I built iatroX because I wanted a clearer route between having a clinical question, finding the relevant information and understanding it well enough to use again. My experience of dyslexia and dyspraxia is part of that motivation. It is not evidence that every neurodivergent clinician wants the same interface, or that my own preferences should become a rule for everyone else.
I am Dr Kola Tytler, an NHS GP and the founder and developer of iatroX. The personal background described here is also set out on the iatroX About page, checked on 3 October 2026. The design argument is straightforward: remove unnecessary work around learning while preserving the effort required to understand the medicine.
The difficulty is not always the clinical concept
In the account already published on the About page, I describe information scattered across many tabs, dense guideline documents and search results that make the relevant material difficult to locate. My medical education at King's College London and subsequent GP training in London sit alongside experience of building digital products.
Those experiences create a practical question for product design. When someone stops during a learning task, what actually stopped them? It may be the concept. It may also be the work required to find the right page, reconstruct the context or decide what to do next.
A product should not assume those are the same problem. A clearer route into a question will not automatically resolve a misconception. Equally, a longer explanation will not help if the learner cannot work out where the explanation belongs in the task.
Make the next action visible
For clinical learning, I want the sequence to be understandable: attempt something, examine the reasoning, identify the unresolved point and return to it appropriately. The learner should not have to design a complete knowledge-management system before beginning.
In iatroX's September 2026 product description, the Socratic Tutor opens on the attempted question. That connection is important to the design: the discussion begins with a specific learning problem rather than requiring the learner to reconstruct the entire context in a new conversation.
This is a product decision, not a measured claim that every user starts faster. A proper usability evaluation would need to observe the task, record the product version and compare the work required under defined conditions.
The same discipline should apply to the phrase "easy to use". It should eventually mean something observable, such as finding the next action or returning without losing the task, rather than functioning as a reassuring adjective.
Help with the misconception, not just the answer
A learner can recognise an explanation without being able to generate the reasoning independently. That is why the educational workflow described in September 2026 includes targeted Tutor follow-ups rather than simply another block of answer text.
The design intention is to ask what led to the attempted answer and explore the distinction that was missed. It should remain possible for the learner to challenge the explanation, inspect a source and recognise when further verification is needed.
I would not describe that design as proof that every tutoring exchange is correct or that the platform prevents overreliance. Those would be different claims requiring different evidence.
A useful product can make active learning more convenient without pretending that convenience itself establishes learning. The learner still needs to explain the principle and apply it to another question.
Keep the sources close without making them decorative
Ask-iatroX's September 2026 product description identifies NICE, CKS, SIGN and emc SmPC information as its clinical-reference foundation, with linked sources and a published methodology. Retrieval, ranking, citation grounding, output checking and uncertainty handling are parts of that method.
The design challenge is to make the answer approachable while keeping the evidence inspectable. A source link should help the clinician examine the basis of a statement, not merely make the page look authoritative.
The distinction between the source's existence, its support for the claim and its applicability to the question still matters. A genuine citation does not turn every summary into a correct patient-specific decision.
That is why less navigation must not mean less access to qualifications, exceptions or uncertainty. The complexity of the interface and the complexity of the medicine are separate design problems.
One professional goal can need several learning methods
A clinician may need to understand a concept, practise applying it, rehearse a conversation and retain a learning record. Those activities are connected by the professional goal, even though they are not interchangeable.
As described in September 2026, iatroX brings question banks, Socratic Tutor, study planner, simulations and CPD tools into the paid learning subscription. Ask-iatroX and free question access remain free without a trial expiry or verification gate.
The useful value argument is not that a learner receives access to unrelated examinations. It is that the methods relevant to their own goal can be used without unnecessary transfers between separate learning systems.
That does not eliminate the need for other tools. Local protocols, specialist references, supervised practical teaching and conversations with colleagues have roles that a combined platform should not pretend to replace.
My preferences should be challenged by other users
A founder's experience can identify a problem worth solving. It cannot establish how common the problem is or whether the proposed solution works equally well for people with different needs.
Some learners may prefer a dense reference view. Others may prefer a visual overview before reading, or written instructions before voice interaction. A design that suits my own way of working may need to offer a different route for them.
The appropriate response is not to label one preference more advanced. It is to test the task with a range of users and distinguish necessary differences from avoidable complexity.
W3C's cognitive-accessibility guidance is a useful reference for that work. It is a design resource, not a certificate that iatroX already meets every accessibility need.
What should count as progress
I would distinguish three questions in a future evaluation. Can the learner begin and return to the task? Can they understand and apply the clinical concept? Can they recognise the limits of the answer and seek the right further help?
A product might improve the first without yet demonstrating the second. It might also make an answer feel reassuring without helping the user judge whether it applies. Those distinctions should remain visible when describing results.
This article reports no new usability trial, comparative timings or accessibility scores. It sets out the rationale for design choices and the questions that should challenge them.
The standard I want the product to meet is not "it works the way I think". It is that the software asks less unnecessary work of the learner while leaving them better placed to do the thinking that matters.
Frequently asked questions
Was iatroX built only for neurodivergent clinicians?
No. My own experience informed the problem, but the platform is intended for clinical reference and medical learning across users with different needs and preferences.
Does the founder's experience prove that iatroX is ADHD-friendly?
No. Personal experience can inform design, but a specific accessibility or usability claim needs transparent criteria and appropriate testing with users.
Is reducing cognitive friction the same as removing the need to reason?
No. The aim is to reduce unnecessary searching, organisation and navigation while preserving the work of understanding, checking and applying the medicine.