CEO, Albus Health
For decades, respiratory medicine has measured what patients do during the day—spirometry in clinic, symptom diaries filled out imperfectly, exacerbations counted after the fact. What it has largely failed to measure is what happens to patients while they sleep: the eight hours when, for asthma, COPD, heart failure, and a growing list of chronic diseases, the physiological signals of worsening are most legible and least observed. Mikesh Udani, CEO of Albus Health, argues this is not a gap in ambition, it is a gap in technology.
Albus Health has built a contactless, radar-and-acoustic sensor system designed to passively monitor nocturnal symptoms over months and years without requiring patients to wear, swallow, or do anything. The device is currently deployed to collect nocturnal parameter data in clinical trials with top-ten pharma partners across respiratory and oncology indications. An FDA 510(k) submission is expected later this year. Udani sat down with Vanguard to explain why the night shift may be the most undervalued data window in drug development.
Moe: Why did you believe the eight hours patients spend asleep held more diagnostic signal than anything clinicians were already measuring?
Mikesh Udani: Before we started Albus, before we decided which clinical signals to measure, we did quite a bit of primary and secondary research—and the evidence was compelling. For asthma and COPD, clinicians have known for decades that nighttime symptoms—increased cough, respiratory rate, heart rate—are known markers of worsening. The more severe the disease, the more prevalent these symptoms are. They predict exacerbations, and they measure recovery and response to treatment. So what we figured was that we had a real opportunity to accurately and robustly measure symptom burden, and the changes related to clinical events such as early worsening or treatment outcomes, if we could accurately measure nighttime symptoms.
That started with asthma and COPD, but as we got deeper, we learned that nighttime changes and sleep are critical in most chronic diseases—heart failure, rare diseases, most neurological conditions, even in lung cancer and breast cancer, where drug-induced toxicities show up in respiratory parameters, some of them at night.
The second point was that nighttime is eight hours—a third of an individual’s life—and what is unique about this specific period is that it’s quite standardized and ideal for long-term monitoring. People are in the same place, and there are no confounding sources of noise that cloud the diagnostic signal. We also learned that any remote monitoring technology asking patients to do something or wear something has adherence challenges. The moment you rely on patients to do something, there is a high risk you will stop getting data for two to three weeks. So to measure reliably over long periods of time, the technology needs to be passive—and for passive monitoring, nighttime is ideal. It was these two factors which were really quite compelling, and we’re now seeing it in the data.
Moe: Why was contactless radar-based sensing the right architecture when bed sensors and camera-based systems were already in the field?
Mikesh Udani: We tested many products on the market and spoke to a lot of people before we built this—patients, their families, clinicians, payers, even regulators. What we learned was that for a technology to work for all of these stakeholders over the long term, it must meet a few requirements.
First, it should measure a range of symptoms. You can have the world’s best heart rate monitor, but heart rate alone doesn’t give you enough information across a range of diseases to take clinical decisions. Second, it must show clinical accuracy in a real-world setting. Clinical validation against gold standards is very important, and it must happen in real-world conditions where patients spend most of their lives. Third, it shouldn’t rely on patients to do or wear something, or to have certain digital or physical infrastructure at home—the moment it has those requirements, it reduces the number of patients who can use it. And finally, privacy must be at its core. It cannot be an afterthought. It has to be designed in.
With camera-based systems, privacy is a significant concern. Once patients are not comfortable with that, it’s very hard to convince them to use it over time. With under-the-bed sensors—technologies using ballistocardiography—they do a reasonably good job of measuring heart rate, and some have validation against gold standards, but only in ideal conditions. For example, they’re not very reliable with different bed or mattress types; if there are two people in the bed, or a pet, there is little to no validation. And the range of symptoms they can measure is quite limited. In asthma and COPD, nighttime cough becomes very important, which bed sensors or cameras don’t measure. It’s not feasible to ask patients to put something under their bed, wear something, and have a separate cough monitor on the side. Device fatigue happens in clinical trials and in clinical care. We felt there was a gap—no suitable technology that met clinicians’ requirements and was acceptable to patients for the long term. That’s why we went into building something new.
