Sunday, September 21, 2014

Cosmos and the Power of Knowledge

Fun fact: the 'C' and 'S' are the first letters to come out of the nebula/iris on the show's introduction and stand for 'Carl Sagan,' the host of the original 1980 series
Some of you may have heard of Cosmos: A Spacetime Odyssey, a recent remake of Carl Sagan's popular PBS science series from 1980 hosted by Carl's protege, Neil deGrasse Tyson.

It's literally the best thing I've ever seen on TV (or a screen of any kind - thank you Michael Faraday!)  I consider myself fairly well read when it comes to science and technology (and space!  Heck, I even earned my Astronomy merit badge as a Boy Scout many years ago).  But with each passing episode of Cosmos, not only have I learned multiple new concepts, but I'm often left with tears in my eyes at the simplicity and power of the ideas, as well as the emphasis on openness and accessibility of knowledge.

Despite my man-tears, the best part of Cosmos for me is that I've been able to share it with my two 8 year old daughters, Miriam and Catherine.  We've watched the first 10 episodes together, and they can't wait until we can sit down again to watch the next episode.  In their minds, Cosmos > Scooby Doo, and that's saying something.

After finishing episode 10 today ("The Electric Boy" about Michael Faraday), Miriam was particularly touched.  Her reaction prompted me to send a short thank-you note to Dr. Tyson himself, which I did via the Hayden Planetarium website.  I'm not sure Dr. Tyson will ever receive that note, so I decided to post it publicly:
Dear Dr. Tyson, 
I just wanted to take a moment to thank you for your work to popularize science in so many ways.  In particular, one of my daughters, Miriam, has immensely enjoyed watching Cosmos.  Miriam is 8 years old and is on the Autism spectrum, although fairly high functioning. 
Earlier today Miriam was asking me questions like, "How fast is the fastest race car?"  I responded, "About 300mph."  Then she asked, "How fast is the speed of light."  "670 million mph," I replied. 
She then proceeded to write out those numbers and subtract one from the other, getting answers like 669,999,700.  I asked her what she was doing, and she said, "I want to be a genius like Newton.  Do you have any of his books?"  My jaw just about dropped to the floor, then my wife and I looked at each other and started laughing in disbelief.  Of course, she was completely serious.  I told her I'd find one of Newton's books for her.  Because I wasn't fast enough (this was literally just a few hours ago), she found my 3-inch thick AP Chemistry textbook (it's been laying on a bookshelf for years) and after a brief scan of the table of contents, immediately started reading the chapter on "Electrochemistry" (probably because we had just finished watching the Faraday episode earlier today). 
So congratulations!  You've just made Sir Isaac Newton my 8 year old Autistic daughter's role model.  Mission accomplished, sir! 
Warm regards, 
Ricky Bloomfield (and his daughter, Miriam)
Dr. Tyson (and the rest of the Cosmos team), thank you for sharing your love of science and knowledge.  If nothing else, you've at least touched the life of an 8 year old little girl (and her father).

Monday, September 15, 2014

Innovating with Epic

Today I had the opportunity to present some of my recent mobile work at the 2014 Epic User Group Meeting (UGM) in Verona, WI.  This is a gathering of over 10,000 physicians and employees of hospital systems across the country who use the Epic EHR.  Duke has been live on Epic ambulatory for a little over two years, and live on Epic inpatient just over a year.  We've learned quite a bit since going live, and the goal of UGM is to share that knowledge with other hospitals, as well as to learn from them.

Taking a collaborative approach to the improvement of EHRs is not only smart medicine, but, in my opinion, essential to reaching our goals of improved clinical care, enhanced patient satisfaction, reduced costs and streamlined research.  Every hospital system brings a unique perspective owing to their own unique strengths and challenges.  They will confront and solve problems similar to our own, but likely in different, and hopefully even more effective, ways.

In March of this year, I led the roll-out of Epic's Haiku and Canto mobile applications across Duke University Health System.  These applications provide access to the EHR from iOS and Android mobile devices.  While the feature set is robust and continually improving with each iteration, there will inevitably be requests for feature improvements from discerning physicians, myself included.  Rather than wait for these features to be implemented, I decided to dig a little deeper to see what changes we might be able to make on our own.  Due the flexibily of the platform on which Haiku and Canto are built, I was pleasantly surprised by what we could accomplish in a very short time.

My UGM presentation, entitled, Hacking Haiku: When You Just Can't Leave Well Enough Alone, is a summary of this work.  Please log in to the UserWeb to view the full details of the presentation, including the presentation slides and demonstration videos.  When creating this presentation, I started with the underlying principle that anything I shared with the group should be able to be replicated by those in attendance.  To that end, I also included a detailed step-by-step guide for these modifications (or "hacks," if you will), also available on the UserWeb.

