Diabetes Meets Machine Learning, Part 2

Continuous glucose monitoring takes center stage

John Halamka, M.D., president, Mayo Clinic Platform, and Paul Cerrato, senior research analyst and communications specialist, Mayo Clinic Platform, wrote this article.

In February, we discussed the benefits of using machine learning (ML) to improve screening for diabetes and for managing the disorder. In some situations, ML can more accurately detect the presence of prediabetes, for instance. Similarly, there’s research to show that the right algorithms can improve the treatment of Type 1 diabetes.  But there’s also mounting evidence to suggest ML can benefit Type 2 patients.

One of the problems clinicians and patients have in managing Type 2 diabetes is the fact that medical care is episodic.  Patients may see their physician or nurse practitioner once every few months, which means most of the time, they are on their own. And while it is true that glucose meters let patients monitor their blood glucose (BG) levels on a daily basis, those readings have their limitations. Many patients hesitate to take enough readings because finger pricks hurt. It can also be difficult to determine the need for changes in one’s diet, medication, and physical activity based on a few daily BG readings. And while glycated hemoglobin A1c provides a picture of one’s long-term metabolic control, it doesn’t provide immediate feedback on how to respond to day-to-day changes in food intake, stress levels, exercise and the like.  

Mounting evidence suggests that continuous glucose monitoring (CGM) is a game changer that will profoundly improve a Type 2 diabetic’s ability to control their condition — at least for those patients willing to commit the time and energy needed. Advances in hardware technology and artificial intelligence have ushered in a new approach to metabolic control that takes advantage of CGM. As the name implies, CGM provides patients with continuous, immediate readings of their BG levels, about once every 5 minutes. This is accomplished with the help of a subdermal sensor and a wireless connection to a glucose meter. Some systems will also send this data to a tablet or smartphone, which can then be used to make more informed decisions on how to adjust one’s carbohydrate intake, for instance, or to alert a patient about a particular food or beverage that’s spiking their BG levels. The system will also alert users about unexpected hypoglycemic episodes before they become dangerous. 

The Endocrine Society and the American Diabetes Association consider CGM standard of care for insulin-dependent Type 1 diabetes,¹ but until recently, CGM platforms were considered too expensive and not efficacious  enough to justify their use in type 2 diabetes. That thinking is changing, however. Tejaswi Kompala, MD, and Aaron Neinstein, MD, with the Division of Endocrinology, University of California, San Francisco, point out: “High costs and uncertainty over efficacy and necessity have kept CGM from widespread use in people with T2D. However, the newest CGM models, the Abbott Freestyle Libre and Dexcom G6, have begun to overcome many of these technical barriers to use of CGM systems. The sensors are inserted painlessly, are small enough to fit easily under clothing, can remain in place for 10 to 14 days.” ²

Devices like the Freestyle Libre really are an important advance in patient self-care because they eliminate the need to fingerprints. * Combining this technology with machine learning and other types of AI is  helping clinicians and patients interpret a massive influx of data into actionable insights. Several vendors, including Livongo, Canary Health, and Omada Health,* have entered this space, providing patients with easy-to-navigate digital tools that take advantage of the new technology. Livongo, for instance, combines clinical decision support with patient support that includes customized glucose meters and nurse coaches who send personalized messages to patients in need of advice. The Livongo program uses 4 technologies it refers to as AI+AI, which represents Aggregate, Interpret, Apply, and Iterate. It aggregates data from a variety of sources, including its custom-built devices, as well as a patient’s age, gender, zip code, medical claims, and pharmacy claims. The aggregated data is interpreted to create a unique data set it calls Health Signals, which are derived from its applications, devices, coaches, and other sources. It uses these signals to build relevant healthcare messages and outputs for its apps. The Apply in AI+AI refers to the applications linked to its glucose meter, BP cuff, and digital scale, as well as the “human applications,” namely, its coaching system and care coordination team. Finally, the system folds the signals it has generated back into the AI engine to make the system smarter.³

The recent entry of UnitedHealthcare, the largest healthcare insurer in the US., into this area should send a strong signal to clinicians and Type 2 patients alike: This approach has merit!  Major 3^rd party payors don’t invest their dollars without doing their due diligence to determine that their offerings will likely be cost-effective. Brian Thompson, CEO of the company’s Medicare and Retirement unit, summed up the potential for ML-enhanced diabetes care: “Continuous glucose monitoring can be a game changer for people enrolled in our Medicare Advantage plans, as the data can be translated into personalized information that can be acted upon in real time.”

