A recent Wall Street Journalarticle pointed to a biased algorithm widely used in hospitals that unfairly prioritized white patients over black ones when determining who needed extra medical help.
While AI has been cited as a data-driven technology that doesn’t make decisions based on emotions, but on actual facts – the reality is that the facts can be misleading.
In the above example, race wasn’t a deliberate factor in how the AI algorithm reached its decision. It actually appears to have used predictive analytics based on patients’ medical spending to forecast how sick patients are.
Yet, the problem is that black patients have historically incurred lower healthcare costs than white patients with the same conditions, so the algorithm put white patients in the same category (or higher) than black patients whose health conditions required much more care
Bias is inherent in a lot of things we do and often, we just don’t realize it. In this case, the data assumed that people who paid more for services were the sickest. As illustrated, we have to be considerate of the data we use to train algorithms, Cost of services or amount paid shouldn’t be information we use to determine who is sicker than another.
In another example, if skin-cancer-detection algorithms are typically trained on images of light-skinned patients, they would be less accurate when used on dark-skinned patients, and could miss important signs of skin cancer. The data must be inclusive to provide the best results.
While AI can accelerate disease diagnoses, bring care to critical patient populations, predict hospital readmissions, and accurately detect cancer in medical images, the example illustrates the caveat: AI bias –whether because of a lack of diverse data, or the wrong type of data – exists in healthcare and it can lead to social injustice, as well as harm to patients.
In addition to racial bias, unchecked algorithms can cause other types of bias as well, based on gender, language or genealogy. In fact, according to IBM research, there are more than 180 human biases in today’s AI systems, which can affect how business leaders make their decisions.
As an example of gender bias in healthcare, for many years cardio-vascular disease was considered a man’s disease, so information was available based on data collected from men only.
This could be fed into a chatbot and lead a woman to believe that pain in her left arm was less urgent – possibly a sign of depression – with no need to see a doctor right away. The consequences of this oversight could be devastating.
By Ken Perez, vice president of healthcare policy, Omnicell, Inc.
Discussions about the application of artificial intelligence (AI) in healthcare often span multiple areas, most commonly about making more accurate diagnoses, identifying at-risk populations, and better understanding how individual patients will respond to medicines and treatment protocols.
To date, there has been relatively little discussion about practical applications of AI to improve medication management across the care continuum, an area this article will address.
The Significance of Medications
What’s the first thing that comes to mind when someone mentions prescription drugs in the United States? In poll after poll, the high and rising costs of medications are American voters’ top healthcare-related issue.
This concern is well founded. The U.S. spends almost $400 billion a year on medications–$325 billion on a retail basis and about $75 billion for inpatient and outpatient use.
To put the $400 billion in perspective, it is equal to about 11% of total U.S. healthcare expenditures, and it’s one of the top reasons why the U.S. spends much more on healthcare than other industrialized countries.
Medication Management Shortcomings
Unfortunately, there are a lot of issues with the medication management system, broadly defined.
It’s estimated that 20-30% of prescriptions are not even filled, not even picked up at the retail pharmacy. According to the Centers for Disease Control and Prevention (CDC), each year, adverse drug events result in 1.3 million visits to the emergency department, and of those ED visits, over a fourth, 350,000, result in hospitalizations, which result in significant costs.
Over the past 50 years, much legislation has been passed to regulate and reform the U.S. healthcare system, and this has significantly increased the administrative burden on healthcare provider organizations. As a result, according to data from the Bureau of Labor Statistics, the National Center for Health Statistics, and the United States Census Bureau’s Current Population Survey, the number of administrators has grown by 3,200% since 1970, while over the same period, the number of physicians has been relatively flat, in line with population growth. Correspondingly, per research funded by the Physicians Foundation, it is estimated that the average physician and/or his or her staff spends 785 hours per year on quality reporting.
The administrative burden also falls heavily on pharmacists. According to a national survey by the American Society of Health-System Pharmacists (ASHP), pharmacists spend over three-fourths of their time on non-clinical activities—mostly manual, administrative processes.
In spite of the massive amount of spending on medications, the medication management system is fraught with errors at multiple steps in the medication-use process, prescriptions are often not filled, and over one-fourth of all hospital readmissions are potentially preventable and medication related.
