Dana M. Lewis | #OpenAPS 🤖 Profile picture
Nov 9, 2019 19 tweets 11 min read Read on X
THREAD:

(from my presentation at #DData19 regarding what we're learning in the real world with DIYAPS systems like #OpenAPS.)
First, some context - it was 5 years ago at one of the early DData meetings that @scottleibrand & I shared about the smart, louder alarm system (#DIYPS) that we had created & I had used in the last year...and what we planned to do next, like trying to close the loop. #DData19
And we did! We closed the loop for me in Dec. 2014, and launched #OpenAPS in February 2015.

Why was it worth doing? Automated insulin delivery has so many benefits compared to old-school "manual" diabetes. #DData19
But one early piece of push-back we got was: "well, it clearly works for you, but how do you know it works for anyone else?"

We didn't, but we felt it was important to share design (openaps.org/reference-desi…) & code & let others build from it & learn from what we learned. #DData19
To help answer the question if it worked for other people, though, we did a survey of the early adopters in the community and presented it at ADA Scientific Sessions in 2016, and also followed up with a publication about it in a journal. #DData19 Image
And in the last 5 years, we now know that there's not just one, not just a few, but many thousands of people globally using DIYAPS systems.

And here's what we've learned in the real-world community as a result:

#DData19
1. It works well for a lot of people.

There's oodles of research now published in the literature & being presented at scientific conferences. They all generally have the same results; A1c goes down, TIR goes up, time spent hypo and hyper goes down. #DData19 Image
And, there have not just been retrospective real-world data studies, there have also been observational studies, prospective studies (including this one at EASD: bit.ly/2NVfOmC), and there are RCTs of DIYAPS in the works. #DData19
2. We also are learning *why* people are choosing to DIYAPS.

The @OPENDiabetesEU project (of which I am a part) did a survey last year called "DIWHY". This survey also showed TIR & A1c improvements, as well as revealed many of the reasons why people are using DIY. #DData19 ImageImage
3. DIYAPS may use "simple" algorithms, but they're effective. "Machine learning"/"AI" is not required to get best-in-world results.

For example - with slightly faster (~45min peak time) insulin and algorithm improvements in DIYAPS, I no longer have to bolus for meals. #DData19
So no matter what life throws my way - like my nephew giving me + 16 other people norovirus a few years ago at Thanksgiving, where I didn't eat for 3 days 🤢, or falling off a mountain and breaking my ankle in 3 places - I do less work, and get fantastic results. #DData19
4. We also know we can quantify different diabetes behaviors, and separate the impact of these behaviors.

For example: what is the difference for A1c + TIR in someone who stops meal bolusing and announcing meals? Or mostly announces but doesn't bolus? etc. #DData19 Image
5. And we now have data & capability to answer questions we never thought possible for individuals...like understanding when pump sites become ineffective on average; where does someone get the best accuracy based on CGM body placement; best methods for exercise; etc. #DData19
6. But there's still a LOT of low-hanging fruit left to improve QOL for PWD. It's not enough to have these capabilities if we don't use them. This is a call to action for you, industry: do more. Do better. Go faster. We can do more to improve QOL for PWD. #DData19
For example, a lot of researchers think they've reached the frontier, the bleeding edge, and are studying DIY (something). However..... we're over here. We haven't stopped in the last 5 years. #DData19
I recognize, though, researchers typically study the way the world 'is'. But we patients are designing our future, the future we need diabetes care to be, the way things 'ought' to be. And that's sometimes why there is a gap. #DData19
We ought to have real-time access to all our data (all device types) on the device of our choosing. We ought to have reliable, accurate, affordable, and accessible devices, too. #DData19
We ought to have interoperability in all of these devices; and we ought to have the flexibility to do *less* work and achieve *better* outcomes. We also should be able to define our priorities, for our real lives, rather than these being defined by industry. #DData19 Image
So at the end of the day - we are (still) not waiting. We've all (in the DIY and diabetes communities, and industry), made progress. But it's not enough. We can, and should, do more. Please join us. #WeAreNotWaiting

</end> #DData19 Image

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with Dana M. Lewis | #OpenAPS 🤖

Dana M. Lewis | #OpenAPS 🤖 Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @danamlewis

Jun 22
1/ What if there was a tool to help identify who might have exocrine pancreatic insufficiency (EPI/PEI)?

EPI is a significant issue for many people with diabetes (likely more common than gastroparesis or celiac).

