Diabetes technology teardown, 2019 edition

The tools and services I use to manage my blood sugar

Please do not interpret this post as medical advice. Everyone is different, and the tools I use might not work, or may even be dangerous, for you. Consult your healthcare provider before making any changes to the way you manage your diabetes.

Last year, I wrote a post detailing the different tech, gadgets, and apps I use to manage my diabetes. Given how fast the world of diabetes is changing, I thought I’d update this post every year to go through any changes I’ve made. I use all the products and services below almost every day and I’m not paid to write about them. Well, maybe except for Steady which I founded 🙂

I’ll cover a few updates in this 2019 post: The InPenDexcom G6 CGMHealthKit, and Steady.

InPen — a connected insulin pen and app

The InPen connected insulin pen

First up is the InPen from Companion Medical that I now use for my fast-acting insulin Humalog. I switched to InPen mainly because it’s connected to my phone and reports my doses to its own app and into HealthKit. It’s useful for several reasons like when I’m not sure if I actually took the dose or not, or, trying to remember how large the dose was or when I took it. Even fast-acting insulin is actually not that fast and needs about 1 hour to reach its peak effect. Sometimes it can be really helpful to remember how long it’s been since your last dose to help make a decision about taking extra insulin to prevent taking too much.

The InPen app is pretty straight forward but includes a few nifty features like a bolus calculator (which I don’t use due to not counting carbs), how much insulin I have on board and my total dose per day.

Finally, I’m also happy to be back to using cartridges instead of single-use pens as it reduces the plastic waste quite a bit.

Dexcom G6

The G6 from Dexcom

At the beginning of the year, I switched from using Libre + MiaoMiao + Spike to using Dexcom’s new G6 system as my CGM. There were a few reasons I made the switch. First of all, its fully integrated which means the sensor talks directly to my phone, no need for an extra device that does that part. Also, I had been having a bunch of issues with Spike where it drained my battery, became unresponsive, etc. The G6 is also calibration free so no need to poke your finger to ensure its accurate which is one less thing to worry about every week.

These improvements are great but the biggest difference took me a few weeks to realize. All of the numbers and information that Spike was giving me was actually causing a fair amount of stress for me. I constantly checked my phone and since it was much more sensitive, I found myself overcorrecting more which led to lows. With the G6 I have far fewer lows on a weekly basis.

With the G6 I have far fewer lows on a weekly basis.

Now, the G6 is not better on all dimensions. The alarm system can be super annoying, especially at night, and it can be frustrating to not being able to see delta numbers sometimes but all in all, it’s worth it for me. I know Spike also supports the G6 but for now, I’m going to stay with the official app.

HealthKit

This is not new per se but I’ve been using HealthKit much more recently. My favorite feature is that you can quickly see how much insulin you’re using per day and compare it with previous weeks. Why is that helpful? Well, it allows me to understand how my insulin sensitivity is affected. Or, like in the case below, discovering when the insulin in my pen has gone bad. Switching out the reservoir helped me come down from 18U per day to 11.

The difference in dosing between stale and fresh insulin

Storing all data in HealthKit is great because it functions as an interoperability layer between different apps and services. For example, the InPen automatically writes all doses to HealthKit which allows me to review data without any manual input. And, Steady then reads the data, importing it to their system, and lets my doctor see all my doses. Mind=blown!

Steady

Last, but not least, I’ve switched to a new endocrinologist, Steady Health. Not surprising as I also founded the company but it does feel really good to be able to say that I’m now, together with a great team of course, responsible for my own care.

First of all, I look forward to every visit.

So what is it like being a member? First of all, I look forward to every visit. No more general advice like “eat healthier” or “exercise more” but instead we look at real data from my life trying to understand what can be improved. And the team has access to not only my CGM but also food, exercise, and insulin doses which enables them to give me actionable advice that’s personal.

But the thing that is most valuable to me that I now have someone that understands the disease as well as my relationship to it, someone that holds me accountable. I was recently in a tracking session and had been struggling with some overnight lows when my doctor sent me this message:

Continuous coaching by the Steady team

I can’t tell you how much better that message made me feel and how much less anxiety and worry I experience knowing that someone is actually looking at my data with the intention of helping me improve.

I’ll cover Steady more in the coming weeks and months but if you live in California and want to check it out, you read more about us on https://steady.health.

Hope you enjoyed this 2019 diabetes technology update, if you have any questions don’t hesitate to reach out!

Start Analyzing your own CGM data in 5 Easy Steps

A Beginner Friendly Tutorial

In our previous blog post, we presented some visualizations of CGM data that was gracefully provided by Henrik Berggren, the founder of Steady and Type 1 since 20 years.

But, what’s even more interesting is of course to analyze your own data!

The goal of this article is to provide a follow along guide to everyone using a CGM to get started tinkering their own data, visualizing it and produce plots like these. Plots that can be useful extrapolate new insights about your lifestyle.

