<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Benjamin Vandendriessche on Medium]]></title>
        <description><![CDATA[Stories by Benjamin Vandendriessche on Medium]]></description>
        <link>https://medium.com/@benvddriessche?source=rss-bb22181f7c06------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*nrO8oCxtEofME5Pm-uPKGw.jpeg</url>
            <title>Stories by Benjamin Vandendriessche on Medium</title>
            <link>https://medium.com/@benvddriessche?source=rss-bb22181f7c06------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 12:03:23 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@benvddriessche/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Cardiotoxicity in Cancer Treatment]]></title>
            <link>https://medium.com/byteflies/cardiotoxicity-in-cancer-treatment-64a707e71b0e?source=rss-bb22181f7c06------2</link>
            <guid isPermaLink="false">https://medium.com/p/64a707e71b0e</guid>
            <category><![CDATA[digital-health]]></category>
            <category><![CDATA[cardiotoxicity]]></category>
            <category><![CDATA[wearables]]></category>
            <category><![CDATA[remote-patient-monitoring]]></category>
            <category><![CDATA[oncology]]></category>
            <dc:creator><![CDATA[Benjamin Vandendriessche]]></dc:creator>
            <pubDate>Sun, 19 Mar 2023 23:47:04 GMT</pubDate>
            <atom:updated>2023-03-19T23:47:04.706Z</atom:updated>
            <content:encoded><![CDATA[<h4>Can remote patient monitoring help diagnose and prevent cardiotoxicity resulting from cancer treatment?</h4><p><em>Arno Bossaert and Benjamin Vandendriessche authored this article. It is a quick primer on cardiotoxicity in cancer treatment programs, and how remote cardiac monitoring solutions could improve the standard of care and drug development in oncology. It is in no way intended to provide an exhaustive overview of the field.</em></p><p>Remarkable progress has been made in the diagnosis and management of many types of cancer. Nowadays, patients have access to advanced diagnostic tools and increasingly more targeted treatment options, resulting in improved disability-free life expectancy. Consequently, we are also starting to understand the importance of managing cardiovascular damage that can be caused by many established and novel cancer treatments. In fact, this has led to the new subspeciality of cardio-oncology.</p><h3>What is Cardiotoxicity?</h3><p><a href="https://www.hearthousenj.com/learning-center/diet-nutrition/what-is-cardio-oncology-cardiotoxicity/">Cardiotoxicity</a> is cardiovascular damage caused by certain cancer treatments. It can develop days, weeks, or even years after treatment (<a href="https://www.ecrjournal.com/articles/cardio-oncology-focus-cardiotoxicity">ref</a>). Several cardiac problems such as cardiac ischemia (reduced blood flow to the heart leading to heart muscle damage), heart failure, arrhythmias (abnormalities in electrical conduction in the heart), pericarditis (inflammation around the heart), coronary artery disease (reduced blood flow in the vessels that supply the heart muscle with blood), and embolisms (blood cloths) are all potential manifestations or causes of cardiotoxicity. In many cases, cardiotoxicity presents as cardiomyopathy, a catch-all term for conditions that make it harder for your heart to pump blood. Certain types of cancer drugs and treatments are associated with a higher cardiotoxicity risk. Some important but non-exhaustive examples are (<a href="https://my.clevelandclinic.org/health/diseases/16858-chemotherapy--the-heart-cardiotoxicity">ref</a>, <a href="https://www.sciencedirect.com/science/article/pii/S2666087322001120?via%3Dihub">ref</a>, <a href="http://www.cancerresearchuk.org/about-cancer/treatment/immunotherapy/types/checkpoint-inhibitors">ref</a>):</p><ul><li><strong>Anthracyclines</strong>: Chemotherapy that is often used to treat leukemia, lymphoma, breast cancer, sarcoma, or multiple myeloma.</li><li><strong>Trastuzumab</strong>: Targeted therapy drug commonly used to treat breast, stomach, and esophageal cancer.</li><li><strong>Radiation therapy</strong>, especially when targeted at the chest: Often used to treat breast cancer or certain leukemias.</li><li><strong>Fluoropyrimidines</strong>: Often used in the treatment of colorectal, breast, and aerodigestive tract cancers.</li><li><strong>Immune Checkpoint Inhibitors</strong> (ICIs): Immunotherapy used for melanoma skin cancer and lung cancer.</li></ul><h3>Telemonitoring Tools to Reduce Cardiotoxicity Risks</h3><p>Cardiotoxicity risk is a dynamic variable, and the risk changes throughout the care pathway, depending on factors such as age, cancer history, pre-existing cardiovascular disease, and cardiotoxic events. The risk of cardiotoxicity is important to understand and balance against the benefit of the cancer treatment, before, during and after treatment. The risk can be reduced by several mitigation strategies, including (<a href="https://academic.oup.com/eurheartj/article/43/41/4229/6673995">ref</a>):</p><ul><li><strong>Primary prevention</strong>: Cardioprotective treatment strategies focused on preventing cardiovascular damage from occurring, evaluated on a patient-by-patient basis.</li><li><strong>Secondary prevention</strong>: Similar to primary prevention but focused on those with a history of cardiovascular disease and/or prior cardiotoxic events.</li><li><strong>Cardiovascular surveillance</strong>: Careful and frequent clinical evaluation during cancer treatment to detect early signs and symptoms.</li><li><strong>Early cardioprotection</strong>: If surveillance identifies (sub)clinical cardiotoxicity, cardioprotective treatment should be started promptly.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*BpstVHdCClYenXo3" /><figcaption>Figure adapted from the <a href="https://academic.oup.com/eurheartj/article/43/41/4229/6673995">2022 ESC guidelines on cardio-oncology</a>. It depicts how: 1) at baseline (pre-cancer treatment), primary and secondary prevention strategies can reduce the occurrence or effect of cardiotoxicity (BLUE arrow); and 2) during treatment early cardioprotection should be initiated as soon as possible (GREEN arrow) to bring the increased risk for severe cardiotoxicity back down (DASHED line), based on results from surveillance activities. It also shows how long-term surveillance after completion of cancer treatment is recommended to minimize chronic cardiovascular events (ORANGE arrow).</figcaption></figure><p>Patient surveillance with remote patient monitoring (telemonitoring) tools is a potentially powerful method to keep a close eye on patients in these care trajectories without generating much additional hassle such as repeated physician visits.</p><h3>The Underutilized Role of ECG Monitoring in Cardiotoxicity</h3><p>Several cardiology and oncology professional societies have emphasized the effectiveness of an electrocardiogram (ECG) and other non-invasive monitoring solutions in managing cardiotoxicity. The European Society of Cardiology (ESC) suggested that all cancer patients should undergo ECG monitoring before treatment to identify cardiotoxicity risk factors. Nowadays, many user-friendly ECG monitoring tools are available that can remotely follow-up patients for longer periods of time, typically multiple weeks.</p><p>The relevance and type of ECG monitoring depends on the specific cancer treatment. For example, in Immune Checkpoint Inhibitors (ICIs), remote ECG monitoring is a recommended strategy (<a href="https://academic.oup.com/eurheartj/article/43/41/4229/6673995">ref</a>, <a href="https://www.frontiersin.org/articles/10.3389/fcvm.2022.927769/full">ref</a>). Myocarditis is a rare (up to <a href="https://pubmed.ncbi.nlm.nih.gov/31517036/">1.14%</a> or <a href="https://pubmed.ncbi.nlm.nih.gov/35267453/">2.1%</a>, dependent on the source) ICI adverse event but one with extremely high mortality (up to <a href="https://pubmed.ncbi.nlm.nih.gov/31517036/">46%</a> or <a href="https://pubmed.ncbi.nlm.nih.gov/35267453/">50%</a>), making continuous surveillance to identify early warning signs crucial.</p><p>As increasingly advanced cancer therapies are developed, more patients survive. Because of this phenomenal evolution we are now starting to understand acute and especially long-term side effects of these treatments in much more detail, including cardiotoxicity, which is a rapidly evolving field. Fit-for-purpose remote monitoring strategies can improve the standard of care for patients and inform clinicians, researchers, and drug developers about the cardiotoxic effects of new therapies.</p><h3>Byteflies Remote Patient Monitoring</h3><p>Byteflies has extensive expertise in developing and deploying remote patient monitoring solutions. Our wearable and medically certified Sensor Dot technology has been used in clinical trials, real-world data studies, and in state-of-the-art telemonitoring applications in clinical practice.</p><p>As an example of the latter, Byteflies <a href="https://byteflies.com/cardiocarehome">CardioCare@Home</a> provides a comfortable and integrated cardiac patch solution for continuous ECG monitoring at home up to several weeks, including advanced arrhythmia detection analytics. OncoCare@Home is developing continuous remote monitoring features that are relevant during and after cancer treatment, including for new classes of immunotherapies that can cause cytokine release syndrome (CRS) and neurotoxicity. At the intersection of these two diagnostic fields, we can assist innovators, clinical researchers, and physicians with innovative solutions to tackle new challenges in cardio-oncology!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ZbzdxJIG3mFIGgjx" /></figure><p>Byteflies is working towards a future where healthcare is fully personalized and preventive. Our diagnostic solution allows high quality, long-term and synchronized measurement of vital parameters and digital biomarkers with user-friendly patches and powerful data analytics, independent of where you are and optimized for your specific needs.</p><h3>Interested to Learn More?</h3><p>At Byteflies, we move mountains to enable wearable health. With <a href="https://byteflies.com/epicarehome">EpiCare@Home</a>, we are monitoring seizures to improve the standard of care for people with epilepsy. With <a href="https://byteflies.com/cardiocarehome">CardioCare@Home</a>, cryptogenic stroke patients are screened for cardiac risk factors much faster and with a user-friendly tool. Our technology allows pharmaceutical companies to monitor patients easily and objectively in clinical trials.</p><p>Interested in working with us to reduce cardiotoxicity due to cancer treatment? Curious to know what we can offer for your patients, project, or clinical trial? Contact us at <a href="/service/mailto:hello@byteflies.com">hello@byteflies.com</a> or via <a href="http://www.byteflies.com/">www.byteflies.com</a>. We look forward to hearing from you!</p><p><em>The information contained in this article represents the views and opinions of the writer(s) and does not necessarily represent the views or opinions of other parties referenced or mentioned therein.</em></p><p><em>The article is not intended to be a substitute for professional medical advice, diagnosis, or treatment. Always seek the advice of your physician or other qualified healthcare professional with any questions you may have regarding a medical condition. Never disregard professional medical advice or delay in seeking it because of something you read in this article.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=64a707e71b0e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/byteflies/cardiotoxicity-in-cancer-treatment-64a707e71b0e">Cardiotoxicity in Cancer Treatment</a> was originally published in <a href="https://medium.com/byteflies">Byteflies</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What has Byteflies built for people with epilepsy (so far)?]]></title>
            <link>https://medium.com/byteflies/what-has-byteflies-built-for-people-with-epilepsy-so-far-48bb7189d13a?source=rss-bb22181f7c06------2</link>
            <guid isPermaLink="false">https://medium.com/p/48bb7189d13a</guid>
            <category><![CDATA[virtual-care]]></category>
            <category><![CDATA[seizures]]></category>
            <category><![CDATA[epilepsy]]></category>
            <category><![CDATA[digital-medicine]]></category>
            <category><![CDATA[remote-patient-monitoring]]></category>
            <dc:creator><![CDATA[Benjamin Vandendriessche]]></dc:creator>
            <pubDate>Wed, 15 Mar 2023 23:10:11 GMT</pubDate>
            <atom:updated>2023-03-23T12:32:13.517Z</atom:updated>
            <content:encoded><![CDATA[<h3>What has Byteflies Built for People with Epilepsy (so far)?</h3><h4>EpiCare@Home — An end-to-end seizure monitoring service</h4><p>Back in 2021, we <a href="https://medium.com/byteflies/what-is-byteflies-building-for-people-with-epilepsy-and-why-9b3c55f56dfc">wrote a post</a> with the subtitle “A personalizable seizure monitoring wearable for home use that extends the capabilities of specialized hospital equipment”. That post focused on the “why”, meaning we defined several unmet clinical needs in epilepsy and discussed innovations that we considered essential to meet those needs.</p><p>We identified a need for improvements in:</p><ol><li>The delivery of <strong>state-of-the-art care</strong> for people with epilepsy.</li><li>The development of <strong>new treatment strategies</strong>, especially for treatment resistant epilepsy.</li><li><strong>Reducing</strong> the <strong>unpredictability </strong>of seizures.</li><li><strong>Managing stress</strong>, treatment <strong>side</strong> <strong>effects</strong>, and epilepsy <strong>comorbidities</strong>.</li></ol><p>Since then, we have stayed passionately committed to the mission of tackling them on two fronts:</p><ol><li>Leveraging the suite of Byteflies remote patient monitoring (RPM) tools to build a fit-for-purpose seizure monitoring service for use in clinical care and drug development.</li><li>Identifying like-minded people and organizations. RPM and virtual care solutions in neurology are still very much a “zero to one” story despite some important innovations, such as neurostimulation and -modulation devices. In contrast, fields like cardiology have a well-established history of using innovative technology to monitor patients (ultra)long-term. In neurology and epilepsy, it will take a village as the -somewhat tired- cliché goes. Consequently, we also focus on identifying the inhabitants of that village.</li></ol><p>For a more in-depth discussion on the unmet needs in relation to the current standard of care, have a look at the <a href="https://medium.com/byteflies/what-is-byteflies-building-for-people-with-epilepsy-and-why-9b3c55f56dfc">original post</a>.</p><p>This post will focus on what we have built since then, and how it is being used in practice to help people with epilepsy and their families right now.</p><h4>Table of Contents</h4><ul><li><a href="#f543">Crucial technological enablers</a></li><li><a href="#57e1">Which patients and seizure types?</a></li><li><a href="#f665">Clinical acceptance and reimbursement</a></li><li><a href="#6b13">The village</a></li></ul><h3>Crucial Technological Enablers</h3><p>Byteflies Care@Home products are bundles of hardware and software medical devices, combined with data annotation and logistical services, which are purposefully put together to cover a specific disease indication or care path. The figure presents that same idea more schematically:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qTjP3F0uafzFPKyRjWrZtA.png" /><figcaption>A Byteflies Care@Home product is prescribed, after which any hardware components are shipped to the patient or medical practice. During the monitoring period, the user follows a specific monitoring protocol and data is uploaded daily to be annotated and converted into a report for the healthcare professional.