Review of Mobile Devices and Health by Ida Sim in the NEJM

Mobile Devices and Health

Mobile Devices and Health was published yesterday in the New England Journal of Medicine. Authored by Ida Sim, M.D., Ph.D who is a Primary Care Physician, Professor at UCSF & coFounder at Open mHealth (follow her on Twitter @IdaSim).

It’s exactly the type of papers we need to see more of so please read it at the NEJM website and share that link with your colleagues and social networks:

mHealth Insights

“Mobile health — the application of sensors, mobile apps, social media, and location-tracking technology to obtain data pertinent to wellness and disease diagnosis, prevention, and management — makes it theoretically possible to monitor and intervene whenever and wherever acute and chronic medical conditions occur. With 81% of North American adults owning a smartphone,1 this frontier could be reached in the foreseeable future in the United States and is particularly relevant to the management of chronic diseases. More than 40% of U.S. adults have two or more chronic conditions2 and chronic conditions now account for 71% of all U.S. health care spending,3 so the promise of mobile health is especially attractive.”

While I clearly have a conflict of interest (I coined the term mHealth more than a decade ago and define it as ‘the leverage of mobile – the newest mass media – for health’) I think it’s rather obvious that this is not what Mobile Health means and by thinking it does you limit your capacity to understand the scale of opportunity that comes from taking a Mobile First approach to healthcare.

An obvious example of how such limited thinking comprises your ability to see the opportunity is demonstrated by how Dr Sim has presented the concept that mHealth is somehow limited by adoption of smartphones. We all should know by now that the most kind thing a carer can say to a Patient in 2019 with a chronic condition is “Here’s my mobile number” and that obviously doesn’t require anyone to have a smartphone as you can benefit from this if you just both have even the most basic mobile phones.

I realise that not everyone is like me and had a smartphone BEFORE they had a camera phone but it’s well worth medics appreciating that in 2019 even the most basic feature phones are more powerful and smarter than the latest smartphones from just a few years ago.

“Mobile health is at the swirling confluence of remote sensing, consumer-facing personal technologies, and artificial intelligence (AI). Data from smartphone applications (colloquially known as “apps”) and an ever-growing range of wearable and environmental sensors can be processed with the use of machine learning and other AI techniques to support medical decision making. Here, I review the current state of sensing, digital biomarkers, and digital therapeutics (the use of online technologies in the treatment of behavioral and medical conditions); discuss the challenges of integrating mobile health into clinical care; and describe regulatory, business, and ethical issues confronting mobile health. I do not discuss sensors and apps intended solely for use by health care professionals in health care settings. Because mobile health is a nascent technology and rigorous evidence of clinical validity is generally lacking, rather than presenting a review of existing systems, I present an overview for practitioners and policy makers to understand key aspects of this rapidly evolving field (see video)”

We live in a strange time when even popular tech journalists still think “we have no idea if Mobile Healthcare apps work’ but it surprises me that a lot of frontline medics still seem to think that ‘Mobile health is a nascent technology and rigorous evidence of clinical validity is generally lacking” because it’s not like these same medics would want to live their own lives without mobile technology.

I think this is largely because they think mobile health is synonymous with ‘healthcare apps on smartphones’ but also because even the most technical young medics have very little concept that there is anything inherently wrong in the outdated administrivia heavy IT systems they use everyday. Remember: even if no one is paying to publish papers in expensive journals stating this it is a fact that if Patients see you’re uncomfortable using mobile technology they may lose confidence in how current you are with your medical knowledge.

“Of passive sensors, the smartphone is the most ubiquitous. It has a nine-axis inertial motion sensor that tracks motion and position in three-dimensional space. A three-axis accelerometer measures acceleration in the x, y, and z axes; a three-axis gyroscope senses rotation around each axis; and a three-axis magnetometer compensates for magnetic drift to maintain position accuracy. These sensors enable physics-based capabilities, such as detecting the number of steps that a person takes during a day. Most smartphones can also sense geographic position, atmospheric pressure, ambient light, voice, and touchscreen pressure. Creative uses of these sensors and a built-in camera can turn the smartphone into a fall detector,4 spirometer5 (by sensing air pressure on the microphone), or heart-rate sensor.6″

I love made for mobile medical sensors and the smartphone medical is still an ever fascinating opportunity that’s still surprising me every week but it’s important to realise that the mobile has a personal secure browser and that is probably THE best way to collect a medical history from a Patient.

