CGM Data: Around the World and Back Again – to my Pebble Watch

After attending the DiabetesMine Innovation Summit last November and being inspired by the Nightscout crew (and pretty much everyone else there as well!), I finally decided to take the time to set up a Nightscout website. I’ve had it up and running since Dec 2014. Being an adult with Type 1, I monitor my own blood glucose levels. I don’t have anyone that would be interested in following my data on a daily (much less hourly) basis, so I wasn’t sure if the functionality of Nightscout would be worth it for me. After 8 months, I can say with certainty that it has been…and then some!!!

I don’t actually use my NightPebbleWatchscout website or phone app very much but I LOVE wearing my real-time data on my Pebble watch. I’m probably 1000 times more likely to glance at my watch to check on my blood sugar than I am to dig in my purse (through all the snacks and toddler/kid gear) and haul out my Dex receiver or my phone. The setup to get my data to my watch was somewhat complicated but with the addition of the Dexcom G4 receiver in May, my data is almost 100% constant without much involvement from me other than occasionally restarting a few apps on my iPhone.

Last week, however, it suddenly quit working and I got emails written in what may as well have been a foreign language, saying things like “your Heroku app on free dynos needs to recharge.” I’m pretty techno-savvy and am generally an early adopter of new technology but I’m not a programmer and certainly don’t fully understand all of the ins and outs for each step I followed in the pipeline to ultimately setup Nightscout. Going back through them to figure out which accounts I used in my setup needed changes and what else it might affect wasn’t quite like starting at the beginning for me…but it was close.

There are so many different ways to setup Nightscout, including what hardware you’re using, what service providers you chose, what account parameters you chose with some of those service providers, etc., that it’s not an easy task for me to filter through the Nightscout setup guides and Facebook group looking for clues related to updating each piece of the specific pipeline that I selected to use. So, after I finished getting my data stream to my Pebble working again (took me about 3 hours), I decided to make a little diagram to help me understand how all the pieces I’m using fit together. Disclaimer: I’m pretty sure I have my pipeline figured out correctly but wouldn’t bet my house on it : )


My hope is that next time I need to make a change, I’ll be able to get it corrected more quickly…and with a little less terror! My current pipeline configuration prior to last week was free and now, with the required changes, looks like it will cost me $14/month, but it works for me and I wouldn’t trade it for the world.

MeandKidsI can no longer imagine not having my data on my Pebble. I feel so much safer. EVERY time I’m with my kids at the park, or the market, or driving in the car, and I glance at my Pebble, I’m thankful. Quite simply: I want to be the best mom I can be; I want to stick around to watch my kids grow up…and Nightscout helps me do that.

Big House : Small CGM Transmitter Range

So, we finally moved into our new house about a month ago. One of the strangest diabetes-related things I’ve had to get used to is the size and shape of my new house when it comes to my CGM signal. The Dexcom G4 is supposed to be in range as long as the receiver and transmitter are within 20 feet from each other. Our old house was kind of a 50s bungalow, where I could set my receiver in a central location and still have my signal picked up all over the house. Only our master bathroom stretched the range, so showering was the only time I had to remember to pick up my receiver and take it with me. Our new house is bigger and also more spread out. When we first moved in and I was unpacking, I missed hours of data at a time (and I missed a lot of lows and highs).

Caller-ID Big House:Small CGM Transmitter RangeI’ve really had to get adjusted to carrying my receiver with me from room to room in the house. It has turned into a bit of a ball-and-chain but my thinking is if I’m not going to have the receiver collecting a signal 99.9% of the time, then what’s the point of wearing this additional, ugly (yet not inexpensive!) hunk of plastic on my belly all the time, right? I’d love to take a vacation from my CGM occasionally but for me diabetes devices are like the original Caller ID box (remember those days?!) once you have it, you really can’t ever go back to being without it.  I hope to look back at my tummy devices and signal ranges one day just like I look back at the Caller ID box now and say “Wow, can you believe I used to use one of those things all the time? And I thought it was so high-tech!”

