Connecting with diabetes patients through data

A doctor’s first meeting with Continuous Glucose Monitoring

After six years of practicing internal medicine at a Swedish County Hospital, I moved to San Francisco to build a personalized, tech-enabled diabetes clinic. This story is about how patients helped me identify the next step in my career.

Four years ago, one of my patients with Type 1 Diabetes brought a continuous glucose monitor (CGM) to my office for the first time. A CGM is a device that attaches to your skin, and a thin filament is inserted a couple of millimeters to measure the glucose level between cells. The sensor transmits data wirelessly to a smartphone app that calculates the corresponding level of glucose in your blood.

A CGM sensor on the back of my arm

My patient loved the device since she no longer had to prick her finger every time she wanted to measure her blood sugar. It also showed if her blood sugar was trending up or heading down, which helped her make smarter decisions about how much insulin to take. After some research and help from our diabetes nurse, I figured out how to hook the CGM up to my computer to view the data, and I was blown away by the depth of information! Until that day, I had done my best to give people advice around their insulin treatment after looking at a few spot measurements per day, but now I had a continuous curve that showed me exactly what had happened after each meal and insulin dose.

Learning from patients

There was just one problem: I had no training in analyzing this type of data, and it was hard to find any documentation on how to go about it. So I did what every doctor should when encountering a new problem: I learned from my patients. After all, they make hundreds of treatment decisions every day!

The first thing we reviewed was the Ambulatory Glucose Profile (AGP). AGP is a visual representation of average blood glucose by the time of day, overlaid with glucose variability where we could see just how much their glucose fluctuates.

An AGP showing blood sugar trends for the past month (source: Dexcom Clarity)

Understanding trends in CGM-data

The AGP is an excellent tool to quickly grasp how you have been doing over the past couple of weeks. It also provides some clues if there is a particular time of day where blood sugar management is more challenging.

Let’s pretend that the graph above is yours. We can see that your average blood sugar is highest in the afternoon and comes down around midnight. The obvious next question is: Why does your blood sugar increase? To answer this question, we have to take a closer look at a single day, and this graph is from last Tuesday.

A graph showing blood sugar variations during one day (source: Dexcom Clarity)

As you can see, blood sugar increased in the afternoon. So the obvious question is: What did you eat for lunch, and how might you have avoided this increase in blood sugar? Let’s say that you had a lunch containing 30 g of carbs and took 3 Units of fast-acting insulin to cover it. A couple of more questions come to mind.

  • Was there really 30 g of carbs in that meal, or was that just a guesstimation?
  • Is it right for you to take 1 unit of insulin for 10 g of carbs?
  • Was the meal a rice bowl with carbs that are absorbed quickly, or a salad with lots of fiber that delay the gastrointestinal uptake?
  • Did you take your insulin 20 minutes before eating (a pre-bolus), when you started eating, or even after eating?
  • Did you exercise after your meal, or did you sit down in your car, stuck in traffic, stressed out over going late for a meeting?

These factors all have a profound impact on blood sugar levels. In fact, there are up to forty-two factors that can affect blood sugar, and that’s why those daily decisions can be so tricky.

You need to log more than just blood sugars

Now, I want you to stop for a short moment here to ask yourself: “What did I have for lunch Tuesday last week?” Do you remember all the details? Most people don’t! So I asked my patients to log their food and insulin and bring their diary to their next appointment.

The resulting discussions were inspiring! I started to appreciate how challenging it is to manage blood sugar levels, but I also got lots of insight into why some people found it overwhelming to figure out how to handle situations that others seemed to manage so easily. The question started to rise within me: Could I empower those struggling patients to learn about their bodies, improve glucose control, and reduce the risk for complications like heart attacks, kidney failure, and even blindness?

I had started to build some skills when it comes to interpreting data and talking about it with my patients, but a few hurdles remained. First of all, I didn’t have access to my patients’ data until they were in my office. Most patients found it very difficult to set up the software needed to upload data to our system from home, and to juggle the cgm-data and hand-written notes took a lot of time from us talking about blood sugar management. Using the tools we had, our meetings were not long enough to dive deep into the details of their blood sugars. I continued to work with a few very motivated patients logging their behavior, but when it came to my dream of empowering patients that struggled, I had come to a dead end.

The solution to my problems — Steady Health

And that was when I received a call from my friend Henrik. He’s an engineer living with diabetes who had learned a lot from analyzing his CGM-data and wanted to help more people with diabetes to go on the same journey. And it turned out we had both identified the same fundamental problem, that the tools available to patients are fragmented and poorly integrated into health care. So many months of research, we decided to build a system that allows people to share their data with health care and helps doctors and nurses analyze the results. But we didn’t want to stop there. We knew that traditional healthcare organizations are ill-equipped to provide the personalized and continuous care needed to support people with chronic conditions, and we wanted to do something about that too!

So we decided to build Steady Health, San Francisco’s first clinic specialized in helping people to learn about blood sugar management based on CGM-data. We have a fantastic care team with years of experience of coaching people with diabetes to reach their goals. And they work in tandem with our engineers to create tools that help them analyze data and communicate smoothly with patients between visits.

We are currently wrapping up our first pilot project, where we have worked closely with 12 people living with type 1 diabetes. They have used our app to track their blood sugar and the behaviors that affect it. I’d love to tell you what we have learned and describe the coaching model that they helped us develop, but that is another story that deserves a post of its own. Stay tuned!

To learn more go to or email us at

Thanks to Henrik Berggren and Aimée José. 

Leave a comment

Leave a Reply

%d bloggers like this: