acharya, PhD, Manager of Chemistry and Referral Testing at the Laboratory Alliance of Central New York, in an exclusive interview with Dark Daily. “If the situation comes that you have to do the testing, you have to be ready for it.”
The current criteria for determining PUIs include clinical features, such as fever or signs of lower respiratory illness, combined with epidemiological risks, such as recent travel to China or close contact with a laboratory-confirmed COVID-19 patient. The CDC notes that “criteria are subject to change as additional information becomes available” and advises healthcare providers to consult with state or local health departments if they believe a patient meets the criteria.
Test Kit Problems Delay Diagnoses
On Feb. 4, the FDA issued a Novel Coronavirus Emergency Use Authorization (EUA) allowing state and city public health laboratories, as well as Department of Defense (DoD) labs, to perform presumptive qualitative testing using the Real-Time Reverse Transcriptase PCR (RT-PCR) diagnostic panel developed by the CDC. Two days later, the CDC began distributing the test kits, a CDC statement announced. Each kit could test 700 to 800 patients, the CDC said, and could provide results from respiratory specimens in four hours.
However, on Feb. 12, the agency revealed in a telebriefing that manufacturing problems with one of the reagents had caused state laboratories to get “inconclusive laboratory results” when performing the test.
“When the state receives these test kits, their procedure is to do quality control themselves in their own laboratories,” said Nancy Messonnier, MD, Director of the CDC National Center for Immunization and Respiratory Diseases (NCIRD), during the telebriefing. “Again, that is part of the normal procedures, but in doing it, some of the states identified some inconclusive laboratory results. We are working closely with them to correct the issues and as we’ve said all along, speed is important, but equally or more important in this situation is making sure that the laboratory results are correct.”
During a follow-up telebriefing on Feb. 14, Messonnier said
that the CDC “is reformulating those reagents, and we are moving quickly to get
those back out to our labs at the state and local public health labs.”
Serologic Test Under Development
The current test has to be performed after a patient shows
symptoms. The “outer bound” of the virus’ incubation period is 14 days, meaning
“we expect someone who is infected to have symptoms some time during those 14
days,” Messonnier said. Testing too early could “produce a negative result,”
she continued, because “the virus hasn’t established itself sufficiently in the
system to be detected.”
Messonnier added that the agency plans to develop a serologic test that will identify people who were exposed to the virus and developed an immune response without getting sick. This will help determine how widespread it is and whether people are “seroconverting,” she said. To formulate this test, “we need to wait to draw specimens from US patients over a period of time. Once they have all of the appropriate specimens collected, I understand that it’s a matter of several weeks” before the serologic test will be ready, she concluded.
“Based on what we know now, we believe this virus spreads
mainly from person to person among close contacts, which is defined [as] about
six feet,” Messonnier said at the follow-up telebriefing. Transmission is
primarily “through respiratory droplets produced when an infected person coughs
or sneezes. People are thought to be the most contagious when they’re most
symptomatic. That’s when they’re the sickest.” However, “some spread may happen
before people show symptoms,” she said.
The virus can also spread when people touch contaminated surfaces and then touch their eyes, nose, or mouth. But it “does not last long on surfaces,” she said.
Where the Infection Began
SARS-CoV-2 was first identified during an outbreak in Wuhan, China, in December 2019. Soon thereafter, hospitals in the region “were overwhelmed” with cases of pneumonia, Dr. Acharya explained, but authorities could not trace the disease to a known pathogen. “Every time a new pathogen originates, or a current pathogen mutates into a new form, there are no molecular tests available to diagnose it,” he said.
So, genetic laboratories used next-generation sequencing, specifically unbiased nontargeted metagenomic RNA sequencing (UMERS), followed by phylogenetic analysis of nucleic acids derived from the hosts. “This approach does not require a prior knowledge of the expected pathogen,” Dr. Acharya explained. Instead, by understanding the virus’ genetic makeup, pathology laboratories could see how closely it was related to other known pathogens. They were able to identify it as a Betacoronavirus (Beta-CoVs), the family that also includes the viruses that cause SARS and Middle East Respiratory Syndrome (MERS).
This is a fast-moving story and medical laboratory leaders are advised to monitor the CDC website for continuing updates, as well as a website set up by WHO to provide technical guidance for labs.
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.
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.”
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.
Clinical laboratory leaders interested in positioning their labs to be paid for added-value services will get knowledge, insights, and more at upcoming third annual Clinical Lab 2.0 Workshop in November
It’s a critical time for medical laboratories. Healthcare is transitioning from a fee-for-service payment system to new value-based payment models, creating disruption and instability in the clinical lab test market. In addition, payers are cutting reimbursement for many lab tests.
These are among the market factors leading some pathologists
and clinical lab leaders to seek new or alternative sources of revenue to keep
the lights on and the machines running in their laboratories. Some might say,
it’s a dark time for the lab industry.
“This is not the time to be shy or timid,” he declared. “The
quantitative value of medical laboratory domain is significant and will be lost
if not exploited or leveraged.”
