CALM (Close-loop Algorithm based Longevity Medicine): Improving Healthcare with Digital Health, AI and Longevity goal

?

CALM (Close-loop Algorithm based Longevity Medicine): Improving Healthcare with Digital Health, AI and Longevity goal

By

Ehud Baron MD D.Sc

25 Feb 2022

?

In an excellent article “Evidence and Efficacy in the Era of Digital Care” by Suhas Gondi, Brian W. Powers, and William H. Shrank, they bring up that Healthcare today does not benefit from the huge data with clinical value collected by digital health. As reasons for that they mention the lack of evidence-based ?data about Digital Health programs, and lack of incentives from pers and providers. In this article I would like to continue to develop these arguments and suggest ways how medicine can greatly benefit from wearables data. AI algorithms to process it and suggest treatment and setting as ultimate goal Healthy Longevity instead of treating each disease and symptom separately. For that, I would like to propose a new paradigm for future medicine:

CALM: Close-loop Algorithm based Longevity Medicine.

I believe that such an approach will lead to more cost efficient and lifesaving medicine, as well as improvement in clinical studies, regulatory approval and development of devices and medication, if accepted. Its success depends on its adaptation and contribution of doctors, researchers, providers, payors and of course the users. Currently it is more like a theoretical suggestion, but I try to do my modest contribution to make it happen and hope that others will join.

In a nutshell, it is summarized by the drawing on top: Medicine is about Continuous closing of the loop:

Monitoring – Diagnosis – Treatment – Monitoring….

I,e, instead of waiting until disorder is discovered and fix it, prevent it from happening by continuous monitoring, diagnosing any diversion when it is small and put things back on track by treatment while continuing to monitoring. I believe that now you wonder what is new about that as this is what medicine tries to do today. The main difference is the size of steps between each element in this loop. E CALM suggest very small steps done continuously.

In a way it is like driving a car on a winding road. You do not wait until you hit the safety rail besides the road or almost fall off the cliff, but continuously correct your position on the road. This approach which is highly intuitive can bring to big savings in healthcare cost, as cost of correction increases exponentially with size of diversion. It also leads to better safety and longer healthier life.

I believe that most of the readers will agree to this but wonder how practical it is to implement. I agree that until recently it was not practical, but today it becomes practical following the 3 revolutions below.

?

There are three major new revolutions in the health field that might call for a major change in medicine as practiced today:

1. Wearables - Proliferation of continuous sensing consumer devices like wearables, that enable 24/7 monitoring.

2. A.I. - The major role AI plays in medicine is its ability to integrate huge amount of data and convert it to diagnosis and actionable interventions.

3. Longevity - Cellular biology, genetic therapy, and the goal of healthy longevity as ultimate goal of healthcare.

While these three revolutions might sound like unrelated, I believe that it is the integration of these three that can lead to a more advanced medicine. Such medicine can fit better the health needs of everyone. It can be more cost efficient for the providers and payers and safer for the patients, as well as reducing burden for providers, doctors, and nurses. It can cut cost for pharmaceutical companies, shorter time for FDA approval, and change the way clinical studies are done.

This article shares my views how the integration of the three might pave the way to this new medicine framework that I called CALM.

1.???????Wearables - The introduction of wearables, hearables and Nearables provides medicine with an ability to access huge amount of invaluable physiological and behavioral data, outside of the hospital or Doctor’s office during daily life and in the natural environment of the user. Never before medicine had access to this kind of information. Almost all medical decisions about diagnosis, treatment, prognosis as well as clinical trials for new medications or treatment, are based today on spot measurements and not in the daily life environment. This revolution in data calls for applying the concept of Closed Loop borrowed from Control theory. It is the first element of the 3 that contributes to a a similar revolution in Medicine, to take advantage on this huge amount of invaluable data.