 

*Products and services mentioned are not endorsements.  Mayo Clinic has no financial relationship with any of these vendors.

 

References

1. Peters AL, Ahmann AJ, Battelino T et al. Diabetes Technology—Continuous Subcutaneous Insulin Infusion Therapy and Continuous Glucose Monitoring in Adults: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab 101: 3922–3937, 2016.

2. Kompala T. Neinstein A. A New Era: Increasing Continuous Glucose Monitoring Use in Type 2 Diabetes. Evidence-Based Diabetes Management, March 2019, Volume 25, Issue 4.

3. Cerrato, P. Halamka, J. Reinventing Clinical Decision Support: Data Analytics, Artificial Intelligence, and Diagnostic Reasoning. CRC Press/HIMSS, 2020, p. 38-39.

When Technology, Policy, and the Urgency to Change Converge

Our new book, The Digital Reconstruction of Healthcare, is about to be published by Taylor and Francis, as part of its HIMSS book series. We wanted to give readers a preview of what’s to come so we are posting the Preface of the book ahead of time.

In our last two books, we began the conversation discussing the power of words, including misdiagnosis, cynicism, and optimism.^1,2 In this book, our focus is on reconstruction, and all its implications for healthcare. To some, it might suggest the tearing down of an existing structure, a complete replacement of the healthcare ecosystem as we know it. Neither of us believe that’s warranted. Our goal, instead, is to address the unsustainable situation that we currently face in the United States and around the globe, and the emerging digital tools that are transforming patient care.

These solutions are not intended to demolish the foundation upon which medicine is built, but neither are they designed to patch up crumbling walls or apply duct tape to the ineffective, cost-prohibitive practices currently in place. To extend the metaphor: The foundation of healthcare may remain solid but many of the walls, floors, windows, and doors that sit on this foundation are rotting and need to be replaced. The next eight chapters will provide evidence from multiple sources, including deep learning specialists, consultations with thought leaders and government officials around the world, peer-reviewed studies, unpublished data, and cutting edge initiatives at Mayo Clinic and several other healthcare leaders — in addition to our combined 60+ years of experience working in healthcare. The preponderance of evidence from all these sources makes a compelling argument: Business as usual is no longer an option; the digital reconstruction of healthcare is no longer on the world’s wish list. It’s becoming a sustainable reality — and one that is all the more necessary in light of the COVID-19 pandemic. That reality will include the partial shift from caring for patients in hospitals, clinics, and medical offices to meeting their needs through telemedicine, hospital-at-home programs, and remote patient monitoring.

In Chapter 1, we address the question: “Is digital reconstruction necessary?” and include a review of the evidence on the effectiveness of digital healthcare, the shortcomings of episodic patient care, diagnostic errors, and our inadequate infrastructure.

Chapter 2 looks at the merits and limitations of telemedicine, hospital and home, and remote patient monitoring. It offers advice on making informed telemedicine choices and the impact of COVID-19, and provides a review of the scientific evidence. We also take a closer look at Mayo Clinic’s Advanced Care at Home program.

Chapter 3 discusses the digital assault on COVID-19, including the development of better predictive and diagnostic tools, expanding the knowledge base to address the pandemic, and the importance of taking a holistic approach to the infection.

Chapter 4 once again explores the value of big data, artificial intelligence, and machine learning, a topic we have looked at in several previous books. The discussion analyzes the evidence in diabetes, cardiovascular disease, cancer, gastroenterology, and psychiatry. We also address one of the most difficult issues in medicine: when does correlation imply causality. Finally, we devote a section to advanced data analytics, including a summary of how Mayo Clinic’s Clinical Data Analytics Platform operates.

Chapter 5, “Exploring the Artificial Intelligence/Machine Learning Toolbox,” is a primer on artificial neural networks, random forest modeling, gradient boosting, clustering, and linear and logistic regression. We are working from the assumption that many readers do not have a background in statistics or data science and hope these brief tutorials translate these complex topics into plain English.

Chapter 6 dives into the many conversational technologies emerging in healthcare. We begin with the role of natural language processing and then discuss the potential of voice technology to help diagnose disease and the role of Siri, Google Assistant, Alexa, and other patient-facing tools. Finally, we emphasize the urgent need to fight misinformation — with truth and trust.

