Dear Accessibility Officer,

On [DATE] I was refused boarding flight [FLIGHT NUMBER] at [AIRPORT] with my assistance dog. The stated reason was that my dog is not accredited by Assistance Dogs International or the International Guide Dog Federation.

I am a disabled person under the Equality Act 2010. My dog is owner-trained to perform specific tasks that mitigate the effects of my disability and meets the behavioural standard expected of any working assistance dog in a public setting.

The Equality and Human Rights Commission business guidance confirms that owner-trained assistance dogs have the same access rights as charity-trained dogs. ADI/IGDF accreditation is a private quality mark, not a statutory requirement in UK law.

I therefore consider your refusal to amount to discrimination contrary to sections 20 and 29 of the Equality Act 2010. The safety exception in Schedule 3, Part 7 does not extend to blanket documentation requirements applied by ground staff before any aircraft-specific risk has been considered.

Please confirm in writing within 14 days:
1. The exact policy under which I was refused.
2. Why a behavioural assessment of my dog was not offered as a less discriminatory alternative.
3. Your formal proposal for resolution, including refund of [AMOUNT] and compensation for distress.

A copy of this letter has been sent to the Civil Aviation Authority Passenger Advice and Complaints Team (CAA-PACT).

Yours,
[YOUR NAME]
[YOUR CONTACT DETAILS]

Sarah can take her owner-trained assistance dog into a supermarket. She can take the same dog into a restaurant. She can board a train. She can stay in a hotel. She can visit her GP. She can enter a shopping centre.

This is the story of how UK aviation came to require something the law has never required, and why, if it ever ends up in court, the result is genuinely difficult to predict.

What the law actually says

Under the Equality Act 2010, service providers cannot discriminate against disabled people. Section 29 covers the provision of services. Section 20 imposes a positive duty to make reasonable adjustments. The Equality and Human Rights Commission's own business guidance, published in 2017 and still in force, is explicit on the question of assistance dogs:

"Assistance dogs can also be owner trained and the owner selects their own dog to fit their own requirements."

That single sentence, in the official guidance from the UK's statutory equality regulator, settles the question for every café, every taxi, every shop, every hotel, every hospital, every train, every bus, and every dentist in the country. An owner-trained assistance dog has the same access rights as a guide dog trained by Guide Dogs UK, a hearing dog trained by Hearing Dogs for Deaf People, or a mobility partner trained by Canine Partners. The law does not look at who trained the dog. It looks at whether the dog assists a disabled person, and whether the dog is under control.

This isn't a quirk of British law. The Americans with Disabilities Act, under regulation 28 CFR 36.302(c), is equally explicit: service animals may be owner-trained, and service providers may ask only two questions before granting access. No certification. No registry. No paperwork. Two of the most established disability legal frameworks in the world, both saying the same thing.

Why most disabled people own-train

Charity-trained assistance dogs are extraordinary animals, produced by extraordinary organisations doing genuinely vital work. But for the average disabled person seeking an assistance dog in the UK today, charity training isn't a choice. It's a queue:

For the disabilities that don't fit any charity's eligibility criteria, chronic illness, epilepsy, mental health conditions, certain autoimmune disorders, there is no charity waitlist at all. The choice is between paying a private trainer £15,000 to £40,000, or training the dog yourself for a fraction of that cost.

Owner-training, then, isn't a fringe preference. It is the realistic and often the only path for the majority of disabled people in the UK who need a working dog. The law recognises this. Civil society recognises this. Every UK ground service from the corner shop to the NHS recognises this. And then the customer reaches the airport.

The aviation carve-out, and how it's being stretched

The Equality Act 2010 contains a narrow exception. Schedule 3, Part 7 allows aviation services to treat disabled people less favourably where it is necessary for safety, or required to comply with international aviation agreements, or compelled by the physical limitations of the aircraft.

That exception was written for genuine aircraft-specific safety risks: turbulence, cabin pressure, evacuation procedures, weight limits. It was not written as a blanket licence to demand particular paperwork from particular handlers.

Yet that is what has happened. Most UK-based airlines now require, as a condition of carriage, that an assistance dog be accredited by Assistance Dogs International or the International Guide Dog Federation, two respected private accreditation networks that together cover roughly 100 to 140 member training organisations worldwide. ADI is a US non-profit headquartered in Ohio. IGDF is a UK-registered charity based in Reading. Neither is a government body. Neither sets UK law. Neither has any statutory authority over UK aviation.

What they have, from the airline's perspective, is something more useful: a piece of paper. And the moment an airline accepts that paper as the only acceptable proof of an assistance dog's status, the airline has created an extra-legal entry barrier that the Equality Act 2010 was specifically designed to prevent.

