The term ‘digital health,’ which encompasses a wealth of different technologies, including everything from ingestible sensors to wearables to phone apps and even robotic caregivers,(1) is poised to drastically change how we monitor and manage disease.
In this final instalment of our three-part series, we will discuss the role of the ‘Internet of Things’ (IoT) in the management and research of respiratory diseases with physician and artificial intelligence (A.I.) researcher, Dr. Brian Modena. We will also learn about his hopes for the future of digital health, and what we may come to expect from these technologies in the future.
Dr. Brian Modena (US) is a practicing allergist and researcher at the Allergy & Rheumatology Medical Clinic (La Jolla, CA). He is interested in the application of A.I. when applied to multiple data streams (i.e., IoT technology, genomics, natural language processing) for asthma phenotyping and the development of personalized asthma therapy regimens.
For the longest time, medical decisions have been based a lot on patient recall, whether about medication use or disease symptoms, typically given by the patient during one-off office visits under time pressures. This is rapidly changing with the advent of these new IoT devices that can now track things like a person’s activity, their sleep patterns, or medication uses. This is important for several reasons. People often have a tough time remembering when to take their daily medications.2 Adherence to medications may be improved with reminders, for example on smartphones, but, importantly, adherence to prescribed medications must be accurate and fully appreciated by care providers prior to changing or stopping medications. Alternatively, the medication use can also help to indicate disease flares and prevent downstream complications. For example, in the care of asthma, a sharp increase in ‘rescue’ medications to provide immediate relief of asthma symptoms may be the first warning sign that a patient is experiencing a severe asthma exacerbation. Early intervention and treatment of these patients may prevent dangerous complications or avoid ED visits. All this information is useful to care providers to manage patients and to make sound medical decisions.3
When you have asthma or other respiratory diseases, people know that when your condition worsens, you are less active, and it often wakes you up at night with difficulty breathing.4 Again, you can use IoT devices such as activity trackers for measuring things like sleeping patterns and activity levels. Smartphones can also track activity levels, and their apps may be used to allow for patients to report their symptoms or drug side effects.
We’re also seeing growing interest and investment from Pharma in ‘siteless’ research efforts. Currently, clinical research is extremely expensive, and estimates are that it requires over $1 billion to bring a drug to market.5 Siteless research offers the possibility to reduce costs, improve enrolment times, and enrol more diverse clinic populations into clinical trials by enrolling patients, and collecting clinical data, outside of major medical centres.6 In truth, we’ll likely never go completely ‘siteless’ because you need a ‘site’ for regulatory and legal purposes, but ‘hybrid’ research offers an invaluable way to reduce costs by reducing research coordinator and manager times, facility and overhead fees, and participant transportation costs.
“Digital health technology is only just touching the surface as far as their implementation.“
We are currently using inhaler trackers [to track medication adherence and rescue medication use] in both our clinical and genomic research. These devices have already proven to be immensely useful. We use activity trackers to measure exercise habits, steps-per-day, and sleep patterns as measurables for asthma control. We also use devices that measure lung function and exhaled nitric oxide at home. We use smartphones to administer standardized asthma control surveys. We find all of this very important and useful now in our clinical management of patients as well. We use artificial intelligence to help predict outcomes and identify asthma phenotypes that are more likely to be severe and develop disease complications.
However, I would also say that we’re just touching the surface as far as the adoption and implementation of digital health technologies, and some of that has to do with logistical barriers. For example, if I want to use an IoT device with some of my older patients, we may have to bring them into clinic, help them register the device, and walk them through how to use it properly. This requires time and money, and limits scalability. There may also be an issue of payment for the devices.7 The question remains as to who will pay for IoT health devices, e.g., an activity tracker (e.g., Fitbit) that may be used to track someone’s activity as a surrogate marker for asthma control. Some patients have privacy concerns.7 Physicians may worry that this data may be used by insurance companies to deny coverage to medications.8 If a person is not adherent to inhaler therapy, would this be used as reason to deny payment of the more expensive injectable medications? This is particularly concerning when the device manufacturers form partnerships with the insurance companies. Finally, often it’s simply hard to get people to want to do all of this when there is not always an immediate, observable benefit.
I think one is: who is going to pay for it all? Especially in the US, we need to establish how [and whether] Medicare and private insurance is going to pay for all the possible IoTs that may be used for chronic disease management. Who and how will it be decided on which devices will be reimbursed?