Chapter 7, “Securing the Future of Digital Health,” tackles one of healthcare’s most vexing problems: Cyberattacks. We outline the need for comprehensive risk analysis, staff education to reduce the risk of phishing attacks, along with several basic precautionary steps, including encryption, strong passwords, firewalls, and the like. We also include a section on one of the most vulnerable parts of the healthcare ecosystem: “The Internet of Medical Things.”

Finally, in Chapter 8, we explore international initiatives to digitally reconstruct healthcare. Specific programs in the United Kingdom, China, and the Netherlands are discussed, as are the needs of low-resource nations.

The emergence of the numerous digital health solutions discussed in the following pages does not imply that information technology will singlehandedly rebuild the healthcare ecosystem. Healthcare needs much more than that. Call it “intensive lifestyle management.” Unfortunately, too many IT enthusiasts see technology as a savior and are eager to invest billions of dollars setting up countless initiatives, platforms, and networks in the hope that they will create a more cost-effective system. That kind of magical thinking is doomed to failure over the long-term. If properly deployed, technology will augment other resources much like AI-fueled algorithms are now augmenting the diagnosis of eye disease and cancer. Society will still need to address the underlying cultural, financial, and clinical root causes behind our failing healthcare system — issues that are beyond the scope of this book. We both have the humility to recognize that digital health and all the tools it brings to bear, are only part of the solution. Our experience and research, nonetheless, demonstrate that they are a crucial part of that solution.

 

Paul Cerrato, MA

John Halamka, MD, MS

  

References

1. Cerrato, P. Halamka, J. Reinventing Clinical Decision Support: Data Analytics. Artificial Intelligence, and Diagnostic Reasoning. Taylor & Francis, HIMSS, 2020. Boca Raton, Fl.

2. Cerrato P, Halamka J. The Transformative Power of Mobile Medicine: Leveraging Innovation, Seizing Opportunities, and Overcoming Obstacles of mHealth. Academic Press/Elsevier. 2019, Cambridge, MA.

The Year of Living Dangerously

It's been exactly one year since our lives changed. On March 10, 2020, Governor Charlie Baker declared a state of emergency for Massachusetts, changing the way many of us travel. On March 11, 2020, the World Health Organization declared COVID-19 a global pandemic, its first such designation since declaring H1N1 influenza a pandemic in 2009. On March 15, 2020, I flew to Minnesota and prepared my Rochester apartment for a lockdown. I said my goodbyes to colleagues on March 16 and flew back to Boston. We've run the Mayo Clinic Platform at a distance for the past year.

During the pandemic, those old enough to have overcome adversity have done better than the young and less experienced. As we age, we do what must be done, responding to the unexpected and gaining resilience. Consider how resilience has created many good things over the last 12 months.

What have we gained?

Despite our geographic distance, the Mayo Clinic Platform team has truly become a family. The differences between our personal lives and work lives have melted away because we live and work in a continuous stream, navigating each day's events to care for everyone who depends on us at Mayo, at home and in the external world. We've increased our productivity, agility, and pace, doing more each week than would have been physically possible in person. We've also recognized that it's possible to have too much of a good thing, so we have put guardrails on our schedules, including protected Saturdays. As conferences and presentations have become virtual, the number of opportunities for communication has markedly increased. Lost days due to travel are gone, and a one-hour keynote takes just one hour.

We've mastered new remote working technologies, streamlined processes and shortened turnaround times. We've changed the way we recruit and hire talent, reducing our dependence on geographic proximity. We've also created a cadence that begins each week with decision-making and goal clarity and then empowered Platform staff to do what must be done. As servant leaders, we're always available, but staff working remotely have more independence than working in an office. This has enhanced their self-reliance and confidence. We've also brought every deal, every product launch, and every project to completion on time. Finally, we've all stayed healthy, and none of us have experienced any health-related consequences of COVID.

But what have we lost?

As work moved virtual, we lost the "over the cubicle" effect. We don't have casual conversations by the coffee machine or when crossing the hall.  Humans are a social species. We're used to proximity and communication with facial expressions, posture and the handshakes/hugs that offer reassurance. Without that, virtual-only communications can lack context and create anxiety.

Since everyone is connected all the time, there is an expectation of instant response and same-day issue escalation. That forces multi-tasking because we're doing two jobs — the meeting in front of us and the hundreds of mini-meetings occurring in email, text and calls.