What the airlines would say

Airlines would argue that their policies exist for safety, consistency and operational practicality. Cabin crew are not dog trainers. Boarding decisions often need to be made quickly. Airlines may also point to liability concerns if an animal behaves unpredictably in a confined aircraft cabin.

These concerns are not trivial. An aircraft is not a café. A poorly-behaved dog at 35,000 feet cannot be asked to leave. Cabin crew already manage a substantial workload under safety-critical conditions, and adding individual animal assessment to that workload is a genuine operational question.

The question, however, is whether excluding every owner-trained assistance dog is a proportionate response to those concerns, particularly when the Equality Act requires service providers to consider reasonable adjustments wherever possible, and particularly when comparable industries have found ways to manage exactly the same risk.

The behavioural-assessment alternative

The safety exception in equality law is not a blank cheque. To rely on it, a service provider has to show that the restriction is proportionate, that it is necessary, and that there is no less discriminatory alternative.

There is a less discriminatory alternative. It is the same alternative used by every café, every taxi, every hotel, every train, every hospital, every restaurant, every NHS surgery, and every American airline operating under federal DOT rules: observe the dog's behaviour.

A working assistance dog can be assessed in minutes by a trained member of cabin crew at the boarding gate. Sit. Down. Stay. Settle at the handler's feet. Quiet voice control by the owner. No barking, no aggression, no soiling indoors. Every other industry that hosts assistance dogs uses this assessment, because behaviour is what matters. Behaviour is the actual safety variable. Behaviour is observable, on the spot, by anyone with four hours of training.

A modest training programme for cabin crew would cost airlines an estimated £20,000 to £50,000 to roll out across an entire workforce. That cost is small enough that the courts have repeatedly held similar measures to be reasonable adjustments that service providers must make under Section 20 of the Equality Act.

Whether current airline policies would survive judicial scrutiny remains largely untested. No UK court has yet been asked to directly examine whether blanket ADI/IGDF requirements are a proportionate response to genuine aviation safety concerns. However, disability-rights lawyers may argue that less discriminatory alternatives already exist, particularly where a dog's behaviour can be assessed individually rather than assumed from documentation alone.

That question remains open. But it is increasingly difficult to ignore.

And the discrimination happens on land

There is a further point that legal observers find compelling. The moment of refusal, the moment a handler is told their dog cannot fly, happens at a check-in counter, or a boarding gate, or a service desk. It happens on the ground, before any aircraft is involved. It is a decision made by ground staff, in a building, looking at a dog, applying a written policy.

There is nothing aviation-specific about that decision. The same conditions that apply to a hotel reception apply to a check-in desk. The safety carve-out in the Equality Act was written to cover constraints inherent to the aircraft itself, not the discretion of a member of ground staff applying an internal policy. Whether the carve-out reaches that far is a question the courts have never been asked to decide.

Even the establishment is calling for reform

This isn't a fringe complaint from owner-trainers. Assistance Dogs UK, the umbrella body for the fourteen British charities all accredited by ADI or IGDF, has itself publicly stated:

"ADUK believes that we urgently need clearer definitions in law of assistance dogs, alongside consistent standards for training and welfare that all working assistance dogs can aim to meet, whether trained by ADUK members or otherwise."

When the umbrella body for the airlines' own preferred accreditation networks publicly calls for reform that would explicitly include dogs trained outside that network, the policy position of the airlines has been overtaken by the consensus of the sector. The airlines are now defending a standard that the standard-setters themselves no longer think is acceptable.

Final thought

Three things need to happen.

First, government needs to clarify that the safety carve-out in the Equality Act applies only to genuine aircraft-specific risks, not to ground-staff documentation requirements. The Department for Transport and the Equality and Human Rights Commission could resolve this with a single piece of guidance.

Second, airlines need to do what every other comparable industry already does: train their staff to assess assistance dog behaviour individually, and accept any dog that meets a behavioural standard, regardless of who trained it.

Third, disabled handlers need to know their rights. A refused boarding is not necessarily a verdict. It may be the start of a discrimination claim that, on current legal grounds, has a real chance of succeeding.

Until one of these things happens, the gap between what UK law says about owner-trained assistance dogs and what UK aviation does about them will remain one of the quietest, longest-running pieces of unequal treatment in British disability rights. It is time it ended.

Frequently asked questions

Are UK airlines legally allowed to refuse my owner-trained assistance dog?

Airlines rely on a narrow safety exception in Schedule 3, Part 7 of the Equality Act 2010. Whether that exception genuinely covers blanket ADI/IGDF requirements has never been tested in a UK court. In practice, most refusals are based on policy interpretation rather than settled law. A refused handler with a well-behaved dog and a clear paper trail has a real prospect of bringing a successful discrimination claim.

What's the difference between ADI/IGDF accreditation and UK assistance dog rights?

ADI and IGDF accredit training organisations, not individual dogs. UK assistance dog rights under the Equality Act 2010 apply to the dog and handler regardless of who trained the dog. ADI/IGDF accreditation is a private quality mark, not a legal requirement for assistance dog status in the UK.

Can I claim compensation if a UK airline refuses my owner-trained assistance dog?

Yes, potentially. Compensation under the Equality Act 2010 typically covers injury to feelings (£900 to £49,300 under the current Vento bands), out-of-pocket costs (rebooked flights, accommodation), and in some cases aggravated damages. Claims are usually filed in the County Court within six months of the incident. The Equality Advisory and Support Service offers free guidance.

Will I have problems flying back to the UK from abroad?

Possibly. Many non-UK airlines apply similar ADI/IGDF requirements, and destination country animal-import rules add another layer. The Equality Act 2010 generally applies only to UK-based airlines or to services provided in the UK. For inbound flights, you may need to rely on the carrier's own accessibility policy, the destination country's disability law, or international aviation rules.

Where do I report a refused boarding incident?

Three places. First, the airline's own accessibility complaints process. Second, the Civil Aviation Authority's Passenger Advice and Complaints Team (CAA-PACT), which oversees airline accessibility complaints in the UK. Third, the Equality Advisory and Support Service (EASS) for disability discrimination guidance. Documenting the refusal in writing within 24 hours is essential.

Owner-Trained Assistance Dogs

Epilepsy Alert Dogs UK: Rights, Training and Everything You Need to Know

📖 11 min read·By the ADR Team·Updated May 2026

Alert vs response dogs: the critical distinction every handler needs to understand

The terms "epilepsy alert dog" and "epilepsy response dog" are often used interchangeably in popular conversation. They describe two very different things, and the distinction matters enormously, practically, scientifically and legally.

A seizure alert dog warns its handler before a seizure begins. This is predictive behaviour: the dog detects something in the handler's body, almost certainly biochemical, possibly olfactory, and communicates it through an observable signal, typically pawing, circling, persistent nose nudging or refusing to leave the handler's side. A genuine alert gives the handler time to reach a safe place, call for help, take medication or lie down on a surface where a fall cannot cause injury. The time window varies from a few minutes to over an hour.

A seizure response dog does not predict seizures. Instead, it is trained to perform specific actions during or after a seizure: lying across the handler's body, fetching a phone or alerting device, moving furniture away, staying present through the postictal phase, or activating a medical alert system. These tasks do not require any ability to sense biochemical changes in advance. They are trained responses to observable events, a falling body, a convulsion, stillness, that a dog can reliably learn.

This distinction shapes everything: the training route you can realistically pursue, the type of help you will actually receive, and the expectations you should set with your neurology team and support network.

What tasks do epilepsy dogs perform?

The exact tasks an epilepsy assistance dog performs depend on the dog, the handler's seizure pattern, the handler's living situation and the training approach used. There is no single fixed list. That said, the most commonly trained and documented tasks fall into the following categories.

Pre-seizure alerting. The dog signals to the handler, typically through pawing, barking, circling or body-pressing, that a seizure is imminent. This gives the handler time to reach a safe position, contact someone or take prophylactic medication. This behaviour is the rarest and most complex of all epilepsy dog tasks. It cannot be reliably trained in every dog and is not universally achievable.

Positional assistance during a seizure. Many response dogs are trained to lie across the handler's torso during a tonic-clonic seizure. This serves two purposes: it can help limit the physical spread of convulsive movements, and it provides a grounding physical presence. Some dogs are trained to use their body to prevent a handler from rolling into dangerous positions.

Injury prevention during a fall. Some dogs are trained to position themselves alongside their handler during a detected aura or at the first sign of collapse, acting as a physical buffer against falls onto hard surfaces. This requires extremely precise and individually tailored training.

Fetching help or activating an alert device. A dog can be trained to fetch a phone, press a large-button alarm, activate a medical alert pendant or find a named person in the home. This is one of the most reliable and trainable response tasks, and it is particularly valuable for people who live alone.

Staying through the postictal phase. The period after a generalised seizure, the postictal phase, can last from minutes to hours. Handlers may be confused, physically exhausted, frightened or temporarily unable to move. A trained epilepsy response dog will stay with the handler, provide deep pressure therapy if trained to do so, and remain calm, providing both practical and emotional grounding through recovery.

Deep pressure therapy (DPT). DPT involves the dog applying firm, sustained pressure to the handler's body, usually the lap, torso or legs, on a specific cue or in response to a trained trigger. For some handlers, DPT during or after a seizure reduces distress and supports faster recovery. It can also be used in the lead-up to a known seizure trigger or during periods of heightened anxiety about seizure risk.

