The fact is, a handful of manufacturers currently have a stronghold on our diabetes data. Their systems collect this vital health data for us, but are designed to withhold it so that we rely on their products exclusively to access it, share it, or analyze it.
More and more patients are finding this intolerable, and are fed up with waiting for the Pharma industry to embrace long-overdue open standards.
I’m a bit obsessed with this issue myself, and made it the crux of this year’s DiabetesMine Innovation Summit event. And I keep hearing about patients who are taking matters into their own hands. One of those is Benjamin West, a thirty-something type 1 software developer who lives in San Francisco and works at the networking company Meraki (they do the wifi for Motel 6 and Peet’s coffee). In his free time, he’s creating a web-based “sandbox” and recruiting other tech-savvy PWDs to help him crack the code on diabetes data sharing. No kidding!
“The pump is attached to you and is performing therapy on you. You have the right to know and have access to those medical records. But for some reason, pumps and glucose meters don’t operate that way. The only way to get those records is to go through the vendor. I think there’s a real ethical imperative here, an ethical wrong,” Benjamin says.
Sounds a lot like the ePatient Movement’s rally cry, “Gimme my damn data!”
Several projects Benjamin has running simultaneously include creating firmware to run on Beaglebone (an iPhone stand-in); collaborating on a data mining effort called DUBS (Diabetes Understanding By Simulation); creating Insulaudit, a Python library that tries to collect data from numerous devices; and a “Decoding CareLink” project that uses the CareLink USB stick like a remote control to direct the pump and download data.
Whether or not you agree that the industry is obliged to hand over your data free of charge, I think few of us with diabetes would argue that making products more interoperable and data sharing easier is an imperative to getting better outcomes from these tools. In other words, we need this connectivity to improve care!
“Every day I can’t get data out of these devices is another day that I’m that much more at risk,” Benjamin says. “The technology that’s available everywhere now makes this stuff as easy as lego blocks. I’ve got a program that integrates my pump and meter with my cell phone… Connecting these communication protocols is not hard, it’s just secret.”
Just Who is Benjamin West?
Benjamin has written to the FDA, posted a number of discussions about his projects on TuDiabetes, and even tried to enlist the Society for Participatory Medicine to help him gather a team to enter Sanofi’s Collaborate/Activate crowdsourcing competition last year — but no one was biting.
FDA brushed him aside, and on the TuDiabetes site, other members attacked him for being too focused on raw data. He says it was probably a perception issue: “I think I came off as an overly techy geek that was just annoyed. But I’ve had diabetes for 10 years and I find this intolerable. There’s no way to cross-check to make sure your pump or meter is operating correctly, or merge with other devices – it’s just weird that you can’t do that. My point is that safety is not a feature, especially not with medical records. Safety is public — a public good.”
Benjamin’s actually a New Jersey native who was diagnosed at age 21 while studying music at Arkansas’ Harding University. He intended music to be his life, but discovered an aptitude for computer science and eventually got a dual degree.
After the diagnosis hit (“I spent a couple of bad days in the hospital”) he went to an endocrinologist in Little Rock, AK, who trained him on syringes, but switched him to an insulin pump within a year, citing pumping as the Gold Standard of care.
“At the time I kind of thought someone, somewhere would have software that would model my metabolism – since diabetes is all about metabolism – and I’d be able to look at the data and see how my pancreas was behaving. I would use that data to respond with therapeutic decisions… I was very quickly disenchanted,” he says.
Instead, he started writing down his blood sugars on note cards, which he was soon drowning in. This turned him off logging completely for several years. Without proper data tracking, his glucose control got all out of whack. “I actually went to the hospital few times,” he says.
He realized he could use the principles from his IT work to “go rogue” on creating the diabetes data connections he needed. “We know how to set up a network so that it works for all. We may not have a huge development team (about 6 people), but we have a web-based management console that anyone can figure out, or they can share a link with others to figure it out.”
He has connections with data miners through his father, formerly with Lawrence Livermore labs who now works at AT&T developing analytics and prediction software. And Benjamin’s brother Nathan, who’s finishing up a graduate degree in computer engineering, is also involved. He’s working on firmware for the Beaglebone – an “open board iPhone” that lets developers build accessories. “Nathan plugged in a USB with all my diabetic gear. I can send an SMS message and it’ll attach one of the devices to the internet. You don’t need special vendor support; you can just add other devices!” Benjamin says.