Current Health announced that it has launched a collaboration with Mayo Clinic to develop remote monitoring solutions that accelerate the identification of COVID-19-positive patients and predict symptom and disease severity in patients, healthcare workers and other at-risk individuals in critical service sectors.
Using digital biomarkers collected by Current Health’s FDA-cleared remote monitoring sensors and platform, experts from Mayo Clinic and Current Health will also be able to expedite identification and assessment of treatment efficacy and improve care for patients with or at risk of COVID-19 infection. Through this collaboration, Current Health and Mayo Clinic aim to improve patient outcomes while preserving and optimizing health system capacity worldwide.
Today, more than 40 hospital systems around the globe use Current Health’s remote patient monitoring platform to monitor and manage patient health. These systems are now increasingly using Current Health to monitor and manage patients infected with COVID-19 at home and in the hospital. The next stage is to use digital biomarkers collected by the Current Health solutions, such as temperature, heart rate, oxygen saturation, activity and posture, to develop AI-based algorithms that can detect and predict symptom and disease severity to enable proactive treatment.
This collaboration will leverage Current Health’s existing patient database – which already includes anonymized vital sign data and raw physiological sensor data from hundreds of patients infected with COVID-19 and thousands of uninfected patients – as well as algorithms developed by Mayo Clinic, which will be used to provide individualized care to patients with complex and critical medical conditions. By working together, Current Health and Mayo Clinic hope to scale data analytics, add to Mayo Clinic’s major advancements in accelerating COVID-19 detection and diagnosis, and further efforts to understand and treat this disease.
“Our collective ability to save lives hinges on our ability to understand this virus quickly. COVID-19 has presented in many ways across different people, which has made it very challenging to understand the virus and how it develops,” said Chris McCann, CEO and Co-Founder, Current Health. “By combining our platform with the deep medical and scientific expertise that exists at Mayo Clinic, we seek to explore both known and novel biomarkers, as well as how they manifest in entirely diverse populations. This will be critical to determining how we define, and enable effective treatment of, this disease.”
“Combatting the COVID-19 pandemic is our number one priority,” said Jordan D. Miller, Ph.D., who directs the Center for Surgical Excellence and leads the investigative team at Mayo Clinic.
“Real-world, continuous data – from patients infected and not infected with the disease – is essential to understanding and predicting how the disease presents and evolves,” says Abinash Virk, M.D., an infectious disease expert at Mayo Clinic. “If we are successful in accomplishing our goals, we believe we will improve how patients with COVID-19 are identified, monitored, managed, and ultimately help with their recovery.”
Mayo Clinic will also become an investor in Current Health as part of this collaboration.
The Mount Sinai Health System has partnered with Sana Labs to launch Project Florence, a personalized learning platform to enhance the skills of nurses treating COVID-19 patients in New York City. The group, facilitated by the New York Academy of Sciences, is also making the platform available for free to hospitals around the world to improve medical response and care during the pandemic.
The virtual training platform, available through Sana Labs, provides a curriculum developed by Mount Sinai that includes the latest on industry resources and policies from organizations including the American Association of Critical Care Nurses.
After users complete an AI-powered adaptive assessment that measures their knowledge, the platform recommends personalized content in real time to address individual skills gaps. It can be accessed from any internet-connected device including phones, tablets, laptops, and desktop computers. The project was officially launched at the Mount Sinai Health System on Monday, April 13.
“The profound shortage of intensive care nurses and respiratory therapists will be one of the most significant hurdles facing U.S. hospitals treating critically ill COVID-19 patients,” said Jane Maksoud, RN, MPA, senior vice president and chief human resources officer, Mount Sinai Health System. “Project Florence will be a great benefit to staff preparing to care for critically ill patients. We are grateful for the partnership we have developed with Sana Labs and the work we have done together to assist our nurses on the front line.”
A projected 4.8 million Americans will be hospitalized for COVID-19, according to the American Hospital Association. Of those hospitalized, an estimated 40 percent or nearly 2 million patients will require admittance to the ICU. While there are currently about 550,000 critical care nurses in the United States, tens of thousands of nurses will be in demand in the coming months.