“Active Sensing. Passive sensors collect observable data. Subjectively perceived states of health (e.g., pain and other symptoms) are equally important for informing patient-centered care and, at this time, can be captured only by asking the patient. Until recently, information about such outcomes has been obtained from questionnaires administered at intervals of weeks to months that ask patients to integrate their experiences during some past interval of time (e.g., “in the past 7 days, how often … ?”). The ubiquity of personal devices makes possible an alternative approach called ecologic momentary assessment (EMA) that is well suited to capturing some types of patient-reported outcomes. EMA involves “repeated sampling of subjects’ current behaviors and experiences in real time, in subjects’ natural environments.”18 EMAs are less subject to recall bias than infrequently administered questionnaires and can be administered multiple times a day to capture short-term variations in responses. EMAs range from simple text-message prompts to short one- or two-item app-based questions and have been used to collect information on chronic pain,19 anxiety,20substance-use disorders,21 and many other conditions. Widely used in the social and behavioral sciences, EMA is an emerging method for outcomes assessment in both clinical care and clinical research. Newer image-based EMAs, if designed with cultural sensitivity, offer intriguing opportunities for bridging language, literacy, and numeracy barriers”

To say that mHealth is something that is only just emerging is nonsensical. Back in 2008 at 3GDoctor we started offering the ability for Patients to use their mobiles to provide their own history to Doctors prior to a documented video consult:

Eleven years on to describe this as emerging is plain denial of what’s happening in the USA (where Kaiser were the first big healthcare brand to go Mobile First in 2012 and Walgreens were generating $375Million a month in revenue from their smartphone app back in 2014) other parts of the world that don’t have healthcare systems designed around the needs of insurers eg. in the NHS where honest practical GPs are just doing innovative things and sharing news of what they’re doing (instead of paying to have a paper published in a journal).

“Digital Therapeutics and Diagnostics. Simply monitoring patient outcomes remotely generally does not by itself improve clinical outcomes”

I think this isn’t true. Of course it’s possible to anything the wrong way but Mobile devices can now be embedded inside medical devices that Patients are already reliant on and we’ve seen that with popular FDA cleared medical devices like glucometers for years now and having accurate information on the use is invaluable for Clinicians who can work more effectively when they have more informed, timely, useful and honest consultations with their Patients.

“A major challenge of mobile health is the high drop-off rate in sensor and app usage.40 In one survey, more than half of users of activity trackers stopped using their device, and a third did so in the first 6 months.41Strategies to improve engagement include discussion between patient and provider of goals for tracking and clear plans for attaining them. Joint goal-setting followed by joint review of the data appears to motivate continued tracking and therapeutic engagement.42 However, it is not clear, at present, whether enhanced patient–provider communication and shared decision making improves clinical outcomes”

I’m suprised people are still talking about the reluctance of Patients to use sensors and apps given to them by their carers because for the most part Patients still don’t get to even see their medical record and notes.

Also worth noting: this usability issue with trackers was a simple challenge that they managed to deal with in Japan 10+ years ago when the Raku-Raku (simple to use) mobiles used by seniors began featuring activity monitors (doing away with the need to wear separate motion tracking devices) and displaying the user activity as an icon on the idle screen. For more on this please read “Mobile Health Devices: Will Patients actually use them”.

Integration with Clinical Care. At the front lines of care, two challenges dominate the implementation of mobile health. The first is the vast quantity of data. Ancillary staff can help review and triage data, and visualization tools can mitigate the cognitive burden of interpreting the data. However, the most effective response is to develop and show to clinicians only those digital biomarkers that inform clinical action or clinical understanding (e.g., temporal and severity profiles of patient symptoms). Simultaneously, these biomarkers must be of sufficient direct value to patients to justify their participation in the data-collection effort. None of this is easy. Greater investment in the science of digital biomarkers is needed to evaluate the value of mobile health data for clinical use.

Let’s just start by letting Patients use the tools of our time to share their information with their Doctors, and let Doctors use these tools to move beyond the 2000 year old office visit only model of care.