2.??????AI or Algorithm based – This is an integral part of such revolution as without the computational power and algorithms for processing this huge amount of data the first part of collecting real time data from millions of users cannot be realized. When I speak about AI I mainly refer to time series prediction algorithms that can integrate multi parameters change over time into diagnosis and predicting the appropriate treatment interventions.

3.??????Treatment with Healthy Longevity as a goal – Recent advances in cellular medicine, stem-cells regenerative medicine and genetic medicine brings us to the understanding that most chronic diseases like cancer, Alzheimer, Parkinson are in a way the results of changes in the cellular level influenced by Ageing. Also, that healthcare should look at the ultimate goal of healthy longevity instead of solving specific symptoms.

?

This article is my modest attempts to propose the principles for such new approach to medicine, that will extend the current capabilities to benefit from it. I suggest calling it CALM and hope it can inspire other doctors , medical researchers Payers and providers to elaborate on that and adopt it.

1.????What is CALM ?

?CALM is the abbreviation for: Close-loop Algorithm based Longevity Medicine

Close loop – Closed loop is a term borrowed from Control Theory. All our physiology is based on numerous closed loops that include sensors to sense the current state and ways to move from this state through a long chain of small changes into a desired state. In the CALM model, It means that patient is continuously monitored to progress in the desired direction and any deviation is corrected immediately until goal state is reached. The closed Loop includes 3 components: Continuous monitoring and sensing, A.I algorithms to convert real time data into clinical and actionable directions and treatment with the goal of maximizing healthy Longevity.

Medicine today is mostly Open Loop. I.e. based on spot measurement done in Doctor’s office or hospital, user report and previous medical records, a diagnosis is determined. Based on that, a treatment protocol is determined, and user is instructed to follow it. The doctor does not know what happens after the patients leaves the room. The doctor that set up the path for cure, does not know if the patient follows it, takes medication as instructed, follows the Lifestyle recommendations etc, also, the effect of the medication is unclear until next visit or patients complain. Even for inpatients that are not continuously monitored, there is no data between nurse/doctor’s visits of patient bed.

According to the Closed loop approach, the patient or user will be continuously monitored, and the monitored parameters will be sent in real time for A.I processing. After being automatically processed by algorithms that will detect any deviation from the desired path, a corrective action will be taken. If correction is not achieved, the doctor will be notified to correct diagnosis or treatment. A wearable like a SmartWatch for monitoring pulse rate, Oxygen saturation, Blood Pressure, temperature or CGM device for monitoring Glucose , nearable like sensor under the mattress or a video camera, hearable with sensors in the earphone, all these devices can help in acting as sensors for the Closed Loop.

Algorithm based – the huge amount of data collected by the sensors mentioned above cannot be processed without the help of computerized Algorithms, such algorithms, and especially the class called AI (Artificial Intelligence) are algorithms devised specifically to process huge amounts of real time data and extract clinical valued events. E.g., for a hypertensive patient that gets anti-hypertensive medication, the smart watch can detect excessive ?surge in Blood Pressure towards wake-up time, and alert the doctor to change medication schedule, dosage or even change in medication. It can help also to sees the medication kinetics and help in determining most suitable medication combination. This goal can be even better achieved by considering other Hemodynamics like Cardiac Output.

Longevity Medicine – Is medicine that is targeted toward achieving Healthy Longevity instead of curing specific diseases. Today medicine is mostly directed toward curing specific ailments. E.g., an elderly patient with higher Blood Pressure than normotensive range, might get anti-hypertensive treatment to lower blood pressure without noticing that this might increase risk of dementia and mortality. However, if the goal is healthy longevity, the goal state of the Closed Loop and Algorithms is to consider all factors that can reduce longevity, including reduced perfusion to the brain. To achieve a computational solution, we suggest computing “Biological Age” continuously so we can see if a specific intervention increases or decreases it.