“These companies don’t necessarily interact with each other; there’s no ‘one stop shop’ where you can buy and sync all the IoT devices and software. This adds immense complexity, wastes time, and prevents scalability.“
Getting patients to use IoT devices over the long haul will always be a challenge.9 Some people do, but most people lose interest over time. This is even true for devices that we buy personally; there are likely millions of Fitbits and Apple watches left in drawers and on countertops every day. This is particularly true if the device or app requires a daily activity or work for them to do, such as charging. We are much better off if we can find technology that passively measures data while requiring no or minimal patient input or effort, which is the holy grail of mHealth.9Another major problem is that updates to Apple’s iOS or Google’s Android platforms may crash or cause dysfunction in IoT apps. It will, at times, throw everything off, and if you’re running a clinical trial, that’s a huge problem. For example, you can’t just stop collecting data during a clinical trial for a month while your app development team tries to figure out what changes have been made on Android, so that they can get the app up and running again.
Data security and patient privacy are major and increasing concerns. These are two different things: security and privacy. First, there is the issue of data security. For example, if I have an IoT device, the data may be secure on the device itself initially. But when the device sends the data to my phone, the phone must unwrap it, do stuff with it, and then send it to the cloud for storage. There may be multiple weak points in this system, particularly when the data is unwrapped, or when transmitted over the internet using current TCP/IP network protocols. When we are talking about important health data, this becomes even more critical to have secure and trusted data before making medical decisions.10 There are also retributions to reputation that accompanies data loss.11 There are financial consequences for violating HIPAA laws.12 Privacy, on the other hand, has to do with protecting an individual’s identity in the population. This presents a separate set of issues as well.
But it’s a constant battle with these hackers, to try and keep data secure, and then there’s all these different ways that our people are trying to secure it. I think the best way is to try to encrypt the data at the source, but it’s a very difficult situation, and a really complex question.
Early in the pandemic, there was obviously a large drive to be able to deliver medical care to people remotely, and so I do think that it’s going to be here to stay. It’s good in that regard, and I think that if you add it on some devices, it will only improve these remote interactions between physicians and patients.
Yet, I don’t think the medical care is as good when you’re doing telehealth visits alone for many conditions. Sure, if it’s something simple like treatment of cold sores, then telehealth visits are ideal. But for more complex disease, there is simply no substitution for in-person visits with patients, for several reasons. One reason is the physical exam, but other reasons include the development of patient-physician relationship, and by the gains made with interpreting social cues and signals. Communication is difficult enough to fully understand and appreciate with in-person interactions, and so then when you take it into a telehealth visit, it becomes even more difficult.
“Physicians take a lot of subtle cues from people and interactions during their patients’ physical exams, and they can get ‘Zoom fatigue’ too.”
It’s weird because that’s the reversal of what I’ve always thought telehealth would be like; I thought it would improve patient-physician interactions by allowing people to be more relaxed and comfortable in their own homes.
Number one is the judicious use of medications, especially in the avoidance of ‘over treating’ patients. The wholesale acquisition costs on the asthma monoclonal therapies medications are around $36,000 per patient per year, 13 and we shouldn’t prescribe them when they’re not needed. Sometimes patients get started on these medications with no plan of discontinuation and no end in sight, meaning that they’ll be on them for years. And so, by using IoT data on the individual patient’s condition, we may hopefully improve the decision-making process related to these therapies.
I think the other major role is clinical and translational research. For example, in genomic and genetic research, you need as good data as possible regarding a patient’s disease ‘phenotype.’ This means the biological and clinical characteristics of their disease, like disease severity and response to medications. If you can improve the data coming to you, we will greatly improve the accuracy and reliability of our genetic discoveries i.e., if you have genetic data, you can correlate it with better clinical data, and better determine the genetic causes of disease.
I think a lot of the concerns will work themselves out in the next five to 10 years, and regarding security, it will get better and better. I think the advantages are immense, and we can also think about the amount of money that everybody is likely going to save with clinical trials and research.6
“Siteless research helps to reduce costs and creates unlimited possibilities of enrolling the elderly, patients without means of transportation and patients in rural areas and communities.”
Additionally, people situated all over the world can be enrolled for a much more diverse population in clinical trials.6 This means you are not just dependent on middle and upper classes of people that are currently getting the best medical care at major medical centers.
Dr. Modena’s key takeaways
*Disclaimer: interview has been edited for clarity and brevity. All the interviews in this series reflect the views and opinions of the interviewees and do not reflect any opinions from any other parties, including Teva Pharmaceuticals.
RESP-42427 September 2021