</figcaption></figure><p>EpiCare@Home combines medical wearable devices that acquire a low-montage electroencephalogram (EEG), which can be further augmented with cardiorespiratory and motion data. We have covered this set up and how it compares to a traditional in-hospital video-EEG in a <a href="https://medium.com/byteflies/what-is-byteflies-building-for-people-with-epilepsy-and-why-9b3c55f56dfc">previous post</a>.</p><p>Since then, we added two important technological enablers that make EpiCare@Home fit-for-purpose in its intended dual use environment, namely the patient’s home and in clinical neurology workflows:</p><ol><li>User-friendly adhesive electrode “<strong>patches</strong>”; and</li><li><strong>Data annotations services</strong></li></ol><h4>Electrode Patches</h4><p>Because scalp EEG is a finicky signal to record and prone to technical measurement artifacts, high quality “cup” electrodes are typically used in the hospital or for ambulatory EEG recordings. These electrodes are glued to the scalp and filled with a conductive paste to generate a low resistance path from the scalp to the recording device. Although EpiCare@Home can use these types of electrodes, they come with a major usability downside. No one likes to have electrodes glued to their scalp for long unless absolutely necessary and they are almost impossible to reapply without help from someone trained to do so.</p><p>Therefore, we set out to develop an entirely new EEG “patch”, which we call “<strong>EEG Adhesives</strong>”, with embedded electrodes that use a discreet “behind-the-ear” design so that most of the components can be hidden out of sight, to minimize any visual stigma.</p><p>Back in 2019, Byteflies together with several academic and industrial partners were awarded a research grant (<a href="https://researchportal.be/en/project/plug-n-patch">Plug ‘n Patch</a>) to explore the design, production, and validation of various patch designs with embedded biopotential electrodes.</p><p>When developing these EEG Adhesives, or any type of medical skin adhesive for that matter, it is important to understand how long users will be expected to wear them. Based on conversations with epileptologists, we identified multi-day up to multi-week continuous EEG monitoring as a good trade-off between patient comfort and diagnostic yield (see “<a href="#57e1">Which Patients and Seizure Types</a>” for more details).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/0*Pg3o8dYxEgc1TNFI.png" /></figure><p>In early 2020, that led to the first small scale prototype production of several designs, some of which are pictured below. Volunteers and academic researchers extensively tested each design to assess usability with a particular focus on wear comfort, and -of course- signal quality.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/386/1*xLCBr7IYMjz9ftrusvI95Q.jpeg" /></figure><p>Specifically, we needed the find a balance for the following <strong>quality attributes</strong>:</p><ul><li>Biocompatibility, and in particular breathability that allows a monitoring period of up to 4 weeks, with no or as little skin irritation as possible.</li><li>Inter-electrode distance as it relates to the ability to record a meaningful EEG signal and compatibility with a broad range of ear shapes.</li><li>Electrode technology: dry, solid gel, hydrogel as a function of signal quality and biocompatibility.</li><li>Adhesion: strong enough to provide good EEG signal quality for 2 days while light enough to not impede easy removal without glue residue.</li></ul><p>Although they look deceptively simple, each EEG Adhesive is actually a stack of ten components that must seamlessly come together, and our manufacturing partners need to be able to produce them at scale.</p><p>Based on the collected feedback, we went through eight (!) more production iterations to identify the optimal balance of user comfort and signal quality (hardware is hard … anyone?!). Each iteration went through extensive user testing and signal quality assessments, compared to reference cup electrodes under activities of daily living (sleeping, walking, working, etc.), and had to pass strict medical device requirements.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9rJ0NqIJPA_xLsI0Qzsgkw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*li6c9IGMEFVc0wLT8fQwKw.png" /><figcaption>Example of signal quality assessment test: (left) EEG signal from Byteflies EEG Adhesive; (right) EEG signal from cup electrodes on the same subject. The top graph depicts the EEG time domain, the middle graph the EEG frequency domain (for the alpha, beta, delta, and theta bands), and the bottom graph the accelerometer signal. EEG signal experts reviewed the results of these tests and made recommendations to the hardware developers.</figcaption></figure><p>Although we’ve had great success with dry electrode technology for recording cardiac signals, for the much lower amplitude EEG signal, this turned out to be too complex. Therefore, one of the design decisions made was to use hydrogel electrodes. This common type of electrode has a skin-friendly gel that permeates the surface layer of the skin, thereby lowering resistance, similar to the function of the conductive paste in the cup electrodes.</p><p>In late 2022, we settled on the following hydrogel design:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/375/1*P8ISFNStSzMQc_7EJ2iqXQ.png" /></figure><p>Of course, technology never stops, and we are actively working on next-gen EEG Adhesives that further improve usability by doing away with the wires and snap connectors. That said, our current generation of EEG Adhesives works very well, and we are beginning to roll these out to our users.</p><p>We are grateful to all partners in the <a href="https://www.demorgen.be/nieuws/minister-crevits-pompt-4-miljoen-in-pleister-die-mensen-met-epilepsie-kan-helpen~b31f0ef3">Plug ‘n Patch consortium</a> who tirelessly worked to make this possible. In fact, one of our <a href="https://pubmed.ncbi.nlm.nih.gov/36681896/">research partners</a> attempted to use investigational prototype EEG Adhesives for up to 8 months per patient. This research came with many technical challenges but is nevertheless an interesting data point towards a future where material sciences keep improving skin-compatible materials for ultra long-term usage (multiple months and up).</p><h4>Data Annotation Services</h4><p>A second technological enabler has to do with data annotation. Wearable devices and other RPM solutions are collecting valuable data over increasingly longer time periods. The major benefit is that this should greatly improve the diagnostic value of the data. This is especially true in epilepsy, where seizures can be far and few between (but therefore no less debilitating). With the ability to collect long-term high density clinical-grade data, comes the need to convert that data into <strong>clinically actionable information</strong> <em>and</em> do it in a way that does not increase the data review burden for healthcare professionals, who are already strapped for time.</p><p>If you are now thinking “well, that’s what algorithms are for… duh”, you’d be right but that’s only part of the story. Algorithms to be used in this context are “Software as a Medical Device” (SaMD) and come -rightfully so- with strict performance and safety requirements dependent on their intended use.</p><p>Machine learning (ML) has made great strides in the last half decade or so, including for the processing of physiological signals such as EEG and electrocardiography (ECG). It is a powerful set of tools to develop highly performant algorithms relatively quickly on the condition that you have access to high quality <strong>labeled </strong>data. In a supervised setting, to allow the “machine” to learn from the data, it first needs to know which patterns it is expected to identify. In the case of EpiCare@Home, this labeled data needs to be generated by our system as we use an atypical reduced behind-the-ear EEG montage, augmented with other cardiorespiratory and motion signals. No historical databases are available that we could use retrospectively to train an ML model, at least not one that would perform well in the real world. Generating high quality labeled data is time intensive, immensely costly, and requires a lot of technical expertise but, in this case, there is no way around it.</p><p>That is why we decided to roll out EpiCare@Home combined with a <strong>human-driven annotation service</strong>. This means that data from the wearable devices is reviewed by certified annotators. The events they annotate as “potential seizure events” are shared with the neurologist for final validation. This has two advantages:</p><ol><li>People with epilepsy can benefit from EpiCare@Home now because the data annotation service allows a healthcare professional to use it without increasing their data review burden.</li><li>We continuously generate high quality labeled data based on real-world data collection, which is needed to continue to develop the certified algorithms mentioned above.</li></ol><blockquote>Generating high quality labeled data is time intensive and immensely costly but there is no way around it.</blockquote><p>And of course, we are working on that second step. While trained annotators are very good at identifying relevant events, they are still human and can only process so much data in a day. A performant algorithm can process data exponentially quicker; not with the intention of removing human annotators from the equation completely, but to allow them to focus on the more complex cases. In other words, once our ML model is ready for prime time, it will become the annotator&#39;s best friend.</p><p>Over the years, we have collected data on more than 600 people with epilepsy, accounting for more than 4000 seizures as part of the academic <a href="https://eithealth.eu/spotlight-story/seizeit/">SeizeIT1</a> and <a href="https://eithealth.eu/product-service/seizeit2/">SeizeIT2</a> studies via long-standing collaborations with partners in the EU and UK. In these studies, data was collected in-hospital with video-EEG and behind-the-ear EEG simultaneously, to identify the seizure types that can be reliably detected in this configuration. This data was also used to develop seizure detection algorithms which resulted in multiple publications (<a href="https://pubmed.ncbi.nlm.nih.gov/28359222/">1</a>, <a href="https://ieeexplore.ieee.org/document/8331658">2</a>, <a href="https://pubmed.ncbi.nlm.nih.gov/29295522/">3</a>, <a href="https://pubmed.ncbi.nlm.nih.gov/29027928/">4</a>, <a href="https://pubmed.ncbi.nlm.nih.gov/29747021/">5</a>, <a href="https://ieeexplore.ieee.org/document/8333462">6</a>, <a href="https://www.worldscientific.com/doi/abs/10.1142/S0129065720500355">7</a>, <a href="https://pubmed.ncbi.nlm.nih.gov/32808854/">8</a>, <a href="https://www.frontiersin.org/articles/10.3389/fneur.2020.00145/full">9</a>, <a href="https://onlinelibrary.wiley.com/doi/full/10.1111/epi.16470">10</a>, <a href="https://pubmed.ncbi.nlm.nih.gov/34240748/">11</a>, <a href="https://onlinelibrary.wiley.com/doi/full/10.1111/epi.17061">12</a>, <a href="https://pubmed.ncbi.nlm.nih.gov/34731723/">13</a>, <a href="https://pubmed.ncbi.nlm.nih.gov/36681896/">14</a>) with the following high-level takeaways:</p><ul><li>A behind-the-ear <strong>low-montage EEG</strong> set up can in many cases be used to identify epileptic seizures reliably, especially when a prior diagnosis of epilepsy exists. See for instance references # 3 (focal seizures), 9 (focal temporal lobe seizures), 12 (absence seizures), and 13 (focal and absence seizures).</li><li>The ability to identify seizures (by algorithms and/or human annotators) can be improved by including additional signal modalities, such as heart rate and motion. See references # 6 (cardiac, muscle, and physical activity) and 10 (cardiac and brain activity) for examples. These so-called <strong>multimodal </strong>approaches hold a lot of promise for performant real-world seizure detection with unobtrusive wearables, especially when no video feed is available.</li></ul><p>So … problem solved, right?</p><p>Not quite.</p><p>First, most of these studies developed <em>and </em>validated their algorithm on a single and small dataset which always runs the risk that the results will not <strong>translate </strong>well <strong>to the real-world</strong>. In a <a href="https://pubmed.ncbi.nlm.nih.gov/36681896/">recent example</a>, an algorithm that was developed based on a subset of the in-hospital SeizeIT1 data was used “as is” to identify seizure events on ambulatory home data without accounting for the difference in setting (hospital vs. home) and the use of different electrode technology to acquire the data. The algorithm underperformed, while visual inspection of the same dataset by a trained annotator would have yielded much better results. Similarly, upgrading the original algorithm to “learn” from more relevant data could have improved the results.</p><p>Secondly, recall that we are working in a medically certified space and algorithms that were developed for academic purposes (and thus publication), usually were not designed to meet the patient safety and performance requirements of a SaMD.</p><p>Some work has also been done on “<strong>personalized algorithms</strong>”. The idea is that you start from a base algorithm that is then tweaked with prior knowledge about the patient’s seizure signature to increase detection performance. A major downside of this approach is that it tends to “overfit” an algorithm to a known set of historical data which may be hard to obtain in the real-world. In addition, it may not account for the fact that people with epilepsy often experience changes in seizure types and burden.</p><blockquote>Medical algorithms are “Software as a Medical Device” (SaMD), which comes with strict quality, performance and safety requirements.</blockquote><p>The TL;DR so far is that the usability and clinical performance of EpiCare@Home has improved significantly by introducing EEG Adhesives and adding data annotation services. Combine that with the exploratory and validation research that happened over the last couple of years, and we are ready to make the technology available to a wider range of people with epilepsy and gradually scale up the volume of data it can handle.</p><h3>Which Patients and Seizure Types?</h3><p>The number one question we receive about EpiCare@Home is:</p><blockquote>Which seizure types can it monitor?</blockquote><p>Excellent question. And the answer? Many, but … it depends. Before you get your gallows out of storage, hear us out.</p><p>The prescribing neurologist should consider the following two questions before using EpiCare@Home:</p><ol><li><strong>What do I already know about my patients’ seizures?</strong><br>- Is their seizure signature generalized, focal or a combination?<br>- How frequent are their seizures?<br>- In the case of focal seizures, is the location of the EEG electrodes (behind-the-ear) likely to record them (e.g., consider temporal vs frontal lobe seizures)?<br>- Are there other physiologic or contextual signatures (e.g., cardiorespiratory, motion) that EpiCare@Home may be able to pick up?</li><li><strong>What diagnostic question(s) am I trying to answer?<br></strong>- Am I trying to establish an initial diagnosis of epilepsy? If so, is a reduced EEG montage a good option?<br>- Am I trying to acquire additional diagnostic information to reduce the seizure burden of a refractory patient with a long-documented history of epilepsy?<br>- Am I evaluating a change in treatment? For instance, to reduce side effects or due to a change in circumstances (e.g., pregnancy).</li></ol><p>Epileptologists are taught “<strong>don’t treat the EEG</strong>”, as it may lead to overinterpretation, misdiagnosis, and even unnecessary treatment. Because EpiCare@Home does not include a video feed, its other signals (heart rate, respiratory rate, motion, seizure diary events …) provide important contextual and physiologic information that assist with reading the EEG. Our <a href="https://medium.com/byteflies/what-is-byteflies-building-for-people-with-epilepsy-and-why-9b3c55f56dfc#3012">previous post</a> discusses in more detail how we can compensate for the lack of video.