Health Data Avatar
for Early Disease Detection
True prevention must be accessible to everyone, including the populations routinely missed by traditional screening pathways, such as underserved communities, immigrants and digital nomads. Health Data Avatar democratises early risk detection by shifting the focus from provider-held fragments to a foundation of patient-complete data. We pair this comprehensive data aggregation with plain-language journaling support, allowing users to easily track the daily context, side effects, and lifestyle factors that shape health outcomes but rarely make it into official clinical files. Crucially, HDA empowers users with data amendment and commenting capabilities, providing a long-overdue mechanism for patients to actively validate, annotate, and correct their own medical records. By giving patients the tools to spot gaps and resolve conflicting information, we transform passive data storage into a highly accurate, dynamic health timeline that makes earlier, life-saving detection accessible to everyone
Health Data Avatar: The Foundational Layer for Diabetes Prevention
We cannot prevent what our data cannot see. By unifying fragmented health histories, Health Data Avatar (HDA) powers dynamic risk detection—aiming to identify Type 2 Diabetes trajectories up to 7 years earlier than standard clinical tools.
Type 2 diabetes is highly predictable and largely preventable, yet the UK faces an escalating crisis costing the NHS £10.7 billion annually. The tragedy is that up to 67% of newly diagnosed Type 2 patients have no prior record of pre-diabetes (non-diabetic hyperglycaemia).
​
This is not a failure of clinical programmes; it is a structural failure of fragmented data. Current detection pathways rely on isolated, static snapshots — like a single HbA1c blood test — which frequently misclassify at-risk individuals and fail to engage younger or diverse populations.
​
​
The Solution: A Patient-Complete Data Foundation
​
To accurately predict chronic disease, AI models require a rich, longitudinal, and multi-modal narrative. Because provider-held records only capture what happens inside the clinic, they are incomplete by design.
HDA acts as the missing interoperability layer. We empower patients to centralise their entire health history, building the world's first patient-complete data foundation.
​
Our infrastructure enables:
​
-
Multi-Modal Data Fusion: We securely aggregate and structure core clinical labs, genetics, imaging metrics, and wearable data across providers and borders into a single, unified timeline.
-
Structuring the Unstructured: Real health happens between appointments. HDA allows users to log their symptoms, side effects, and lifestyle context through plain-language conversational journaling. Our advanced engine safely translates these human experiences into structured, clinician-ready data points.
-
Dynamic Risk Stratification: Rather than relying on generic, static risk questionnaires, HDA continuously updates a patient's risk profile in real-time as new clinical or behavioural data is added.
​
​
From Storage to Predictive Intelligence
​
HDA is not just a digital vault; it is a predictive engine. By bridging the gap between clinical records and patient-generated context, we construct a longitudinal data tree that uncovers hidden patterns.
​
Our multi-modal data foundation will allow to use our advanced risk detection models trained on research data targeting early diabetes type 2 detection. By structuring this comprehensive, patient-owned dataset, our goal is to dynamically identify 'invisible' high-risk trajectories up to 7 years earlier than current NHS methods.
​
Once identified, these individuals can be seamlessly directed into targeted lifestyle interventions, such as the NHS Diabetes Prevention Programme (DPP), before costly and irreversible complications arise.
Partner With Us to Transform Prevention
We are actively seeking Integrated Care Systems (ICSs), NHS Trusts, Primary Care Networks, and academic researchers to pilot our platform and validate multi-modal risk detection models using longitudinal, patient-complete datasets.