Automated Insulin Delivery (AID) System Study Day for Type 1 Diabetes

Section 1: The world of AID therapy and how to onboard equitably

Session 1: The three W's

• Why AID therapy: A dive into the daily variability of insulin requirements and how AID systems meet the challenge.

• Where we were: Our honest and hard-to-swallow review of inequitable allocation of AID systems up to 2023.

• Where we are going: Looking ahead to 2024 and beyond to ensure equitable onboarding to meet NICE TA 943 for HCL.

Top three learnings:

1. The flexibility of AID systems meets highly variable insulin needs and highlights just how challenging type 1 diabetes is to manage.

2. Self-assessment to determine whether AID onboarding is equitable is the essential first step. Stop pretending and review the data. It might be painful, but pain brings change.

3. The exciting potential of
   NICE TA 943
   for unprecedented access to AID therapy. The rate-limiting step is the onboarding cadence of diabetes teams. There is a need to think differently and embrace digital technologies to speed up onboarding equitably.

Dr Addala: Type 1 diabetes disparities – awareness, challenge yourself, challenge the systems

Dr Addala fully engages with the issue of inequity and, crucially, offers practical solutions to challenge the status quo. She puts herself on the line to optimise impact and effect meaningful change.

Top three learnings:

1. Look after your own backyard: Start by recognising and addressing disparities in your immediate surroundings. Access to effective treatment, such as CGM and AID systems, is often unequal. Affluent white populations typically have better access to advanced diabetes technology and better glucose outcomes. Non-white individuals from less affluent backgrounds face more barriers. How significant is the inequality in technology provision in your own service?

2. Awareness and action: Inequity is multi-layered. Begin with self-awareness and action at the individual and team level. Recognise personal and team biases. Equity is not the same as equality; tailor care to individual needs, providing more support to those facing learning and language barriers. Meet families where they are. From BCH's experience, families from privileged backgrounds can often manage with virtual video support and self-learning, whereas families with language barriers and lower educational attainment may require around four times as much face-to-face contact with interpreters. That may not be equal, but it is equitable. Would you need a head start if you were racing Usain Bolt in the 100m?

3. Challenge and advocate: Once you have addressed disparities within your own sphere, broaden your focus. Understand power dynamics and work towards equitable access to technology and services. True progress requires both individual and systemic change, but it starts with small personal steps.

By starting with self-awareness and action, then expanding efforts to address wider disparities, you can contribute to a more equitable landscape in diabetes treatment and care.

Session 2: The three C's

• Capacity creation: Strategies to make room in the service to prioritise AID onboarding.

• Creation of pathways: How we switched from face-to-face-only onboarding to a versatile virtual model.

• Checking the results: Insights from our audit.

Top three learnings:

1. AID therapy is a high-value activity that may require deprioritising other work. A hybrid virtual programme can halve educator time and increase onboarding cadence roughly five-fold.

2. Step-by-step teaching guides, Survive and Thrive guides, and school care plans act as templates that save time.

3. Starting settings calculators and AID download assessment tools should be treated as tools for guidance, not gospel. They make suggestions based on average insulin sensitivity. Use them cautiously, keep clinical judgement central, and maintain a continuous audit-and-improve mindset.

Here are all the resources, with the evidence base behind our programme. Reading the
full story
of our service transformation may be valuable for teams wanting to meet the
NICE TA 943 challenge.