</p><p>In conversations with many epileptologists, we have refined its position in the care path for people with epilepsy as follows:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*djFLoEHUtDONpgl5BkCGjA.png" /><figcaption>EMU = epilepsy monitoring unit, Sz = seizure(s), ULT = ultralong-term monitoring (defined as more than 1 month).</figcaption></figure><p>Consider the following two examples, loosely based on EpiCare@Home patients:</p><blockquote>Based on reports from a parent, a pediatric patient is suspected to have typical absence seizures. A brief (1 hour) video-EEG in the hospital does not yield a conclusive result. The epileptologist decides to send the patient home with EpiCare@Home for 3 days. From the data, eight electrographic absence seizures of more than 3 seconds are identified which provides enough information for the epileptologist to start a treatment plan. One month later, the patient is monitored again for 3 days and this time no seizures are identified which matches the reports from the parents.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*B38d5h5gktQKHXcq59_b5g.png" /><figcaption>Example of an absence seizure in the EpiCare@Home clinical dashboard. The top graph is accelerometer, and the bottom graph is bilateral behind-the-ear EEG.</figcaption></figure><blockquote>An adult patient with a decade long history of refractory focal impaired awareness (FIA) epilepsy self-reports a recent increase in seizures. He has spent many weeks in video-EEG monitoring units over the years and is not looking forward to doing it again. Because the epileptologist knows that his seizures are primarily occurring in the left temporal lobe, she suggests EpiCare@Home for one week, followed by a dosage adjustment and another week of monitoring. She does make it clear that if the data is inconclusive, he may need to come into the monitoring unit after all. The first week of monitoring yields three seizures, the second only one. The patient also reports an improvement. The epileptologist suggests maintaining this treatment regimen for a couple more months and then check-in again, potentially with another one week of EpiCare@Home monitoring.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8I919dJHzy0WmKGYW0dUqg.png" /><figcaption>Example of a focal seizure in the EpiCare@Home clinical dashboard. The top left graph is heart rate, and the top right graph is accelerometer. The bottom graph is a local left behind-the-ear and cross (left-to-right) EEG channel.</figcaption></figure><p>These are just a few representative examples, and we have more. In fact, we presented <strong>retrospective </strong>data at the Antiepileptic Drug and Device Trials (AEDD) and American Epilepsy Society (AES) meetings in 2022 that the patients, caregivers, and neurologists using EpiCare@Home, reported a positive clinical or quality of life impact in 93% of cases. The diagnostic data generated by EpiCare@Home was used as follows in 22 case studies (including 8 children):</p><ul><li>~30% reported an existing treatment regimen was adjusted, or a new treatment was started.</li><li>~30% reported confirmation of a known diagnosis or a referral to another specialist.</li><li>~15% maintained a treatment plan.</li><li>~15% reported an improved care experience.</li></ul><p>Two more video impact stories from parents with children with epilepsy who were helped with EpiCare@Home can be found <a href="https://www.youtube.com/watch?v=lxcNPMdNfU0">here</a> and <a href="https://www.youtube.com/watch?v=NylpB4GnkkM">here</a>.</p><p>Of course, we need <strong>systematic </strong>and preferably <strong>prospective </strong>data analysis to assess the value and clinical acceptability of EpiCare@Home at scale in the care trajectory of people with epilepsy.</p><p><strong>So, what has already happened on that front and what is ongoing?</strong></p><p>As mentioned previously, the SeizeIT studies acquired a large body of data. The clearest example to date of a clinical workflow assessment based on that data is a <a href="https://onlinelibrary.wiley.com/doi/full/10.1111/epi.17061">paper</a> that evaluated both algorithmic and blind reading by annotators of 284 absence seizures. The results demonstrated that EpiCare@Home provides a clinically viable method to monitor absence seizures as compared to video-EEG. A major limitation is that all data was recorded in the hospital and not yet at home.</p><p>Similarly, the results of small-scale blind reading studies were presented at the <a href="https://www.sciencedirect.com/science/article/pii/S1525505021006259">Epilepsy Foundation Pipeline Conference</a> in 2020. For absence seizures, 96% sensitivity was reached based on 155 seizures. For focal seizures, 83% sensitivity was reached based on 46 focal seizures. When ECG and electromyography (EMG) data were added, the sensitivity increased to 93%. In both cases, the recorded data vastly outperformed self-reporting via seizure diaries.</p><p>Additional analysis of the SeizeIT datasets by academic partners is ongoing.</p><p>Finally, we have recently started a <a href="https://clinicaltrials.gov/ct2/show/NCT05738226">prospective clinical trial</a> at Thomas Jefferson University (TJU) to focus specifically on focal epilepsies.</p><h3>Clinical Acceptance and Reimbursement</h3><p>In Belgium and Germany, we and epileptologists have witnessed the impact of EpiCare@Home in clinical practice firsthand, in pediatric and adult patients, and in generalized, focal and rare epilepsies (including so-called developmental epileptic encepalopathies or DEEs).</p><blockquote>A retrospective analysis of EpiCare@Home case studies found that in 93% cases, a positive impact was reported by either patient, caregiver, and/or neurologist.</blockquote><p>Now, we are at an important crossroads since EpiCare@Home is not yet <strong>reimbursed </strong>in Belgium or abroad. We are working hard to collect reimbursement evidence, while continuing to roll out EpiCare@Home. This results in a typical digital health “catch 22”: the lack of reimbursement means that it is difficult for healthcare professionals to implement it in their workflows, while we need to be looking for other sources of funding to bankroll evidence collection.</p><p>Fortunately, we are far from alone in realizing this. Health Technology Assessment <strong>(HTA) frameworks</strong> for rapid evaluation of new and promising digital health technologies are popping up everywhere and many regulators are realizing that the dogma of the Randomized Clinical Trial (RCT) does not apply everywhere. Even more encouraging is that <a href="https://eithealth.eu/opportunity/call-for-applications-external-advisory-group-to-the-european-taskforce-for-harmonised-evaluation-of-digital-medical-devices-dmds/">harmonization efforts</a> are under way in the EU to recognize HTAs across member states.</p><blockquote>The digital health catch 22: how to fund the roll out of promising and safe technologies while moving through the slow reimbursement process?</blockquote><p>Nevertheless, these efforts will not result in concrete results overnight and much work remains to be done, so until then finding the right partners is crucial, which brings us to … the village.</p><h3>The Village</h3><p>Over the years, we have built up a diverse network of partners and like-minded individuals who share the mission: <strong>to improve the standard of care for people with epilepsy</strong>, with a particular focus on virtual care solutions that are patient-centric and value-based.</p><p>In neurology and epilepsy, this is imperative as no single solution will “fix” the unmet clinical needs we outlined above. Instead, we must work together to bring a diverse and complementary set of digital health solutions to the patient.</p><p>One concrete example would be to provide several seizure RPM tools to neurologists and epileptologists so they can progressively “ramp up” the sophistication of the monitoring tools, knowing that that usually goes hand in hand with the obtrusiveness of the technology. For instance, solutions like EpiCare@Home could provide a logical steppingstone towards sub-scalp ultralong term EEG monitoring devices.</p><blockquote>We must work together to bring a diverse and complementary set of digital health solutions to the patient.</blockquote><p>We already mentioned some of the partners and it is beyond the scope of this post to mention everyone that has contributed to this mission over the years. We do want to include a shoutout to the following partners, who have gone above and beyond to rally stakeholders:</p><p>The <strong>Epilepsy Foundation</strong>, in particular the Antiepileptic Drug and Device Trials and Pipeline conferences, which always bring together a critical mass of technology enablers. And last but most certainly not least, <strong>UCB Pharma</strong>’s digital health focused teams for their leadership in recognizing the value of EpiCare@Home and similar solutions and their continued efforts to bring them to market and innovate the clinical trial space.</p><p><em>The information contained in this article represents the views and opinions of the writer(s) and does not necessarily represent the views or opinions of other parties referenced or mentioned therein.</em></p><p><em>The article is not intended to be a substitute for professional medical advice, diagnosis, or treatment. Always seek the advice of your physician or other qualified healthcare professional with any questions you may have regarding a medical condition. Never disregard professional medical advice or delay in seeking it because of something you read in this article.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=48bb7189d13a" width="1" height="1" alt=""><hr><p><a href="https://medium.com/byteflies/what-has-byteflies-built-for-people-with-epilepsy-so-far-48bb7189d13a">What has Byteflies built for people with epilepsy (so far)?</a> was originally published in <a href="https://medium.com/byteflies">Byteflies</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[CovidCare@Home – Telemedicine for Covid-19 Patients]]></title>
            <link>https://medium.com/byteflies/covidcare-home-telemedicine-for-covid-19-patients-96d1e7ad46ea?source=rss-bb22181f7c06------2</link>
            <guid isPermaLink="false">https://medium.com/p/96d1e7ad46ea</guid>
            <category><![CDATA[digital-health]]></category>
            <category><![CDATA[covid19]]></category>
            <category><![CDATA[telehealth]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[wearables]]></category>
            <dc:creator><![CDATA[Benjamin Vandendriessche]]></dc:creator>
            <pubDate>Wed, 13 Jan 2021 14:26:57 GMT</pubDate>
            <atom:updated>2023-03-20T21:07:50.086Z</atom:updated>
            <content:encoded><![CDATA[<h3>CovidCare@Home – Telemedicine for COVID-19 Patients</h3><h4>Rapid roll-out demonstrates healthcare providers are ready to adopt digital medicine technologies on a larger scale.</h4><blockquote>This story was originally published on Jan 13, 2021 and discussed data on the first three months of CovidCare@Home. We updated the story on Aug 8, 2021 to include all data collected from Oct 2020 through Jul 2021.</blockquote><h4>Table of Contents</h4><ul><li><a href="#af77">Introduction: a brief history of telemedicine</a></li><li><a href="#43d7">CovidCare@Home</a></li><li><a href="#ffa3">The need for reimbursement</a></li><li><a href="#0571">Summary</a></li></ul><h3>An exceedingly brief and opinionated history of telemedicine</h3><p>T<em>elehealth </em>and <em>telemedicine </em>describe health care services rendered without face-to-face contact between a provider and patient, and the terms have been around for decades. However, the number of telehealth applications with demonstrable impact have been few and far between, despite the omnipresence of secure telecommunication technologies in most parts of the world. One technological innovation was missing though: the ability to reliably measure clinical parameters at home. More recently, new biometric monitoring technologies (<a href="https://www.nature.com/articles/s41746-020-0260-4">BioMeTs</a>) have entered the market. Think wearable devices that track your health continuously in the background. It is the advent of specialized and user-friendly BioMeTs that promises to transform telemedicine from a niche application into <a href="https://www.sciencedirect.com/science/article/abs/pii/S0378512221001109">routine clinical care</a>.</p><p>Although BioMeTs nowadays are typically internet-connected, devices with similar intended uses already have a long history. The best-known example is the <a href="https://www.heart.org/en/health-topics/heart-attack/diagnosing-a-heart-attack/holter-monitor">Holter ECG monitor</a> which is a staple of any cardiologist looking to diagnose certain cardiac problems while people are going about their usual activities. A typical Holter monitor (see image below, middle) is worn for 1–2 days.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9GAlA4O3G7_WtsLW6w9KxQ.png" /><figcaption>From left-to-right: (1) Model 445 Mini-Holter Recorder circa 1976 (adapted from the <a href="https://americanhistory.si.edu/blog/2011/11/at-the-heart-of-the-invention-the-development-of-the-holter-monitor-1.html">National Museum of American History</a>), (2) example of a contemporary multi-lead Holter Monitor, and (3) an example of the new generation of wearable ECG monitors that can monitor heart rate, respiratory rate, and activity (Byteflies Sensor Dot ECG patch, image is used with permission from copyright owners Stad Antwerpen and Frederik Beyens).</figcaption></figure><blockquote>It is the advent of specialized and user-friendly wearable devices that promises to transform telemedicine from a niche application into <a href="https://www.sciencedirect.com/science/article/abs/pii/S0378512221001109">routine clinical care</a>.</blockquote><p>For a long time, the <strong>usability </strong>of these early ambulatory devices was lacking. The miniaturization of electronics and the rise of the <a href="https://en.wikipedia.org/wiki/Internet_of_things">Internet of Things</a> (IoT) made it possible to fit a medical device into the size of a consumer electronics gadget <em>and </em>include functionality that users are now expecting such as cloud-connectivity and interaction via mobile apps. In the healthcare sector, a second and more fundamental factor is at play: <a href="https://www.accenture.com/us-en/insights/health/leaders-make-recent-digital-health-gains-last">patients expect <strong>guidance</strong></a> from their healthcare provider on the use of digital health applications. However, in many cases, <a href="https://npjdigitalmedcommunity.nature.com/posts/65821-why-v3-verification-analytical-validation-and-clinical-validation">trust</a> in these novel devices and technologies is lacking with healthcare professionals. It should come as no surprise that someone who does not trust a device or technology will not be able to provide proper guidance on its usage.</p><p>And then came along COVID-19 …</p><p>During global health crises, such as the SARS-CoV-2 pandemic, healthcare systems are put under enormous pressure, even leading to <a href="https://www.nejm.org/doi/full/10.1056/NEJMms2009984">patients requiring life-saving care to be turned away</a> as resources are stretched too thin. Healthcare professionals embraced various <a href="https://www.jmir.org/2020/11/e22280">digital health</a> and <a href="https://c19hcc.org/telehealth/">telehealth</a> applications at breakneck speed to help deal with the overload. At <a href="https://www.byteflies.com/">Byteflies</a>, we also decided to do our part and launched <a href="https://www.covidcareathome.com/">CovidCare@Home</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Z2BgS5Fn4ZJ4Kxa9XvPK9w.png" /><figcaption><a href="https://trends.google.com/trends/explore?date=2018-01-01%202021-01-01&amp;q=telemedicine,telehealth">Google Trends data</a> for the terms “telehealth” (orange) and “telemedicine” (blue) expressed as relative search interest (100 = peak popularity) over the last 3 years. The peak coincides with the rapid spread of the novel coronavirus.