Shotorbani has reason to be positive. In recent years the Project Santa Fe Foundation (PSFF) has emerged to advocate for, and teach, the Clinical Lab 2.0 model. Clinical Lab 2.0 is an approach which focuses on longitudinal clinical laboratory data to augment population health in new payment arrangements.
Earlier this year, PSFF filed for 501(c) status, according to a news release. It is now positioned as a nonprofit organization, guided by a board of directors whose mission is “to create a disruptive value paradigm and alternative payment model that defines placement of diagnostic services in healthcare.”
Progressing Toward Clinical Lab 2.0
At the 24th Annual Executive War College on Lab and Pathology Management held in New Orleans last May, the nation’s first ever Clinical Lab 2.0 “Shark Tank” competition was won by Aspenti Health, a full-service diagnostic laboratory specializing in toxicology screening.
“This project, as well as all of the other cases that were presented, were quite strong and all were aligned with the mission of the Clinical Lab 2.0 movement,” said Shotorbani, in a news release. “This movement transforms the analytic results from a laboratory into actionable intelligence at the patient visit in partnership with front-liners and clinicians—allowing for identification of patient risks—and arming providers with insights to guide therapeutic interventions.
“Further, it reduces the administrative burden on providers by collecting SDH [social determinants of health] predictors in advance and tying them to outcomes of interest,” he continued. “By bringing SDH predictors to the office visit, it enables providers to engage in SDH without relying on their own data collection—a current care gap in many practices. The lab becomes a catalyst helping to manage the population we serve.”
Aspenti Health’s Shark Tank entry, “Integration of the Clinical Laboratory and Social Determinants of Health in the Management of Substance Use,” focused on the social factors tied to the co-use of opioids and benzodiazepines, a combination that puts patients at higher risk of drug-related overdose or death.
The project revealed that the top-two predictors of co-use
were the prescribing provider practice and the patient’s age.
“They did an interesting thing—what clinical laboratories
alone cannot do—the predictive value of lab test data mapped by zip code for
patients admitted in partnership with social determinants of health. This helps
to create delivery models to potentially help prevent opioid overdose,” said
Shotorbani, who sees economic implications for chronic conditions.
“If clinical laboratories have that ability to do that in
acute conditions such as opioid overdose, what is our opportunity to use lab test
data in chronic conditions, such as diabetes? The cost of healthcare is in
chronic conditions, and that is where clinical lab data has an essential role—to
support early detection and early prevention,” he added.
Clinical Laboratory Data is Health Business Data
One clinical laboratory working toward that opportunity is TriCore Reference Laboratories in Albuquerque, N.M. It recently launched Diagnostic Optimization with the goal of improving the health of their communities.
“TriCore turned to this business model,” Shotorbani
explained. “It is actively pursuing the strategy of intervention, prevention,
and cost avoidance. TriCore is in conversation with health plans on how its lab
test data and other data sets can be combined and analyzed to risk-stratify a
population and to identify care gaps and assist in closing gaps.
“Further, TriCore is identifying high-risk patients early
before they are admitted to hospitals and ERs—the whole notion of facilitating
intervention between the healthcare provider and the potential person who may
get sick,” he added. “These are no longer theoretical goals. They are
realizations. Now the challenge is for Project Santa Fe to help other lab
organizations develop similar value-added collaborations in their communities.”
Renee Ennis, TriCore’s Chief Financial Officer, told American Healthcare Leader, “Women go in (to an ER) for some condition, and the lab finds out they are pregnant before anyone else,” she said, adding that TriCore reaches out to insurers who can offer care coordinators for prenatal services.
“There is definitely a movement within the industry in this
direction [of Clinical Lab 2.0],” she added. “But others might not be moving as
quickly as we are. As a leader in this transition, I think a lot of eyes are on
what we are doing and how we are doing it.”
Why Don’t More Lab Leaders Move Their Labs to Clinical
Lab 2.0?
So, what holds labs back from pursing Clinical Lab 2.0?
Shotorbani pointed to a couple of possibilities:
A lab’s traditional focus on volume while not
developing partnerships (such as with pharmacy colleagues) inside the
organization; and
Limited longitudinal data due to a provider’s
sale of lab outreach services or outsourcing the lab.
“The whole notion of Clinical Lab 2.0 is basically connecting the longitudinal data—the Holy Grail of lab medicine. That is the business model. Without the longitudinal view, the ability to become a Clinical Lab 2.0 is extremely limited,” added Shotorbani.
New Clinical Lab 2.0 Workshop Focuses on Critical ‘Pillars’
Project Santa Fe Foundation will host the Third Annual Clinical Lab. 2.0 Workshop in Chicago on November 3-5. New this year are sessions aligned with Clinical Lab 2.0 “pillars” of leadership, standards, and evidence. The conference will feature panels addressing:
C-suite Drivers: moderated by Mark Dixon, President of The Mark Dixon Group;