Resources	Description
Step-by-step teaching guides	The educational content was developed using methods outlined in the ATTD consensus on AID technologies (1) and the UK Diabetes Technology Network's Best Practice Guidelines (2). These resources cover AID expectations, hypoglycaemia and hyperglycaemia management, infusion site care, nutrition advice, CGM management, and analysis of downloaded data. Exercise advice is based on the ISPAD 2022 exercise guidelines (3) and a detailed practical guide (4). The materials incorporate GAME-SET-MATCH strategies for optimising TIR (5–7) and a Mealtime Insulin Guide for type 1 diabetes. To cater to different learning styles, we included manufacturers' PowerPoints, YouTube videos, and various explanatory tools. We also produced short educational videos (5–10 minutes) hosted on screencast platforms to ensure clarity and consistency. Each AID system (780G, CamAPS, Control-IQ, and Omnipod 5) has a custom teaching guide, reviewed by the manufacturers' education teams. The diabetes team and CYP taking part in education sessions refined these materials over six months. Download links: 780G Omnipod 5 Control-IQ CamAPS (under construction)
Survive and Thrive guides	The customised “Survive and Thrive” PDF guides are created for each AID system using JavaScript coding and Adobe Acrobat Pro. They generate a personalised hypoglycaemia treatment plan (5,6) and calculate carbohydrate needs for exercise based on user weight, using ISPAD algorithms (3). QR codes link to short videos covering key management techniques. The “Survive” section focuses on hypo/hyperglycaemia management, accurate CGM readings, infusion site care, and essential kit. The “Thrive” section details nutritional and exercise management strategies, GAME-SET-MATCH approaches, and the Mealtime Insulin Guide. Each AID system (780G, CamAPS, Control-IQ, Omnipod 5) has its own guide, reviewed by the manufacturers' education teams. Download links: 780G Omnipod 5 Control-IQ CamAPS
Interactive school care plans	We transformed AID school care plans into interactive PDFs with embedded videos and an 80% competency score requirement, reducing direct HCP teaching time (8). Download links: 780G Omnipod 5 Control-IQ CamAPS with Dana CamAPS with Ypso
AID starting settings calculator	We adapted the AID starting dose calculator from University College London's diabetes team. It follows ISPAD 2022 insulin setting guidance (9) and paediatric bolus calculator research (10), with adjustments informed by ATTD consensus recommendations. The calculator generates tailored initial settings for multiple AID systems to support smooth onboarding. Download links: AID starting settings calculator Algorithms behind the calculator
AID download assessment tool	The starting settings calculator led to a download assessment clinical tool. We integrated its algorithms with our understanding of factors influencing automated and manual modes (1). The tool helps identify which settings affect each mode and provides suggestions for optimising glucose control, based on TIR, total daily insulin, target basal percentage, and specific AID system. Download link: Download assessment PDF tool
Google Forms	We developed Google Forms for AID system selection and pre-initiation education. These incorporate teaching-guide materials, current insulin settings, setup guides, links to AID simulators, and quizzes with an 80% pass requirement before onboarding. The forms also streamline organisation of onboarding sessions. Examples: AID system selection form prototype 780G pre-AID education form prototype Omnipod 5 pre-AID education form prototype Control-IQ pre-AID education form prototype CamAPS FX (under construction)

Session 3: AID system selection

Thanks to Anne-Marie Frohock from Oxford for helping to put these slides together.

We made a slight mistake in one of the talks: Control-IQ predicts 30 minutes into the future, not 10.

Top three learnings:

1. Each AID system is like a unique supercar; suitability varies by individual physiology, preferences, and context.

2. Online tools such as Google or Teams forms are practical for system selection and pre-AID education.

3. Our role is to guide rather than dictate AID system choice.

---

Here is a
guide on choosing an AID system
for people living with type 1 diabetes.

Section 2: In-depth analysis of AID systems

Session 1: Control-IQ

You will need the
AID tool, the
Survive and Thrive guide for t:slim, and the T:Simulator app for
Apple or
Android.

• Understanding how it works and how to assess downloads

• Case study analysis and AID tool introduction

• Note: Control-IQ predicts 30 minutes into the future, not 10, as was mistakenly suggested in an earlier version.

This next video shows one approach to creating and assessing a Control-IQ download. It is not “the” way, just how we currently do it. Take what is useful, ignore the rest.

Now it is your turn.

Case study 1 and the
AID tool for case study 1 – work through and bounce ideas off colleagues.

Case study 2 and the
AID tool for case study 2 – again, work through and compare notes.

Here are our thoughts on the two
t:slim case studies. Do the work first, then cross-reference. You will almost certainly spot things we missed.

Top three learnings:

1. Keeping basal rates and correction factors updated is critical. A single flat basal can be completely fine for simplicity; use the correction factor to tune the algorithm. Those with clear dawn or dusk phenomena may need different basal rates.

2. The correction factor is the most important setting to optimise in Control-IQ for improving time in range.

3. Control-IQ has huge flexibility, which is powerful but demands responsibility. In children and young people, Control-IQ requires relatively frequent updates to basal rates and correction factors by families or the diabetes team.

Session 2: Omnipod 5

You will need the
AID tool, the
Survive and Thrive guide for Omnipod 5, and the Omnipod 5 simulator for
Apple or
Android.

• The algorithm in view and how to work through a download

• Case study exploration and AID tool usage

This next video is our approach to creating and assessing an Omnipod 5 download. Again, this is not prescriptive, just a practical guide.

Now it is your turn.

Case study 1 and the
AID tool for case study 1 – work through and discuss.

Case study 2 and the
AID tool for case study 2 – same drill.