“Integration of Sensor, Smartphone, and Electronic Health Record (EHR) Data for Patients and Clinicians. The second (and related) challenge is how the inclusion and presentation of data will fit into an already complicated and overstretched workflow. Clinicians cannot be expected to log in to separate websites for every sensor or app their patients are using. That said, integration with the electronic health record (EHR) is currently extremely challenging and costly. A recent development may offer hope: the federal government’s “meaningful use” requirements for EHR certification are calling for greater interoperability through an emerging data-exchange standard called Fast Healthcare Interoperability Resources (FHIR). FHIR allows external third-party apps to integrate into the EHR workflow. For example, Apple enables data to flow from FHIR-enabled EHRs to the Apple Health app on iPhones and from there to other apps in the Apple ecosystem (). With this new ability to combine EHR and mobile health data, iOS mobile health apps may become more useful and effective. Absent an Android equivalent for EHR access, Android apps may over time become systematically less effective than iOS apps, which raises deep ethical concerns, as discussed below. A recently started open-source project called CommonHealth ( aims to mitigate this concern by bringing EHR data integration to Android smartphones”

I notice Apple gets a lot of criticism for this but the reality needs to be appreciated by medics that Android smartphones really aren’t technically equivalent to Apple iPhones.

Apple is structured in a way that it owns the whole ecosystem and so can reassure the regulators, hospitals, insurers, Medics (who are putting their careers on the line entrusting these devices) and Patients (who are entrusting Apple with keeping their information private/secure) that the customer is using a specific device with the latest up to date software. In sharp contrast even brand new Android devices vary considerably in terms of their performance, specs, operating system, etc. On top of that they can install apps from a variety of app stores, have native software on them placed there by the mobile operators and manufacturers and Google. This makes it much more challenging environment in which to innovate and assure partners to have trust because there are many ways that data can be compromised if you don’t have a means of trusting the devices in Patients/Clinicians hands.

“FHIR also supports integrating the output of third-party digital health tools directly into the EHR workflow without requiring that the output be written into the database of the EHR, since such a requirement would trigger often prohibitive legal and security concerns. With this SMART-on-FHIR integration approach (SMART stands for Substitutable Medical Applications and Reusable Technologies), a clinician who is signed into the patient’s record can view that patient’s third-party app or sensor data in an embedded window without the need for a separate log-in. This approach opens up the workflow and “screen real estate” of the EHR to essentially unlimited innovation. However, SMART-on-FHIR integration is currently very limited”

Sometimes while these innovations seem great and easy to work, the reality is they often fail because they’re an attempt to run before you can walk.

We must be mindful that Doctors have a responsibility to stay current with technology. The Royal College of General Practitioners is a world leader in championing tech and yet in 2019 a Doctor is still failed for bringing a mobile into an exam room for the entrance exams despite the fact that many of the consultations they provide wouldn’t need to happen if they just used the tools of our time and they’ll probably not ever consult with a Patient who hasn’t already used their mobile to research their symptoms, diagnosis and medications.

“Health care organizations that wish to implement mobile health cannot overlook the logistics and legal implications of providing patients with sensors and apps. Additional staffing must be considered for assisting patients with technology setup, providing technical support, and responding to patients’ questions and concerns. For example, the Ochsner Health System has assigned dedicated staff to act as “geniuses” in their O Bar, a retail-like space modeled after Apple’s Genius Bar that carries Ochsner-approved apps and devices that patients can “test drive.” As patients independently bring mobile health data to clinicians (e.g., Apple Watch and Fitbit data) and as digital health vendors try to sell to health care organizations, health care leaders will need to grapple with how best to support mobile health and the use of patient-generated health data.

I think that you’re taking the wrong approach in wanting to be implementing mHealth innovations that require ‘Additional staffing’ needs. Instead of always trying to expand the healthcare workforce we should be trying to use technology to better utilise the teams we already have. Around the world healthcare professionals waste their lives doing arduous administrivia and we have the opportunity with mHealth to consign this to the history bin. Let’s seize that opportunity not try and build genius bars so that we can do even more complex things that we’ve no experience of ever having actually done before.