While some of these hacks involve integrations with or modifications to Epic IP and can only be shared via the UserWeb, a significant part of the presentation discussed a custom app I developed, called the Duke Medicine Mobile Companion.  The app is extremely simple, and performs one essential function: allows users to view additional content external to the Epic apps while providing a one-tap method to return to those apps.

For more information on the app and its functionality, please check out the code on GitHub.



The app is currently iOS-only, and was written in Apple's new Swift programming language, now officially available for apps released in the App Store.  I thought this would be a good opportunity to learn the new language, while at the same time ensuring that the code is modern and robust.  Suggestions and improvements are welcome, obviously.

Why no Android?  It's actually not necessary.  On iOS, the app works by accepting URLs prefixed by the dukemedmc:// URL scheme, a few of which we've added to the Epic apps.  When these URLs are tapped, the Duke Medicine Mobile Companion is opened and information is passed to the app in the form of URL parameters.  A button in the app then provides an easy way to return to the Epic apps.  On Android, a back button is built into the OS natively that not only works within apps, but also between them.  So on Android, we can simply open these URLs in other apps directly (such as the Chrome browser or Google Maps app) and then the user taps the back arrow to return diretly to the Epic app.  There's no need to reinvent to wheel when it's not necessary!

I hope these ideas along with the implementation details will be helpful to other healthcare systems.  The more we collaborate and share, the stronger we'll all be.  With the explosion of mobile health apps in the past few years - with more surely to come - it's also as important as ever to ensure that we protect and simplify our provders' workflows so they can spend even more time caring for patients, rather than simply treating diseases.

Friday, September 12, 2014

 Watch - Relevant to Healthcare?



The anticipation in advance of the Apple event this week was similar to that prior to the announcements of the original iPhone in 2007 and the iPad in 2010.  It would have been an extreme disappointment if the event ended without the announcement of a new product category.

Of course, everyone figured it would be a watch, but just like the previous announcements in 2007 and 2010, no one had any idea what it would look like or what it would do.  Apple is very good at secrecy when mass manufacturing for a product hasn't yet commenced, or when the product includes new software (as evidenced by the surprise announcement of the Swift programming language in June).  Both were true in case.  Prior to the announcement, the  Watch, as it is now known, had been produced in only limited test quantities, and the new version of iOS that it runs had never before been seen publicly.

The  Watch was announced in the context of the prior announcement of HealthKit in June.  Special emphasis has been placed on the health and fitness features of the watch, with Apple going as far as producing a separate promotional video highlighting this use.

That said, the expectations for what this watch would do for health and fitness bordered on magical.  Sources as reliable as the Wall Street Journal predicted that the device would "include up to 10 sensors."  With current technology, I couldn't imagine how that many high-quality sensors could be included in a 1st generation product while still maintaining a reasonable size, cost and battery life.  Sure enough, the actual product was more firmly grounded in reality, with just a single medical sensor to measure pulse, in addition to the activity tracking of the M8 chip and the new barometric functionality to measure elevation.

It's quite likely that these rumors have some basis in fact, however.  The  Watch is the most intimate product ever created by Apple, and I'm not just referring to the ability to share your heartbeat with that special someone.  It's a device that is in constant contact with your skin.  Just that fact alone opens the door for many sensors beyond simple measurement of heart rate.  In fact, I'd wager that the  Watch could open the floodgates to R&D into transcutaneous sensors, with the potential to transform how we see ourselves.  The quantified self has just gotten much closer to reality.

So what can the  Watch do for healthcare right now?  At the very least, the profile of wearable electronics will get a huge boost in the consumer sector.  They will become more socially acceptable, and users will come to rely on them and even trust them to make meaningful differences in their lives.  Laying this foundation now will be critical to encourage the adoption of more advanced wearable sensors in future devices.

How useful the watch ends up being for consumer health and fitness will ultimately come down to the combination of hardware and software.  The fact that all of this data will be saved to HealthKit is very attractive, because the data can then be shared with any other HealthKit-enabled health or fitness app the user chooses.  Giving this power to the user is a smart move, and will likely foster a more open and diverse ecosystem of useful health and fitness apps moving forward.

With respect to use of the  Watch in the clinical setting, I love learning about and using new technology, but I’m also cognizant that the modern reality of electronic medical records has resulted in tension within the physician-patient relationship.  During a typical visit, physicians often find themselves staring at a computer screen rather than interacting with the patient.