Photo-a-Day: Advocate

Photo-a-Day-AdvocateI use each of these items every day to help manage my diabetes.  Each one tracks a necessary variable.  In order to optimize my treatment, I need to analyze the data I collect, and yet none of these devices talk to one another or have a common analysis platform.  I advocate for simplifying diabetes data collection and analysis, while empowering patients.  (See my proposal, Pixels: Your Personal Diabetes Big Picture).

This post is one in a series for the National Diabetes Month of November.  Kerri from initiated the Photo-a-Day idea and prompts and lots of other diabetes bloggers have chosen to follow suite.

Pixels: Your Personal Diabetes Big Picture

Over the last several weeks my husband and I spent our evenings working on a proposal (below) we submitted to the Target Simplicity Challenge.  The whole point of our idea is to simplify and customize diabetes analysis and management for patients and doctors.  Check it out!  Let us know if you’d use a site like this, or what features you’d like to see to manage your diabetes.


This is an initial mock-up to help convey the idea. Click for a larger pdf version.


PROBLEM:  Diabetes management requires continuous collection, tracking, and analysis of vast amounts of data, however current digital interfaces for diabetes data have four crucial faults.

1)   Web-based programs that pull data directly off of diabetes devices such as glucometers, insulin pumps, and continuous glucose monitors are often designed by the device manufactures and are largely proprietary.  Therefore data is siloed, increasing both access and analysis difficulty.  For healthcare professionals, this means patients come in with data either manually entered on paper, printed from the multitude of proprietary web-based platforms, or still stored on the diabetes device and requiring download assistance.

2)   They usually track only two or three variables, medication (e.g. insulin), blood glucose (BG) levels, and sometimes carbohydrate intake.  This limited number of variables is grossly lacking in context for treating individual patients, thus lending itself only to very generalized treatment recommendations…a one-size-fits-all approach.  In reality many other factors play a major role in optimizing control such as monthly cycle, fat and protein content, glycemic index, physical activity, remembering medication, stress level, etc.

3)   They simply display the data but offer very little or nothing in the way of analysis assistance.  While graphs are helpful, they are insufficient to continuously analyze long-term patterns across multiple variables.

4)   They lack any means of patient engagement and empowerment.  Data entry is time consuming and often complicated.  Analysis is focused on patient error rather than rewards.

PROPOSAL:  In order to optimize diabetes management, we need to simplify and unify diabetes data analysis, while empowering and engaging patients.  Start by creating a single platform for integration of data from multiple diabetes devices. Then leverage the current Quantified Self, and Lifelogging trends to incorporate important variables via smartphone apps and wearable fitness devices that patients already use (many of which utilize an open API), allowing patients to create a custom diabetes management mashup of apps displayed alongside diabetes device data.  The individually tailored scope will have broad patient and doctor appeal, e.g., one patient may focus on testing BG 4 times a day, need a medication reminder, and passively track walking with a pedometer, while another may test BG 10-15 times a day, use a CGM, avidly track running, nutritional information, geolocation, mood, sleep patterns, and monthly cycle.

Enable doctor and patient data mining for potential meaningful and actionable correlations by applying prevailing 3D data visualization and comparative analysis techniques and algorithms (already in practice to mine big genomics data) e.g., compare several data sets to come up with key talking points for limited doctor-patient interface time such as, “75% of your nighttime lows occur during the week following your period in your menstrual cycle” or “On days when you walk after lunch, your BG is an average of 20% lower than on days when you don’t” or use a passive geolocation app to see what places you frequent when your sugars are higher/lower.

Finally, engage users in multiple ways.  The mashup format provides individual context for the data and will enable patients to set small, attainable goals that make sense for their lifestyle.  Incentivize both with rewards and positive feedback, e.g provide coupons and create a gamification aspect where you receive Target rewards as positive reinforcement, or earn badges for meeting your goals, or receive a text saying “Great job checking your BG 5 times today!”.  Integrate the site with social media, so users can share positive feedback and receive/provide support.   Integrate healthy meal suggestions with Target shopping lists.  Provide a Target kiosk to assist patients with uploading device data or setting up an account, etc.

Check out these related (less professional and more personal) posts:

My Fitbit Flex: A Delightful Surprise

Big Diabetes Data Requires Big Analysis