Little Help from Above
Benjamin queried all the leading pump and meter companies to grant open access to their communication protocols (like the user manual for programmers). So far, only JNJ LifeScan has been very cooperative. Medtronic has been particularly resistant, which frustrates Benjamin as a Medtronic pumper.
One guy he’s collaborating with, Will Johnson in Berkeley, CA (creator of DUBS), has apparently developed a tool for glucose predictions that could be a boon to patients.
“We’ve got these predictive algorithms ready to go – we’re just waiting for the data,” Benjamin says.
As an example: one “hacking” project they’re working on is using the CareLink USB stick like a remote control, to download data and perform custom analysis, and then create a therapeutic plan for the next 6 hours, taking into consideration your plans to eat or exercise. Wow!
It turns out Medtronic actually markets its own remote control device, but the one Benjamin is crafting would have fewer buttons, more functions, and be more customizable by the user.
On the ethical argument over data property rights, Benjamin provides this link to a geeky but relevant presentation on “the coming civil war over general purpose computing.”
As a touchpoint, I asked respected diabetes technology guru Scott Hanselman about Benjamin’s work (he actually tipped me off to the story!). He simply says that what Benjamin is doing is “super innovative.”
D-Hackers Wanted!
In short, Benjamin is looking for more collaborators to make auditing these devices transparent. “The only way I know how to do this is to make a library that can be peer
-reviewed. We need collaborators and/or vendors to open documents to protocols.”
“We need coders, IT problem-solvers, documentors – and vendors who are really excited about innovation. You don’t have to be a professional developer – we’re looking for anyone who’s good at solving tech puzzles.”
“As soon as we get the data decoded, we will need to know how to program the visualization. I would challenge your readers to help me design the perfect set of visualizations,” he says, adding that Glucosurfer might provide one good example.
On the vendor side, Hanselman believes that opening protocols will only come as a result of market pressure. “If the company wants to keep it closed, then let folks stop buying their stuff,” he says.
“In other words, we patients (their customers) need to put pressure on these companies to play fair?” I asked.
“Yes, I agree that pressure makes the difference — 100%,” Hanselman writes back.
Interested parties can contact Benjamin West here.
[...] Amy Tenderich writes: The fact is, a handful of manufacturers currently have a stronghold on our diabetes data. Their systems collect this vital health data for us, but are designed to withhold it so that we rely on their products exclusively to access it, share it, or analyze it. More and more patients are finding this intolerable, and are fed up with waiting for the Pharma industry to embrace long-overdue open standards. I’m a bit obsessed with this issue myself, and made it the crux of this year’s DiabetesMine Innovation Summit event. And I keep hearing about patients who are taking matters into their own hands. One of those is Benjamin West, a thirty-something type 1 software developer who lives in San Francisco and works at the networking company Meraki (they do the wifi for Motel 6 and Peet’s coffee). In his free time, he’s creating a web-based “sandbox” and recruiting other tech-savvy PWDs to help him crack the code on diabetes data sharing. No kidding! “The pump is attached to you and is performing therapy on you. You have the right to know and have access to those medical records. But for some reason, pumps and glucose meters don’t operate that way. The only way to get those records is to go through the vendor. I think there’s a real ethical imperative here, an ethical wrong,” Benjamin says.Sounds a lot like the ePatient Movement’s rally cry, “Gimme my damn data!” [...]
I think the challenge for Benjamin is that most people don’t understand the importance of data interoperability as a foundation for more powerful reports, applications, etc. I keep referring to this as the diabetes data bus, If it’s fully interchangeable then any device can plug into this bus and add data to it, or read data from it.
This would enable data interpretation software (subject to FDA approval) to analyze the results and help endos and patients deal with the increasing flood of data.
If the standard is fully open (meaning it doesn’t cost $$$ to get a copy of it) I will bet the mobile market for diabetes-related apps explodes. I also know that we’ll see much more interesting reports that really help people understand what’s going on now and how to adjust things so their control and quality of life is better in the future.