</figcaption></figure><h3>CovidCare@Home</h3><p><a href="https://www.covidcareathome.com/">CovidCare@Home</a> (CC@H) is a Belgian consortium of technology and healthcare providers who joined forces after the first COVID-19 wave in the spring of 2020 to provide a fast and pragmatic home monitoring solution for COVID-19 patients, thereby freeing up valuable hospital beds (<a href="https://www.tijd.be/tech-media/technologie/ziekenhuizen-omarmen-telemonitoring-coronapatienten/10260760.html">press</a>).</p><p>The solution consists of a couple of BioMeTs, a patient app, a healthcare provider dashboard, and a telemonitoring team. You can find some more details on the <a href="https://www.covidcareathome.com/">project website</a>. On the provider side, a dashboard lists all active and past patients, their vital sign measurements and survey answers, and associated trends ranked according to the National Early Warning Score (<a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6542226/">NEWS2</a>). The dashboard can integrate with the electronic medical records (EMR) of the hospital partners, and when someone’s vital signs deteriorate, the hospital or family doctor can contact them.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VDPzzBA9io8UjFUtkMNqFg.png" /><figcaption>CC@H dashboard for healthcare professionals. A <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6542226/">NEWS2 score</a> is calculated from a patient’s vital signs. Users can visualize all historical measurements and associated trends for each patient in their hospital. Family doctors can also access the data for their patients.</figcaption></figure><p>Just like patients can move through various hospital departments depending on the level of care they need, telemonitoring applications require a clear definition of “paths” (or “patient journeys”, if you will). The primary group of patients that CC@H was used for are on the so-called <strong>posthospitalization path </strong>(black arrow in the diagram below), i.e. they are discharged from the hospital and sent home with CC@H to complete their recovery under close surveillance. They are in the comfort of their own home but in case their health deteriorates, a swift intervention is possible. A second path is <strong>prehospitalization </strong>(orange arrows in the diagram): people with a suspected or confirmed coronavirus infection may show up at the emergency department. Before CC@H, the choice was binary: admit that person or send them home. With CC@H, a third option is available when an admission is not yet warranted. Home monitoring can be used until a hospital admission does become necessary. Similar to the second path, <strong>assisted living facilities </strong>can use a form of prehospitalization monitoring (green arrow in the diagram). Considering these facilities have a high concentration of high-risk individuals, CC@H can be used to monitor a large group of people continuously, to facilitate a quick intervention when social distancing measures are in effect.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*djTGo8rV5M8QbOwSFsvK_A.png" /><figcaption>CC@H is compatible with various patient care paths. The primary one is posthospitalization monitoring (black arrow). Prehospitalization monitoring can be initiated for people in the emergency department but for whom an admission is not yet warranted (orange arrows) or assisted living facilities (green arrow). The data collected by the CC@H includes vital signs and questionnaire data, which is shared with the telemedicine team and the family doctor (blue dashed arrows).</figcaption></figure><h4>What did we achieve so far?</h4><p>CC@H kicked off late September 2020 in the Flanders region (6.6M inhabitants) of Belgium and enrolled around ten hospitals per month until the end of the year. Between October 2020 and July 2021, more than 850 COVID-19 patients were monitored at home, accounting for almost 20,000 vital sign measurements. Instrumental for this fast roll-out was the fact that CC@H is provided for free to the participating hospitals. <a href="https://www.byteflies.com/">Byteflies</a> is bearing the bulk of the costs with grant support from <a href="https://www.vlaio.be/">VLAIO</a> and the <a href="https://www.stadantwerpen.be/">City of Antwerp</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/861/1*CxuAx9BY45wgv0VqqQlV0g.png" /><figcaption>Green dots are CC@H hospitals, red dots are Flemish hospitals not yet enrolled, and light blue dots were regionally out of scope for the first phase.</figcaption></figure><blockquote>In just 3 months, CC@H was rolled-out to 28 Flemish hospitals and by February 2021, more than 500 COVID-19 patients had been discharged early to continue their recovery at home.</blockquote><p>Enrollment of hospital partners happened while the <a href="https://apnews.com/article/virus-outbreak-belgium-europe-43286270306b9d0dcb0d36a0b9ed5f6f">second wave was ramping up in Belgium</a>. The blue plot in the chart below displays the cumulative number of COVID-19 patients that were sent home with CC@H between October 2020 and July 2021 (left y-axis). For reference, the orange plot displays the new hospitalizations in Flanders per 100,000 inhabitants over that same period (right y-axis). Relatively speaking, CC@H enrollment was growing month over month as more hospitals started integrating it into their clinical workflows. The acceleration from February to March can be explained by the availability of a medical reimbursement path (see <a href="#ffa3">further</a>).</p><blockquote>Error correction: In the first version of this story published on January 13, 2021, the source of epidemiological data for Flanders was not correctly used leading to an overestimation of the actual number of COVID-19 cases, and thus an underestimation of the impact of CC@H and an overestimation of the potential cost savings. In the current version, that error was corrected.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xNckjmCAWSP7l1KAmiHyWA.png" /><figcaption>Total CC@H recruitment in Flanders between October 2020 &amp; July 2021 (blue plot, left y-axis) and total hospitalizations in Flanders per 100,000 inhabitants in the same period (orange plot, right y-axis). The COVID-19 statistics are taken from <a href="https://datastudio.google.com/embed/u/0/reporting/c14a5cfc-cab7-4812-848c-0369173148ab/page/ZwmOB">Sciensano’s dashboard</a>.</figcaption></figure><p>Factors to consider when interpreting the CC@H enrollment numbers:</p><ul><li>By January 2021, approximately 40% of Flemish hospitals were enrolled and that number remained stable until July 2021.</li><li>Not all patients admitted for COVID-19 are eligible for pre- or posthospitalization home monitoring. It is not straightforward to estimate the real eligibility rate as no standardized protocol was in place across all participating hospitals. Based on our data, 10–25% of hospitalized patients were eligible for the CC@H pre- or posthospitalization paths.</li></ul><p>To put the absolute CC@H recruitment and new hospitalization numbers in perspective, the chart below shows the “Relative Reach”. This is defined as the total number of CC@H patients over the total (new) hospitalizations in Flanders that could be eligible for CC@H. The eligibility rates were estimated at 10% or 25% as explained above. At 10% eligibility, the maximum relative reach in a single month was 57% (blue plot). For the higher eligibility rate of 25%, the maximum in a single month was 23% (orange plot). Over the entire monitoring period, between 14–35% of all potentially eligible CC@H patients in Flanders were home monitored, and a larger number of patients was reached during wave peaks.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jrUMCwCk9HRGRgZgGZswcg.png" /><figcaption>Reach is the relative number of hospitalized patients that were enrolled in CC@H through the participating hospitals, compared to the total number of hospitalizations in Flanders that are expected to be eligible for CC@H. The relative reach for a 10% eligibility rate is shown in the blue plot and a 25% rate in the orange plot. The COVID-19 statistics are taken from <a href="https://datastudio.google.com/embed/u/0/reporting/c14a5cfc-cab7-4812-848c-0369173148ab/page/ZwmOB">Sciensano’s dashboard</a>.</figcaption></figure><p>Considering:</p><ol><li>The extremely short onboarding period for a large group of geographically dispersed hospitals …</li><li>Of a new and unproven technology …</li><li>Based on in-kind contributions from technology and healthcare providers (i.e. without structural reimbursement support) for the first four months …</li><li>During a raging pandemic …</li></ol><p>Byteflies and all our CC@H partners are immensely proud of this achievement. Of course, with what we have learned and built, the objective is to ensure that in similar future scenarios, 100% of patients eligible for telemonitoring can be reached virtually instantaneously.</p><blockquote>The objective of our remote patient monitoring efforts is to ensure enough capacity to monitor all eligible patients in Belgium.</blockquote><h4>What is the potential economic impact?</h4><p>A final key factor is the potential cost savings. CC@H was made available for free because we wanted to do our part in supporting our healthcare providers, but a telemonitoring application obviously costs money to build and maintain. And it is not only the technology platform, the “chain of accountability” needs to be ensured as well. In case of deteriorating vital signs or concerns from the patient, a rapid and appropriate follow-up by qualified healthcare professionals needs to be triggered. Luckily, the Belgian authorities responsible for medical reimbursement acted swiftly and by December 2020, temporary reimbursement codes for COVID-19 related telemonitoring were available. By February 2021, all CC@H provider partner could start using these codes which had a further net positive effect on CC@H recruitment during the third wave.</p><p>Which begs the question: can a solution like CC@H in fact lead to a reduction in health care costs? From the collected data and feedback from healthcare providers, we estimate that patients are discharged from the hospital at least 1–2 day earlier with CC@H (posthospitalization) or never require an admission (prehospitalization). A <strong>hospital day</strong> for a COVID-19 patient in Belgium costs around <a href="https://www.departementwvg.be/wat-wie-en-hoeveel">€800</a> (non-intensive care and excluding any procedures). A <strong>telemonitoring week</strong> with CC@H is roughly four times less. In addition, the amount of time a patient is under close observation is extended by a full week! In other words, CC@H is not only reducing health care spending, it may also improve the quality of care.</p><p>Assuming two hospitalization days were avoided on average across the pool of CC@H patients, we estimate the achieved healthcare cost savings in Flanders for each month CC@H was active in the graph below. During wave peaks, this reached up to €250,000 per month, for a total cost saving of €1.2M over 10 months.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EiXspi4gwxVJscVbsj5W3g.png" /><figcaption>Assuming two hospitalization days were avoided on average for all CC@H patients, the plot displays the achieved healthcare cost saving in Flanders for each month CC@H was active. This considers the average cost of two hospitalization days and the cost for a full week of telemonitoring with CC@H.</figcaption></figure><blockquote>CC@H may have directly prevented €1.2M in healthcare costs in Flanders in just 10 months&#39; time while extending the observation period.</blockquote><p>If we assume that CC@H was deployed across all Flemish hospitals during this same period, we can also estimate the total potential cost savings in Flanders using the same eligibility rates of 10 and 25% as explained earlier. During a wave peak, CC@H could have reduced health care spending by €0.8–2.0 million per month (blue and orange plot below). For the CC@H active period, this amounts to a total potential cost saving of €3.8–9.4 million in just 10 months <em>with</em> a longer observation period of the patient. Like before, this estimation assumes two hospitalization days are avoided (either via the pre- or posthospitalization path), and only takes those estimated direct costs into account. Whether or not the prolonged monitoring period has additional cost benefits remains to be determined, for instance because we can intervene early for people who are getting worse, potentially preventing a critical care admission.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sznY5tK7elZScsqJykZhwg.png" /><figcaption>For the same eligibility rates as discussed previously (10%, blue plot and 25%, orange plot), the cost saving can be estimated when replacing two hospital days with one week of home monitoring. Potential cost savings are calculated for each week with data available. The COVID-19 statistics are taken from <a href="https://datastudio.google.com/embed/u/0/reporting/c14a5cfc-cab7-4812-848c-0369173148ab/page/ZwmOB">Sciensano’s dashboard</a>.</figcaption></figure><blockquote>One <strong>week </strong>of telemonitoring with CC@H is eight times cheaper than two hospitalization days.</blockquote><h4>What did we learn?</h4><p>The main objective of the first phase was to rapidly develop an application that would assist our healthcare providers in these unprecedented times, while demonstrating the feasibility of CC@H and the willingness of Belgian hospitals to adopt telemedicine applications. We believe the feasibility is clear, and the willingness and drive of our hospital partners to implement CC@H into their clinical workflows has far exceeded our expectations.</p><blockquote>The willingness and drive of our hospital partners to implement CC@H into their clinical workflows has far exceeded our expectations.</blockquote><p>We have received feedback from CC@H users, both healthcare professionals and patients. Although some are anecdotal in nature, the takeaways were clear:</p><ul><li>On average, 1–2 hospitalization days can be replaced by one week of CC@H monitoring in the <strong>posthospitalization </strong>path. In the prehospitalization path, the potentially avoided number of hospitalization days could be higher, but we currently do not have enough information to estimate that.</li><li>CC@H patients have been <strong>recalled </strong>to the hospital based on their telemonitoring data; we are working with our hospital partners to get a clearer picture on the impact of these recalls.</li><li>The main <strong>demographic </strong>open to using CC@H is born after 1960.</li><li>The average monitoring <strong>duration </strong>was 6–7 days per patient and <strong>adherence </strong>was well over 95%.</li><li>Patient users overwhelmingly communicated that CC@H gave them <strong>peace of mind</strong>, especially the combination of vital sign monitoring and the communication channel to their healthcare provider. <a href="https://www.janpalfijn.be/sites/default/files/public/azjp_palfijnnieuws_l21-web.pdf">Example</a> (Dutch).</li><li>In the first 3 months of CC@H, Byteflies built <strong>more than 30 new features </strong>based on feedback from the hospital partners, demonstrating that agile co-development and safe medical devices can go hand-in-hand.</li></ul><h3>The need for reimbursement</h3><p>For most digital medicine application, regulatory clearance and reimbursement by a payer are critical to drive adoption by healthcare providers. Encouraging steps have been made by regulators over the last few years. High profile examples are the <a href="https://www.nature.com/articles/s41746-020-0306-7#:~:text=Germany%27s%20new%20Digital%20Healthcare%20Act,will%20pay%20for%20their%20use).">Digital Healthcare Act</a> (DVG) in Germany and the <a href="https://www.fda.gov/medical-devices/digital-health-center-excellence">FDA Digital Health Center of Excellence</a> in the US. In Belgium, the <a href="https://mhealthbelgium.be/">mHealth pyramid system</a> is being developed.