This a rather long and detailed post. To make it easy, all of the steps are described in detail to provide a thorough understanding of what we’re doing. The main segments (feel free to jump around to the ones that interest you) are:


What’s CGM and why should I care?

Continuous Glucose Monitor (CGM) sensors are wearable devices that measure blood glucose concentration at frequent intervals throughout the day, usually every 1–5 minutes. These devices have revolutionized the data available to diabetes patients and caregivers compared to finger pricking, the status quo of glucose measuring today. If you want a deeper background on the impact of CGMs, check out this post: The wearable that changed my life.

This new stream of glucose data enables a bunch of new applications like real-time predictive alerts for high/low blood sugar and closed-loop systems where insulin is delivered based on the current glucose levels. It’s also a gold mine for tracking change over time, behavioral patterns and, in short, data science!


Prerequisites

Skills

Some programming experience will be handy but even without it, you will be able to follow along. Most of the process here is a matter of copy-and-paste. Should any issues come up, don’t hesitate to write in the comments or send me an email.

Software

So what do you need? Amazingly enough just a browser and a Google account. We’ll be making use of Google Colaboratory which mean we do not need to install any software on our computer, how convenient!

Data

Of course, we also need some data to play with. I’ll be using a fake, generated, data set so that we can share it freely. Note, the data is a .csv in the format that comes from Dexcom which is what this implementation is designed for, but all sensor types are able to fit this guide with minor adjustments. With that out of the way, let’s get started!


Setup — Quick version

  1. View code used in this post here, Click “open in Colab” then “File” and “Save to Drive”.
  2. Export your Dexcom data following these 2 steps and upload the .csv file to the same folder as your code in Google Drive.
  3. Open the Colab file and change the filename (second cell) to the name of your CGM data file.
  4. Click “File” then “Run all”
  5. See all plots and visualization in this post appear before your eyes!

That’s is! If you’re more curious about what all this means and how to read all the plots, keep reading.

Setup — Detailed version

Make a new folder in your Google Drive, upload your CGM data into it. Then make a Google Colaboratory file by hitting “New → More → Colaboratory”.

If you haven’t used it before, you need to press “Connect more apps” and search for it. Now open the Colaboratory file.

For the first lines in your Colab file, enter the following

then click the “Play button” to the left of the cell in order to execute. This prompts you to authorize the Colab file to access your Google Drive files.

With our new access privileges, we can now read the CGM data we placed in our folder. We’ll do this by using the Pandas library like this

import pandas as pd
df = pd.read_csv('drive/My Drive/my-cgm-exploration/Example_CGM_data.csv')
df.head(20)

The head command previews the 20 first rows of our newly imported data frame which should be similar to this one (cut down to 5 for readability)

We notice that the first 10 rows lack time stamps and hold information like name, device info, etc. We’re not interested in those so let’s trim them out of our data frame by

Looking at our preview the data now looks more consistent, great! Although, we’re not going to use some of these columns, let’s pick the ones we’re interested in, namely blood sugar and what time it was measured. Also, let’s rename those long clunky names to something easier to type out

Alright, this seems more manageable. Now we’ll do one last thing before actually starting to inspect the data, most times with time series such as this we want to index the samples by, you guessed it, time. Pandas enable this with a few lines of code, make sure to switch the time zone ‘US/Pacific’ to the one which applies to you, list of time zones.

Okay, now we have successfully imported all the data, selected the data we’re interested in and done some basic formatting to make out lives easier.

We can now continue to quality assure the data then, at last, derive some insights from it!


Quality Assurance of the Data

If we’re going to produce some trustworthy insights from our data we need to make sure the data itself can be trusted. Measurement errors, missing values or similar anomalies can skew our results and cause trouble down the line so we better make sure we’ve dealt with them beforehand.

To keep things simple we’re not going to wrangle the data in any way here, just observe any anomalies and quirks with the data that might be good to have in mind later on.

Let’s check for a common issue, missing values. First, we’ll check at what frequency glucose sample values are to be expected

From this description, it’s clear that we expect a measurement every 5 minutes. However, some anomalies seems to occur as at least one interval is only three seconds and a period of missing data of a little more than 3 days are also present.

Repeated measurements are an issue that has a tendency to occur when syncing the device. Let’s take a closer look at those repeated measurements

We can tell that some measurements that are only seconds apart have glucose values that differentiate almost 15 mg/dl which is a lot more than is to be expected. We notice about 10 data points, in a 90-day period, have this issue.

Now, let’s review the frequency of samples that’s far apart, where we seem to be missing a lot of data.

Hmm, at two places we’re missing more than three days of consecutive data. We’ll need to keep that in mind going forward. Okay, we now have a better understanding of the frequency of our data, before we start visualizing the data we should get a grasp of how our glucose values are distributed.