“I’m very excited to bring this innovative approach to Mount Sinai hospitals to help advance the skill set of our nurses,” said Diane Adams, MS, Chief Learning Officer of Mount Sinai Health System. “Not only are we advancing the essential skills of our staff, but we are also meeting the needs of our community during a particularly critical time across New York City, the United States, and the rest of the world.”
As hospitals shift priorities from other departments to ICUs, the two-day curriculum is tailored to each individual and suitable for nurses, as well as other medical professionals who are called to assist and may require an update on their understanding of ICU equipment and procedures.
Future technology is changing the world of health. As a result, new ways on how health research is conducted and performed are beginning to emerge. Major Contract Research Organizations or CROs are starting to employ AI in pre-clinical tests, thus revolutionizing the role of technology in healthcare.
Artificial intelligence is a type of intelligence displayed by machines and computer systems. Nowadays, there are several ways how pre-clinical CROs use AI in their studies. But, first, what are pre-clinical CROs?
Pre-clinical CRO Defined
Pre-clinical CROs, otherwise known as Pre-clinical Contract Research Organizations, are companies that provide knowledge, skills, and experience needed to transform a medical or pharmaceutical idea concept into a final product. There are a lot of processes involved before the final product is revealed, which include the discovery and development stage, pre-clinical research stage, the clinical research stage, and, lastly, the FDA review.
The period between pre-clinical tryouts and the unveiling of the product is where the role of a pre-clinical CRO is most critical. Drug ideas and prospective products may fail within this period; hence modern pre-clinical CROs, like Ion Channel CRO, continue to dig deeper into the capacity of future technology to increase efficiency in health research.
Reasons Why Pre-clinical CROS Are Using Future Technology/AI To Conduct Health Research
Reduces uncertainty in pre-clinical experiments – AI is now being used to reduce the improbability that comes with pre-clinical trials. This will go a long way in reducing time spent on research, cutting down financial costs, and optimizing data gathering.
Gathers data and obtains actionable insights – Researchers now use AI to streamline data collection and selection of recipients of pre-clinical tests. Data collection and analysis are an integral part of health research, and keeping up with the zillions of data available is impossible for the human researcher. However, with the aid of AI tools, such as deep learning and machine learning, it is possible to analyze, select patterns, and connect relevant data that can lead to drug discovery.
Researchers also make use of reports generated by AI to gain actionable insights during their pre-clinical studies. AI tools can also improve recipients’ selection by choosing the most appropriate group capable of responding to pre-clinical research and tests.
Last month saw the rollout of the latest upgrades to Amazon’s Echo speaker line: earbuds, glasses, and a ring that connect to Amazon’s personal assistant Alexa. These new products are just three examples of a growing trend to incorporate technology seamlessly into our human experience, representing the ever-expanding frontiers for technology that have moved far past the smartphone.
These trends and others are going to make a big impact in the healthcare space, especially as providers, payers and consumers alike slowly but surely recognize the need to incorporate tech into their workflows to meet the growing consumer demand for digital health tools. At the same time, the data-hungry nature of these innovations is creating its own problems, driving a discussion around privacy and security that is louder and more urgent than ever.
Here are three trends to look out for in the coming year:
Artificial Intelligence (AI) and Machine Learning are growing into themselves
In 2020, we will continue to see AI and machine learning push boundaries, while at the same time mature and settle into more defined patterns.
With the adoption of technologies like FaceID, facial recognition technology will be an important player in privacy and security. It can be leveraged to simplify the security requirements that make multi-factor authentication a time-consuming process for healthcare professionals — on average, doctors spend fifty-two hours a year just logging in to EHR systems. On the patient end, this same technology has the ability to detect emotional states of patients and anticipate needs based upon them, and the success of startups like Affectiva, the brainchild of MIT graduates, shows its tremendous promise..
Meanwhile, FDA-approved innovations from Microsoft and others claim the ability of computer vision for assisting radiologists and pathologists in identifying tumors and abnormalities in the heart. While robotic primary care is a long way off, some view AI as a rival to more niche clinical positions.
Response by Kristin Simonini, vice president of product, Applause.