“Any technology asking patients to do something or wear something has adherence challenges. The moment you rely on them to do something every day, there is a risk you will stop getting data two to three weeks in.”
Moe: Why can’t a smartphone app do what your device does?
Mikesh Udani: That’s an easy answer. To start with, phones don’t have the sensors needed for accurate multi-metric monitoring in real-world settings. There are apps that try to take data from other devices, such as pulse oximeters, and put a diagnostic layer on top, but a phone sitting next to your pillow on its own just doesn’t have enough clinical signal. This is evident from the lack of clinical validation of smartphone apps against gold standards. Then from a patient perspective, we need to remember that there are subpopulations such as elderly and children who may not be using smartphones. Even the ones who do would need to remember to turn on an app, keep the phone charged, install it correctly on the bed before sleeping, etc.—and as we know, all of these add to the risk of missing or noisy data. Finally, there are valid data security and privacy concerns. In our experience, in a regulated environment—for pharmaceutical companies or for healthcare practitioners—unless it’s patient-reported outcomes, stakeholders have not made the shift to bring-your-own-device or phone apps for remote monitoring despite the technology existing for a few decades now.
It is an absolute requirement that there is clinical validation before a clinician, a regulator, or a pharma company makes a claim or a decision based on the data. The onus is on the technologies to show supporting evidence. From what we’ve seen, the combination of radar and privacy-preserving acoustic sensors has a much, much higher clinical signal, reliability, and correlation with clinical gold standards than what we see with apps and smartphones.
Moe: How do nocturnal patterns shift in the days before an asthma attack, and what does that signal ask a clinician to do differently?
Mikesh Udani: What we’ve shown in our preliminary data is that cough and respiratory rate start to increase up to 5 days before an attack, and these symptoms reach their peak during the attack. When patients are given the right medication, the symptoms come down to baseline. We do need to do further validation to assess the precision and sensitivity—that’s an important next step we’re currently working on. What we can say right now is that these signals are clinically valid, meaning they help us distinguish between disease population and healthy population. They’re repeatable—if you measure patients in the same healthy state, the measurement of symptoms will be the same. And they’re responsive: when patients start to get worse, these symptoms get worse, and when they get better, these symptoms get better.
Once this goes through further validation to establish sensitivity and precision thresholds, the first line of action would be to enable patients and their families to self-manage—maybe double their inhaler, double-check inhaler technique, be aware of triggers. The second line of action, if those measures don’t resolve the symptoms, could be to take preemptive clinical input without waiting for it to become an emergency. It makes a big difference between going to a hospital in a state of emergency in the middle of the night versus going a couple of days before and saying, “I’m getting quite worse—if things don’t improve, I might have a major asthma attack.” Outcomes are significantly better if you don’t wait until the last moment. Late recognition of severe asthma attacks can be life-threatening, and every attack leads to irreversible damage to the patient’s lungs.
What’s also important is that some of these interventions are already prescribed to patients. Patients are already given a written asthma action plan—if their symptoms worsen, they’re told what to do. The missing link right now is that patients don’t perceive these early warnings until it’s too late. If we plug in that gap, there’s a lot of value for every stakeholder.
Moe: How are pharma companies using the device today, and what is the realistic pathway toward a primary endpoint?
Mikesh Udani: Large pharma are already using it in their clinical trials globally. We started with respiratory and have now entered oncology. Rare diseases, neurology, and cardiovascular are next. So far it’s been used to measure quality of life, recovery, and response to treatment, and as a tool to trigger clinician assessment in case of symptom flare-ups. We’re now getting into studies where the goal is to progress to label claims and safety measures.