Looking okay here, the min and max values are well within what’s to be expected. To conclude this section, we now have a sense of what the data looks like, we’ve discovered some anomalies to keep in mind, like our missing values.


Analytics and Visualizations

Okay now the fun part, let’s start visualizing the data to reveal some interesting insights. This section will be laid out by asking questions, then trying to let the data and plots give us the answers.

What’s the trend of my average blood glucose?

We’ll resample daily to get a mean glucose every day, then have a look at whether we can see a trend in the data or not.

Here we’ve fitted the daily glucose averages with linear regression to investigate if we can see any significant trend in the data. Clearly, we can’t distinguish any evident trend here, especially considering the wide confidence interval (the shaded area around the regression line).

Although, we can tell that on average the mean blood glucose for a full day is about 115 mg/dl.

What’s the trend for time spent in range?

You neither want your blood sugar to be too high or low for extended periods of time, therefore time in range is a common metric for how diabetes management. We’ll apply these standards thresholds for blood sugar in range then see what that trend looks like

Okay, looking good more time spent in range! But wait, what’s those days where close to 0% were spent in range..?

Yes, it those days where we missed data, let’s ignore them and try again,

So, actually, we notice a slight negative trend when we’ve removed the bad data, not very significant though but nonetheless an indicator of how time-in-range have progressed the last 90 days.

How’s my average day looking?

Time in Range is not a perfect measure, we want to account for variance too. Let’s have a look at an average day and see what’s the common trends.

Here we can see the variability and trends of blood sugar throughout the day. There’s a general trend of night-time highs — where we can see higher than average glucose levels after bedtime.

Does the day of the week influence my blood sugar?

I always look for an excuse to use violin plots, they look so nice! In this case, they actually make a lot of sense, let’s group our data by weekday and check whether the distributions differ depending on the day of the week

Immediately we notice that Thursdays, Fridays, and Sundays display a larger variance than other days.

We can take it one step further and add in how progression looks for each day. We’ll separate the first and second half of the data and review it again. We can then observe that Fridays are getting a lot better but on Sunday’s there’s been more time in high recently compared to before.

Are there any evident daily or periodic trends?

As we saw in the first plot, recurring patterns can be seen during the day. Let’s dig into this idea and observe the Blood Glucose samples in a heatmap.

A heatmap gives us a bird’s eye view on the data and enables easy detection of patterns. For example, scanning across the nighttime portion of the plot, we can see infrequent streaks of red (highs) broken up by chunks of dark blue (lows). Perhaps tighter glycemic control during nighttime/mornings might be worth experimenting with in this case.


Conclusions and Related Content

In conclusion, what we’ve done in this post is to import our data, reviewing the data quality and made some basic visualizations. The beauty is that now you can keep adding content to the Notebook and try new exciting thing to analyze and use your data in whatever way is best for you!

Why do this and step out of the already provided data views by the sensor providers? I’m of the opinion the CGM data entails a lot more than it’s utilized for today, dietary patterns, exercise impact maybe sleep cycles? The point is that we’re not yet aware of all the potential in this data and the way of finding out where the limit lies is by experimenting with it!

If bio-hacking, CGM’s and data science interest you, this might too:

Learn more about Steady

Thanks to Henrik Berggren.

Data Exploration at Steady Health

Using CGM data to answer your questions about diabetes

At Steady Health, data and care is our focus. To understand our members’ data, we started asking some fundamental questions and thought we’d share our early findings. In coming blog posts, we’ll cover the process of how we use technology to empower our clinical staff to provide world-class care.

If you haven’t heard of Continuous Glucose Monitoring (CGM), it’s a wearable sensor that measures blood glucose every 1–5 minutes. It has revolutionized the ability to make data-driven diabetes care decisions. To read more about the impact of CGM check out The wearable that changed my life. The author, Henrik Berggren, has also gracefully provided 90-days of his Dexcom data that we’re about to explore in this post.

Okay, let’s jump in! We have a bunch of questions on our data, let’s start visualizing the data to reveal some interesting insights.

These sections will be laid out by asking questions and try to let the data and plots give us the answers.

What are the daily trends?

Let’s have a look at an average day for a Patient, meaning all glucose values aggregated and shown over 24 hours. Our goal is to uncover any trends and cycles in the Blood Glucose data.

Member glucose data aggregated and Displayed on 24–hour Plot. The shaded area is one standard deviation from the Mean (solid line).

Here we can see the variability and trends of blood sugar throughout the day. There’s a general trend of night-time highs — where we can see higher than average glucose levels after bedtime.

Further, we see higher variability around the same time. Something to revisit.

Time In Range, a well-known success-metric for people living with diabetes, however, is not perfect. It tracks how much time a Patient had spent neither with a too High not too Low blood glucose. However, we want to account for variance too. As we observe, most of the member data fall within the standard definition of In Range. Broadly suggesting good glycemic control. Yet looking at the valleys in the plot, we can see the lowest blood glucose reading occurs pre-breakfast and early in the morning. Showing a cyclic nature shown, which Time In Range fails to surface.