Healthcare has long been looked at as a laggard when it comes to adopting digital services. Part of that is due to the stringent regulations of the industry and the sensitivity surrounding personally identifiable information. Part of the blame, however, falls on healthcare providers themselves. As more and more providers in the industry start to embrace digital innovation, a number of key trends emerged over the past decade including:
The embrace of mobile technology for scheduling appointments and other routine tasks
Telehealth patients accessing doctors for consults, education, and certain outpatient treatments across a variety of fields
The IoT explosion (Fitbits and other wearables) providing customers health information to drive the healthcare they receive
Healthcare’s focus on patient experience means bringing a critical eye to current digital experiences. Ease-of-use and inclusivity must be considered to ensure high-quality digital experiences across all touchpoints, particularly on smartwatches, tablets, and smart speakers
In terms of predictions for 2020, we expect use of voice technology will continue to grow and empower the healthcare industry in new ways, including supporting patients. The benefits that voice brings to healthcare can be seen in medical record transcriptions, chatbots sharing the work, sharing knowledge, voice-user interface, and connecting clinics to customers.
In addition, AI will continue to impact the healthcare industry in numerous ways. As healthcare embraces AI, it will also need to address issues of bias. All types of AI – from virtual assistants learning how different users ask for the same thing, to healthcare apps identifying potential health issues from uploaded photos – have been hampered by the same challenge: sourcing enough data to teach the machine how to interpret and respond, and then testing the output at scale to ensure the results are accurate and human-like when necessary. To mitigate bias concerns, healthcare will need to make AI more representative of patients.
Baystate Health, the premier integrated health system serving more than 800,000 patients in western New England, announced a partnership with Life Image, the largest medical evidence network providing access to points of care and curated clinical and imaging data, to develop novel artificial intelligence tools that would help advance technical innovations in radiology, neurology and oncology.
Specifically, TechSpring, the innovation arm for Baystate, will work with Life Image to evaluate a number of AI solutions including those that promise to improve speed and accuracy in diagnosing blood clots in stroke patients; improve clinical pathways for physicians treating or diagnosing a patient by finding and comparing clinical criteria against a group of de-identified patients with similar clinical characteristics; and identify potential patient matches to oncology clinical trials in order to advance cancer research, as well as give western New England residents better access to potentially life-saving treatments.
Baystate and Life Image began working together 10 years ago when the health system became one of the company’s first customers. Life Image created the image exchange category when it developed solutions more than a decade ago to help solve the many technical and structural barriers that prevented the seamless exchange of medical images.
With its beginnings in image exchange, Life Image is now a global medical evidence network that offers ‘living’ datasets of novel imaging data that’s linkable to other clinical information and provides network access to points-of-care to enable improved care delivery, novel research and innovation.
It’s no secret that cyberattacks are escalating, rising in tandem with the growing sophistication of technology. One industry that has taken a massive hit by cyberattacks in recent years is the healthcare industry. The healthcare industry is increasingly reliant on technology and data connected to the internet, such as patient records, lab results, radiology equipment and hospital elevators. Now imagine if a cybercriminal encrypted an entire hospital’s data with a nasty ransomware. Doctors would be unable to pull up a patient’s medical records, or worse, utilize equipment connected to the internet to make a proper diagnosis.
Unfortunately, this is the reality that healthcare industry professionals are facing today. And while 92% of healthcare organizations are confident in their ability to respond to cyberattacks, there is a plethora of malicious activity that poses a great threat to their networks. Here are the main cybersecurity challenges faced by the industry today:
The Rise of Ransomware
You might recall the WannaCry attack of 2017, the ransomware worm that attacked hospitals as well as other industries by exploiting a weakness in Windows machines. This worm infected thousands of computers around the world and threw the United Kingdom’s National Health Service into chaos. This resulted in the Health Care Industry Cybersecurity Task Force to conclude that healthcare cybersecurity was in critical condition.
Why was the healthcare industry so impacted by this cyberattack? Many hospitals struggle to keep up when it comes to upgrading their operating systems due to the sheer volume of devices on the network. However, much of the software in a medical-specific device is often custom made, making system upgrades difficult. Additionally, manufacturers tend to avoid prematurely pushing out modifications that could potentially impact patient safety. For these reasons, medical machines continue to exist with outdated software, putting them at greater risk of cyberattacks such as ransomware.