That multi-tasking and the focus required to use video-based work tools has caused extreme fatigue.  That fatigue can affect our mood. Societies throughout the world are experiencing an epidemic of depression during COVID. As part of the response to stress, alcohol sales have skyrocketed and that will have its long-term consequences.

We've also lost the ability to decompress, gather for spontaneous conversation, or spend hours immersed in a book in front of a fire. Our mobile devices have become an extension of our brains. Sometimes the best thing you can do is just be together without an agenda. It's what I call the gift of time.

I've described the five stages of COVID as isolation/PPE, testing/contact tracing, therapies/trials, vaccines/passports and transition to a new normal. My wardrobe is now complete with many different mask types. I have my contact tracing apps. I've co-chaired the collection of data for evaluating numerous therapies. I've received my vaccines and have a credential on my phone. That means it's time to transition to a new normal — thoughtfully and incrementally. I return to Rochester the first week of April, following Mayo's guidelines for masking, expected behaviors and Mayo Clinic work patterns. The Subaru I keep in Minnesota has a dead battery after a year of limited use. We'll fix that. I've used the Way.com platform to "rent" a parking space at the Courtyard by Marriott so I can keep the car at the airport for $4 per day, coming and going in alignment with evolving work patterns. I will need to restock my refrigerator since year-old pickles, peanut butter and pepperoncini do not make a balanced meal.   

I'm ready to restart the pattern of travel, resuming daily living activities in Minnesota and rebuilding what we lost during COVID while also holding the gains we made. I can honestly say that I have no regrets about our lives together over the past year. We did what needed to be done. We were strong but still admitted our vulnerabilities. We helped those around us on the journey and never made decisions based on self-interest. 

As we reflect on the past year, I think we can say that it changed our lives and made us more resilient for whatever comes next. This year will be one we tell our grandchildren about.

Vaccine Credentials Done Right

Vaccine credentials could be a safe and convenient way for people to resume a more normal daily life, but it’s critical that we do it in the right way. Public and private sector collaboration is essential, and we must be driven by science and our best understanding of what vaccines mean for immunity under different conditions. It’s reasonable to expect that some businesses, schools and countries may require it – just like yellow fever vaccine certificates.

So if we are moving toward those uses, we need to create a way patients can share their vaccination status if they wish to do so, with tools designed to be voluntary and not discriminate. But first we need to develop international standards for organizations administering the vaccines to make credentials available in a format that’s accessible, interoperable and digital. We’re working on that with other health care providers, EHR vendors, state immunization systems, pharmacies and tech companies. This work will enable people who have received the COVID-19 vaccine to access, store and share their records in a secure, verifiable and privacy-preserving way.

At this time, patients who have been vaccinated at Mayo Clinic can choose to show their vaccination status on their phones using a health app, if they wish to. The app will say they’ve been vaccinated, which vaccine they received and the date they got it. Over the next several months, the Vaccine Credential Initiative, a collaboration of academia and industry, aims to create a standardized way of proving immunization, using a displayable QR code. Very importantly, that standard format, called the SMART Health Card, has been endorsed by stakeholders throughout the U.S. The vaccine credential will be issued by the doctor’s office, hospital or pharmacy that gives the vaccine, and the data can be uploaded to an app on a phone. Name and birth date are included, but no other medical information is shared, and the vaccine credential is always under the control of the person who received the vaccine. The code sits in a digital wallet, just like a ticket to a sports event that gets scanned when you enter. And also importantly, the vaccine credential will be something that can be printed and shown in paper form, for those who don’t have or use smart phones.       

And to be clear, this effort is not about requiring vaccination but offering patients a verifiable way to share their vaccination status if they wish to do so. There is always a choice. Of course, we must address privacy concerns and questions that may arise and always ensure that we have the right protections in place so everybody is comfortable with how their health information is shared and used.

Again, it’s very important to recognize that these tools are designed to be voluntary and must not be used to discriminate. The efforts are well-meaning, but we must follow guidelines for appropriate use and ensure they are not used to limit access to jobs or essential needs like getting groceries or seeking health care. 

So, a vaccine credential could help reopen the economy, while not creating another divide as millions of people don’t have smart phones. The digital divide is real, and all our solutions must work on a variety of phones as well as on paper for those without technology. Tech tools may be preferable for many consumers, but we also have to consider how we’re going to handle those who don’t have ready access to a smartphone or other device. 