Do Weekends differ from Weekdays?

I always look for an excuse to use violin plots, they look so nice! In this case, they actually make a lot of sense too. Let’s group our data by day-of-week and observe how the distributions differ. Et violá.

Glucose distribution for a given day of the week

Immediately we notice that Thursdays and Friday display a larger variance than other days. (Which can be fine, if the member applies the appropriate correction Insulin dosages.) Definitely, something to revisit with this member.

Other days look pretty solid, with a well-bounded distribution. Monday, Tuesday, Wednesday and Saturday all look to be In Range.

From these insights, we can uncover behavioral triggers, like going for runs on Wednesday, or Sunday brunches. We want to make actionable items with the member based on these insights.

We can even take it one step further.

Let’s add in data after a coaching session with these behavior goals. The left-side (green color) of the plot below is pre-coaching and right-side (salmon color) is post-coaching.

Violin Plot split with two batches of member Glucose Data (Green = Pre-coaching, Salmon = Post-coaching)

We can then observe that Fridays are getting a lot less variable, but on Sunday’s there’s been more time in high recently compared to before. Cool!

Are there any daily or periodic trends in the blood glucose data?

As we saw in the first plot, recurring patterns can be seen during the day. Let’s dig into this idea and observe the Blood Glucose samples in a heatmap.

Blood Glucose samples for 90 days. Dates on X-axis / Time of Day on Y-axis

A heatmap gives us a bird’s eye view on the data and enables easy detection of patterns. For example, scanning across the nighttime portion of the plot, we can see infrequent streaks of red (highs) broken up by chunks of dark blue (lows).

Perhaps tighter glycemic control during nighttime/morning might be worth experimenting with.

Looking at the big picture, apart from the short episodes mentioned above, no major periodic trends are evident for the 90-day period. In cases where the member is moving cities, starting a new job, or training for a marathon, these birds-eye view plots is a useful tool to uncover larger trends.


Summing up

This was just a brief introduction of how member data is our ground truth and how we let our focus narrow until we can draw conclusions. With this process, we can build amazing things like meal detectors, predictive systems, queryable data and so on.

If these kinds of questions interest you, drop us a line at hello@steady.health. We are hiring!

Stay tuned for future blog posts, including how to start exploring CGM data yourselves!

Learn more about Steady

Thanks to Sid Ghodke and Henrik Berggren. 

A new, Steady, path

Why a seasoned Diabetes Educator decided to leave the comfort of a large, established system, and join a startup

The winding roads of Mount Tamalpais, California

The simple answer

Diabetes changes frequently. Research unveils new learnings about this disease regularly. Management tools are continually being introduced. I believe Diabetes care needs to be as fluid as the individual living with Diabetes and there is a huge opportunity for improving how care is delivered.

The complex answer

Diabetes is a disease that requires a team. A team to support one on a regular basis through a daily, weekly, monthly rollercoaster ride. While many people have great personal support systems, they likely have limited medical support when they most need it; when they are in the middle of troubleshooting.

Today, care is often limited, inconvenient, dependent on unreliable, fragmented technology and is, ironically, not patient-centered.

  • Access to providers is limited by insurance plans or overbooked schedules with far too short appointment times.
  • Getting to an appointment (travel time, finding parking, waiting, waiting, waiting in a cold and unwelcoming office for something that will never address all your concerns and issues) is incredibly inconvenient.
  • Providers are dependent on unreliable, fragmented, technology to access critical data for making safe decisions and recommendation.
  • With so many daily hurdles, ironically our system of care is becoming less and less Patient-centered.

However, far and beyond the limitations of current diabetes care, it is painfully obvious that tools to manage diabetes are being developed at a pace which far exceeds what current systems can support. The infrastructure of traditional medical practices are just not able to respond at the pace solutions are being introduced. Because of this, I found myself spending more and more time helping people with technical support, which ultimately resulted in less time doing what I was professionally trained to do…help people better manage their diabetes.

A more personal perspective

Change is never easy nor something I seek out. However, it is part of life and a particularly large part when living with diabetes. Having lived with T1D for over 35 years, I understand that every day presents a new challenge.

Having lived with T1D for over 35 years, I understand that every day presents a new challenge.

So when I was presented with an opportunity to contribute to something really different, I hesitated, assessed, stalled, questioned, and ultimately avoided making a decision. Too many specifics about this care model raised questions for me. While I knew my current job was not perfect, it was an established place of work with clear processes and procedures to follow if I veered off track. This was comforting to me.

As we all know, this disease is unpredictable and changes so frequently that obtaining details to events which happened 3 months ago is virtually impossible to recall at tradition quarterly provider appointments. And we also know that logging is one of the toughest things to do when living with diabetes! If someone sees a pattern or has time-sensitive questions, they should be able to easily document/record it and then discuss with a provider in a timely manner. These type of interactions simply do not require a face to face appointments. They require a brief discussion and a review of data. This can all be easily accomplished with the model we are building at Steady Health.