Lack of Investment
Many organizations within the healthcare industry suffer from a lack of investment in cybersecurity solutions. Despite the number of breaches that occur, healthcare is behind other sectors when it comes to taking security measures. Only 4-7% of healthcare’s IT budget is allocated to cybersecurity, while other sectors allocate about 15% to their security practices. However, the finances associated with a cyberattack if these solutions aren’t put in place can take an even greater toll on an organization. Some hospitals and healthcare insurers see estimates of over $5 billion in costs as the result of cyberattacks on their systems. On top of the costs incurred finding a solution to fix these breaches, healthcare organizations then have to deal with fines from the Department of Health and Human Services Office of Civil Rights.
Securing Connected Devices
With the growing adoption of IoT, more and more devices are being connected and used in healthcare systems. However, as connected medical devices become more powerful and widely adopted, they become greater targets for malicious actors to exploit. According to the Cybersecurity in Healthcare report, over 16% of IT professionals can’t patch their own operating systems, leaving the network wide open for attack. Now imagine if a cybercriminal gained access to just one medical device on the exposed network. This could lead to the theft of sensitive patient data or even unauthorized access to an implanted device that could cause physical harm to the user.
Artificial intelligence (AI) has two faces in healthcare.
One face sings the praises of AI as the tonic that will enable healthcare to
deliver better clinical outcomes at a lower cost and the second face is full of
skepticism and raises barriers to adoption at every turn. It is heartening to
see that a third face is emerging, the thoughtful and appropriate use of AI to
predict adverse health events; to identify and stratify patients in need of
health, social, and human services; and the application of AI in the automation
of tasks, activities, and processes.
To understand the likely evolution of AI-based automation,
it’s important to evaluate the interaction of humans and machines across these
five levels. At each level of automation, the following questions must be asked
Who produces insights? – Does the human
or the machine (AI) analyze data and deliver insights from such analysis? Does
the human or the machine describe what something is, how it trends, why
something is happening, and what might happen next?
Who decides and how? – Once all relevant
analysis has been conducted, does the human or the machine make the decision
based on the derived insights?
Who acts based on the decision? – Finally,
a decision should lead to an action by either a human or a machine? The action
can be in the digital or physical environment.
Based on the responses to these questions, IDC has
identified the following five levels of AI-based automation:
Human Led – At the first level, it is
the human who analyzes the data using limited technology, such as tools for
only descriptive analytics; it is the human who makes the decision based on the
analysis (or experience); and it is the human who acts based on the decision.
Human Led, Machine Supported – At the
second level, the human continues to lead data analysis, decision making, and
action steps but is now more reliant on the machine across these steps.
Machine-led, Human Supported – At the
third level, it is the machine that is using a wide range of analytic and AI
techniques to conduct the analysis and produce insights. These insights are
reviewed by humans. The human still makes the decision based on machine’s
recommendations, and it is the human who acts based on the decision. However,
at this level, the machine acts to provide oversight over human decision making
Machine Led, Human Governed – At the
fourth level, the machine analyzes data and produces insights without the need
for human review. At this level, the machine decides based on the analysis of
all available data and a framework of human-developed governance policies and
procedures. At this stage, it is also the machine that acts based on the
decision under the governance of humans.
Machine Led – At the fifth level, the
world has likely achieved general AI. At this stage, there is a full AI-based
automation without the need for human involvement. At this level, we need to
think of machines that set their own goals and understand all mathematical,
economic, legal, and other external constraints. Most AI academics and experts
in labs of commercial enterprises predict this level of AI to arrive no sooner
than in about 50 years.
recent years, one of the shortcomings in the commercial sphere of AI has been
the misrepresentation of the scope of possible automation. Too often, we hear
claims of AI systems automating end-to-end processes and predictions of massive
labor losses, this does a disservice to organizations trying to plan for the
appropriate level of investment in AI. There is a need for a pragmatic
framework that decision makers across industries can use to assess
opportunities and risks of AI-based automation. The levels of AI-based
automation must also be viewed in the context of the scope of automation. We
define this scope where:
Task is the smallest possible unit of work
performed on behalf of an activity.
Activity is a collection of related tasks to be
completed to achieve the objective.
Process is a series of related activities that
produces a specific output.
System (or an ecosystem) is a set of connected
IDC’s AI automation framework was developed to help wade through the hyperbole associated with AI. Our goal is to help provide a planning tool and key piece of vendor evaluations processes to fully understand the role AI is playing in software and guide strategic decision making.