Although we don’t expect that most of the world to be vaccinated until the end of 2021, I wouldn’t be surprised if some businesses require proof of vaccination in the future – either on paper or digitally – just like some countries now require proof of a negative COVID-19 test to enter. Or people may have to agree to on-site testing or proof of recent quarantine. Our work in health care is to ensure that patients have the tools they need to support whatever approach they choose.

Listen Better, See Deeper

Combining Medical Attentiveness with Artificial Intelligence

John Halamka, M.D., president, Mayo Clinic Platform, and Paul Cerrato, senior research analyst and communications specialist, Mayo Clinic Platform, wrote this article.

Embracing an “ecology of attention” will significantly improve patient care, according to Mark Kissler, MD, at the University of Colorado.¹ Kissler and his colleagues point out that clinicians spend much of their time multi-tasking and navigating around interruptions. While such juggling acts are often unavailable, it’s important to occasionally step back and ask: Is this the best use of my time? Equally important: Do the distractions cause “lapses in judgement, insensitivity to changing clinical conditions and medication errors.”¹ If so, there are practical solutions that can help refocus our attention.

Kissler et al. offer several recommendations.  Initially, we need to recognize the difference between reachability and deep interpersonal availability. Most clinicians want to be reachable to help solve problems within their scope of practice, but spreading oneself too thin can jeopardize one's communication quality. Designing the physical spaces in which we interact with patients and colleagues is another area where we can build better attentiveness. For many years, the business and tech worlds believed that bullpens and shared office space would foster communication. Still, we are now beginning to realize that all the distractions may impede creativity and productivity. Finally, the University of Colorado team emphasizes the need to build attentiveness into one’s organizational culture: “Provide clinicians with the tools and language to prioritize attention in their daily practice.” That can be accomplished by developing a culture that encourages staffers to listen with curiosity, communicate with empathy, and remain open to others' perspectives, even when that perspective contradicts our understanding of the facts.

Of course, as every clinician knows, even the most attentive listener can still miss things. The medical interview can only uncover so much, necessitating a careful physical exam and diagnostic testing when appropriate. While imaging studies have always been a part of the diagnostic process, machine learning has taken these procedures to a new level, with companies like Zebra Medical, GE, Siemens, and AIDOC introducing useful services. AIDOC, for instance, has created a suite of services that combines three layers: an algorithmic layer, a product layer, and a clinically viable solution layer. All three are combined and implemented directly into the workflow. According to AIDOC, the platform reduces “turnaround time and increases quality and efficiency by flagging acute anomalies in real-time. Radiologists benefit from state-of-the-art deep learning technology that is "Always -on," running behind the scenes and freeing them to focus on the diagnosis itself.”

Surveys suggest a need for an always-on service to help radiologists cope with the unrealistic workload that many face daily. One study found that: “ Based on 255 uninterrupted eight-hour workdays per year, radiologists need to review one image every three to four seconds to meet workload demands.” The hectic pace likely contributes to misdiagnoses and loss of life. The diagnosis of lung cancer with imaging is one of the most challenging issues to contend with. It is estimated that misinterpreted chest X-rays are responsible for 90% of presumed errors in diagnosing pulmonary tumors.²  Mounting evidence suggests that ML-enhanced imaging data analysis may catch the disease at a much earlier stage, reduce hospital length of stay and health care costs, and save lives. For example, a prospective, randomized clinical trial that evaluated AIDOC-assisted CT scanning during the intracranial hemorrhage diagnosis found that algorithms gave clinicians an earlier heads-up.  Specifically, the researchers looked at 620 consecutive head CT scans. They collected the turnaround times (TAT) for positive ICH findings, i.e., how long it took to complete the CT scan to report the findings to clinicians who needed the results to make a treatment decision.³ Wismuller and Stockmaster compared TAT when CT results were flagged in radiologists’ worklists to CT results that were not flagged. When radiologists were told about the potentially dangerous findings early on, TAT was 73 minutes +/- 143 minutes, compared to 132 minutes +/- 193 minutes when they were left in the dark early on.

These prospective results were supported by retrospective analysis of a much larger data set. A study presented at the 2019 Society of Photo-Optical Instrumentation Engineers conference analyzed over 7,000 head CT scans from urban academic and trauma centers. Using convolutional neural networks, AIDOC generated a specificity of  99%, the sensitivity of 95%, and overall accuracy of 98% in diagnosing intracranial bleeds when compared to ground truth from expert neuroradiologists.⁴

Reports like this certainly don't imply that machine learning-enhanced algorithms will someday replace physicians. High-quality patient care will always require clinicians who are empathetic listeners. Nor do they suggest that AI will replace experienced radiologists. But they suggest that those who ignore digital medicine innovations will eventually be replaced by those willing to combine traditional approaches with emerging digital techniques that augment human decision-making.