Merging technology/data management and care are buzz phrases right now. Everyone is talking about how much better health(care) should/can/will be with access to the right technology. And while I do believe technology will contribute to improved health outcomes, this is not what drove me to join. Rather, I was drawn to the idea of a high touch service that provides remote care to patients between face to face appointments as I believe Not to mention the amazing talent of the diverse team of clinicians and engineers, all working toward a single goal.

Ultimately, I asked myself

“How will you feel when you learn that Steady Health has hired a Diabetes Educator?”

That was it. That’s all I needed to make my decision. I believe so strongly in Steady’s model for providing care to persons with diabetes, that I did not want to miss out on such a unique and impactful opportunity.

If you feel the same way I do about diabetes care, I’d encourage you to give Steady Health a try. We’re currently accepting members in our first San Francisco location. Head to our website to sign up. We look forward to helping you manage your diabetes in a whole new way!

How I used experimentation to improve my diet

Please do not interpret this post as medical advice. Everyone is different and the changes I made might not work, or might even be dangerous, for you. Consult your healthcare provider before making any changes to your diet, medication or any other part of your routine.

If you are interested in improving your diabetes management using data and technology, head over to steady.health to find out more about my new company.

Introduction

Last year I published a post about how I had started to use experimentation to improve the management of my diabetes. It’s been a few months so I thought it was time to follow up with a progress report. Back then I saw pretty dramatic improvements to my overall A1C, lost weight and was feeling better than ever. I can proudly say that I’m still on an improvement path several months later, and I’m eager to share how I’m doing it. Let’s start with how I used to approach it and why I believe it’s broken.

The traditional way

Before CGMs people relied on a basic method to figure out how much insulin to take for each meal: carb counting. Carb counting is what it sounds like, a manual effort to assess carbohydrates by estimating the carb density of each ingredient on the plate. You quickly realize that for some meals like casseroles, soups etc. its almost impossible to figure out unless you made them yourself from scratch. But, even for meals with fairly distinct ingredients, it turns out to be surprisingly hard. For many years I struggled with this method, not understanding why some meals impacted my blood sugar more than others.

In a study done by researchers at Wingate University, they found that the average accuracy for carb counting was only 59%. What I find remarkable about this is that the average length these subjects had diabetes was 26 years(!). Let me put those facts together, after 26 years of practice, the average accuracy is less than 60%. Ouch. That seems like a pretty poor approach as taking the right amount of insulin for meals is key to good management, if you don’t you’ll end up high, or even worse, low, and have to course correct. What is even more surprising is that even with these results, carb counting is still one of the key methods being recommended by doctors and the ADA for improved management.

The alternative

I decided to take a different approach: experimentation. After getting a CGM I realized that if I could remember what I had eaten and the insulin dose I could use the blood glucose response to figure out if I had taken enough for that specific meal. And next time I eat the same dish I can look back at last time and make a more informed decision.

I recently moved into a new office and that creates some challenges in managing my levels as I’m exposed to a set of new restaurants and take-out places. After a few weeks of trial and error, I’ve found a bunch of lunch options that I like and that are easy to manage with insulin

  • Mixt salads are great but the difference in insulin need between them can be significant, for the Cobb salad I only take 2 units, but for Orchard, I need 3
  • The local taco truck is good but only if they serve corn tortillas, flour ones are harder to manage
  • Chipotle bowls are surprisingly easy if you skip the rice, neither guacamole or beans require much insulin for me
Mixt salad, Tacos, and a Chipotle bowl

There are some challenges with this approach as well:

  • Remembering to log meals
  • A tool to look at food pictures, insulin dose, and your CGM result together doesn’t exist today which makes it hard to piece the data together
  • There is no simple way to share the data with your endocrinologist

But, all in all, this method has proven easier, more accurate and consistent for me which brings me to…

Long(er) term results

I’ve been tracking my own data using the Steady platform over the last 6–7 months and I’m pretty excited about my progress.

First of all, let’s look at my average blood glucose on a weekly basis. As you can see, its pretty stable around 112 mg/dL.

Average numbers only capture part of the picture so a better metric is how much time I’m spending inside of my range of 70–140 mg/dL. This, coupled with my average being stable means less variability which is great.

But, what about lows? This might be the progress I’m most excited about. I set a goal for me a few months ago: No more than 3 urgent lows per week, and it seems like I’ve been making strides. Keep in mind that this progress is coupled with stable averages and a reduction in variability 🙌

Learnings

Experimentation is definitely worth trying if you’re struggling with diabetes, or, just want to get your blood sugar under control. It has significantly improved my health and at the same time made my diabetes easier to manage. If you’re interested but don’t know where to start, head over to steady.health and sign up for more information, or, send me an email at henrik at steady.health

Thanks to David Kjelkerud. 

