Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
23 changes: 15 additions & 8 deletions docs/EN/Getting-Started/Overview.md
Original file line number Diff line number Diff line change
@@ -1,16 +1,23 @@
# Overview

## What is iAPS?
iAPS is an open source artifical pancrease system based on the OpenAPS algorithm. Using your inputted settings, carbohydrates and historical data, it aims to automate insulin delivery to reduce the time spend managing your diabetes. Before starting with iAPS, you should consider alternative commercial options such as the Tandem IQ and Omnipod 5, or other open source applications like Loop and AndroidAPS. iAPS is not approved by any health care authority. You are building and running this system at your own risk.

Powered by the OpenAPS algorithm, **iAPS is an open-source Artificial Pancreas System** designed to simplify diabetes management. By leveraging your specific settings, carbohydrate information, and historical data, it aims to **automate insulin delivery** and significantly **cut down on the daily effort** you put into managing your diabetes.

We encourage you to explore all your options before committing to iAPS. Consider commercial alternatives like the **Tandem IQ** and **Omnipod 5**, or other well-known open-source solutions such as **Loop** and **AndroidAPS**.

Important: iAPS is not approved by any healthcare authority. **You are responsible for building and operating this system at your own risk**.

## Getting Started
Before starting with iAPS, you should have a basic understanding of what ICR, ISF and basal rates are. If you do not have a clear understanding, or require some help identifying your settings, please read the appropriate documentation.

To use iAPS, you are required to build the application from the source code. This does not require substantial technical know-how but is a time consuming process. You may need to carry this out through several sessions on your first attempt.
Before you begin with iAPS, it's helpful to have a basic grasp of concepts like **ICR (Insulin-to-Carbohydrate Ratio)**, **ISF (Insulin Sensitivity Factor)**, and **basal rates**. If you're unsure about these or need help figuring out your personal settings, please refer to the relevant documentation.

To use iAPS, you'll need to **build the application from its source code**. Don't worry, this doesn't require advanced technical skills, but it is a time-consuming process. You might find yourself needing several sessions to complete your first build.

Once installed, you'll configure your settings. By default, iAPS will act much like your current pump, though it might suggest temporary basal adjustments occasionally. The real power of iAPS unlocks when you enable **"Closed Loop"**, enabling **automatic bolus features**, and **autotune**. As you get more confident with the app and your settings, these are generally the first three features you'll want to configure:

Upon installation, you will need to configure your settings appropriately. By default iAPS acts no different than that of your pump, with the exception that it may recommend temporary basals from time to time. The magic happens by turning on "Closed Loop", enabling automatic bolus features, and turning on autotune. In general, these are the first three settings you will want to configure as you gain confidence in the app and your settings:
* **Enable Closed Loop for automation**.
* **Increase Max IOB (Insulin on Board)** by setting it to "average meal bolus + 3x max daily basal."
* **Enable SMB (Super Micro Bolus) and UAM (Unannounced Meals)** for automatic bolusing (but make sure your ISF is optimized before turning these on).

- Enable Closed Loop for automation
- Increase Max IOB via "average mealbolus + 3x max daily basal"
- Enable SMB and UAM for automatic bolusing (ensure your ISF is optimized before enabling this)
See Configure for more information on iAPS configuration.
For more detailed information on configuring iAPS, please see the "Configure" section.
28 changes: 9 additions & 19 deletions docs/EN/Getting-Started/ProjectStatistics.md
Original file line number Diff line number Diff line change
@@ -1,36 +1,26 @@
# Project Statistics

## What is it?
## Showing the Impact of DIY APS

One of the hardest advocacy efforts any project has is to so both the adoption, and impact, of that project. DIYAPS systems are no different, as even though there could be thousands out there using it, from a vendor perspective, we are a fringe group. And from a professionals perspective, our results are questionable.
It's tough for any project to demonstrate its reach and effectiveness, and **DIY Artificial Pancreas Systems (APS)** are no exception. Even with potentially thousands of users, from a vendor's viewpoint, we're often seen as a niche group, and healthcare professionals might question our results.

The intent with the statistics server setup is to show both vendors, and professionals, that not only are we out there, but we are getting results equal to, if not better, then non-DIYAPS solutions.
The goal of our statistics server is to change that perception. We aim to clearly show both vendors and professionals that not only do **DIY APS users exist in significant numbers, but we're also achieving results that are equal to, if not better than, commercial non-DIY APS solutions**.