The science of seizure prediction: what we know and what we do not

The question of whether dogs can genuinely predict seizures before any observable change occurs in their handler, and if so, how, is one of the most contested and carefully studied questions in the field of medical assistance animals.

The most widely accepted theory is olfactory. A number of studies, including research conducted by Medical Detection Dogs in collaboration with the University of Birmingham, have found that the human body produces detectable volatile organic compounds (VOCs) during seizure activity. If these compounds are released in the pre-ictal phase, before the seizure begins, a dog with a well-developed olfactory system trained to associate a specific scent with an alert behaviour could, in theory, detect the coming seizure before the handler is aware of it.

A 2021 study published in Scientific Reports by Catala et al. found evidence that seizures produce a distinctive odour across different seizure types, and that trained dogs could identify it with high accuracy in controlled conditions. This provided some of the strongest empirical support to date for the biological basis of seizure scent detection.

However, several significant caveats apply in real-world settings.

First, not all seizure types produce the same olfactory signature, or any detectable one. Absence seizures, focal seizures without obvious motor involvement and certain forms of non-convulsive status epilepticus may not generate the same chemical profile as a generalised tonic-clonic seizure.

Second, even in dogs that clearly appear to alert before seizures, researchers cannot always determine whether the dog is detecting a genuine biochemical pre-ictal signal or responding to very subtle behavioural or physical changes in the handler, changes so early in the seizure process that neither the handler nor observers have noticed them, but which are still technically post-ictal in origin.

Third, and most importantly for anyone considering this route, the ability to alert before a seizure cannot be trained to order. It appears to be a capacity that some dogs develop, sometimes spontaneously, sometimes through targeted scent training. It cannot be guaranteed in any individual dog, even one from a reputable charity training programme specifically focused on this task. This is why Medical Detection Dogs and others working in this area are careful in how they describe what they can and cannot promise prospective handlers.

What this means in practice: if you are considering an epilepsy assistance dog, do not base your entire safety plan on the expectation that your dog will alert. A well-trained response dog whose tasks begin at the moment of seizure onset is both more reliably achievable and provides life-changing support in its own right.

The ADUK route: what is available and why the waiting list is so long

If you are interested in a charity-trained epilepsy alert dog in the UK, the landscape is narrow. Medical Detection Dogs is the only ADUK-accredited charity in the UK currently training epilepsy alert dogs. Based in Great Horwood, Buckinghamshire, their programme involves rigorous scent training and extensive assessment of both the dog and the handler's seizure profile.

Their waiting list regularly exceeds three years. Selection criteria are strict: candidates typically need a confirmed diagnosis from a specialist neurologist, a documented seizure pattern that is frequent enough to provide reliable training data, sufficient cognitive and physical capacity to handle and care for a trained working dog, and a living environment suitable for a working dog. Many genuinely epileptic people with a real need for assistance do not meet all of these criteria simultaneously, or face the wait time as an insurmountable barrier.

For people seeking a seizure response dog rather than a true alert dog, a small number of other assistance dog organisations offer programmes, though none currently hold ADUK accreditation specifically for this task type. The broader assistance dog charity sector in the UK is under significant capacity pressure, and demand for all types of medical alert dogs substantially outstrips what accredited programmes can provide.

Is owner-training realistic? What handlers need to know

The answer depends sharply on what you are asking the dog to do.

For seizure response tasks, owner-training is realistic. Teaching a dog to fetch a phone, activate an alert device, lie across your body on cue, stay with you during and after a seizure and move through postictal recovery by your side, these are achievable training goals for a suitable dog with a capable handler and good professional support. They require time, consistency and ideally the input of a qualified clinical animal behaviourist (one registered with the Animal Behaviour and Training Council, or ABTC), but they are not beyond the reach of a motivated and prepared owner-trainer.

For seizure alert, genuine biochemical pre-seizure detection, the picture is more complex. The trained element of alert work involves conditioning the dog to perform a specific alert behaviour in response to a seizure scent sample. This is technically an owner-trainable task, and some individuals have worked with clinical behaviourists and specialist scent trainers to attempt it. However, the fundamental limitation is not the training: it is whether the individual dog has the olfactory sensitivity and stability to detect and respond to the scent reliably in the chaos of real life, under varying conditions and across different seizure types. Many dogs that undergo scent training do not develop a reliable alert, or develop an alert that is inconsistent in the field. This is not a failure of the handler or the trainer. It is a reflection of biological variation.

What does a realistic owner-training journey look like? It typically involves: a period of careful breed and individual dog selection; foundation obedience and public access training (essential before any task work begins); engagement with a clinical animal behaviourist who has experience with medical alert dogs; close liaison with your neurology team to document your seizure pattern and inform training decisions; and a realistic timeline of 12 to 24 months before the dog is ready to work reliably in public settings.