Diabetes technology teardown — The tools I use to manage my blood sugar

Please do not interpret this post as medical advice. Everyone is different, and the tools I use might not work, or may even be dangerous, for you. Consult your healthcare provider before making any changes to the way you manage your diabetes.

Update: I’ve decided to build a new company focused on improving diabetes care using data and digital tools. Sign up to be an early member here!

Introduction

After publishing my last post about how I used data from a continuous glucose monitor to dramatically improve the management of my diabetes, I got a lot of questions about which tools I use. Getting everything set up was not straightforward and required a fair amount of research, trial, and error. So I thought I’d contribute my experience and hopefully help others.

Medicine, Devices, Sensors & Apps

I’ve divided my tools into four overall categories to make it easy to overview:

1. Medicine

Let’s start with the basics. To help the body process glucose I use insulin like most diabetics. I use standard injection pens and typically inject 4–7 times a day depending on what my blood sugar level looks like.

Injection pens with Humalog & Tresiba, and glucose tablets

If my levels drop too low, there is a simple fix: snacking. However, there are challenges with eating whatever I have lying around my house or office as I typically don’t know what effect it will have on my levels. Instead, I always carry glucose tablets that I know, through testing, will increase my level about 20 mg/dL. And, the increase is almost instant; I don’t have to wait for the tablets to go through my digestive system.

2. Devices

I’m an Apple person and have been for the past ten years. I have Apple devices exclusively in my home except for my Amazon Echo. These devices are used for many more things than just helping me manage my blood sugar, but they all play a crucial role. I use a MacBook Air as my primary work computer and spend about 10–12 hours on it every day. My phone is an iPhone X, and the watch I currently use is a first generation Apple Watch. The watch is helpful as it enables me to glance at my blood sugar level without having to pull out my phone. It also helps me track physical activity more accurately, and I use the data it collects to better understand how being active effects my blood sugar.

Note: Some of the tools mentioned work for Windows as well and for Android there are alternatives so don’t be discouraged if you’re a non-Apple user.

3. Sensors

Now to the fun stuff. For continuous glucose monitoring, I use the Freestyle Libre sensor from Abbott. It’s a so called “Flash Glucose System ” that uses near-field communication to enable the wearer to scan it with a standalone reader to see his/her latest data. I like the Libre sensor because of its low profile and how easy it is to apply to your arm — which you have to do every ten days. And for a while, I was carrying the standalone reader device that comes with it. In Europe I believe you can also use your phone to scan the sensor but that feature is not available in the US, yet.

MiaoMiao device & Freestyle Libre sensor

Carrying an additional device that I had to keep charged etc. was fine for a while but ideally I wanted everything to be on my phone. Then I came across a small device called the MiaoMiao reader. It’s essentially a little add-on to the Libre that you place on top which automatically scans the sensor every 5 minutes and sends the data to your phone. This add-on transforms the Libre to a fully-featured continuous glucose system. You attach the MiaoMiao reader on top of your Libre sensor using adhesive stickers, and even though it does make it slightly bigger, it still sits comfortably on my arm without being in the way too much. It is also rechargeable, which is something most other Bluetooth add-ons are not.

4. Apps

Now that we’ve covered the medicine I take, the sensors that measure my levels and the devices that receive the data let’s go through the different apps I use to bring it all together.

First of all, the app that I use more than any other app on my phone is called Spike. Spike displays my blood sugar data as a graph over the last 24 hours in 5-minute intervals. It also gives me the direction my level is going and how fast. I can’t overstate how helpful this is as I try to figure out if I need to make a treatment decision. Spike is the product of a thriving hacker ecosystem that exists around diabetes management tools. Since many of these tools require approval by the FDA to be official and published in an App Store, many developers have opted to use alternative paths like TestFlight or by merely allowing users to compile them on their own.

Spike has tons of functionality like alarming on highs and lows, allowing you to input your insulin and carb intake, and much more. It’s the most essential app on my phone by a significant margin.

Spike on iPhone X and Apple Watch

Lastly, Spike also has an app for my Apple Watch that allows me to glance at my wrist to check my level. It’s incredibly valuable in situations when your phone might not be convenient to pull out such as when I’m riding my bike or driving.

But tracking and managing my levels on a day-to-day basis is only one of the benefits of having access to continuous data. Many of the lessons I’ve learned have come from analyzing data in aggregate, allowing me to compare days and weeks with different routines, diet and trying out different types of insulin. To do this, I use a system called Nightscout. It’s more of a platform and is also created and maintained by a community of open source developers. Nightscout collects all of the data generated both by my sensor and Spike and makes it available through a web interface as well as an API. This is great because through it you can allow other people to “follow” you and your curve in almost real time. It’s especially valuable for parents and others that are guardians of people with diabetes.

For me, Nightscout provides two key features: reports which allows me to view more substantial amounts of data in an aggregate format and that I can share with my endocrinologist, and an API that allows other applications, that I authorize, to get access to my data.

An example of an app that uses the API is something called Nightscout Menubar. It’s exactly what it sounds like, an app that displays my blood sugar level in the menubar of my Mac, letting me easily glance at it while I’m writing posts like this for example 🙂

Nightscout menubar

Nightscout has a ton of features and is backed by a very active community so if you’re interested in learning more this is a good start.

In my previous post, I briefly mentioned using Excel for tracking experiments that I ran on myself. It’s quite straightforward and allows me to keep track of the things I eat & drink and activities like running and playing basketball. Since I have all the data saved it’s easy to go back and fill out data after the fact should I forget to enter it when it happens. The sheet with the experiments doubles as a tool to help me remember how much insulin to take for the salad at my favorite café, Reveille.

Excerpt from my test sheet in Excel

Lastly, I recently started using an app on my phone to track my sleep called SleepCycle. Lots have been written about sleep quality, the amount of sleep we get, and it’s impact on blood sugar. Anecdotally, I feel like my levels are harder to manage on less sleep. I look forward to tracking for a while longer and will publish any findings I uncover.

Why all of these tools?

As you can see, it’s not the simplest of setups. But these tools enable me to do two crucial things:

  1. Tightly manage my blood sugar levels on a daily basis, correcting highs with extra insulin and lows with glucose tablets.
  2. Analyze my data, develop insights and find patterns that are otherwise hard to spot.

Both of these helps me to learn more about how the disease is affecting my body. And, the more I learn, the more personalized my strategies for management become.

What’s missing?

There is still room for improvement. Many of these devices are still too expensive, not accurate enough and the data is hard to bring together and make sense of. I think the ecosystem today is missing two important and valuable features. First is being able to set clear goals for what you want to work on. Lifestyle changes are hard to sustain and focus is essential. And there are always tons of things you could be working on, but the key is to concentrate your efforts on a few clear goals and learn from them.

Which brings me to the second one — testing and iterating. For building websites for example, many services help you test ideas and analyze the results. How come there isn’t something like that for our bodies? I hypothesize that we’ll discover that making a few critical changes in things like diet and medication can have tremendous impact on our overall well-being.

I hope you enjoyed this follow-up post. I’m always curious to hear what setups other people with diabetes use so please reply in the comments or send me an email at henrik -at- henrikberggren.com

Thanks to David Kjelkerud. 

The wearable that changed my life — How a continuous glucose monitor got my A1C down to 5.8%

Please do not interpret this post as medical advice. Everyone is different and the changes I made might not work, or might even be dangerous, for you. Consult your healthcare provider before making any changes to your diet, medication or any other part of your routine.

Update: I’ve decided to build a new company focused on improving diabetes care using data and digital tools. Sign up to be an early member here!

Introduction

As of this year, I’ve been a Type 1 diabetic for 18 years. And I’ve never felt as good as I do today. What got me here was made possible by a small circle-shaped sensor placed on my arm called a continuous glucose monitor, but let me back up and start from the beginning.

When I turned 20 years old, I was doing military service in the Swedish navy and for a couple of months, I had been feeling pretty bad. I was dehydrated, I had to go to the bathroom ten times a day on average and I had terrible mood swings. I thought I was suffering from an extended cold until I met the navy doctor. After listening to my symptoms, he decided to prick my finger and check my blood sugar level. I think I remember it being around 500mg/dL (normal is 90) and the doctor said: “We have to take you to the hospital.” I was diagnosed with diabetes.

The Freestyle Libre sensor from Abbott

The complexity of diabetes

Diabetes is often mischaracterized as a single disease when it really describes a set of symptoms that can have multiple different causes. Traditionally we’ve divided patients into two categories: Type 1 and 2, or juvenile and adult, but some new research has emerged recently suggesting there are five different clusters with slightly different characteristics. In this new categorization, I’d be in the “Severe Autoimmune Diabetes” camp which is the most severe form of the disease.

Diabetes is a disease that, due to the lack of insulin, prevents your body from processing glucose (a simple form of sugar used as fuel in your body) which results in elevated levels of it in your bloodstream. Over time, high glucose in your blood will damage your organs and lead to cardiovascular disease, stroke, blindness, kidney failure, and heart attack. For people with the condition, this means having to inject insulin on a daily basis to try and make sure your body gets enough to cover things like food.

As a diabetic, you have a wide range of tools that help you manage your blood sugar levels at any given moment: syringe pens filled with injectable insulin, a blood sugar meter, glucose tablets and more. Managing your levels is not only about preventing highs; as a Type 1 diabetic, you also risk taking too much insulin resulting in dangerously low levels which can lead to a coma and even death. It’s a delicate balance that every person with diabetes has to manage closely on a daily basis.