Here are our thoughts on the two
Omnipod 5 case studies. Do not look first; do the thinking, then cross-check. You will spot plenty we missed.

Top three learnings:

1. Tailoring target levels at different times of day allows the algorithm to match individual insulin requirements more closely.

2. Maximising connectivity between Pod and sensor – particularly maintaining line of sight – matters, especially for swimmers.

3. Using activity mode for exercise and alcohol is essential, as the algorithm's aggressiveness is roughly halved.

Session 3: MiniMed 780G

You will need the
AID tool, the
Survive and Thrive guide for 780G, and the
MiniMed 780G simulator.

• System mechanics and assessment strategies

• Interactive case studies and AID tool deployment

The next video is our approach to creating and assessing a 780G download. Again, not prescriptive – just a worked example.

Now it is your turn.

Case study 1 and the
AID tool for case study 1 – work through and debate.

Case study 2 and the
AID tool for case study 2.

Here are our thoughts on the two 780G case studies (document title reused from t:slim case studies):
780G case study notes. Do the work first, then cross-check.

Top three learnings:

1. Time in range can be optimised by using a short active insulin time (2–2.5 hours) and a tight target (5.5 mmol/L). However, for people with erratic or frequent activity, aggressive auto-corrections can create a lot of insulin on board, which can backfire if exercise is unpredictable.

2. Managing exercise and alcohol using the Temp Target is essential.

3. Manual basal rates and correction factors still matter and should be updated regularly in case manual mode is needed.

Session 4: CamAPS FX

You will need the
AID tool, the
Survive and Thrive guide for CamAPS FX, and, if you have an Android phone, the Amazon App Store and CamAPS FX app.

Full disclosure: this is the AID system we currently have least experience with.

• System exploration and data analysis

• Practical case studies and AID tool utilisation

Now it is your turn.

Case study 1 AID tool and
case study 1 graphic.

Case study 2 and the
AID tool for case study 2.

Here are our thoughts on the two CamAPS FX case studies (again, reusing an older document title):
CamAPS FX case study notes. Do not skip the work – the learning is in the graft.

Top three learnings:

1. Adapting target levels for varying intra-day insulin requirements is powerful; CamAPS allows targets from 4.4 mmol/L up to 11.0 mmol/L.

2. Using the “Add meal” functionality supports better handling of hypos and high-fat meals.

3. Inaccurate carbohydrate counting can blunt algorithm performance. The more accurate the carb count, the better the system behaves.

Section 3: Specialised insights

Session 1: Nutritional wisdom with Francesca Annan

• Expert advice on getting the basics right: adequate energy, balanced meals, carb counting, and pre-meal bolusing.

• Francesca's experience in understanding both systems and people is immense.

Top three learnings:

1. Major in the major: Kids need adequate energy and carbohydrates to grow and thrive. Structured days, balanced meals, and pre-meal bolusing remain the core of good time in range.

2. Minor in the minor: High-fat and protein meals with AID systems are important, but do not overcomplicate until you see a problem. As Francesca says, find out before you fiddle – the algorithm may cope just fine.

3. The art is individualising advice: meet the person where they are, know your AID systems, and learn through trial and error.

Session 2: Exercise strategies with Dessi Zaharieva

Dessi combines research, clinical, and educational expertise in a way that is relentlessly practical.

This talk on exercise and the four AID systems is all killer, no filler.

Top three learnings:

1. Start activity modes 1–2 hours before exercise and consider reducing the carbs entered into the bolus calculator by around 25% if eating 1–2 hours pre-exercise.

2. The Omnipod 5 can work underwater if the sensor and Pod remain very close together.

3. After exercise, be ready to replace insulin quickly. Prolonged insulin suspension during activity can lead to glucose “explosions” after the next meal. Each AID system has specific ways to blunt this.

I also share my thoughts on exercise and AID systems in a longer presentation (~40 minutes). It is not the same calibre as Dessi's, but may be useful for the nerdier crowd:

• The first 10 minutes cover why exercise is arguably the best longevity drug.

• The next ~15 minutes walk through risk factors for hypoglycaemia during exercise in order of importance.

• The final ~20 minutes cover AID case studies, including open-loop use during exercise and watch-outs for getting drunk after team sports when using AID.

Top three learnings:

1. AID and exercise may require rethinking traditional strategies. T25/T25 (25% carb reduction, 25% basal or bolus tweak, depending on system) can be a pragmatic starting point, but rapid adaptation is often needed.