I suggested to someone in the FDA that lack of a standard was a safety issue because we can’t get a true overview of what’s happening. I figured it was worth a try but they didn’t think it could be sold as that.
What we need is a big market player who will push this issue. if OneTouch, Minimed, or Dexcom published their data formats they might be able to build a de facto standard around that. The other alternative would be for all the diabetes device makers to have a consortium for this. As far as I know they really dont cooperate in any way.
And in the meantime we all live with the consequences.
Thanks Amy for continuing to push this issue.
Forgot to mention that I’m happy to volunteer in this effort. I’m proficient in Java, C, Objective-C. Understand client/server, remote APIs, threading etc.
Please pass my email on to Benjamin.
link in with me too Bernard!
I wish I was competent to work on this. It is something that has been just wrong for so long. It is our data, it belongs to us, and it should all be able to be looked at in one place.
important story. thanks for writing it, amy. good luck to benjamin.
I would happily buy Benjamin and Bernard a Beer (BB&BaB!) and listen to them figure out how to right this problem.
I would think that Meaningful Use of EHR maybe something we should be exploring and talking about in the DOC as a means of leveraging conversations about our data.
Meaningful use includes patient empowerment
http://www.healthit.gov/policy-researchers-implementers/meaningful-use
What about Continua Health Alliance?
There is a program on sourceforge that captures all of the USB data sent back and forth from a USB enabled device, which is particularly useful for reverse engineering. I do have a driver to the CareLink USB if anyone is interested.
I already programmed the entire AP project (sliding mode controller-no glucagon feedback loop) in Matlab in Spring 2011, with Minimed CGM and pump data as the inputs. Any undergraduate engineering student is capable of doing it, because the students have up to calculus 3, some linear algebra, some differential equations, and concrete programming experience. Having type 1 diabetes, makes it easier to figure it out, though.
It is really sad to think that when the first “official” AP is FDA approved and released (by the way, the term AP is really a misnomer, and it should be ditched by the JDRF), it will probably cost at least $30,000 for the person wearing the device.
Anyways, when it comes to the medical device companies, they should be focusing on refining the design and the efficacy of the physical (and only needed) components of medical devices. Interfaces, frequently unused features (Area under curve (AUC) function on the MM pumps, anyone?), could be programmed by anyone with math and programming exposure, including basically any sort of engineer. When medical device companies make a big deal about functions such as AUC, it is really a gimmick.
As for a data standard, the US government should have a mandate for all medical records, insurance claims, and consumer medical devices adopt a data standard. In addition, all data generated by these devices should be readily accessible by the government. The companies already have to submit 2,000 pages to the FDA for insulin pumps in order for clearance it is not a big deal for them to submit more data under a universal standard such as IEEE 11073, which is also an International Standards Organization (ISO) standard, which several countries accross the world agreed upon.
If the US government cannot agree on adopting an ISO standard for the storage of healthcare data, I don’t know what to say. I guess we will have an upwards battle in improving the lives of anyone who has health issues–and also in the reduction in the cost of care.
Type1Engineer is that USB Sniffer for Windows? If not can you post a URL?
https://github.com/medevice-users/diabetes
Howdy all, I’m organizing everyone’s technical efforts (there are many!) to try and point in the same direction. There is a medevice google group for anyone interested.
For those interested in just the medtronic protocol, here it is: https://github.com/bewest/decoding-carelink/tree/rewriting
[...] MT @HugoOC: The ethical imperative of #diabetes interoperability and our right to our "damn data" http://t.co/ybcYkHy5 #HealthIT #ONC #MU [...]
Ben, You are a badass. Please Link in with me, facebook me, find me on tu diabetes, I share your thoughts and enthusiasm and I am interested to help you further your efforts.
I am a physician, mobile health hacker, startup advisor and founder, and dmom.
let’s talk…
Bernard, here is how dex is dealing with the interoperability issue…
have y’all seen this?
http://www.qualcommlife.com/ecosystem-partners
I contacted dex about the qualcom 2net partnership and they did put me in the till as an interested physician to view patient data on my emr, I use hellohealth
All roads to dex data may be through 2net…
Thanks Natalie, I’ll add this info to our medevice-users wiki.