</p><p>Specifically to battle COVID-19, the FDA provided <a href="https://www.fda.gov/medical-devices/emergency-use-authorizations-medical-devices/coronavirus-disease-2019-covid-19-emergency-use-authorizations-medical-devices">emergency use authorizations</a> for certain medical devices. Although similar broad emergency uses never materialized in the EU, many countries did pass temporary COVID-19 specific reimbursement policies. Likewise, Belgium’s single payer system introduced a lot of <a href="https://www.riziv.fgov.be/nl/covid19/Paginas/ziekenfonds.aspx">support measures</a> to deal with the pandemic. In December 2020, the National Institute for Health and Disability Insurance (RIZIV/INAMI) launched a <a href="https://www.inami.fgov.be/nl/covid19/Paginas/zorg-afstand-covid1-patienten-thuis-telemonitoring.aspx">pilot study</a> to monitor COVID-19 patients at home which was drafted in collaboration with Byteflies and other stakeholders*. This pilot study defines reimbursement codes for telemonitoring of pre- and posthospitalization COVID-19 patients, and its availability led to a marked acceleration in CC@H enrollment as we showed above.</p><p>*The link is the official press release in Dutch and French, a non-official English summary is available <a href="https://mtrconsult.com/news/inami-pilot-project-telemonitoring-covid-19-patients-home-belgium">here</a>.</p><blockquote>The RIZIV pilot study is a textbook example of diverse stakeholders coming together to solve a unique problem.</blockquote><p>The pilot reimbursement allowed CC@H to continue to provide a pragmatic telemonitoring solution for our physicians and nurses in the frontlines of the pandemic at a larger and more coordinated scale, while also collecting the evidence that supports a more permanent roll-out for this type of telemonitoring reimbursement codes.</p><p>We commented on the potential cost savings of CC@H. However, the real take home message is that for many chronic conditions, such as heart failure, chronic obstructive pulmonary disease, epilepsy, multiple sclerosis, and Alzheimer’s disease, to name just a few, the potential health care cost savings are orders of magnitude larger if we succeed in decentralizing a significant portion of care to the home environment. If done right, this will also improve the standard of care.</p><h3>Summary</h3><p>In the three first months of CC@H and with limited financial means, we focused on an incredibly quick onboarding of hospital partners. Conservative estimates indicate that we reached up to 57% of the COVID-19 patients eligible for home monitoring in months with a high number of cases. This was only possible because our hospital partners recognized the potential of CC@H.</p><p>Despite encouraging vaccine news, COVID-19 is unfortunately not yet a thing of the past. Our primary objective in the short-term is to keep supporting the Belgian healthcare system. Of course, we have longer-term goals as well. Although no one wants to think about future pandemics, they will happen, and we will be better prepared, hopefully with permanent reimbursement codes in place. But more importantly, the promise of digital health applications for chronic conditions is finally ready to materialize.</p><blockquote>CovidCare@Home was rolled out at an unprecedented speed because our hospital partners saw the need and recognized its potential.</blockquote><p>Despite all the hardship COVID-19 has and is causing for a lot of people, we hope that some silver lining will emerge, including the understanding that patient-centered digital health applications are here to stay and are quickly becoming an essential part of routine clinical care. Now is the time to ensure that these roots take hold permanently. Not for the sake of technology adoption, but for the sake of a <a href="https://www.nature.com/articles/s41746-019-0206-x">holistic</a>, <a href="https://www.forbes.com/sites/joeharpaz/2019/12/09/5-ways-technology-will-enable-value-based-care-in-2020/">value-based</a> and <a href="https://nam.edu/compassionate-patient-centered-care-in-the-digital-age/">compassionate</a> healthcare system that puts the needs and well-being of the patient first.</p><p><em>The information contained in this article represents the views and opinions of the writer(s) and does not necessarily represent the views or opinions of other parties referenced or mentioned therein.</em></p><p><em>The article is not intended to be a substitute for professional medical advice, diagnosis, or treatment. Always seek the advice of your physician or other qualified healthcare professional with any questions you may have regarding a medical condition. Never disregard professional medical advice or delay in seeking it because of something you read in this article.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=96d1e7ad46ea" width="1" height="1" alt=""><hr><p><a href="https://medium.com/byteflies/covidcare-home-telemedicine-for-covid-19-patients-96d1e7ad46ea">CovidCare@Home – Telemedicine for Covid-19 Patients</a> was originally published in <a href="https://medium.com/byteflies">Byteflies</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What is Byteflies Building for People with Epilepsy and Why?]]></title>
            <link>https://medium.com/byteflies/what-is-byteflies-building-for-people-with-epilepsy-and-why-9b3c55f56dfc?source=rss-bb22181f7c06------2</link>
            <guid isPermaLink="false">https://medium.com/p/9b3c55f56dfc</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[digital-health]]></category>
            <category><![CDATA[epilepsy]]></category>
            <category><![CDATA[wearables]]></category>
            <category><![CDATA[neurological-disease]]></category>
            <dc:creator><![CDATA[Benjamin Vandendriessche]]></dc:creator>
            <pubDate>Wed, 06 Jan 2021 13:11:17 GMT</pubDate>
            <atom:updated>2023-03-20T21:06:24.783Z</atom:updated>
            <content:encoded><![CDATA[<h4>A personalizable seizure monitoring wearable for home use that extends the capabilities of specialized hospital equipment.</h4><p>Epilepsy is a chronic neurological condition characterized by recurrent unprovoked <strong>seizures</strong>. Although epilepsy can sometimes be linked to a specific cause such as brain injury or certain genetic syndromes, often its origin (<a href="https://www.epilepsydiagnosis.org/aetiology/epilepsies-etiology-groupoverview.html">etiology</a>) is unknown.</p><p>Epilepsy is more common than you may think. An estimated 65 million people of all ages have epilepsy and in the US <a href="https://n.neurology.org/content/76/1/23.abstract">one in 26</a> people will develop epilepsy at some point in their lifetime. About <a href="https://www.epilepsy.com/learn/about-epilepsy-basics">one in 3</a> people in countries with high access to expert care are treatment resistant. Low to middle-income countries account for up to <a href="https://www.who.int/news-room/fact-sheets/detail/epilepsy">80% of all cases</a>, the large majority of whom lack access to appropriate care.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ukacoeDR8DdcqOzXuGcf-A.png" /><figcaption>Summary epilepsy demographics: <strong>†</strong> See references in the text; <strong>‡</strong> Incidence denotes the number of new cases worldwide per year (<a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5272794/">reference</a>); <strong>#</strong> Prevalence denotes the number of people worldwide with active epilepsy (<a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5272794/">reference</a>).</figcaption></figure><p>Seizures are the hallmark symptom of epilepsy (even though having seizures does not necessarily mean you have epilepsy!). They are caused by an imbalance in electrochemical signaling in the brain. In turn, this can trigger additional symptoms that range from hardly noticeable to extremely disabling. Although various antiepileptic strategies exist, in most cases, they do not cure epilepsy but merely suppress the occurrence of seizures.</p><p>Because epilepsy is a chronic condition, because seizures are unpredictable and can be very disabling, because antiepileptic drugs have side-effects, epilepsy can have a profoundly negative impact on someone’s quality of life. In addition, many people must deal with stigma and discrimination in addition to their disease.</p><h4>Table of Contents</h4><ul><li><a href="#3012">What is the unmet need in epilepsy?</a></li><li><a href="#dfdd">How can Byteflies Sensor Dot meet this need?</a></li><li><a href="#6050">Summary</a></li></ul><h3>What is the unmet need in epilepsy?</h3><p>If it was not clear from the introduction, there are quite a few unmet medical needs for people with epilepsy, ranging from:</p><ol><li>Delivering <strong>state-of-the-art care</strong> to a broader demographic and larger geographic areas, especially where <a href="https://www.naec-epilepsy.org/about-epilepsy-centers/what-is-an-epilepsy-center/">comprehensive epilepsy centers</a> are not available.</li><li>Developing<strong> new treatment options</strong> for <a href="https://www.ilae.org/files/ilaeGuideline/Definition-of-Drug-Resistant-Epilepsy-2009-1528-1167.2009.02397.pdf">treatment resistant epilepsy</a>.</li><li>Reducing the <a href="https://application.wiley-vch.de/books/sample/3527407561_c01.pdf"><strong>unpredictability</strong> of epilepsy</a> and the associated stress caused by not knowing when the next seizure will occur.</li><li>Managing <a href="https://www.sciencedirect.com/science/article/pii/S1059131116303247">psychosomatic stress</a>, <a href="https://www.epilepsy.com/learn/professionals/resource-library/tables/common-and-rare-aed-side-effects">treatment side effects</a> and <a href="https://www.cdc.gov/mmwr/preview/mmwrhtml/mm6243a2.htm">epilepsy comorbidities</a>.</li></ol><p>The last bullet point may seem like a grab bag of issues, but the overarching theme is to maintain or improve someone’s <strong>quality of life</strong> when developing a care path.</p><p>We did not identify these unmet needs in a vacuum. <a href="https://www.byteflies.com/">Byteflies</a> interacts on a nearly daily basis with a network of experts, including people with epilepsy, neurologists, and pharma and academic researchers. Recently, we organized the <strong>Byteflies Epilepsy Day</strong> which included multiple panel discussions that touched on these topics. You can find the full event recording <a href="https://youtu.be/kRnclUe-zjQ">here</a> and a teaser video below.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FGIX3ZFK7CZY%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DGIX3ZFK7CZY&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FGIX3ZFK7CZY%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/ee6d24b78cba2a282279a0a92f701114/href">https://medium.com/media/ee6d24b78cba2a282279a0a92f701114/href</a></iframe><p>Byteflies develops wearable health solutions. In other words, medical applications that derive data from small wearable sensors and turn that long-term data into actionable clinical insights for healthcare professionals and patients. We are a <a href="https://www.dimesociety.org/index.php/about-us-main/defining-digital-medicine">digital medicine</a> company that takes pride in <strong>creating patient value</strong>. That means we let our technology development (the <em>What</em>) be driven by the need of the patient and healthcare professional (the <em>Why</em>).</p><blockquote>The “Why” should always guide the “What”.</blockquote><p>In that spirit, before diving into some of the specifics of Byteflies’s technology in the next section, let’s focus first on these unmet needs.</p><h4>State-of-the-art care</h4><p>Most people with a suspected or confirmed diagnosis of epilepsy will at some point undergo a video-electroencephalography (video-EEG) observation. This test requires admission to a specialized epilepsy monitoring unit (EMU) for 1–7 days. During that time, the following data is typically recorded continuously:</p><ul><li>Twenty or more channels of electroencephalography (<a href="https://en.wikipedia.org/wiki/Electroencephalography">EEG</a>), i.e. brain activity as measured on the scalp</li><li>Video-audio feed</li><li>Electrocardiography (<a href="https://en.wikipedia.org/wiki/Electrocardiography">ECG</a>), i.e. heart activity to derive heart rate and sometimes <a href="https://www.ahajournals.org/doi/full/10.1161/01.cir.93.5.1043">heart rate variability</a></li><li>Electromyography (<a href="https://en.wikipedia.org/wiki/Electromyography">EMG</a>), i.e. muscle activity</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*njVaXPfztEDFzEA2O2UvHA.jpeg" /><figcaption>Video-EEG system with scalp electrodes being set up for data capture.</figcaption></figure><p>All this biopotential data is sampled at a high frequency (typically 250 times per second and higher) and then reviewed by EEG readers to identify potential epileptic events.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*tsiSnB0QDEf_qmI-MtLSlA.jpeg" /><figcaption>Example of an EEG trace with video feed for identification of epileptic events. Vertical markers indicate one second of data. Biopotentials, such as EEG, EMG and ECG, are generated due to the electrochemical activity of groups of cells. EEG measures biopotentials generated by brain cells (neurons), EMG originates from skeletal muscle cells, and ECG from heart muscle cells (cardiomyocytes). Specialized biopotential electrodes and sensitive amplifiers are needed to record these signals on the skin surface.</figcaption></figure><p>Since epileptic seizures originate in the brain, EEG is obviously an important signal to evaluate. However, it is often not enough. Seizures come in many shapes and sizes and detecting them on the EEG signal is not always straightforward. They may originate in deep brain structures that are not readily picked up by a scalp EEG measurement or the signal could be obscured by motion artifacts<strong>*</strong>. Because many seizures will provoke involuntary and rhythmic movements patterns (known as <a href="https://www.epilepsydiagnosis.org/seizure/motor-overview.html">motor patterns</a>), the video feed is an important secondary signal, as is the EMG signal. Additionally, some seizures may evoke certain involuntary stereotypical behaviors, such as lip smacking, chewing, or pacing, known as <a href="https://www.epilepsy.com/learn/types-seizures/focal-onset-impaired-awareness-seizures-aka-complex-partial-seizures">automatisms</a>. These automatisms are also easy to recognize on video. Finally, the seizure may also affect the <a href="https://en.wikipedia.org/wiki/Autonomic_nervous_system">autonomic nervous system,</a> which is responsible for regulating many unconscious bodily functions, including the heart and respiratory rate. By recording the ECG, any changes in heart rate during the seizure can be evaluated.</p><blockquote><strong>*</strong>Entire books have been written on this subject and it is beyond the scope of this post to go into more detail. In case you are interested, the <a href="https://www.sciencedirect.com/book/9781455712670/atlas-of-sleep-medicine">Atlas of Sleep Medicine</a> and <a href="https://connect.springerpub.com/content/book/978-1-6170-5278-1">Ambulatory EEG Monitoring</a> are useful sources.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*jlGbLQSkWYMX91FQvMXwkQ.jpeg" /><figcaption>The most common signals captured during a video-EEG observation.</figcaption></figure><p>When applied correctly, video-EEG can be a powerful diagnostic tool but –as always– there are some downsides too:</p><ul><li>A (potentially long) hospital admission is required.</li><li>It requires access to a hospital with a specialized epilepsy monitoring unit and it can be expensive.</li><li>There is no guarantee someone will have a seizure during their stay.</li><li>The person is monitored under artificial conditions that do not reflect their usual environment and activities.</li></ul><p>After that moment of intense observation, the patient goes back home, and all that fancy equipment is replaced by self-reporting via a <strong>seizure diary</strong>, which can be a phone application or simply pen-and-paper. Seizure diaries require an active effort to track seizures and symptoms, and despite people’s best efforts, they are known to be inaccurate (some examples: <a href="https://pubmed.ncbi.nlm.nih.gov/22652423/">1</a>, <a href="https://jamanetwork.com/journals/jamaneurology/fullarticle/794768">2</a>). Consequently, many neurologists will not pay much attention to them.