Organisations that support owner-trainers in the UK, such as Support Dogs, some regional assistance dog training groups, and independent clinical behaviourists with medical assistance dog experience, can provide varying levels of guidance. The quality and availability of this support varies significantly by region, and there is no single national body governing owner-trainer support in the way ADUK governs its member charities.

Working with your neurology team and a clinical behaviourist

An epilepsy assistance dog, whether charity-trained or owner-trained, works as part of a wider management plan for a condition that is medically complex. The involvement of your neurology team is not a bureaucratic requirement: it is genuinely useful.

Your neurologist or epilepsy nurse can provide a detailed written description of your seizure type, frequency and pattern. This information shapes training in concrete ways. A dog being trained to respond to generalised tonic-clonic seizures needs to learn very different cues and tasks from one being trained to assist a handler whose seizures begin with focal onset and involve primarily absence-type presentations. A dog that has been trained to a specific seizure profile and then placed with a handler whose seizures present differently may not reliably perform the tasks it has been trained for.

Your neurologist can also help you communicate the nature of your condition to employers, housing providers or schools when questions arise about whether your dog is genuinely medically necessary. While no UK law requires you to provide such documentation as a condition of access, and landlords and service providers cannot demand a letter from your doctor as proof, having clear documentation available can de-escalate difficult situations quickly.

A clinical animal behaviourist registered with the ABTC brings a different set of expertise. They can assess whether your dog has the temperament, drives and learning capacity for assistance work; design a training programme that maps your dog's developing skills to your specific medical needs; advise on the progression of public access training; and help you troubleshoot if trained tasks break down or become inconsistent in real-world settings. Clinical behaviourists are distinct from general dog trainers: their qualification involves university-level study of animal behaviour and is regulated through a professional register.

The combination of neurological input on the medical side and behaviour science expertise on the training side gives owner-trainers the best realistic foundation for success. Neither alone is sufficient.

What epilepsy qualifies under the Equality Act 2010?

The Equality Act 2010 defines disability at section 6 as a physical or mental impairment that has a substantial and long-term adverse effect on the person's ability to carry out normal day-to-day activities. "Substantial" means more than minor or trivial. "Long-term" means 12 months or more, likely to last 12 months, or likely to last for the rest of the person's life.

The vast majority of people with epilepsy will meet this definition, but it is worth being precise about why. The test is not whether you have a confirmed epilepsy diagnosis, it is whether the condition substantially affects your daily life. For many people with epilepsy, the impact is not only the seizures themselves but the restrictions they impose: the inability to drive, the need for supervision during activities that would otherwise be safe, the impact of antiepileptic medication on cognition and alertness, the anxiety around unpredictable seizure occurrence, and the postictal fatigue that can follow a seizure event.

Even people with well-controlled epilepsy, whose seizures are infrequent or currently suppressed by medication, may still qualify if the underlying condition has a substantial long-term effect on how they live and work. The Equality Act 2010 Schedule 1 makes clear that the effect of an impairment is to be assessed without the benefit of measures taken to treat or correct it, with one specific exception for spectacles and contact lenses. This means that if your medication stopped working tomorrow, the question is whether your epilepsy would then substantially affect your daily life, not whether it does so today while your medication is effective.

There is no minimum seizure frequency required. A person who has two tonic-clonic seizures per year but cannot safely shower, cook, drive or walk near traffic without risk may well meet the legal definition. A person with dozens of brief absence seizures daily whose activities are substantially restricted will almost certainly meet it.

Seizure response dog tasks: visual guide

Public access rights: schools, workplaces, transport and hospitals

An epilepsy assistance dog, regardless of whether it is charity-trained or owner-trained, carries full public access rights under the Equality Act 2010. This covers a wider range of settings than many handlers realise, and it is worth being specific about each.

Schools and educational settings. Part 6 of the Equality Act 2010 covers schools, further education and higher education. A child or student with epilepsy whose assistance dog is a reasonable auxiliary aid is entitled to bring that dog to educational settings. The school or institution must make a reasonable adjustment under section 20 of the Act. A blanket no-dogs policy applied without individual assessment is almost certainly unlawful. Head teachers and SENCO teams that are uncertain should consult the EHRC's technical guidance on schools.

Workplaces. Part 5 of the Act covers employment. An employer has a duty to make reasonable adjustments for a disabled employee. Where an employee requires their assistance dog as part of their daily functioning, allowing the dog into the workplace is likely to be a reasonable adjustment, unless the employer can demonstrate a genuine, proportionate justification for refusing it. A blanket pet policy is not a proportionate justification. An employer who refuses should expect the matter to proceed to an employment tribunal.