“Other Challenges: Validation and Regulation. Higher-risk mobile health technologies are considered medical devices under the Food, Drug, and Cosmetic Act. The FDA regulates medical devices under processes that were designed for medical-grade hardware devices, such as hip implants, with known physical properties that change little after market release. Mobile health technologies, in contrast, often combine hardware (e.g., a glucometer) with software (e.g., algorithms for the management of type 1 diabetes) and can be extremely dynamic, with frequent hardware and software updates. A different regulatory approach is needed. The FDA proposes to regulate these technologies (now termed “software as a medical device”) through a new Digital Health Software Precertification Program.”

I think this paragraph makes a statement that is factually inaccurate. Everyone and their dog knows that the FDA does a terrible job of regulating medical devices and we’ve seen that with countless examples eg. the collossal failure of their regulation of the hip implants from JnJ DePuy.

It is more than ambitious to think they will be able to effectively manage the enormous complexity of the mHealth app market without killing innovation through their bureaucracy and overreach (but I’ve written more than enough about that back in 2012!).

“This “Pre-Cert” program, currently under pilot, proposes to precertify companies that demonstrate a “culture of quality and organizational excellence” for streamlined review of their applications. Products of precertified companies do not have to be associated with improved clinical outcomes before market release but will instead be subject to postmarketing performance monitoring to support the claims of safety and effectiveness by the company.

Many details of the program remain to be determined, including how companies can gain or lose precertification status, how different levels of risk of harm will be determined and handled, and how real-world performance will be assessed through postmarketing monitoring and with what consequences. Trust in mobile health technologies and the extent of their adoption will depend on how these and other details are resolved. Thus, the stakes for success of the Pre-Cert program are very high. The desire to support innovation must be balanced against grave concerns about insufficient or delayed oversight”

How is the pre-cert program any different from a situation in which the FDA gets paid by mHealth innovators to learn how to do its job and introduce more regulation and administration?

Surely Dr Sim is aware of the mess the FDA have made with it’s efforts to regulate the ‘e-cigerette/vaping’ market?

In addition to these technologies, the accuracy of digital biomarkers also requires validation. For example, commercial sensors are relatively consistent with each other on step count but not on sleep duration or sleep cycles or duration of physical activity. Because most commercial devices restrict access to their raw data and algorithms, independent verification and validation of the majority of current digital biomarkers is not possible. Greater transparency and accountability, the setting of metadata standards (standards on how to describe data, such as their provenance), and external validation will facilitate evaluation of digital biomarkers proposed for use in clinical care. This can be achieved by making available reference data sets and through publication of tests of validation. An illustrative example is Sage BioNetwork’s Parkinson’s Disease Digital Biomarker DREAM Challenge that made available data from studies of digital biomarkers of tremor and dyskinesia.

These are all desirable things but isn’t it obvious that this is where innovative startups are being motivated to innovate and build their value? For example innovators like SleepTracker help coach you to better sleep without the need for medications for a price that is insignificant compared to sleeping medications – all without side effects. But they wouldn’t be interested in this market if it wasn’t financially viable because regulations made them have to help everyone understand and potentially make a copy of the tech that they use to enable you to do that for yourself. 

“Clinicians, patients, and payers would benefit from digital health formularies that list sensors and apps that are vetted for clinical use, much like medication formularies for drugs. The latest of several attempts to launch such a digital health formulary is by Express Scripts. Other efforts include the establishment of principles and guidelines for app development and rigorous evaluation, but best practices for screening, integrating, and appraising apps remain to be established.”

Amazing that after calling for a need for more “independent verification and validation… Greater transparency and accountability” it’s seen to be a good thing for the USA’s largest pharmacy benefit management organisation (with 26000 employees and revenues of +$100B that is itself a subsidy of the health insurer Cigna) to provide the resource for Doctors to work with Patients that could be all about better control of and more appropriate clinical use of the medications that drive profits in the parent companies. 

While I think it’s obvious that everyone would like the lucrative gate keeper role when it comes to mHealth app quality the reality is it’s a very challenging area and vested interests often detract from a corporations ability to do a good job eg. remember how the American Medical Association CEO called out mHealth as Snake Oil and told us they were going to be the arbitrator of what’s good and what’s not?