 

References

1. Kissler MJ, Kissler K, Burden M. Toward  medical “ecology of attention.” New Engl J Med. 2021; 384: 299-301.

2. Del Ciello A, Franchi P, Contegiacomo A et al. Missed lung cancer: when, where, and why? Diagn. Interv Radiol. 2017;23:118-126.

3. Wismuller A, Stockmaster L. A Prospective Randomized Clinical Trial for Measuring Radiology Study Reporting Time on Artificial Intelligence-Based Detection of Intracranial Hemorrhage in Emergent Care Head CT. Presentation at SPIE Medical Imaging 2020 Conference, Houston, TX, February 15-20, 2020. https://arxiv.org/pdf/2002.12515.pdf

4. Ojeda P, Zawaideh M, Mossa-Basha M, et al. The utility of deep learning: evaluation of a convolutional neural network for detection of intracranial bleeds on non-contrast head computed tomography studies. Proceedings SPID Medical Imaging 2019. https://www.spiedigitallibrary.org/conference-proceedings-of-spie/10949/2513167/The-utility-of-deep-learning--evaluation-of-a-convolutional/10.1117/12.2513167.short

Diabetes Meets Machine Learning, Part 1

John Halamka, M.D., president, Mayo Clinic Platform, and Paul Cerrato, senior research analyst and communications specialist, Mayo Clinic Platform, wrote this article.

Of all the disorders that have responded well to artificial intelligence and machine learning, diabetes mellitus probably tops the list. The evidence supporting a role for Machine Learning (ML)-enhanced algorithms in managing the disease is persuasive and applies to several components of patient care, including screening, diagnosis, treatment, and prognosis.

Let's start with screening: As most clinicians know, there's a difference between screening for disease and diagnosing it. The former casts a much wider net, which means it will include more false positives. Once this larger cohort has been identified, more precise diagnostic testing can be performed to pinpoint patients who have the disorder.  As we explain in our next book, The Digital Reconstruction of Healthcare, screening tools for diabetes have been in existence for many years, including tools to help clinicians and patients identify the presence of prediabetes. Still, new technology has taken the screening process to a new level.

The American Diabetes Association (ADA) defines prediabetes as fasting plasma glucose between 100 and 126 mg/dl, a 2-hour oral glucose tolerance test reading between 140 and 200 mg/dl, or hemoglobin A1c of 5.7% to 6.5%.¹ The Association has developed a risk assessment tool to help clinicians and patients; it asks for a patient's age, gender, family history of diabetes, the presence of hypertension, whether the patient is physically active, their ethnicity, and height and weight. Although this kind of scoring system will detect many persons with prediabetes, it has its limitations and ignores many changes in a patient's medical history over time. An artificial intelligence (AI)-enabled assessment tool that takes advantage of gradient boosting, on the other hand, is capable of factoring in many more risk factors gleaned from a patient’s electronic health record. This type of assessment system can accurately predict which patients will progress to full-blown diabetes, with an area under the curve (AUC) of 0.865.² The gradient boosting approach has proven more accurate than logistic regression and evaluation of standard clinical cutoffs of blood glucose ≥ 110 mg/dl and HbA1c ≥ 6.0%. An analysis by Avivit Cahn from the Hebrew University of Jerusalem and her colleagues found the said ML algorithm incorporated 69 variables to achieve its superior AUC.

There is also reason to believe that AI-enhanced algorithms can help identify individuals before they become prediabetic. The Centers for Disease Control and Prevention (CDC) has a screening test to help predict prediabetes that asks one’s age, family history of diabetes, the existence of hypertension, race or ethnicity, whether one is physically active, gender, height and weight.

But logistic regression, artificial neural networks, random forest analysis, and gradient boosting have uncovered numerous other risk factors to predict prediabetes. Among the predictors that may help refine the risk stratification process: food security, citizenship, monocyte count, uric acid level, ALT, RBC count, and serum calcium and potassium.³ Whereas the CDC screening tool returned an area under the receiver operating characteristic (AUROC) of 64.4% in a large cohort from the National Health and Nutrition Survey, several AI-based algorithms generated AUROCs at or above 70%.