Transport. This is the one area where the specific definition in section 173 of the Act applies. Under Part 12, only dogs trained by named ADUK-accredited charities are explicitly referenced for taxi and private hire vehicle purposes. However, this does not mean that transport providers can freely refuse other assistance dogs. The general service provisions under Part 3 still apply to transport operators, and a refusal to carry a disabled person and their trained assistance dog may still constitute unlawful discrimination under sections 29 and 20 of the Act. Rail operators and bus and coach companies are covered by specific Passenger Rights regulations that broadly require them to accommodate assistance animals.

Hospitals and healthcare settings. NHS settings and private healthcare providers are service providers under Part 3 of the Act. They cannot routinely exclude assistance dogs from clinical areas. Infection control considerations may apply in specific circumstances, operating theatres, sterile environments, intensive care units, but these must be assessed individually and proportionately, not applied as a blanket rule. A patient who depends on their epilepsy assistance dog and is admitted to hospital has the right to be assessed individually, not refused automatically.

Shops, restaurants, hotels and other services. All are covered by Part 3 of the Equality Act. A business that refuses entry to a handler with a registered assistance dog, demands proof of ADUK accreditation, or asks a disabled handler to leave their dog outside is likely committing unlawful discrimination.

Frequently asked questions

Key terms explained

Sources

• Equality Act 2010, section 6, definition of disability
• Equality Act 2010, section 20, duty to make adjustments
• Equality Act 2010, section 29, provision of services
• Equality Act 2010, section 173, assistance dogs (transport provisions only)
• Equality Act 2010, Schedule 1, determination of disability (effect of medical treatment)
• Catala A et al. (2021). Dogs demonstrate the existence of an epileptic seizure odour in humans. Scientific Reports.
• Medical Detection Dogs, epilepsy alert dog programme
• ADUK: Assistance Dog Registration and Proof, official position on ID requirements
• EHRC Technical Guidance on Further and Higher Education (2015)
• EHRC: Disability discrimination guidance
• Citizens Advice, free guidance on disability discrimination and housing
• Epilepsy Action, UK epilepsy charity and information resource
• The Epilepsy Society, clinical information on seizure types and management

Owner-Trained Assistance Dogs

Diabetic Alert Dogs UK: How They Work, Training Routes and Your Legal Rights

📖 11 min read·By the ADR Team·Updated May 2026

How diabetic alert dogs work: scent, science and alert behaviours

A diabetic alert dog, commonly referred to as a DAD, is an assistance dog trained to detect the physiological changes associated with dangerous blood sugar events and alert their handler before the situation becomes a medical emergency. The mechanism is rooted in chemistry, not intuition.

When blood glucose drops significantly, the body's response triggers a cascade of metabolic changes. One measurable result is a rise in isoprene, a volatile organic compound detectable in breath and sweat. During hypoglycaemia unawareness, a condition affecting many people with Type 1 diabetes where the body no longer produces the usual warning signs of a low, this chemical signal may be the only early indicator available. Dogs, with their extraordinary olfactory sensitivity, can be trained to detect this signal reliably.

The alert itself is a trained behaviour, not a spontaneous reaction. Handlers and trainers work to shape a specific, consistent response: commonly persistent pawing at the handler's hand or leg, nose-nudging, staring, barking, or retrieving a specific object such as a glucose testing kit or phone. The goal is an alert that is unambiguous, repeatable, and performed at the earliest point of scent detection, ideally before a CGM device would have triggered an alarm.

Type 1 vs Type 2: who benefits most from a diabetic alert dog?

While any person with diabetes could theoretically benefit from a trained alert dog, the evidence for meaningful impact is strongest in three groups: people with Type 1 diabetes, people with brittle diabetes, and people with hypoglycaemia unawareness.

Type 1 diabetes involves total dependence on exogenous insulin, which creates an inherent risk of hypoglycaemia, particularly overnight, during physical activity, or during illness. The unpredictability of insulin sensitivity makes low blood sugar events difficult to anticipate. A DAD adds a biological early-warning layer on top of whatever technology a handler uses.

Brittle diabetes refers to a particularly unstable form of Type 1 where blood glucose levels fluctuate wildly and unpredictably despite best management efforts. People with brittle diabetes may experience multiple severe episodes per week. A DAD can provide a degree of safety net that technology alone cannot fully replicate.

Hypoglycaemia unawareness is the condition where the body has lost its natural early-warning response to low blood sugar, the shaking, sweating and anxiety that most people experience. Without these cues, a person can become severely hypoglycaemic with no subjective warning at all. A DAD trained to detect the early scent signature of a hypo can provide that warning in place of the body's lost signalling system.

People with Type 2 diabetes who manage their condition with insulin may also benefit, particularly if they have developed hypoglycaemia unawareness as a result of long-term insulin therapy. For those managing Type 2 with diet or non-insulin medication alone, the risk of severe hypoglycaemia is minimal and the case for a DAD is correspondingly weaker.