“Market Growth and Clinical Value: Despite the regulatory flux, the digital health sector saw a record $8.1 billion in investments in 2018. A recent change in the Medicare Physician Fee Schedule that allows physician billing for time spent managing and interpreting data from remote monitoring (e.g., electrocardiographic, blood-pressure, and glucose monitoring) for management of chronic conditions provides an incentive for the use of mobile health data in the clinic. However, no standard models exist for who should pay for mobile health technology that is recommended or prescribed to patients. In clinical research, decentralized clinical trials (see Glossary) are using digital biomarkers as end points and replacing in-person study procedures with virtual and mobile procedures. ResearchKit by Apple and ResearchStack by Android — mobile research platforms that facilitate large-scale virtual recruitment and outcomes assessment — are expanding the reach of clinical studies, analogous to the effect of mobile health on clinical care”

I think it’s important to put VC investments into perspective. In the USA alone $350 Billion was spent on prescription drugs in 2018 and the federal government spent nearly $1.1 trillion. Surely it’s obvious that when the financial incentives are this stark in comparison you shouldn’t be optimistic that the US sick care system wants to drive the adoption of mHealth.

“Mechanisms of Mobile Health and Potential Benefits and Harms. Underlying these market developments is a persistent question: to what end is mobile health? Tracking and reporting data are means to an end, not the end itself. Achieving clinical value, the ultimate goal, may occur through the use of mobile health data as cognitive aids to patients and clinicians (helping people understand or think through an issue), decision aids to patients and clinicians (helping people decide on an action), or motivational aids for patient engagement and activation (). More collaboration is needed among clinicians, patients, and technologists to drive the development of clinically useful mobile health technology and to imagine clinically useful applications of novel sensors, while remaining cognizant of potential harms”
I don’t agree that the ‘ultimate goal’ of mHealth should be to achieve “clinical value”. Yes that’s desirable for the healthcare industry but societally we should be looking for much more. Mobiles are already having a huge impact on the psychology of young people. Most are learning about sex and developing their personal lives through watching ‘porn’ (AKA videos of sexual abuse) and becoming addicted to mobile content and struggling to concentrate, mature or develop with their lives. Instead of looking for clinical value maybe the much bigger opportunity is for mHealth to help these young people and their families to avoid the need for clinical intervention altogether. With Dr Sim’s focus I fear we’re just going to be  trying to fight fires with (eg. manage the obesity, anxieties, narcissism, sexual abuse, depression, stress, disease, etc, etc) rather than prevent them happening.
“Clinical researchers need to develop new evaluation approaches because years-long studies are poorly suited to the pace of change of mobile health technology. Finally, clearer demonstrations of the clinical and business value of mobile health will come when factors far beyond technology itself, such as integration into clinical workflow, payment model, and validation methods, are addressed in tandem with sensor and software development”

Somewhat unusual for me I know but I agree with these great points!

“Ethics. The tremendous promise of mobile health for transforming clinical care and research is tempered by deep concerns about the effect of these technologies on equity, privacy, and patient autonomy. Although there is essentially no digital divide according to race in the United States, Internet and smartphone adoption is lower in lower-income, disabled, elderly, and rural populations. Moreover, Android users, who account for more than half of U.S. smartphone users, have lower average income than iOS users. If iOS mobile health apps are systematically more effective than Android apps at improving health outcomes (e.g., owing to differential access to EHR data, as discussed above), health disparities will worsen. As health institutions increasingly develop a mobile presence with branded apps and other initiatives regarding digital patient experience (see Glossary), they must be careful not to increase health disparities — for example, by offering unequal services to iOS users and Android users”.

I think this is very controversial and it ignores the fact that Android devices are CHEAPER for a reason and that reason is that the software on these devices and the Google Play app store is made for device makers (like Samsung, Motorola, Nokia, Huawei, Xiaomi, etc, etc) so that Google/Alphabet (Android’s parent company) can essentially collect more data on customers and subsequently sell more effective means for advertisers to reach them. To think that you’re going to reduce health disparities by sharing more information with a company that makes its money from  advertising is overly optimistic and somewhat naive.

Android is a product of Googles. For all the positives and negatives it is the same company that makes billions from serving unhealthy junk food ads to customers. It’s the same company that is selling ads from the highest bidder next to the results when Patients search for healthcare information or get a gmail message from their Doctor etc.