Here's how such a tool can help PWD👇🏼🧵

#ADASciSessions #ADA2024
2/ The Exocrine Pancreatic Insufficiency Symptom Score (EPI/PEI-SS) has 15 symptoms, rated by how frequent they are and how bothersome they are (aka severity).

n=324 ppl participated in a real-world survey.

n=118 were people with diabetes (PWD)!

#ADASciSessions #ADA2024
3/ Methods:

EPI/PEI-SS scores were analyzed and compared between PWD (n=118), with EPI (T1D: n=14; T2D: n=20) or without EPI (T1D: n=78; T2D: n=6), and people without diabetes (n=206) with and without EPI.

#ADASciSessions #ADA2024
Read 10 tweets
Jun 6, 2022
📣 Presentation of the primary outcome results from the CREATE Trial, which assessed open source automated insulin delivery (AID) compared to sensor-augmented pump therapy (SAPT) in adults & kids with T1D, at #ADA2022!

@DrMartindeBock1 presenting on behalf of CREATE team.

1/
Shout out to the amazing CREATE study team.

(PS - there’s another publication from our team about HCPs learning experience on open source AID for this study, too! doi.org/10.1111/dme.14…)

2/ #ADA2022
The CREATE trial aimed to study the efficacy and safety of an open source automated insulin delivery system, with a large scale, long term randomized controlled trial.

3/ #ADA2022
Read 15 tweets
Jun 5, 2022
I just realized it's been 3 (!) years since I published my book on automated insulin delivery, with the goal of helping increased conversation and understanding of AID technology for people with diabetes, their loved ones, and healthcare providers!

I'm still very proud that it is available to read for free online, free to download a PDF (both of which have been done thousands of times each: ArtificialPancreasBook.com), or as an e-book, paperback, and now hardback copy. Proceeds from the purchased copies go to Life For A Child.
And, more recently, it has also been translated into French by the wonderful Dr. Mihaela Muresan and Olivier Legendre!

The French translation is available in Kindle, paperback, hardback, or free PDF download formats as well.

Read 6 tweets
Sep 24, 2020
1/THREAD - my presentation is kicking off at #EASD2020 about open source automated insulin delivery.

(You can see a full version of my presentation here: bit.ly/DanaMLewisEASD…, or read the summary below!)
Note we should differentiate between open source (where the source of something is open), and DIY (do-it-yourself) implementations of open source code. Open source means it can be reviewed and used by individuals (thus, DIY or #DIYAPS) or by companies.

/2 #EASD2020
Open source automated insulin delivery (AID) has evolved since the first open source system, #OpenAPS, was made available in Feb 2015!

There are now three open source AID systems (OpenAPS, Loop, AndroidAPS) commonly used by the #DIYAPS community.

/3 #EASD2020
Read 25 tweets
Jun 13, 2020
THREAD:

Poster 988-P at #ADA2020 by Jennifer Zabinsky, Haley Howell, Alireza Ghezavati, @DanaMLewis Andrew Nguyen, and Jenise Wong: “Do-It-Yourself Artificial Pancreas Systems Reduce Hyperglycemia Without Increasing Hypoglycemia”

(PDF available at bit.ly/DanaMLewisADA2…)
This was a retrospective double cohort study that evaluated data from the @OpenAPS Data Commons (data ranged from 2017-2019) and compared it to conventional sensor-augmented pump (SAP) therapy from the @Tidepool_org Big Data Donation Project. #ADA2020
One month of CGM data (with more than 70% of the month spent using CGM), as long as they were >1 year of living with T1D, was used from the @OpenAPS Data Commons. People could be using any type of DIYAPS (OpenAPS, Loop, or AndroidAPS) and there were no age restrictions. #ADA2020
Read 12 tweets
Jun 13, 2020
THREAD:

Poster 99-LB at #ADA2020 by @danamlewis, @azure_dominique, and Lance Kriegsfeld, “Multi-Timescale Interactions of Glucose and Insulin in Type 1 Diabetes Reveal Benefits of Hybrid Closed Loop Systems“

(full content also available at bit.ly/DanaMLewisADA2…) Image
Background - Blood glucose and insulin exhibit coupled biological rhythms at multiple timescales, including hours (ultradian, UR) and the day (circadian, CR) in individuals without diabetes. But, biological rhythms in longitudinal data have not been mapped in T1D. #ADA2020
It is not known exactly how glucose and insulin rhythms compare between T1D and non-T1D, and whether rhythms are affected by type of therapy (Sensor Augmented Pump (SAP) or Hybrid Closed Loop (HCL)). #ADA2020
Read 26 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us!

:(