</p><p>So, what is our <em>Why </em>(unmet need) driver? People with epilepsy and their healthcare professionals are <strong>missing data</strong> when video-EEG equipment is not available. Either because no epilepsy monitoring units are available within a reasonable distance and –if they are– the hassle or expense of repeatedly coming to the hospital is simply too large. That missing data could have been used to:</p><ul><li>More quickly identify an optimal and <strong>personalized treatment </strong>strategy</li><li>Understand if the <strong>dosing </strong>of a drug is right</li><li>Find a balance between<strong> seizure freedom and side effects</strong></li><li>Track changes in <strong>quality of life</strong> and comorbidities over extended periods of time</li></ul><p>The resulting <em>What</em> is our ability to record seizures objectively for prolonged periods of time with unobtrusive and cost-effective equipment outside the hospital, either because advanced equipment is not available or to fill in the gaps between video-EEG observations.</p><p>Several options already exist or are being developed to make this possible:</p><ul><li><strong>Ambulatory EEG</strong>: Similar to the EEG equipment used in the hospital but typically in a more portable format so people can move around at home.</li><li><strong>Subcutaneous EEG</strong>: One or two EEG channels are recorded under the scalp which holds promise for making ultra-long-term EEG recordings but requires a surgical intervention.</li><li><strong>Consumer wearables</strong>: Fitness trackers or wellness devices may in some cases complement seizure diaries.</li><li><strong>Specialized medical wearables</strong>: Medical devices that come in a wide variety of shapes, sizes, and functionality that are purposely built for people with epilepsy for mid-to-ultra long-term use.</li></ul><p>Each of these innovations have their pros and cons and we may dedicate a future post to exploring those in more detail. Byteflies Sensor Dot fits in the “specialized medical wearables” category. We aim to fill the gap between the hospital video-EEG and home seizure diary with user-friendly wearable devices that can record analogous signals compared to the video-EEG reference standard at home. More on that <a href="#dfdd">below</a>!</p><blockquote>Recent improvements in cost-effective and unobtrusive seizure monitoring devices are extending state-of-the-art epilepsy care to the home environment and geographic areas without access to specialized facilities.</blockquote><h4>New treatment options</h4><p>Many different types of epilepsy and seizures are known but even in the most recent effort to <a href="https://onlinelibrary.wiley.com/doi/full/10.1111/epi.13670">improve their classification</a>, an “unknown” category remains. As discussed previously, seizures vary in their location of onset and migration patterns in the brain, as well as the symptoms they elicit and their duration. Many occur without warning, but they can also be triggered by environmental factors, such as sleep quality, stress, alcohol, medication, hormonal changes, and the list goes on. Although most seizures only last for a few seconds to minutes, their effect on a person’s well-being can linger for hours (known as the <a href="https://www.ilae.org/journals/epigraph/epigraph-vol-22-issue-2-spring-2020/after-the-wave-subsides-post-ictal-effects">post-ictal phase</a>).</p><p>Suffice it to say, epilepsy is complex and can be extremely hard to treat effectively because of interindividual but also intraindividual differences (e.g., driven by lifestyle factors). Despite massive investments in the development of new antiepileptic drugs, as well as therapies such as vagal nerve stimulation (<a href="https://www.epilepsy.org.uk/info/treatment/vns-vagus-nerve-stimulation">VNS</a>) and responsive neurostimulation (<a href="https://www.epilepsy.com/learn/treating-seizures-and-epilepsy/devices/responsive-neurostimulation-rns">RNS</a>) devices, the <a href="https://onlinelibrary.wiley.com/doi/full/10.1111/j.1528-1167.2009.02042.x">last four decades have brought about little meaningful improvement</a> in our ability to control seizures for people with epilepsy.</p><p>Echoing this from the patient’s perspective, one of the most common remarks noted in a <a href="https://www.epilepsy.com/sites/core/files/atoms/files/community-survey-report-2016%20V2.pdf">survey conducted by the Epilepsy Foundation in 2016</a> was “<em>Frustration of trying different medications and their side effects with no improvement</em>”.</p><p>Where does that leave us when we repeat our <em>Why → What </em>exercise? Pharma and academia are still investing considerable resources in the identification of novel drugs and mechanisms of action. The latter is especially important because our understanding of what specifically causes epilepsy (down to the neurochemical signaling level) is still incomplete. These efforts should continue, but two important improvements can be made right now:</p><ul><li>In many cases, treating epilepsy is more than treating seizures, especially for treatment resistant individuals. Put differently: existing care paths can be made more <strong>holistic</strong> with an increased focus on long-term outcomes and quality of life.</li><li>Collection of <a href="https://www.fda.gov/science-research/science-and-research-special-topics/real-world-evidence">real-world data</a> (i.e. RWD, data that is collected outside the lab or clinical trial) has a lot of potential to improve our ability to efficiently match existing treatments to the right individuals. Put differently: existing care paths can be made more <strong>personalized</strong>.</li></ul><p>Continuing with the <em>What</em>, these ideas are not unique to epilepsy and similar efforts are underway for other chronic conditions and rare diseases (background reading: <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6708660/">1</a>, <a href="https://www.nature.com/articles/s41573-019-0049-9#Sec33">2</a>, <a href="https://www.annualreviews.org/doi/10.1146/annurev-publhealth-040617-014158">3</a>). Here, we can keep it brief: to collect RWD on seizures, triggers, side effects, and quality of life, on a sufficiently large scale, the same technological innovations as mentioned in the previous paragraph apply. Currently, there is no hard evidence that this will lead to immediate improvements in the standard of care but considering the advantages discussed in the previous section, there appears to be little downside exploring these strategies further. A secondary more long-term effect may be the identification of novel therapeutic targets based on the collection of all this data.</p><blockquote>To understand how to make existing care paths more personalized and holistic with the treatment strategies currently available, we need to improve our ability to capture relevant long-term data unobtrusively in the real-world.</blockquote><h4>Reducing unpredictability</h4><p>In that same survey mentioned <a href="https://www.epilepsy.com/sites/core/files/atoms/files/community-survey-report-2016%20V2.pdf">previously</a>, the <strong>unpredictability </strong>of seizures is one of the main factors identified as impacting quality of life by people experiencing frequent seizures.</p><p>Predicting when a seizure is about to happen is often considered the “Holy Grail” because it would return some control to people with epilepsy. It could also make certain acute forms of treatment more plausible: they would be administered just before a seizure hits, either manually or automatically, in the hope of preventing it from occurring altogether or at the very least reducing its severity (<a href="https://www.ucb.com/stories-media/Press-Releases/article/UCB-announces-NAYZILAM-midazolam-nasal-spray-now-approved-by-FDA-to-treat-intermittent-stereotypic-episodes-of-frequent-seizure-activity-in-people-living-with-epilepsy-in-the-U-S">example</a>). Finally, it could help reduce or even prevent <a href="https://www.epilepsy.com/learn/early-death-and-sudep/sudep">SUDEP</a>, sudden mortality associated with uncontrolled seizures.</p><p>You may have noticed the overuse of conditionals in that last paragraph. Attempts to predict seizures date back to the 1970s and recently have gone through a revival due to the availability of larger datasets and machine learning methods to process them (examples: <a href="https://www.kaggle.com/c/seizure-detection">1</a>, <a href="https://ai.kuleuven.be/news/201cbiomed-irregulars201d-won-the-first-place-in-neureka-challenge">2</a>). Active large-scale efforts include <a href="https://pubmed.ncbi.nlm.nih.gov/29291239/">My Seizure Gauge</a>, <a href="https://www.radar-cns.org/">RADAR-CNS</a>, <a href="https://eithealth.eu/project/seizeit2/">SeizeIT2</a>, as well as improvements in our understanding of cyclical seizure patterns based on ultralong-term EEG monitoring (examples: <a href="https://pubmed.ncbi.nlm.nih.gov/28899023/">1</a>, <a href="https://pubmed.ncbi.nlm.nih.gov/32219856/">2</a>, <a href="https://pubmed.ncbi.nlm.nih.gov/32852091/">3</a>). These recent efforts are important, but it is unlikely they will lead to a rapid breakthrough in our ability to forecast seizures for a large group of people with epilepsy.</p><p>The <em>Why</em> is again straightforward: not knowing when a seizure is about to hit leads to <strong>anxiety </strong>and <strong>stress</strong>, which in turn may increase people’s chances of having a seizure. Having uncontrolled seizures may cost someone their driving license and limit their professional opportunities. Knowing that clinically applicable seizure prediction is still quite far off and may only work for specific people or seizure types, does it even make sense to discuss it now?</p><p>It is a typical chicken-egg problem. Recent advances in wearable sensor technology and data analytics may improve our seizure prediction capabilities but to do so, we need to collect lots and lots of high-quality data. However, more short-term benefits may arise. Many people with epilepsy will experience certain telltale symptoms (“auras”) that alert them a seizure may be coming. Similarly, a lot of information about seizure triggers and risk factors may be hidden in daily activities. The only people qualified to identify those are people with epilepsy or their most trusted relations, but technology can help. Providing user-friendly seizure and symptom cataloguing apps and devices may allow people to uncover a lot of these patterns themselves while creating trust in the healthcare community that the information is backed up by objective verifiable seizure monitoring data.</p><blockquote>The unpredictability of seizures is an especially insidious part of living with epilepsy. New seizure monitoring technology can play a crucial role in removing some of that unpredictability on the level of the individual. Over time, these data collection efforts may turn into more generalizable takeaways and pave the way to more robust seizure predicting capabilities.</blockquote><h4>Improving quality of life</h4><p>Last but not least, a lot of the devices discussed in the <a href="#6992">state-of-the-art care section</a> can also capture other objective <a href="https://www.karger.com/Article/FullText/500413">digital measures</a> such as sleep quality, stress levels, and even environmental parameters such as exposure to loud noise or light patterns. Overlaying this information on top of seizure patterns and other epilepsy symptoms may play a key role in tackling the three other unmet needs. Exactly how that information can be collected as unobtrusively as possible and integrated into care pathways, remains to be seen, and it will take an active effort from healthcare providers to not only reward the treating of core symptoms, but also long-term outcomes, including quality of life.</p><blockquote>Adding objective quality of life measures to routine clinical care for people with epilepsy will require an active effort from the entire community, including regulators and payers.</blockquote><h3>How can Byteflies Sensor Dot meet this need?</h3><p>We set the stage on some of the unmet clinical needs in epilepsy that could be met by advances in digital medicine and specifically remote monitoring technology. The number of devices and applications is diverse, and many already have or hopefully will soon have a positive impact on the management of epilepsy.</p><p>At Byteflies, we develop medical wearable devices that bring the quality of equipment traditionally used inside the hospital to the home environment in a user and home-friendly package. Over the last couple of years, we have been hard at work to build a wearable solution that replicates relevant video-EEG capabilities for use at home.</p><h4>How it works</h4><p>Sensor Dot is our multipurpose wearable that can record biopotential and motion (<a href="https://en.wikipedia.org/wiki/Inertial_measurement_unit">inertial</a>) signals. It is cleared in the EU (CE) and US (FDA 510(k)) as a prescription medical device.</p><p>Sensor Dot is compatible with the biopotential electrodes that are used for hospital EEG but we are also developing disposable <em>patches </em>(adhesives with embedded biopotential electrodes) that are easier to use at home and allow longer monitoring periods.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*n5PXmL_i_J3nX3H2sKxQzA.png" /><figcaption>Byteflies Sensor Dot placed in the neck with bilateral EEG patches. This configuration also measures motion.</figcaption></figure><p>The setup shown in the image above is recording two channels of EEG, one behind each ear (but other configurations are possible), as well as motion via the inertial sensors. A second Sensor Dot can add ECG, EMG, and an additional set of motion sensors.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aA_OeJKFyPo71x8whq_ybg.jpeg" /><figcaption>Byteflies Sensor Dot placed on the chest with a cardiorespiratory patch to record heart rate, respiratory rate, and motion. Image is used with permission from copyright owners Stad Antwerpen and Frederik Beyens.</figcaption></figure><p>The setup with two Sensor Dots measures signals that are the same or equivalent to what is recorded during a video-EEG observation, with the notable difference that the number of EEG channels is lower than what is used in the hospital. In other words, Sensor Dot is not meant to be a replacement for video-EEG equipment. On the contrary, it is meant to complement it by:</p><ol><li>Making the recording of relevant medical-grade signals possible in a form factor that is <strong>small </strong>enough for use in daily life.</li><li>Making a <strong>cost-effective</strong> device available that can be used in locations or under circumstances where video-EEG equipment is not available.</li><li>Increasing the number of day-to-week long seizure <strong>observation periods </strong>that can be made.</li><li>Recording data that is compatible with existing <strong>diagnostic </strong>video-EEG data review workflows, as well as future automated data processing pipelines.</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lZXeZwEycDJOA5P8Z6v9vw.jpeg" /><figcaption>Signal comparison of video-EEG (hospital) and Sensor Dot (home) observations. <strong>#</strong> Sensor Dot can record two channels of EEG data, typically behind-the-ears but other electrode placements are possible. <strong>‡</strong> Measuring ECG and/or EMG requires a second Sensor Dot.</figcaption></figure><p>The current generation of Sensor Dot is ready to tackle these goals in the real-world. At the time of writing (January 2021), we have monitored over 150 people with this set up (and counting), accounting for more than 2000 focal and generalized seizures. Of course, we continue to improve our technology and work with all stakeholders to make sure the <em>Why </em>keeps fitting the <em>What</em>. For instance, with increased data recording capabilities comes the need to automate the data review process to avoid overloading neurologists with datasets they do not have the time to review.</p><blockquote>The current generation of Sensor Dot is ready to tackle these goals in the real-world. Sensor Dot has monitored more than 150 people with epilepsy accounting for more than 2000 seizures.