Clinical research intelligence
Clinical research, connected by intelligence.
ClinQSphereX connects clinical-trial discovery, candidate screening, explainable AI and experimental quantum machine learning in one human-centered research workflow.
Screening support only. Results are potentially eligible shortlists and require researcher review.
The research challenge
Finding eligible participants is one of the biggest operational challenges in clinical research — and it is still largely manual.
Every trial needs the right participants
Recruitment shortfalls delay or end studies that were otherwise ready to run.
Eligibility lives in prose
Protocol criteria are written for humans, then checked by hand against scattered records.
Data grows faster than review capacity
Clinical records accumulate continuously; manual chart review does not scale with them.
Decisions leave no trail
Screening reasoning ends up in spreadsheets, inboxes and phone calls instead of the study record.
Explore research by disease
Each area has its own eligibility problem. Start with the one you work in.
Heart & Cardiovascular
Recruitment for heart failure, coronary and rhythm studies
- Registry trials
- Searched live
- Research data
- Synthetic research data
Cancer
Oncology trial discovery across subtype, stage and prior treatment
- Registry trials
- Searched live
- Research data
- Synthetic research data
Diabetes
Metabolic studies driven by laboratory thresholds
- Registry trials
- Searched live
- Research data
- Synthetic research data
Stroke
Time-sensitive criteria and event history
- Registry trials
- Searched live
- Research data
- Synthetic research data
Kidney Disease
Function-staged eligibility from laboratory values
- Registry trials
- Searched live
- Research data
- Synthetic research data
Brain & Neurological
Discovery across progressive and episodic neurological conditions
- Registry trials
- Searched live
- Research data
- Synthetic research data (limited)
Respiratory
Function tests, exacerbation history and inhaled therapy
- Registry trials
- Searched live
- Research data
- Synthetic research data (limited)
Rare Diseases
Small eligible populations spread across many sites
- Registry trials
- Searched live
- Research data
- Synthetic research data (limited)
Trial status
Registry data
Candidate review
Human decision
Clinical evidence
Source linked
Research activity
Audit recorded
Cardiovascular focus
Heart disease research intelligence
Flagship research area
Heart Disease Research Intelligence
Cardiovascular research is the primary demonstration: routine measurements, active trial discovery and eligibility criteria translated into traceable structured checks.
Conditions
- Coronary artery disease
- Heart failure
- Atrial fibrillation
- Hypertension
- Cardiomyopathy
Research variables
Synthetic data- Age
- Blood pressure
- Heart rate
- BMI
- Relevant laboratory values
- Medical and medication history
Three-state screening
Missing information stays UNKNOWN. It is never silently converted into a pass, and a candidate is only ever described as potentially relevant.
Clinical trial discovery
Live dataRecords are retrieved from the public ClinicalTrials.gov registry when you search. Nothing on this panel is a determination about any individual.
From a disease to a research decision
- 01Disease area
- 02Current clinical trials
- 03Eligibility criteria
- 04Synthetic candidate population
- 05Candidate screening
- 06Classical ML + experimental quantum kernel
- 07Prediction
- 08Feature-contribution explanation
- 09Human researcher
- 10Research action
The chain always ends with a person. A model can rank and explain; only a qualified researcher decides what happens next.
Classical AI vs quantum ML
ExperimentalBoth approaches are compared on the same dataset, the same features, the same split and the same evaluation protocol. Where a value has not been measured, the workspace shows Not evaluated — never a placeholder number.
| Metric | Classical baseline | Quantum kernel |
|---|---|---|
| Accuracy | Not evaluated | Not evaluated |
| Precision | Not evaluated | Not evaluated |
| Recall | Not evaluated | Not evaluated |
| F1 | Not evaluated | Not evaluated |
| ROC-AUC | Not evaluated | Not evaluated |
| Training time | Not evaluated | Not evaluated |
| Inference time | Not evaluated | Not evaluated |
| Memory | Not evaluated | Not evaluated |
| Scalability | Not evaluated | Not evaluated |
Measured values appear in the workspace Quantum Lab once an experiment has been recorded. Scaling behaviour (dataset size against training and inference time) is reported from the same recorded runs. ClinQSphereX does not assume quantum advantage.
Read the research method →Modern biomedical research runs at massive scale
These are figures about the public research ecosystem, not about ClinQSphereX. ClinQSphereX demonstrates how selected clinical and research features can be turned into structured trial intelligence.
Clinical trials
Reference statisticHundreds of thousands of studies are registered on the public ClinicalTrials.gov registry.
Source: ClinicalTrials.gov · Type: public research resource · Access: open (cohort data is controlled access)
Scientific literature
Reference statisticTens of millions of biomedical citations are indexed in PubMed.
Source: PubMed (NLM) · Type: public research resource · Access: open (cohort data is controlled access)
Sequence archives
Reference statisticPublic sequence archives hold hundreds of millions of records across GenBank and the SRA.
Source: NCBI GenBank / SRA · Type: public research resource · Access: open (cohort data is controlled access)
Research cohorts
Reference statisticLarge cohorts such as UK Biobank and NIH All of Us enrol hundreds of thousands of participants under controlled access.
Source: UK Biobank · NIH All of Us · Type: public research resource · Access: open (cohort data is controlled access)
From genome-scale research to trial intelligence
A quantum model does not process billions of sequences. Large datasets are reduced to a small set of engineered features before any quantum encoding happens — that reduction is the scientifically important step.
- 01Large dataset
- 02Feature engineering
- 03Feature selection / reduction
- 04Small experimental representation
- 05Quantum encoding
- 06Quantum kernel
- 07Classification
Explanations, then a human decision
Explainable output
Each prediction is shown with its confidence level and the features that contributed most strongly, including the direction of each contribution and the evidence value behind it.
Feature-contribution methods such as SHAP describe model behaviour. They do not establish clinical causality.
Human review
Every candidate carries a review state, and nobody is enrolled automatically.
- Pending review
- Reviewed
- Requires more evidence
- Not relevant
- Potentially relevant
Research operations, security and governance
Participants, consent, visits, tasks, sites and documents run in one pipeline, with an append-only record of every action.
- Human oversight
- No model output changes a participant's status without a named reviewer.
- Auditability
- Who acted, what changed, when, and against which protocol version.
- Privacy
- Identity data is separated from research data throughout the workflow.
- Data minimisation
- Models receive only the features their task requires.
- Access control
- Organisation, study and site scoped roles and permissions.
- Honest claims
- No fabricated metrics and no compliance certification is claimed.
ClinQSphereX is a research prototype. It is not intended for production PHI without additional security, privacy, validation and regulatory controls. Read the security overview.
Turn complex clinical data into research action.
Explore trials. Structure eligibility. Screen candidates. Compare models. Understand predictions. Keep researchers in control.
Research intelligence, not automated clinical decision-making.
ClinQSphereX is a research prototype using synthetic / Synthea-style data. External research sources are clearly identified, and final research decisions remain with qualified human reviewers.