In medical school we were taught the Latin phrase, “primum, non nocere” or, “first, do no harm.”  When technology interferes with the relationship we have with our patients, harm is being done.  When technology enhances our work while remaining transparent to the interaction, we can make progress.  Technologies such as the  Watch have the potential to take us a step in that direction by providing another avenue for relatively unobtrusive data access.  As a developer myself, I’m looking forward to the arrival of the WatchKit SDK so I can see more fully how this functionality might be integrated into our medical environment, both from the perspective of a user as well as a physician.

How could the watch be improved?  It'll be easier to see once the watch has been released and the limitations become more apparent.  But for starters, the watch needs to be waterproof.  While there's no official statement on this, David Pogue revealed that it's likely just water resistant.  I'm an avid swimmer, and one of the reasons I love my Pebble Smartwatch is because I never have to take it off (plus, it lasts up to 5-7 days per charge).  This would be critical for any users whose primary form of physical activity involves water sports.  Many of our patients who participate in cardiac rehab are often involved in water aerobics, and they would have to remove the watch to participate, thus nullifying much of the activity tracking benefit.

Battery life also needs to be at least several days.  I've heard some lamentations that the watch doesn't have sleep tracking functionality like some of the other popular wearables.  If the  Watch has a battery life of only 24 hours as rumored, sleep tracking would be an impossibility given that the watch would need to charged at night, every night.  The watch also appears a bit too bulky in its current iteration to keep on while sleeping (in my opinion).

I'm confident all these issues will be resolved over time in future iterations of the watch (or competing watches).  Progress will continue and app developers will surely come up with uses that Apple hasn't even considered.  Mobile health innovators should pay close attention to this new product category.  It will eventually revolutionize how we care for our patients and ourselves.

Thursday, September 11, 2014

Apple's HealthKit - A Flexible Platform for Healthcare Innovation

This week Apple officially released the latest version of iOS, complete with their solution for management of a user's health and fitness data, HealthKit. They also announced the  Watch, which includes several health and fitness features.  I've shared some preliminary thoughts on HealthKit already, but now that I've actually had a chance to use it and integrate it into our EHR via the Epic MyChart app, I have a few more thoughts.

It's really easy to use.


Our patients already have a couple ways to securely share data with us.  First, they can write the information on a piece of paper with their option of a pen or a pencil (we're that flexible) and bring it in to their provider.  Second, they can manually enter the information into our online patient portal, Duke MyChart (when enabled by their provider).

HealthKit will eventually provide a third option, but rather than having to remember to write the values down and bring in the piece of paper, or having to remember to manually enter data in to a website, HealthKit will automate this process.

For example, if patients choose to share their blood pressure with their provider using HealthKit, and this feature is enabled by their provider, they need only give permission to the appropriate app, after which their data is securely saved to their electronic medical record.  Any subsequent blood pressure measurements can be saved directly to the medical record in the same way without any intervention by the patient.

It sets a clear expectation that users are in control of their data.


Unlike personal health records that have come before, Apple has approached this problem differently, and has decided to be a data broker rather than a consumer of the health data.  This is a big deal.  At every step along the way, users are informed that they have control of this data.  No data is shared without their express permission, and at any point in the future the user can easily revoke any app's access to the data. The data itself is only stored locally in the device's secure HealthKit data store.

It's secure


While cloud services and Big Data promise to revolutionize data sharing, HealthKit will have none of it. In fact, Apple recently updated their App Review Guidelines to expressly forbid that any data obtained from HealthKit be stored in iCloud:

"27.3 Apps using the HealthKit framework that store users’ health information in iCloud will be rejected"

The only exception to this of which I'm aware is that the user's encrypted iCloud backup could still contain this data.  Personally, as a consumer of these products, I agree with having the option to back-up this data to iCloud.  While it may be considered sensitive, the risk of losing the data without a convient backup would be more detrimental, at least to me as a consumer.  When I put on my physician and privacy hat, I'd recommend that Apple consider giving users the option to exclude this data from iCloud backups, just as the TouchID fingerprint information is always kept in the device's Secure Enclave, and never leaves the device (which will presumably be the same case with credit card information as part of the new  Pay initiative - healthcare and financial data are usually afforded similar levels of security).

As I've mentioned previously, probably the most exciting aspect of a platform such as HealthKit is that it opens up a new avenue for healthcare innovation.  There will be many new ideas that take root and grow owing to the ease-of-use and scale HealthKit provides.  There will also be many failed experiments.  Ultimately, users will be able to decide which new ideas are relevant to them, and which are worthwhile enough to be entrusted with their healthcare data.