2.????The benefits of CALM in development of new medications or medical devices:

The main contribution to medication development of the CALM approach is to move the emphasize from statistical results about the population to Close loop of a medication for a specific patient with the goal of longer healthy life. The clinical studies today are still based on the work of the Agricultural Statistician Sir Ronald Fisher invented 100 years ago. Ronald Fisher laid the foundation of significance tests, used today by the pharmaceutical companies and regulatory agencies like FDA to determine the effectiveness of a specific medication. These tests are based on spot measurement of “Before” and “After” a treatment compared to control group. The main problem with these tests, invented long before the wearables allowed continuous monitoring, is that they are more geared toward agriculture than medicine. In agriculture, one applies a certain fertilizer or irrigation amount to one field compared to the other field and measure the total harvested product. The farmer does not care which individual plants benefited less or more if the total results are satisfactory. In medicine we care about the effect of the medication on each specific individual over time. The current pharmaceutical companies and FDA regulation are trying to answer the question: Is the medication is statistically significantly better than no medication. I.e. is it helpful to most of the patients?

According to the CALM approach the question should be: Is it good for me ?

Instead of assuming that all people are the same and has a Gaussian distribution, the CALM approach will test by a Closed Loop, if the medication is helpful to me to improve my Longevity, regardless if it is helpful to others or not. In a way, the Cloose loop approach can be considered as a sequence of hundreds or thousands small experiments within a specific subject. While the current significance tests answer the question if a specific medication, under specific protocol benefits significantly more patient, the Close loop experiment answers how the medication benefit specific individual over time. It can turn, in principle, the clinical treatment of billions of people to continuous clinical trial, where knowledge is continuously accumulating.

Similar implications can be seen in development of medical devices. Today it is not uncommon that monitors that gained regulatory approval after short statistical test in the lab, fail when used in daily life in real life situation.

3.????Benefits of the CALM approach in the hospital

Today patients that are monitored are mainly in the OR or ICU, while most inpatients and outpatients are checked only from time to time by spot measurements. According to the CALM model, every patient admitted to the hospital will get a smartwatch, there will be sensors in his bed and video cameras that derive physiological parameters like Pulse Rate, Respiration rate, Oxygen saturation, temperature, blood pressure etc. continuously. Such sensors and AI systems will not increase cost but might even reduce it by detecting early signs of deterioration and proposing intervention that will save money. It will also reduce the burden on doctors and nurses as all physiological parameters are sent in real time to the hospital server, processed by AI algorithms and clinical events that needs attention will be sent automatically to the nurse station monitors.

?

4.????Benefits of the CALM approach in daily life

Today there is a distinction between Medical and wellness devices, but these distinctions start to blur as some consumer products like Apple watch get FDA clearance for detecting Atrial Fibrillation. Still, it is very rare that data collected with a consumer device will be used in medical diagnosis or treatment. According to the CALM model, wearables and similar consumer product will be part of the Closed Loop. Today many AI application are still considered as decision Making Support rather than part of the diagnosis and treatment process. According to the CALM model, Algorithm based systems will be part of the medical process. Today consumer get many general advice-like kinds of foods, supplements or exercise that are “good for health”. Consumers does not have any way to check if a certain supplement is really helping them towards healthy Longevity.

5.????Benefits for Life and Health Insurers and Hospitals – The Payors and Providers???

The main goal of the Payors is to attract more members and income and pay less. It looks like their interest, unlike those of the providers, is to reduce payment to providers so it looks like zero-sum game. In a CALM model, both can win: if diversions from health are minor and constantly corrected, there is big saving that both sides can share. If we plot the diversions from health as a graph along the timeline, where amplitude is the diversion, the smaller the RMS (Root Mean Square) or Quadratic Mean, which is the cost, is reduced when amplitude (diversions from health) decreases. E.g. By-Pass surgery cost more than keeping lifestyle that reduces the need for it. While everyone preaches for that, medicine today does not implement this continuous monitoring and loop, suggested by the CALM model.