Several research projects suggest machine learning can improve the treatment of diabetes. In our book on mobile medicine, we discussed DreaMed,*  an Israeli vendor that developed DreaMed Diabetes to provide clinical decision assistance to clinicians managing Type 1 diabetes in patients who are using an insulin pump and continuous glucose monitoring (CGM). Its Advisor Pro algorithms collect several data types, including the patient’s daily blood glucose readings, meal carbohydrate data, insulin dosing records, and physical activity level. The software then uses event-driven machine learning and adaptive technology to create a more individualized regimen, adjusting its insulin dosing recommendations and offering behavior-modification recommendations accordingly. DreaMed Diabetes relies on a clinical AI program called MD-Logic. It uses fuzzy logic and adaptive learning algorithms to approximate how diabetes specialists reason as they adjust an individual's insulin regimen. The company claims that the software provides “faster analysis and deeper insights” than would normally be available with routine endocrinologist care. There is research support for this automated approach to Type 1 diabetes management.⁴

This coordinated interplay between an insulin pump and CGM sometimes referred to as closed-loop control or an artificial pancreas, has also been a focus of Mayo Clinic researchers. Yogish Kudva, MBBS, along with colleagues Sandeep Gupta and Ayan Bannerjee from Arizona State University, has been analyzing continuous glucose monitoring data from people with type 1 diabetes on the Medtronic MiniMed 670G closed-loop system* to develop software that would help each patient and clinic provider improve closed-loop control. They have also developed a system for measuring heart rate, respiratory rate, and self-report symptoms of viral infections in individuals with diabetes. The aim is to investigate the role of glycemic control on the progression of signs and symptoms of infectious disease in both at-home and hospital settings.

Similarly, Dr. Kudva participated in a recent investigation on the role of a closed-loop system to improve glycemic control in adolescents and young adults, part of the International Diabetes Closed-Loop Trial.⁵

Unfortunately, until recently, CGM has been out of reach for most Type 2 patients. But with the introduction of new, state-of-the-art technology, it is now available to them as well. We’ll explore these innovative solutions in part 2 of our series on machine learning-enhanced diabetes.

*Products and services mentioned are not endorsements.  Mayo Clinic has no relationship with any of these vendors.

 

References

1. American Diabetes Association. Classification and diagnosis of diabetes: standards of medical care in diabetes 2019. Diabetes Care. 2019;42(Suppl. 1):S13–S28. doi .org/10.2337/dc19-S002

2.  Cahn A, Shoshan A, Sagiv T el al. Prediction of progression from pre-diabetes to diabetes: development and validation of a machine learning model. Diabetes Metab Res Rev. 2019; e3252. doi.org/10.1002/dmrr.3252

3. De Silva K, Jonsson D, Demmer RT. A combined strategy of feature selection and machine learning to identify predictors of prediabetes. JAMIA. 2020;27: 396–406.

4. Nimri R, Muller I, Atlas E, et al. MD-logic overnight control for 6 weeks of home use in patients with type 1 diabetes: randomized crossover trial. Diabetes Care. 2014;37:3025_32.

5. Isganaitis E, Raghinaru D, Amber-Osborn L et al. Closed-Loop Insulin Therapy Improves Glycemic Control in Adolescents and Young Adults: Outcomes from the International Diabetes Closed-Loop Trial. Diabetes Technol. Ther. 2021l Jan 21. doi: 10.1089/dia.2020.0572.  Online ahead of print.

High-Quality Hospital Care — Minus the Hospital

John Halamka, M.D., president, Mayo Clinic Platform, and Paul Cerrato, senior research analyst and communications specialist, Mayo Clinic Platform, wrote this article.

In a time of COVID, the current care models just can’t remain in place. We need innovative ways to address the high cost of acute care. One solution that is taking a front seat is telemedicine.  Telecare has accelerated from 3-4% of visits in January 2020 to 90% in April to a new normal of 20% in 2021. Cultural change has modified patient expectations for the care they can get at a distance, which can be more convenient, less likely to result in COVID exposure, and more patient-centric.

Research has demonstrated that hospital-at-home programs for patients with specific acute medical conditions can reduce complications and reduce the cost of care by 30% or more.[1] One of the most progressive programs to focus on this transition was spearheaded by Johns Hopkins Hospital in 1994. Bruce Leff, MD, and his colleagues have tested this program with 455 elderly patients from three Medicare-managed systems and a VA medical center.[2] They found that the home model met the quality of care standards comparable to those expected of in-hospital programs. Also, "On an intention-to-treat basis, patients treated in hospital-at-home had a shorter length of stay (3.2 vs. 4.9 days) (P =0.004), and there was some evidence that they also had fewer complications. The mean cost was lower for hospital-at-home care than for acute hospital care ($5081 vs. $7480) (P < 0.001).”