Commercial vendors out there are driven mostly by, of course, profits. If there is anyone out there trying to encourage a vendor to improve support for DIY systems, being able to go at them with "this is the market you are currently giving to your competitor" can be compelling, and this is one of the aspects we are trying to be able to push / promote, is loss of marketshare that they could be tapping if they weren't so closed.

The core part to the statistics is they stay up to date, and this is accomplished on of two ways:

* for those using Nightscout servers, we are able to pull the needed statistics directly
* for those not using Nightscout servers, they are able to configure iAPS to push data into us

In both cases, the process is meant to happen once a day, so that information is updated regularly.
Commercial vendors are primarily driven by profit. If we want to encourage them to better support DIY systems, being able to say, **"This is the market share you're currently conceding to competitors"** can be a powerful argument. This data helps us highlight the missed market opportunity for vendors who remain closed off to DIY integration.

## Where is it?

We have a central statistics server setup at https://iaps-stats.hub.org/
You can view the collected statistics on our central server at https://open-iaps.org/

## What data is collected?

We pull three pieces of information into the system:
We collect four main types of information:

* preferences.json
* statistics.json
* profile data
* profile data consists of your schedule basal, insulin sensitivities and carb ratio data
* preferences.json
* Profile data, which includes your scheduled basal rates, insulin sensitivities, and carb ratio data.

## Who controls the data?

The project is a private contribution to the iAPS project, funded and managed by one person: me

The only outside access to any of the data in the system is via the web pages at the above URL, nobody has access to the raw data.
This project is a private contribution to iAPS, funded and managed by a single individual (me). **The only way to access any data in the system is through the public web pages** at the URL mentioned above. No one has direct access to the raw data.

47 changes: 29 additions & 18 deletions docs/EN/settings/configuration/autotune.md
Original file line number Diff line number Diff line change
Expand Up @@ -4,36 +4,47 @@
- Enable autotune unless you have full confidence in your current profile settings.
- Autotune adjustments are limited by autosens max/min.
:::
---
## What is Autotune?
Autotune makes iterative adjustments to your basal rates, ISF, and carb ratio based on your last 24 hours of data. The adjustments made by autotune are small, with a maximum of 10% change to the current ICR and ISF, or 20% change to the current basal rates per run, but add up over days. Autotune is run nightly by default unless its run manually by the user.

Autotune is designed to work over weeks to slowly improve the accuracy of your initially set ISF/ICR/basal rates. The output of autotune formulas produces a ratio, similar to that of autosens. This ratio is also limited by the autosens max and min preferences.
Autotune is a feature that refines your insulin pump's settings—**basal rates**, **insulin sensitivity factor (ISF)**, and **carb ratio (CR)**—by making small, iterative adjustments based on your last 24 hours of glucose data. Think of it as a long-term fine-tuning tool.

## How does it work?
Autotune is not Machine Learning. It is not AI. It is a statistical regression attempting to compare what happened over the last 24 hours with what profile settings are, and adjust the profile settings accordingly. For basals, it divides the day into hour long increments. It calculates the total deviations for that hour increment and calculates what change in basal would be required to adjust those deviations to 0. It then applies 20% of that change needed to the three hours prior (because of insulin impact time). If increasing basal, it increases each of the 3 hour increments by the same amount. If decreasing basal, it does so proportionally, so the biggest basal is reduced the most.
By default, Autotune runs nightly, but you can also run it manually. Each adjustment is minor—up to a 10% change for ISF and CR, and up to a 20% change for basal rates per run. These small changes accumulate over days and weeks, gradually improving the accuracy of your initial profile settings. Autotune's output is limited by the same maximum and minimum ratios as Autosens, ensuring adjustments stay within safe bounds.

For ISF, it calculates the 50th percentile (median) deviation for the entire day and determines how much ISF would need to change to get that deviation to 0. It applies 10% of that as an adjustment to ISF.
---
## How Does Autotune Work?

For CR, it calculates the total deviations over all of the day’s mealtimes and compares to the deviations that are expected based on existing CRand the known amount of carbs entered, and applies 10% of that adjustment to CR.
Autotune doesn't use machine learning or artificial intelligence. Instead, it employs **statistical regression** to compare your recent glucose data with your current profile settings and then suggests adjustments.