2. Large carbohydrate loads immediately before exercise can trigger the algorithm to drive insulin in during activity, increasing hypoglycaemia risk. Smaller, frequent doses of fast-acting carbs every 20–30 minutes are often safer.

3. Switching to manual or open-loop mode ~90 minutes before planned exercise can standardise insulin conditions and make responses more reproducible. This is mainly for athletes chasing performance and predictability – not everyone needs it, and it requires tuned basal settings and trial and error.

Session 3: Young children and AID with Elizabeth Jelleryd

• Elizabeth's practical approach to managing T1D in young children is superb.

• She draws on her experiences and PhD mentorship with Dr Carmel Smart.

Top three learnings:

1. Pre-meal insulin of at least ~30%, and as much as possible, is usually necessary to prevent post-meal hyperglycaemia.

2. Managing sporadic activity in young children requires extra carbohydrates, especially with AID therapy on board.

3. Structured, balanced meals have a double benefit for children with T1D: they support both nutrition and more predictable glucose control.

Session 4: Thriving with AID and managing alcohol

• The session blends theory, anecdotal experience, and emerging evidence.

• It focuses on practical strategies for managing alcohol intake safely.

Top three learnings:

1. Do not forget the basics. Accurate CGM readings and healthy pump sites are foundational for any algorithm to work properly.

2. Strategic overnight target adjustments can reduce inconsistent post-breakfast spikes after nights out.

3. Activity or similar modes can be used for alcohol management. Bespoke plans rely on understanding how each AID system reduces insulin. Sometimes, manual mode is safer if sugary alcoholic drinks drive high glucose and aggressive algorithm responses. Above all, safety first – use followers and support when drinking.

Session 5: Overcoming skin problems from diabetes devices – Dr Anna Korsgaard Berg

• Defining the main types of skin problems

• Preventing device-related skin issues

• Treating skin complications when they arise

Top three learnings:

1. Types of skin issues and their causes (“MARSI” style framework):

   • Lipodystrophies: Rare, likely autoimmune-related; switching insulin type can sometimes help.
   • Lipohypertrophy: Anabolic effect of insulin causing poor absorption and hyperglycaemia.
   • Contact dermatitis: Eczematous reactions to adhesives; patch testing may be needed.
   • Infections: Less common with good hygiene but still important to recognise.
   • Skin injuries: From mechanical removal, scars, wounds. They compromise the barrier and increase risk of dermatitis and infection. Itching can be an early warning sign.

2. Prevention strategies:

   • Structured skin-care programmes can reduce issues. A programme such as that described by
   Berg et al. (2023)
   can minimise wounds and mechanical damage.
   • Choose adhesives and tapes carefully, avoid unnecessary extra taping, and minimise long-term continuous taping.
   • Avoid alcohol-based skin prep if possible, rotate sites, and use moisturiser after removal.
   • Use adhesive remover and remove devices “low and slow”. Barrier creams, patches, and topical steroids have roles but should not be overused.

3. Treatment principles:

   • Avoid affected sites while they are healing.
   • Use topical steroids for limited periods and lipid-based lotions to restore barrier function.
   • Build and maintain a prevention-focused skin-care plan.

Clinician recommendation: assess skin systematically, prevent where possible, treat early, and follow an agreed skin-care programme.

Session 6: Big picture to small picture – interpreting downloads with Dr Julia Ware

Julia communicates with clarity and deep understanding of AID systems, translating into practical, usable advice for young people and families.

Top three learnings:

1. Structure is key: Start with the big picture then move to the small picture, using the CARES framework from the
   Panther Programme. Use cheat sheets from the
   CYP diabetes network
   or Panther materials to ensure coverage.

2. Big picture: focus on how close the person is to agreed targets:

   • Less than 4% time below range (<3.9 mmol/L or <70 mg/dL).

   • At least 70% time in range (3.9–10.0 mmol/L or 70–180 mg/dL).

   • Coefficient of variation <36%, ideally <32%.

   • >90% time in automode (where applicable).

   • Behavioural patterns: carb entries, number of meals, basal/bolus split (e.g. ~30/70 in young children, ~50/50 in adults), understanding that different systems classify basal/bolus differently.

   • AGP patterns: identify times of day prone to highs or lows and prioritise these.

3. Small picture: focus on meal boluses, carb counting, missed meals, suspensions for showers and disconnections, over-treatment of hypos, early set changes, and overuse of manual corrections. Understand how settings influence the algorithm and manual mode, compare weekdays to weekends, and always meet the person where they are. Adjust target levels and settings in a way that is both physiologically sensible and tolerable in real life.