The choice of primary endpoints in drug trials, as you know, is indication-specific. Albus’ first deployment to support a primary endpoint will be in indications where the sleep-related parameters are already established primary endpoints, and Albus will enable their accurate monitoring at home over longer periods. In fact, given how fundamental sleep is to quality of life over long-term health outcomes, it is surprising that it is not a required endpoint for most chronic indications.
Next will be indications where existing endpoints are unsuitable—they exclude large populations, miss out on meaningful aspects of health, or have over-dependence on patient-reported outcome measures that are very hard to collect reliably over long periods. For example, in respiratory diseases such as asthma and COPD, FEV1 and exacerbations have been used as primary endpoints for decades. The traditional definition of an “exacerbation” relied on subjective reporting which is known to be unreliable, underreported, and variable—patients, families, and clinicians all know this is a real problem. They also fail to represent the disease burden in populations with mild and moderate disease, who have a high symptom burden that lowers their quality of life but does not reflect in these traditional endpoints. This is preventing drugs from reaching these populations. There is a lot of interest from pharma companies in objective measurements of physiology and capturing meaningful aspects of health.
The avenue we think will be most effective is incorporating nocturnal monitoring as part of the drug program from Phase I and Phase II. This data can be taken to the end-of-Phase-II meetings with the regulators to get their feedback on primary and secondary endpoint use in Phase III trials. A couple of our pharma partners are already adopting that pathway—that’s where we’re seeing a lot of interest. We’re also proceeding with FDA clearance of the technology to further support this.
“The missing link right now is that patients don’t perceive these early warnings until it’s too late. If we plug in that gap, there’s a lot of value for every stakeholder.”
Moe: Why should a family with an asthmatic child care about this technology right now, given how much validation work remains?
Mikesh Udani: It’s interesting you ask that. Every patient and family member we have interacted with in our studies or validation work have been overwhelmingly enthusiastic about Albus Home. Some of them ask if they can keep the device! It’s the same response in external pharma-sponsored trials too. A top-10 global pharma did an independent qualitative study to evaluate this. The patient group were told about a hypothetical study in which Albus was an optional component, and then asked if they would choose to participate in the study with or without it. 100% of the participants said they would use the device in a study, even if it was optional. Their key motivation to do this was to support research and bring life-changing drugs to other patients faster. In fact, 30% of participants, completely unprompted, said that the use of Albus Home—which they saw as novel and innovative—cast the study sponsor in a positive light. They felt that the sponsor was using cutting-edge technology to get information that could improve people’s lives.
On the validation part, the technology is already validated to show high accuracy in monitoring clinically meaningful signals in real-world environments. This is an enormous win for patients, as we can now robustly measure their symptom burden and changes related to clinical events such as worsening and response to treatment, without burdening them. Children and elderly will be among the biggest beneficiaries. We finally have technology that works for the patients, rather than asking patients to work for the technology. This is improving patient experience in trials, and will expand drug development efforts to those patients who have historically remained unheard and whose symptoms have so far remained unnoticed. The objective data collected in trials will finally give pharma companies the evidence they need to convince regulators and payors to bring drugs to all patients who need them.
Of course, we’re only just scratching the surface of what this technology can do for patients, and we have no plans to stop here. While the adoption in pharma clinical trials grows, we’re furthering our R&D to support the adoption into routine clinical care. We’re building on the preliminary data we published earlier, progressing to the next stages of validation, and particularly in asthma, we plan to bring this to children as young as 1 year old. There’s a lot of exciting work happening, so stay tuned.
Mikesh Udani is CEO of Albus Health, a UK-based digital health company developing passive nocturnal monitoring technology for use in clinical trials across respiratory, oncology, and other chronic disease indications.
Moe Alsumidaie, MBA, MSF, is founder and Chief Editor of Vanguard Publications, which publishes Clinical Trial Vanguard, Pharma Vanguard and BullScope, and Head of Research at CliniBiz. He has two decades in clinical trial operations and data science, with earlier roles at Genentech, Abbott Vascular and Stanford University Medical Center, and is a guest lecturer in clinical trial sciences at Rutgers University.