My personal journey

During the last 18 years, I’ve been managing this balance decently. The most well-understood metric for diabetes management is A1C. Simply explained, it’s the average blood sugar level over a three month period. My A1C has hovered around 7% for many years, and I’ve been struggling to understand what changes I need to make to improve it. 7% is not bad, but it’s not good either, and it might lead to complications down the line.

A1C levels and their corresponding glucose level

A few months ago that all changed. Over the last couple of years, more than a few diabetics had been telling me to get a Continuous Glucose Monitor (CGM) and last year I finally decided to try one out. A CGM is a sensor that you place on your body that continuously measures your blood sugar throughout your day giving you many data points on how your levels are changing. The sensor is not connected directly to the bloodstream, but instead measures interstitial fluid and then uses an algorithm to predict the blood sugar level. Having a sensor continuously measuring your blood sugar levels without you having to think about it has several advantages:

  1. No more pricking your finger several times a day with bulky equipment
  2. Measure when it was previously impossible, like when working out or sleeping
  3. More frequent data making it easier to connect highs and lows to specific actions like a meal or work out
12-hour CGM curve from my Nightscout app

Enabling experimentation

As a product manager, I’ve always believed in the old cliché “You can’t improve what you can’t measure” and CGM entering my life has been a testament to that. After trying it out for about a week, I started realizing that I now had a real feedback loop in front of me. A feedback loop that would, almost in real time, tell me how food, physical activity, medication, sleep, and other daily inputs affect my blood sugar. So I started doing what product managers always do when they have a feedback loop set up and know what success looks like: run experiments.

The feedback loop enabled by continuous monitoring

To run these experiments I mostly used Excel. I entered my starting blood sugar value, what I was eating, and how much insulin I was taking. After 2 hours I entered what my resulting value was. Using this method, I could not only see what was impacting my sugar levels but also how well I was assessing the amount of insulin needed to cover a meal. I kept going, uncovering more insights. For example, playing basketball raises my blood sugar quite a bit, likely due to cortisol released into my body from stress. Since I play every week, I started developing strategies for mitigating some of its effects. When I play nowadays, my blood sugar still rises but not even close to where I used to end up. To be honest, I would likely not even have found out basketball was the reason for my elevated blood sugar levels because I never would have tested blood sugar during a game!

Screenshot from my Excel sheet filled with experiments

The results

After about four months of testing, iterating, analyzing and making changes, I went to get my first A1C. I had seen pretty dramatic improvements in the data produced by my CGM, but honestly, it seemed too good to be true. At the same time, I had seen my insulin needs reduced by more than 50% which in itself is a pretty remarkable change. And as a result of my improved diet and lowered insulin dose I had lost 7 lbs, which is something I had been trying to do for a while. Then I got the results: 5.8%. Wow. I was feeling better than ever, on less insulin, had more stable energy levels and had a fantastic A1C. I had no idea this was even possible. Now, some people would ask, does this improvement also mean that you have more lows? It’s hard to tell as there isn’t an as clear metric for hypoglycemia, but anecdotally my feeling is that I’m actually having fewer lows now compared to before.

Why did it take this long?

After processing the changes I had made and the impact it had on my life I started to wonder, why hadn’t I done this earlier? I came to the conclusion that before I had ever tried out a CGM, I wouldn’t have known it would unlock this much value. The behavior it enabled was learning. Learning about how food, physical activity, and medication were affecting my blood sugar levels, in an almost real-time way and making it possible to run many experiments in a short amount of time. It also taught me that some of the advice my care provider had given me was not right for me. It turns out that many parameters are different between each person with diabetes. Things like body mass index, metabolism and climate all affect how your body reacts, and even the same kind of food can impact the blood sugar of two diabetics differently. These parameters are today mostly ignored when diabetics get advice on how to manage their condition. Even though we now have the data to understand each person on an individual level.

I suspect that one of the reasons that more people haven’t seen the type of results that I have is that getting there took a lot of time. I was fortunate enough to have most of last year off from work so had the opportunity to spend time on my health, whereas most people don’t. And even though I’m an engineer and product person, getting to the insights that were buried deep down in the data from my CGM was not easy. I had to do lots of manual reporting in Excel and repeat tests several times due to high variance and unexpected effects from previous decisions. It also required some knowledge of data analysis and basic script writing. Simply put, I don’t expect many people, even though they get a CGM, to be able to achieve the results I did, which is unfortunate as it has had a tremendous impact on my life.

If you have diabetes or know someone who does, I suggest you ask your doctor about getting a CGM. And if you want to know more about how I achieved my results, or if you just get excited by the promise these types of sensors and the data they collect, send me an email, henrik -at- henrikberggren.com

Also, check out my technology rundown for an explanation of all the different tools I use to manage my blood sugar: https://medium.com/@henrikberggren/technology-diabetes-manage-blood-sugar-1c1f5ae4fcc1