## Important note on Dynamic Settings
Because Autotune uses Deviations to calculate basal variation, it has to calculate Deviations. It does this using a profile based ISF value.
Here's a breakdown of how it works for each setting:

This means that if you enable Autotune, but use Dynamic functions within iAPS, it will not be using the ISF that was in operation at the time of the loop running in the past, and basal adjustments are likely to be incorrect, as well as ISF adjustments.
* **Basal Rates:** Autotune divides the day into hourly segments. For each hour, it calculates how much your glucose deviated from your target. It then determines the basal rate change needed to correct these deviations. Because insulin has a delayed effect, it applies 20% of this calculated change to the current hour and the two hours prior. If increasing basal, the increase is applied equally to all three hours. If decreasing, the reduction is proportional, meaning the largest basal rates see the biggest decrease.

It can also only run with a single daily ISF and CR. As a result, it doesn't work for most people.
* **ISF (Insulin Sensitivity Factor):** Autotune identifies the median glucose deviation for the entire day. It then calculates how much your ISF would need to change to eliminate this deviation and applies 10% of that adjustment to your ISF.

Simply put, Autotune results can be unreliable if dynamic settings are enabled.
* **CR (Carb Ratio):** Autotune analyzes glucose deviations during all mealtimes throughout the day, comparing them to the expected deviations based on your current carb ratio and logged carb intake. It then applies 10% of the calculated adjustment to your CR.

## How does it differ from Autosens?
---
## Important Note on Dynamic Settings

Autosens is designed to make rapid changes to your profile settings every loop cycle based on your last 8 or 24 hours of data. It accounts for biological changes to insulin sensitivity that occur over the day and night, as well as changes from pump placement.
Autotune relies on a **profile-based ISF** value to calculate glucose deviations. This is crucial: if you use dynamic functions within iAPS (which adjust your ISF and CR in real-time based on current conditions), Autotune will not use the actual ISF that was active when your loop was running in the past. This can lead to inaccurate basal and ISF adjustments.

Autotune by contrast works on a longer time scale, making slow but sustained changes to your profile to get your set baseline settings more accurate. It is run every 24 hours.
Furthermore, Autotune can only operate with a **single daily ISF and CR**. As a result, it may not be suitable for individuals who have varying ISF and CR settings throughout the day. In simple terms, Autotune results can be unreliable if dynamic settings are enabled.

## Should I enable Autotune?
If you believe your profile settings are accurate as is, you can choose to keep autotune disabled; it may worsen control in certain scenarios such as after extended periods of sickness.
---
## How Does Autotune Differ from Autosens?

If you decide to enable it, you should note that autotune is limited by autosens max and min ratios. You can alter these values to give autotune more flexibility but this also affects autosens, dynamic ISF, dynamic CR, and adjust basal.
* **Autosens** makes rapid, real-time adjustments to your insulin delivery settings with every loop cycle. It's designed to account for immediate biological changes in insulin sensitivity due to factors like time of day, night, or pump site changes. It works on a shorter timescale, typically using data from the last 8 or 24 hours.

* **Autotune**, in contrast, operates on a much longer timescale. It focuses on making slow, consistent adjustments to your baseline profile settings to improve their overall accuracy over weeks. It runs once every 24 hours.

---
## Should You Enable Autotune?

Consider these points before enabling Autotune:

* **If your current profile settings are already well-tuned and accurate,** enabling Autotune might actually worsen your glucose control, especially in scenarios like recovering from illness.
* **If you decide to enable it,** be aware that Autotune's adjustments are limited by your Autosens maximum and minimum ratios. While you can modify these values to give Autotune more flexibility, doing so will also impact Autosens, dynamic ISF, dynamic CR, and basal adjustments.

Instead of continuously relying on Autotune for daily adjustments, a better approach is to **periodically review your settings** after a few weeks of Autotune running. Note the new values it suggests and then manually update your main profile settings to match them. This gives Autotune a fresh starting point for any future fine-tuning.

Instead, it is best if you review your settings in a few weeks, note the new autotune values, and change your profile settings to match them, giving autotune a new starting place to move from.