A meta-analysis of 61 randomized clinical trials that looked at hospital-at-home projects found that among 42 trials, which included almost 7,000 patients, this approach reduced mortality (odds ratio, 0.81). Similarly, they decreased readmission rates by 25% (odds ratio, 0.75) and lowered costs. The same analysis revealed that treating every 50 patients in such a program saved one life.[3] Realizing the potential advantages of hospital-at-home programs, several large U.S. provider organizations have entered this space in the last few years, including Mayo Clinic, Partners Healthcare/Brigham and Women’s Hospital, and Mount Sinai Health System in New York. Across the globe, there are also significant programs in Australia, South Wales, and Spain.

Advanced Care at Home, a partnership between Mayo Clinic and Medically Home., tracks heart rate, blood pressure, pulse oximetry, temperature, and respiratory rate in its patient population, using Bluetooth-enabled devices wirelessly connected to the Mayo/Medically Home system. It also uses tablets, a back-up battery system and a Wi-Fi phone. There are, however, critical differences between many home-care programs and the Mayo Clinic system. Many hospital-at-home programs are targeted and designed for low-acuity hospital patients. They use physician house calls as the clinical delivery model. They have a short patient engagement period (2-4 days).

The Medically Home affiliated setup is designed to handle an extended length of stay that includes acute, post-acute and preventative care. It uses a scalable “decentralized” model for high-acuity care and can manage a broad set of diverse use cases and support an extensive patient census. The program uses screening, training, contracting, quality management, and technology and converts "post-acute" community-based supply providers into “acute-level” providers, bringing goods and services to high-acuity patients at home while focusing heavily on the role of paramedics as the centerpiece of its ability to provide “rapid-response” capabilities. In practical terms, that means paramedics and other providers go into the home while being virtually connected with a centralized medical command center staffed by physicians who guide the care for decentralized patients and the decentralized providers that care for them.

Advanced Care at Home has made measurable progress within a relatively short period, going from a speculative pilot project about a year ago to a business plan that will likely prove profitable in 2022. During a recent Zoom call with Ajani (AJ) Dunn, Administrator for the program, he emphasized, “It’s a story about volume.  As we looked at the model, we asked ourselves: Will it be effective? And we found that by scaling it up to the point where we can take out the fixed cost of a traditional hospital stay and replace that with the small variable costs of each service we deliver in the home, we can have a sustainable program.” Dunn explained that by working through a centralized command center staffed by physicians and getting buy-in from third-party payors, this approach is slowly turning the corner financially. Since most insurers do not have a hospital at home plan built into their policy, that has required the Mayo team to negotiate with payors one by one, explaining the cost-effectiveness of the new model.

The next step in the program’s growth will require finding ways to reinvent the existing system. “The off-the-shelf system we use to administer Advanced Care at Home is well-calibrated for traditional medical admissions, including COPD, CHF, pneumonia and the like. That approach involves drawing labs, titrating medications, etc. But we need other ways to use the system. Our plan now is to embed teams into other disease states, including oncology, cardiovascular disease, and transplantation, to understand the natural progression of each disease and the necessary interventions. Then we can create clinical protocols and administrative logistics that replicate in-hospital care in the home, but customized for each disease state.”

Undoubtedly, concerns about high costs and unexpected complications will continue to dissuade patients from seeking in-hospital care for many years. But as this new model expands, it's likely more patients will see the advantages of seeking high-quality hospital care — minus the hospital.

 

References

1. Klein S. “Hospital at home” programs improve outcomes, lower costs but face resistance from providers and payers. The Commonwealth Fund. Accessed January 15, 2020. https://www.commonwealthfund.org/publications/newsletter-article/hospital-home-programs-improve-outcomes-lower-costs-face-resistance

2. Leff B, Burton L, Mader SL, et al. Hospital at home: feasibility and outcomes of a program to provide hospital-level care at home for acutely ill older patients. Ann Intern Med. 2005;143:798–808.

3. Caplan GA, Sulaiman NS, Mangin DA, et al. A meta-analysis of “hospital in the home.” Med J Aust. 2012;197:512–519.