</blockquote><h3>Summary</h3><p>We hope this post helps in clarifying why we are developing Sensor Dot for epilepsy the way we are. A seizure monitoring tool that extends the capabilities of the traditional video-EEG set up beyond the walls of the hospital a few weeks at a time.</p><p>We are incredibly excited to continue our work with the epilepsy community to get Sensor Dot in the right hands. Please follow us on this <a href="https://medium.com/byteflies">blog</a>, <a href="https://twitter.com/byteflies">Twitter</a>, <a href="https://www.linkedin.com/company/byteflies">LinkedIn</a>, or any of our other social media channels to stay up to date, and feel free to reach out via our <a href="https://www.byteflies.com/contactform">website</a>!</p><h4>Further reading</h4><p>The introduction of this post only touches briefly on some important topics. If you are interested in learning more, we recommend the following sources:</p><ul><li><a href="https://www.epilepsy.com/learn/about-epilepsy-basics">The basics of epilepsy by the American Epilepsy Foundation</a> (and basically all other educational materials on their website).</li><li><a href="https://www.who.int/news-room/fact-sheets/detail/epilepsy">The Epilepsy Fact Sheet by the World Health Organization</a> (WHO).</li><li><a href="https://www.ilae.org/education">The Education section on the website of the International League Against Epilepsy</a> (ILAE), including <a href="https://www.epilepsydiagnosis.org/">https://www.epilepsydiagnosis.org</a>.</li><li><a href="https://www.epilepsy.com/living-epilepsy/seizure-first-aid-and-safety/first-aid-seizures-stay-safe-side">The Seizure First Aid training organized by the American Epilepsy Foundation</a>. Everyone at Byteflies has completed this course and it is a fantastic way to learn more about epilepsy and become seizure first aid certified in the process!</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/175/1*AXrlHNJ9z-wG5v49_TxurQ.jpeg" /></figure><p><em>The information contained in this article represents the views and opinions of the writer(s) and does not necessarily represent the views or opinions of other parties referenced or mentioned therein.</em></p><p><em>The article is not intended to be a substitute for professional medical advice, diagnosis, or treatment. Always seek the advice of your physician or other qualified healthcare professional with any questions you may have regarding a medical condition. Never disregard professional medical advice or delay in seeking it because of something you read in this article.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9b3c55f56dfc" width="1" height="1" alt=""><hr><p><a href="https://medium.com/byteflies/what-is-byteflies-building-for-people-with-epilepsy-and-why-9b3c55f56dfc">What is Byteflies Building for People with Epilepsy and Why?</a> was originally published in <a href="https://medium.com/byteflies">Byteflies</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Continuous Cardiorespiratory Monitoring for People with Cystic Fibrosis]]></title>
            <link>https://medium.com/byteflies/continuous-cardiorespiratory-monitoring-for-people-with-cystic-fibrosis-52236a66d7a5?source=rss-bb22181f7c06------2</link>
            <guid isPermaLink="false">https://medium.com/p/52236a66d7a5</guid>
            <category><![CDATA[respiratory-disease]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[wearables]]></category>
            <category><![CDATA[cystic-fibrosis]]></category>
            <category><![CDATA[digital-biomarkers]]></category>
            <dc:creator><![CDATA[Benjamin Vandendriessche]]></dc:creator>
            <pubDate>Thu, 19 Nov 2020 17:04:49 GMT</pubDate>
            <atom:updated>2020-11-19T17:09:37.853Z</atom:updated>
            <content:encoded><![CDATA[<p>Researchers from <a href="https://www.byteflies.com/">Byteflies</a> and <a href="https://www.uhhospitals.org/rainbow">Rainbow Babies and Children’s Hospital</a> at University Hospitals Cleveland Medical Center set out to answer the question:</p><blockquote>Can a Byteflies Sensor Dot monitor relevant cardiorespiratory parameters in people with <strong>Cystic Fibrosis</strong> (CF) and could this lead to improvements in the standard of care?</blockquote><p>That question is not an easy one to answer as we need to determine if Sensor Dot:</p><ol><li>Can measure relevant sample-level¹ physiologic and behavioral signals;</li><li>Is user-friendly enough for home use <em>and </em>is compatible with clinical workflows; and</li><li>Can derive meaningful physiologic and behavioral patterns (<a href="https://www.karger.com/Article/FullText/500413#scrollNav-3-5">digital measures</a>) that correlate with relevant outcomes in CF.</li></ol><p>Most of these questions were treading uncharted territory so we ran a pilot study to obtain preliminary answers, which will hopefully lead to a more formal validation study in the future.</p><p>This article summarizes the main findings of that pilot study but first a couple of housekeeping notes:</p><ol><li>Our pilot study was not intended to generate a definitive answer to any of these questions. It was an expedition to evaluate if we were even asking the right questions. Therefore, this work was <em>not </em>peer-reviewed but rather <a href="https://www.medrxiv.org/content/10.1101/2020.07.27.20151589v1">published as a preprint</a> on medRxiv. A full-length paper is embedded <a href="#a7fa">at the end of this post</a>.</li><li>The study protocol is available on <a href="https://clinicaltrials.gov/ct2/show/NCT04489186">clinicaltrials.gov</a>.</li><li>In addition to <a href="https://www.byteflies.com/">Byteflies</a> and <a href="https://www.uhhospitals.org/rainbow">Rainbow Babies and Children’s hospital</a>, researchers from <a href="https://case.edu/medicine/">Case Western Reserve University</a>, <a href="https://medicine.iu.edu/">Indiana University</a>, <a href="https://www.uhasselt.be/UH/lcrp-en/Scientific-platforms/Mobile-Health-Unit.html">Hasselt University</a>, and <a href="https://www.cf-europe.eu/">Cystic Fibrosis Europe</a> participated in analyzing the data and authoring the paper.</li></ol><p>¹ Sample-level denotes the raw data we can extract from a sensor, such as electrocardiogram data which was sampled at 125 times/second.</p><h3>Table of Contents</h3><ul><li><a href="#e5fd">Why are wearable devices important for people with CF?</a></li><li><a href="#aa05">Is Byteflies Sensor Dot the right tool for the job?</a></li><li><a href="#13fa">Conclusion</a></li><li><a href="#a7fa">Full Paper</a></li></ul><h3>Why are wearable devices important for people with CF?</h3><p>Cystic Fibrosis (CF) is a genetic disorder that affects multiple organs but primarily the lungs (the <a href="https://en.wikipedia.org/wiki/Cystic_fibrosis">CF Wikipedia page</a> is a good place to learn more). Unfortunately, a cure does not exist (yet!) and people with CF require lifelong specialized care.</p><p>The <a href="https://www.cff.org/">CF Foundation</a> recommends that people with CF visit their care center at least four times a year and perform at least two pulmonary function tests (PFTs) per year (more on that later). And that is assuming they are otherwise in good health. Between these visits, as with many chronic conditions, little objective data on people’s lung function, other symptoms, and quality of life is available. More specifically, <a href="https://www.cff.org/Care/Clinical-Care-Guidelines/Respiratory-Clinical-Care-Guidelines/Pulmonary-Exacerbations-Clinical-Care-Guidelines/">pulmonary exacerbations</a> (PEx) are a frequent source of hospitalization. Early detection and treatment of PEx are especially important to prevent long-term problems.</p><p>Knowing that, having an unobtrusive wearable home-use device to monitor for PEx or other signs that indicate the need for a hospitalization could lead to earlier intervention, and – over time – a better understanding of that person’s <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3131847/#:~:text=A%20health%20trajectory%20describes%20the,as%20a%20function%20of%20time.">health trajectory</a>, which is incredibly valuable information for any care team to have.</p><h3>Is Byteflies Sensor Dot the right tool for the job?</h3><p>For this study, we used a prototype version of Sensor Dot that can record the following sample-level signals:</p><ul><li>Electocardiography (ECG)</li><li>Bio-impedance (bioZ)</li><li>Accelerometer in 3-dimensions</li></ul><p>From those sample-level signals, we could derive the following <a href="https://www.karger.com/Article/FullText/500413#scrollNav-3-5">digital measures</a>:</p><ul><li>Heart rate (HR) in beats-per-minute</li><li>Respiratory rate (RR) in breaths-per-minute</li><li>Activity index as a measure of total activity</li><li>Number of coughs</li></ul><p>Sensor Dot was placed on the chest and connected to two standard ECG electrodes as shown in the figure below. After a measurement, a Sensor Dot is placed back on the Docking Station which sends the collected data to the Byteflies Cloud.</p><blockquote>If you are interested in the technical details of how the signals were recorded and the digital measures derived, please <a href="#a7fa">see the full paper</a> at the end of this post.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*26ihT7NpLva0oRuDg20meA.png" /><figcaption>The Byteflies Sensor Dot system, as used in the pilot study.</figcaption></figure><p>In total, 26 people with CF who had a scheduled clinic appointment participated in the study. They all performed a <a href="https://www.uhhospitals.org/health-information/health-and-wellness-library/article/tests-and-procedures/pulmonary-function-tests">pulmonary function test</a> (PFT) which assesses lung function. Based on the outcome of the PFT, participants were divided into three severity categories (MILD, MODERATE, and SEVERE).</p><p>After that necessary setting the stage, back to our original questions:</p><h4>1. Can Sensor Dot measure relevant signals?</h4><p>The figure below shows a 22 h trace of the digital measures highlighted above (heart rate, respiratory rate, activity index, and number of coughs) for a 53-year-old male in the MODERATE severity category. We can make a couple of observations:</p><ul><li><strong>Heart rate</strong> (top graph) goes up with increases in activity index (bottom graph), as expected, and reached a maximum value of 132 BPM.</li><li><strong>Resting HR</strong> appears to be elevated (~ 80 BPM during the sleep phase).</li><li><strong>Resting respiratory rate </strong>is around 16–17 BrPM and goes up to 20–21 BrPM during the day.</li><li>Two <strong>coughs </strong>were detected (dashed marker on middle graph; more on that later).</li><li><strong>Activity </strong>is low throughout the day because this person was staying overnight in the hospital.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_YgsQomOPuLcPlDo_K6UeQ.png" /><figcaption>A trace from a participant in the study, portraying heart rate in beats-per-minute (BPM, top), respiratory rate in breaths-per-minute (BrPM) and cough counts (middle), and activity index (bottom) as recorded by Sensor Dot.</figcaption></figure><p>These longitudinal overviews are interesting, especially if you consider the possibility to track these vital signs over extended periods of time to evaluate long-term trends but that was beyond the scope of our pilot. What is even more interesting is to examine if some of these vital signs correlate with CF severity as used in routine clinical care. More on that in section 3!</p><h4>2. Is Sensor Dot user-friendly enough for home use <em>and </em>is it compatible with clinical workflows?</h4><p>Any evaluation of an innovative technology would be grossly incomplete without also checking its future users are happy with the form factor and functionality. Especially during the initial stages of development so that changes can still be made to the design if necessary.</p><p>Importantly, <em>users </em>in this context not only refers to patients, but also the professionals in their care team. They have differing but complementary requirements that need to be satisfied to make a digital medicine tool successful.</p><p>This pilot study was run inside the hospital, which gave us an opportunity to evaluate Sensor Dot in the hands of healthcare professionals. And of course, the patients themselves were asked for feedback (24 people filled out the survey after using Sensor Dot). This feedback was overall very positive (see figure below), but it is important to note that we only assessed usability over short stretches of time (1 day at the most) and in the hospital.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Y5ds-BdxS4c6j5uxGYgyJw.png" /><figcaption>Patient feedback survey completed after using Sensor Dot for the first time.</figcaption></figure><h4>3. Can Sensor Dot derive meaningful digital measures that correlate with relevant outcomes in CF?</h4><p>As mentioned earlier, the group of participants (26) were divided into MILD, MODERATE, and SEVERE categories based on their PFT results. We evaluated if the digital measures that we recorded with Sensor Dot correlated with these severity categories. We made the following observations (see Figure 4 in the full paper for details):</p><ul><li>The <strong>heart rate</strong> of individuals in the SEVERE group was increased, especially when people were active.</li><li>Some participants were admitted to the hospital for a PEx (inpatients), whereas others came in for routine check-ups (outpatients). When evaluating the average <strong>δHR</strong>, which is the difference between the heart rate during activity and at rest, the δHR in the inpatient group is higher than the outpatients for all severity categories. This indicates that we may be able to capture the cardiorespiratory effects of a PEx in the HR response to physical activity.</li><li>At rest, <strong>respiratory rate</strong> in the SEVERE group was increased.</li><li>We counted <strong>coughs </strong>with a very experimental technique. The details are in the <a href="#a7fa">full paper</a> but briefly: we exploited the fact that the accelerometer in a Sensor Dot is placed on the chest. Consequently, a cough results in a transient pattern on the accelerometer. Because this is a proof-of-concept technique, we asked participants to self-report the number of times they coughed, and these numbers correlated quite well with the calculated numbers. In addition, the inpatients coughed on average ~10 times, while for the outpatients this was ~5 times. Since many PEx are associated with an increase in coughing, this is not a surprising observation but the fact that we can detect this trend with Sensor Dot is potentially interesting for longer-term home monitoring where self-reported monitoring is simply not practical.</li></ul><p>Summarized, our observations are as follows:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RT0nXzJGIkFy7Y3jkU24sg.png" /><figcaption>The observed changes in vital sign in SEVERE in- and outpatients at rest or during mild activity, compared to the same vital signs in the MILD and MODERATE categories (left) and the difference in δHR and number of coughs in inpatients vs outpatients (right).</figcaption></figure><h3>Conclusion</h3><p>We ran a pilot study with Sensor Dot to assess if it can monitor relevant cardiorespiratory parameters in people with CF, and if this could potentially lead to improvements in the standard of care. Although our results are preliminary, they provide us with a first guideline on which parts of the wearable system – from hardware to data processing and visualization – need to be improved to make it fit-for-purpose for CF and other chronic respiratory conditions, such as <a href="https://en.wikipedia.org/wiki/Chronic_obstructive_pulmonary_disease">Chronic Obstructive Pulmonary Disease</a> (COPD).</p><p>Development processes for medical wearables can be long as they require very thorough <a href="https://www.nature.com/articles/s41746-020-0260-4">validation</a>. The more stakeholders that are involved early in the development process, the better all performance and usability requirements can be considered.</p><p>Case in point, the wired electrodes used for this pilot study have since been replaced with next-gen cardiorespiratory patches (stay tuned for more information!).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*r-X3jCOj1zN7K0KztsSQcw.jpeg" /><figcaption>Byteflies Sensor Dot with a disposable Cardiorespiratory Patch. Image is used with permission from copyright owners Stad Antwerpen and Frederik Beyens.</figcaption></figure><p>Because this was a pilot study, it has a couple of limitations:</p><ul><li>The recordings were short (23 h long at most) and not yet done at the patient’s home.</li><li>The study was exploratory and therefore not powered to produce conclusive evidence for the digital measures and their potential correlation with CF severity.</li><li>Self-reported number of coughs are known to be inaccurate. To develop an accurate algorithm for cough counting, a much better reference standard should be used.</li></ul><h3>Full Paper</h3><iframe src="https://drive.google.com/viewerng/viewer?url=https%3A//cdn.nimbu.io/s/80lq45i/assets/1605644218815/cardiorespiratory-monitoring-CF_white-paper.pdf&amp;embedded=true" width="600" height="780" frameborder="0" scrolling="no"><a href="https://medium.com/media/5e49d34ea5230f02d4d704df6d1bf7f5/href">https://medium.com/media/5e49d34ea5230f02d4d704df6d1bf7f5/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=52236a66d7a5" width="1" height="1" alt=""><hr><p><a href="https://medium.com/byteflies/continuous-cardiorespiratory-monitoring-for-people-with-cystic-fibrosis-52236a66d7a5">Continuous Cardiorespiratory Monitoring for People with Cystic Fibrosis</a> was originally published in <a href="https://medium.com/byteflies">Byteflies</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Introducing the Byteflies Publication]]></title>