The science behind diabetic alert dogs: what research shows

Scientific study of diabetic alert dogs has grown substantially over the past decade. The picture that emerges is encouraging, though researchers are careful to note that study designs vary and that real-world performance depends heavily on training quality and the individual dog.

A 2016 study published in PLOS ONE by researchers at the University of Cambridge confirmed that trained dogs could reliably detect breath samples from people experiencing hypoglycaemia at significantly above-chance accuracy. The researchers identified isoprene as the key VOC associated with hypo detection, establishing for the first time the likely chemical basis for the behaviour that DAD handlers had been reporting anecdotally for years.

A 2019 systematic review published in Diabetic Medicine examined evidence from multiple countries. It found that handlers of medical alert dogs generally reported reductions in severe hypoglycaemic episodes, improvements in HbA1c levels, and improvements in quality of life, including reduced anxiety and greater confidence in undertaking daily activities. Limitations noted included reliance on self-report data and small sample sizes.

Research into hyperglycaemia alerting is less advanced. Dogs appear to detect high blood sugar events through acetone-related VOCs associated with ketosis, though the reliability data for hyper alerts is less consistent than for hypo detection. Most training programmes prioritise hypo alerting for this reason, treating hyper alerts as a secondary and useful addition rather than a primary function.

The ADUK reality: why owner-training is the only realistic route for most handlers

Assistance Dogs UK (ADUK) is the voluntary coalition of UK assistance dog charities that have achieved accreditation through Assistance Dogs International or the International Guide Dog Federation. In the context of diabetic alert dogs specifically, ADUK accreditation represents an extraordinary bottleneck.

Only two ADUK-accredited charities in the UK currently train diabetic alert dogs. Their combined annual output is fewer than 20 dogs. Against a UK Type 1 diabetes population of around 400,000, this makes the charitable route effectively inaccessible for the vast majority of people who might benefit. Waiting lists at these organisations run to three to five years, and applicants must meet strict eligibility criteria that many people with well-managed diabetes may not satisfy.

This is not a criticism of those charities. Training a reliable medical alert dog takes two to three years, requires specialist expertise, and costs between £30,000 and £50,000 per dog. The economics of scaling that model to meet demand simply do not work without a fundamental change in funding. That change has not happened.

The result is that owner-training, either completely independently or with the support of a professional dog trainer or clinical animal behaviourist, has become the primary pathway for handlers who need a diabetic alert dog. This is not a compromise or a workaround. It is a legitimate, legal and well-established route that carries the same rights under UK law as a charity-trained dog.

Under the Equality Act 2010, the question is not who trained the dog. The question is whether the handler is disabled within the meaning of the Act, and whether the dog is an auxiliary aid that mitigates the effects of that disability. A person with Type 1 diabetes and hypoglycaemia unawareness qualifies as disabled under section 6. A dog trained to detect and alert to blood sugar events is an auxiliary aid. Training organisation is irrelevant to both questions.

Training a diabetic alert dog: scent imprinting, alert shaping and public access

Training a diabetic alert dog is a structured, multi-stage process. It is not simply a matter of exposing a dog to blood glucose changes and hoping it learns what to do. Each stage builds on the last, and the process typically takes 18 months to three years to complete to a reliable standard.

Stage 1: Scent sample collection. The handler collects saliva or breath samples during confirmed blood glucose events, ideally at the point of a hypo or hyper, before treatment. These samples, usually collected onto gauze and frozen in airtight containers, form the training material. Getting this right is critical: the scent must be collected at the point of the event, not after glucose has been taken.

Stage 2: Scent imprinting. The dog is introduced to the scent samples using positive reinforcement. The dog learns to identify and select the correct scent from an array, initially in controlled conditions. This phase can take several months, depending on the dog's aptitude and the consistency of training sessions.

Stage 3: Alert behaviour shaping. Once the dog reliably identifies the target scent, a specific alert behaviour is introduced and reinforced. The behaviour must be one the dog can perform clearly regardless of the setting, at home, in a supermarket, in a car, during the night. Common choices include pawing, a specific nose-touch, or retrieving a designated object.

Stage 4: Proofing in real life. The dog must learn to perform the alert behaviour in every environment it will encounter, busy streets, shops, transport, offices, schools. This is the longest and most demanding stage. The dog must maintain its alerting behaviour reliably regardless of distractions.

Stage 5: Public access training. Any assistance dog working in public spaces must behave impeccably. This means walking calmly on a lead, settling quietly in public places, ignoring food and other dogs, and not approaching strangers without permission. Public access behaviour is not optional, it is what distinguishes an assistance dog from a pet in a public space.

Working with a professional dog trainer or clinical animal behaviourist throughout this process is strongly recommended, even if the handler is doing most of the day-to-day training themselves. A professional can identify problems early, assess progress objectively, and ensure the alert behaviour is being shaped correctly. Organisations such as the Association of Pet Behaviour Counsellors (APBC) and the Animal Behaviour and Training Council (ABTC) maintain registers of accredited practitioners.