Google has been very effective at ensuring you don’t read bad news about it’s brand as they have a great legal team so if you’re a normal medic and have no idea about this read here about how they are collecting NHS Hospital Records without consent or go further back to their founding story and you’ll realise they were [practically seed funded by illegal drug ads.

In my opinion health institutions should not preoccupy themselves with how unequal the iOS and Android platforms are. Google should be taking on the responsibility of investing in this space just like Apple have been with Apple Health, HealthKit, ResearchKit, etc, etc and hospitals should opt for the best platform for them and their Patients (as clearly the cost differential between an iPhone and Android Smartphone is an insignificantly inexpensive cost compared to the cost of healthcare services

“There is an even more profound digital divide. Use of mobile health and Internet of Things technologies requires digital literacy skills. Patients are essentially being asked to install and maintain their own medical devices and to be adept at managing their own data deluge, a tall order when the majority of U.S. adults are in the lowest three of six proficiency levels for literacy and numeracy and more than 60% are in the lowest two of four proficiency levels for problem solving in technology-rich environments (defined as “using digital technology, communication tools, and networks to acquire and evaluate information, communicate with others, and perform practical tasks”). This gap is further compounded by the paucity of mobile health technologies in languages other than English.”

I see a lot of insurance companies taking advantage of this and craftily pushing apps in their marketing messages to help them resonate with younger smartphone users who would have more profitable insurance needs but I think it’s important to realise what mobile SMS has done for literacy rates and that mobile services in public systems (like the NHS etc) are helping save resources when they aren’t needed because Patients can look up their own healthcare information, results, watch videos, organise and collect their own prescriptions, etc.

If you need a really obvious example think about the benefits of SMS appointment reminders (from 2011).

“Actions for Stakeholders in Mobile Health. With respect to privacy and autonomy, the potential threats are particularly worrisome. Mobile health technologies will increasingly connect to the Internet of Things, in which, like a “one-way mirror,” our virtual bodies and behavior will be visible on a grand scale for purposes to which we have not directly consented. When personal health and nonhealth data comingle in the cloud, companies and governments may access physiological biomarkers to monitor employee stress in the workplace, or marketers may offer us only certain products at differential prices based on our health history. Coupled with algorithms that are not in the public domain, these approaches could deliberately or inadvertently reinforce and entrench existing biases against disadvantaged groups, and incautious deployment of mobile health technology could potentially result in loss of privacy and autonomy amounting to net harm to patients. lists actions that patients, practitioners, researchers, and policymakers can take to guide the evolution of mobile health”

To be fair I think these potential risks are already materialised. Google already has your full search/viewing history and if you gave that information to a good psychiatrist they could probably tell more about you than you know about yourself and that’s before you go into the myriad of incredibly sensitive personal data that’s being collected just through normal use of a smartphone (think eye scans, fingerprints being used to unlock/lock a phone, etc, etc.)

“Conclusions. Mobile health technologies are evolving from descriptive monitoring tools to digital diagnostics and therapeutics that synergize tracking with behavioral and other interventions to directly affect health outcomes. Major challenges include the discovery and validation of meaningful digital biomarkers, regulation of and payment for mobile health technologies, and their integration into frontline care. Clearer articulations are needed of how mobile health technology can concretely affect clinical outcomes, along with more rigorous evaluations of clinical effectiveness”

Interesting ideas but I prefer the idea of mHealth being used to create a new frontline of care rather than just trying to integrate it into today’s frontline of care.

“Networked mobile health technologies have the potential to harm. Concerns about digital surveillance are not unique to mobile health, but health-related risks can be reduced through improved digital literacy among patients, ethical codes of conduct for developers and regulators of mobile health, and transparency and accountability in how health care organizations adopt mobile health technology. The transformative potential of mobile health compels clinicians to take an active role in ensuring that this new frontier will be safe, fair, and just for all patients”

Another very important message in this excellent paper.


*** Update 8 September 2019 ***

Please click here to join a discussion of this NEJM paper over in the Linkedin mHealth group:

Screenshot 2019-09-08 at 10.46.06

Screenshot 2019-09-08 at 10.46.27

About David Doherty
This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s