            <link>https://medium.com/@benvddriessche/introducing-the-byteflies-publication-175cbaea67b5?source=rss-bb22181f7c06------2</link>
            <guid isPermaLink="false">https://medium.com/p/175cbaea67b5</guid>
            <category><![CDATA[digital-health]]></category>
            <category><![CDATA[digital-medicine]]></category>
            <category><![CDATA[wearables]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Benjamin Vandendriessche]]></dc:creator>
            <pubDate>Thu, 19 Nov 2020 17:01:19 GMT</pubDate>
            <atom:updated>2020-11-19T17:01:19.911Z</atom:updated>
            <content:encoded><![CDATA[<p>This is the official Medium Publication for <a href="https://www.byteflies.com/">Byteflies</a>, focused on data- and digital health-minded people.</p><p>In case you never heard of us:</p><blockquote>Byteflies envisions a future where healthcare is truly personalized and preventive. Our solution enables continuous monitoring of synchronized vital signs and digital biomarkers with user-friendly wearable patches, no matter where you are and tailored to your specific needs. Byteflies is developing wearable products for people with epilepsy, sleep disorders, cardiorespiratory and neurodegenerative conditions, and acute viral infections.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/854/1*y3FczoqRLkeej9gj0OemNQ.png" /><figcaption>Byteflies Sensor Dot, a small multimodal wearable device</figcaption></figure><p>Now that we have that squared away, <strong>why should you care?</strong> (Beyond Byteflies being an awesome company on a mission to change healthcare for the betterment of all mankind… of course!)</p><p>We are active in the field of Digital Health, and more specifically <strong>Digital Medicine</strong>. For an excellent primer on Digital Medicine, I urge you to check-out the <a href="https://www.dimesociety.org/index.php/about-us-main/defining-digital-medicine">About Page</a> of the Digital Medicine (DiMe) Society or this more in-depth <a href="https://www.karger.com/Article/FullText/500413">Article</a>.</p><p>Because we pride ourselves on our close interactions with clinicians, patients, and researchers to make sure we build solutions they actually need, we generate <em>a lot</em> of data. Some of that data makes it into scientific papers, but much of it is left behind after it served its primary purpose. Not because it is not interesting to the community, but it might be too early-stage for a full publication, or we simply didn’t have the time (start-up life can be that way). And let’s face it, the data that does make it into a scientific paper isn’t always the most accessible.</p><p>Without further ado, introducing… the <strong>Byteflies Pub</strong>(lication)!</p><p>A place to discuss interesting Sensor Dot datasets, proof-of-concept studies, and TL;DR summaries of papers we published or would like to give a shout-out to. It is our hope this will develop into a <strong>data-focused resource</strong> for the Digital Medicine community. Please subscribe if interested and feel free to reach out!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=175cbaea67b5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Modular Machine Learning in Healthcare]]></title>
            <link>https://medium.com/thelaunchpad/modular-machine-learning-in-healthcare-9674d53225a8?source=rss-bb22181f7c06------2</link>
            <guid isPermaLink="false">https://medium.com/p/9674d53225a8</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[wearables]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Benjamin Vandendriessche]]></dc:creator>
            <pubDate>Thu, 08 Nov 2018 20:01:32 GMT</pubDate>
            <atom:updated>2018-11-08T20:01:32.174Z</atom:updated>
            <content:encoded><![CDATA[<p>It’s hard to imagine developing software that isn’t modular. It would be extremely difficult to debug, troubleshoot, and implement. Modularity is also an important consideration when deploying machine learning (ML) models, though for different reasons. The main benefits of modular ML over end-to-end systems are that the troubleshooting process is more straightforward, the modules are reusable across different use cases, modular ML systems generally require less data, and their interpretability can be a key feature in certain applications.</p><p><a href="https://www.byteflies.com/">Byteflies</a> is a startup that generates medical insights by processing the raw data obtained from <a href="https://www.byteflies.com/product">Sensor Dots</a>, small wearable devices that can record clinical-grade physiologic signals. Traditionally, this type of waveform data is processed with explicit rule-based approaches but medical wearables are ushering in an era where it is becoming normal to record thousands of data points <em>every second</em> on a single person, day and night. Consequently, there is much more noise in the data compared to short-term recordings under controlled circumstances and more robust approaches are needed to process this data. This is exactly where ML techniques can be incredibly powerful if used correctly. This prompts the question: What data processing architecture to use?</p><p>In this piece, we cover how Byteflies has leveraged modularity in data pipelines and ML models, some of the benefits of using a modular approach, and the process of integrating these modular components.</p><p>Let’s start with a practical example. The continuous assessment of lung function is invaluable for chronic respiratory disease patients to ensure they receive the best care possible. If we record cardiorespiratory vital signs, such as heart rate, respiratory rate, oxygen saturation (a measure of how effectively the lungs manage to oxygenate the blood), as well as measuring physical/behavioral activity, we can generate a very accurate measure of respiratory function without the need for expensive medical equipment or a visit to the hospital. But several factors complicate this process.</p><p>First, we record a signal that varies as a function of air in the lungs and stretch of the chest (bioimpedance for the experts). From this signal, which will be confounded by motion and speech artifacts, we calculate respiratory rate. Next, we derive heart rate from the electrocardiogram, oxygen saturation using optical techniques, and activity type classification from the inertial measurements. Each of these signals has its own set of idiosyncrasies and we did not even discuss the integration of these parameters into a single measure of lung function.</p><p>Traditionally, these processing pipelines were long, error-prone, and difficult to maintain. Knowing that, it sounds tempting to harness the power of deep learning by using the raw signals as input data for an end-to-end model that outputs “stable,” “worsening,” or “improving” lung function. If done well, this approach might well yield accurate predictions. However, we will argue why a modular approach makes much more sense for (wearable) healthcare applications.</p><h3>Troubleshooting</h3><p>The first benefit we saw in breaking the data pipeline into modular components was the ease of troubleshooting. Raw physiologic waveform data is incredibly feature-rich, which is both a blessing and a curse. Take for instance the diagnosis of cardiac arrhythmias (abnormalities in the electrical signal that makes the heart muscle contract) on electrocardiogram (ECG) data. When training an end-to-end model to classify raw ECG data, a very large dataset, high-quality labels, and careful feature engineering are necessary since arrhythmias are relatively rare events. If done well, the model can encode the relevant features of the ECG waveform and the relationships among them, which would result in a relatively deep neural net. Alternatively, we can train a simple pattern recognizer (convolutional neural nets, CNNs, are the method of choice) to focus on the most marked ECG feature (R-peak for the experts). Knowledge of the location of this feature allows us to segment the data and make educated guesses about the location of other features as their timing is bounded by physiologic rules. Consequently, the topology of those models can be even simpler.</p><h3>Less (Labeled) Data</h3><p>The hybrid approach (modular models chained together by a number of explicit rules) introduced above, significantly reduces the amount of labeled training data needed while preserving many of the benefits for using ML models in the first place. Labeling of medical and physiologic data generally requires highly trained experts which can be costly, especially for a company like Byteflies that is building a platform that needs to accommodate a wide range of applications.</p><p>Continuing the ECG example, arrhythmias are simply a combination of changes in timing and morphology of specific features that are very well known in medicine. Therefore, we can uncouple the problem of arrhythmia detection from the precise structure of the training waveform data, and focus on the timing of these specific features. Once we trained reliable and simple detectors for basic ECG features, which have neither massive data nor complex labeling requirements, we end up with a set of features per cardiac cycle. These feature vectors can be recombined to generate additional “artificial” training data for every possible arrhythmia known to cardiology.</p><p>Another way to think about this is to imagine again an end-to-end ML system. In that case, enough expert-labeled data is needed for the model to be able to “learn” many of the physiologic dependencies effectively from scratch. Medical science already has a great deal of knowledge about specific pathologies, and a modular structure increases the ability to pass this knowledge to the models.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*W3MZk-q0uI5N1DWG_Esh4Q.png" /><figcaption>ECG and respiration data recorded with a Byteflies Sensor Dot displaying <a href="https://en.wikipedia.org/wiki/Cheyne%E2%80%93Stokes_respiration">Cheyne-Stokes respiration</a>, another example of a specific pathological pattern in physiologic waveform data.</figcaption></figure><h3>Reusability</h3><p>A third major advantage is that the separate modules can be chained together in new analytics pipelines. Consider the example where we trained a simple model to detect the most marked ECG feature (R-peak), from which we can measure the heart rate. Heart rate is a vital sign with many different applications; retraining end-to-end models to incorporate a continuous notion of it would become cumbersome very quickly. Remember, although heart rate on its own is an important vital sign, its true value comes from integration with other vital signs and behavioral patterns to generate digital biomarkers (measurable indicators of health or disease, measured via digital means, that respond to a therapeutic intervention).</p><p>Secondly, consider a rare cardiac pathology that dramatically changes the morphology of one or more ECG features. The feature engineering balancing act when trying to train or improve an end-to-end model is likely to take up a lot of time and might also increase the likelihood of overfitting. The modular approach allows us to improve the existing simple feature detector model. It will still require careful consideration of the training data but evaluating the model performance will be very straightforward. This relates to the next section: interpretability.</p><h3>Interpretability</h3><p>To drive acceptance in the medical and regulatory community for these type of analytics pipelines, clinical validation and understanding how the classification result was generated, is crucial. If the output of a model that consists of multiple modules chained together does not match the expertise of a medical professional, she can simply trace back down the chain to understand how the model arrived at that conclusion in the first place. Returning to our respiratory function example, a downstream integrator model may trigger a warning for a decline in function because that person was exercising. Although that in itself is valuable medical information, the modular approach allows her to verify the outputs of the individual modules (heart rate, respiratory rate, oxygen saturation, and activity) to note that behavioral context. That level of granularity will not be provided by an end-to-end model.</p><p>Specifically for wearable health, we have found that a more modular data/ML pipeline has several advantages over an end-to-end approach, at least for the foreseeable future. This debate is far from settled, as discussed in this <a href="https://medium.com/predict/the-end-to-end-false-dichotomy-33a2bf02eda6">excellent post</a>. While a modular ML approach has worked well for Byteflies, our goal is to give the reader the context to decide what makes the most sense for his/her specific application. The healthcare sector is going through dramatic changes that will only accelerate over the years to come. Modular solutions will be one of the tools to drive adoption by the medical community to create real, positive patient impact.</p><p><a href="https://www.byteflies.com/">Byteflies</a> is a Belgian-American, AI-first medtech company on a mission to make healthcare truly personalized, predictive, preventive and participatory. We built a platform to instantly set up wearable health solutions that consists of our versatile and multimodal wearable device, <a href="https://www.byteflies.com/product">Sensor Dot</a>, infrastructure to manage and deploy wearable devices to thousands of patients, and analytics to transform raw physiologic data into digital biomarkers and clinical endpoints.</p><p>We work with pharmaceutical companies that are embracing clinical trial innovation and the value-based healthcare model, researchers looking for cutting-edge technology to monitor patients, and tech companies wanting to quickly push new wearable technology to the market.</p><p>Byteflies participated in the first class of <a href="https://developers.google.com/programs/launchpad/studio/">Google Developers Launchpad Studio</a>, a cohort focused on the applications of ML in healthcare and biotech.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9674d53225a8" width="1" height="1" alt=""><hr><p><a href="https://medium.com/thelaunchpad/modular-machine-learning-in-healthcare-9674d53225a8">Modular Machine Learning in Healthcare</a> was originally published in <a href="https://medium.com/thelaunchpad">The Launchpad</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>