6.????Benefits to Pharma companies and regulation –

?It looks again that there is a Zero-Sum game between pharmaceutical companies who want to charge more for their medications, and providers and Payors that want to pay less. However, if continuous monitoring can prevent adverse side effects that are built up over time, it might increase sales and reduce recalls. A famous example is the drug Vioxx, that was voluntarily recalled by Merck in 2004. The Vioxx recall occurred based on finding rom APPROVe trial that showed that patients that took more than 25mg of Viox has increased risk of CVD. This study did not show such increased risk for patients that took Vioxx less than 18 months. Vioxx was very profitable drug, and many patients would be happy to continue to use it if they could monitor themselves and get warning if they are among the very small percentage that can suffer from CVD or if there are some warnings, as CVD does not develop overnight.

This means that during the approval stage, a pharma developing new drug should not ask if it is good on average but if it is helpful for specific patients.

7.????Benefits to FDA, MDR,..

Today it happens many times that a device or a drug approved by FDA, turns out to not function as expected. E.g. it turned out that 70% of the Blood Pressure monitors approved by FDA, does not meet FDA requirements of accuracy in real life and after some period of usage, It will be of great advantage if devices and drugs are continuously monitored and FDA will get automatic notifications of poor performance.

Also, regulatory approval should be directed to specific physiological profile of patients that can change over time. This can make approval easier as it can be revoked over time if device or drug does not fulfill its expected promise. In a way, every treatment is a continuous clinical study.

8.????Benefits to Investors and Start-ups

A big advantage to Investors and entrepreneurs is that agreement and developing the CALM can help define the value of any invention to the healthcare field. Such evaluation can be done in the framework of CALM, and the complete loop of Monitoring-Diagnosis-treatment and its contribution to Healthy Longevity. Since Longevity is long term goal, we can use as surrogate the concept of biological age that can be monitored continuously. As example from what we are doing, a monitoring watch, as depicted in the figure above, derives its value from how helpful it is for diagnosis of certain Cardiovascular and respiratory disease and assessment of its treatment. A treatment that I am involved in, that helps in targeted delivery of genetic material and enhancing mitochondrial function, will be judge as part of closed loop to fight with ageing. I hope that readers can add many other examples from the work they are doing.

9.????Longevity monitoring

This article we assume Healthy Longevity as the ultimate goal of Healthcare. Healthy Longevity can be achieved by reducing diseases or direct attempts to slow aging by lifestyle and even cellular biology of rejuvenating cells (as attempted by Altos Labs, backed by Bezos, Calico of Google, and many start-ups today. Many other devices and medications can become pert of a CALM system, if they are combined with other interventions that can help to close this continuous loop and promote healthy longevity. This means that the value of a device, drug, etc. is derived from its contribution to such CALM closed loop of Monitoring-Diagnosis- Treatment-monitoring…. In the optimal way of minimizing diversion, it is important that all monitoring – diagnosis- treatment will be done continuously and in small steps as possible, E.g., Diagnosis changes based on the continuous diagnosis, treatment is made in small steps according to need.

I am aware that such changes will not happen overnight. If at all, but at least we can start thinking in this way. When we develop a device, or medication, or another treatment, we need to think how it fits to the CALM loop.

?

10. Summary

The CALM paradigm suggests major changes to most medicine practices today, based on new advantages conferred on it by new advancements in continuous monitoring by wearables, the computational powers of A.I. and the new advanced in Longevity monitoring. However, as Suhas, et. AL, suggested, it will come into effect only if the regulatory agencies, like FDA, MDR,.. will require evidence based digital health programs and the Payers and Providers will integrate it into their compensation schemes and clinical treatments respectively.

?

References:

1.??????Suhas Gondi, Brian W. Powers and William H. Shrank: Evidence and Efficacy in the Era of Digital Care. J Gen Intern Med, (Feb 2022)

2.??????Guo, C., Ashrafian, H., Ghafur, S. et al. Challenges for the evaluation of digital health solutions—A call for innovative evidence generation approaches. NPJ Digit Med. 3, 110 (2020)