What to expect: success rates, limitations and working alongside CGM technology

A diabetic alert dog is a powerful tool, but it is not a flawless one. Setting realistic expectations before embarking on the training journey is important for the handler, the dog, and the handler's clinical team.

No dog alerts correctly 100 per cent of the time. Miss rates occur, particularly as a dog ages, if its training is not maintained, or during periods of illness or stress in the dog itself. Environmental factors, strong competing scents, changes in the handler's diet, or changes in the specific VOC signature of the handler's events over time, can also affect reliability.

False positives, alerts when blood glucose is actually within range, are common in the early stages of training and should diminish with experience. A high false-positive rate in an established DAD may indicate that training maintenance is needed, or that the dog is alerting to emotional or environmental cues rather than the metabolic scent.

The relationship between a DAD and CGM technology is complementary, not competitive. Most experienced handlers use both. The CGM provides objective data; the DAD may alert earlier and can also alert when a device is not worn, has failed, or is out of range. Many handlers also report that the dog alerts during overnight hypos, a situation where a CGM alarm may be missed by someone who sleeps deeply.

The psychological benefit of a DAD is also significant and should not be underestimated. Many handlers with hypoglycaemia unawareness describe living in a state of chronic anxiety, unable to drive, sleep alone, or exercise without fear of a severe episode. A reliable DAD can restore a degree of independence and confidence that medication and technology alone cannot provide.

Public access rights: what the Equality Act 2010 says

A diabetic alert dog that is trained to mitigate the effects of its handler's disability is an assistance dog under UK equality law. The handler has full public access rights under the Equality Act 2010. This is not a grey area.

Under Part 3 of the Act, service providers, including shops, restaurants, hotels, transport operators, cinemas, leisure centres and any other business or organisation that provides services to the public, must not discriminate against a disabled person. Refusing entry or service to a handler with an assistance dog, without an objective justification unrelated to the dog's trained role, is unlawful disability discrimination.

Crucially, no law in the UK requires an assistance dog to be ADUK-accredited in order to have these rights. The definition of "assistance dog" in section 173 of the Equality Act 2010 applies only to Part 12 of the Act, taxi and private hire vehicle licensing. Outside that narrow transport context, the legal test is functional: is the handler disabled, and does the dog mitigate the effects of that disability? An owner-trained or independently trained DAD passes that test.

A service provider who refuses entry to a DAD handler on the grounds that the dog is not ADUK-accredited, is not wearing a specific vest, or does not carry a particular ID card, may be committing unlawful indirect discrimination. A handler who is refused access can report the incident to the Equality Advisory and Support Service (EASS), seek conciliation, or bring county court proceedings.

Registering your DAD with the Assistance Dog Registry provides documentation that many venues and transport operators find reassuring, a verified online profile with QR code, a professional ID card and NFC tag. This documentation does not legally change your rights, but in practice it often prevents confrontations before they arise.

Working with your diabetes care team

Training and working with a diabetic alert dog is a significant undertaking that should be discussed with your diabetes care team, your consultant endocrinologist, diabetologist or diabetes specialist nurse, from the outset.

Your clinical team can help assess whether a DAD is appropriate for your specific situation. They can document your diagnosis and the impact of your diabetes on daily life, documentation that may be needed when registering your dog, applying for training support, or asserting your rights with a service provider. For people with hypoglycaemia unawareness, clinical documentation of that condition specifically can strengthen any access rights assertion.

Your team should also be aware that a DAD may influence how you manage your diabetes day to day. Some handlers become more confident undertaking exercise or reducing overnight alert thresholds on their CGM in the presence of a reliable DAD. These decisions should be made collaboratively with clinical input, not unilaterally.

Diabetes UK, the charity, has published guidance for people considering an alert dog and can provide signposting to training resources and support organisations.

Frequently asked questions

Key terms explained

Sources

• Hargreaves A et al. (2016). Changes in the Volatile Composition of Breath during a Hypoglycaemic Clamp. PLOS ONE.
• Equality Act 2010, section 173, assistance dogs (transport provisions only)
• EHRC: Disability discrimination guidance
• Diabetes UK: Diabetes and dogs, guidance for people considering an alert dog
• ADUK: Assistance Dog Registration and Proof, official position on ID requirements
• Animal Behaviour and Training Council (ABTC), register of accredited practitioners
• Equality Advisory and Support Service, free helpline: 0808 800 0082

  
  

  Know someone with Type 1 diabetes who could benefit from this? Share it.

  WhatsApp
  Facebook
  X
  LinkedIn
  Email
  DAD handler? Know your full rights.
  Register your dog