For Researchers
Measured, not remembered.
Prospective reproductive exposure, with provenance and consent attached.
Build research questions around recorded change while keeping missingness, analytical limits and approved purpose visible.
The evidence problem
Reproductive exposure is often reconstructed after the outcome.
A 2023 pooled analysis brought together 75 studies and 11 reproductive exposures. Its scale illustrates the work required to harmonise histories collected across different designs, definitions and periods.
Prospective recording does not remove bias or confounding. It can, however, preserve when an observation was recorded, what was missing and which consent and governance conditions applied.
Source: Mao et al., BMC Cancer, 2023.
What Shiora provides
Purpose-bound cohort analysis with provenance and identity-risk controls.
-
Measured exposure
Cycle dates, symptoms and life-stage context can be recorded closer to the event, reducing exclusive dependence on distant recall.
-
Governed analysis
Approved, steward-reviewed and time-bound analysis can be brought to governed data, reducing unnecessary transfer.
-
Controlled identity risk
Minimum-cohort suppression is one control within a wider governance design. The legal basis, de-identification method and access conditions are assessed for each approved use.
How it works
The question is approved before the data is touched.
Define the question
Specify the population, exposure, outcome, variables, analytical approach and why each element is necessary.
Establish governance
Document the lawful and ethical basis, consent conditions, duration, stewardship review and permitted outputs.
Assess feasibility
Review coverage, provenance, missingness, coding consistency and cohort size before treating the dataset as fit for purpose.
Bring analysis to governed data
Approved analysis can be run in a controlled environment to reduce unnecessary transfer of record-level information.
Review and close
Outputs are checked against the approved scope and identity-risk controls. Access ends with the authorised purpose and duration.
Access boundaries
What researchers can assess. What remains protected.
Within an approved use
Defined cohort variables, provenance, missingness, analytical constraints and permitted aggregate outputs under the agreed governance basis.
Outside an approved use
Open-ended record access, unrelated variables, attempts to identify participants and continued access after the approved purpose or duration ends.
Requirements and collaboration
A protocol before a dataset.
A research collaboration begins with a defined question, qualified investigators, an appropriate ethical and legal basis, a data-management plan and a review of minimum necessary variables. Identity-risk controls and output review are selected for the approved use rather than assumed to be universal.
Feasibility, access conditions, duration and commercial terms depend on the research scope and jurisdiction. No dataset or outcome is promised before governance and fitness-for-purpose review.
The difference
Research access ends with the approved purpose.
Cohort definition, permitted variables, identity-risk controls, duration and output review should remain explicit throughout the research lifecycle.
Research FAQ
Questions for investigators and data stewards.
Study design, access boundaries and evidence quality remain explicit.
Can researchers browse individual records?
No. Research access is designed around an approved question, minimum necessary variables and governed analytical conditions.
Does prospective recording eliminate bias?
No. Selection, measurement, missingness, confounding and other limitations still require explicit assessment.
How is consent handled?
The applicable consent or other lawful basis, approved purpose, duration and any restrictions must be established for each use.
What provenance is available?
The model is designed to retain when information was recorded and the context attached to it. Availability depends on the source and approved dataset.
How are small cohorts protected?
Minimum-cohort suppression may be one control, alongside de-identification, access controls and output review selected for the specific use.
Can data be exported?
Any transfer depends on the approved protocol, legal basis, data-minimisation plan, security controls and jurisdictional requirements.
How is missingness represented?
Missing information should remain distinguishable from a negative observation, with coverage reviewed before analysis.
Are outputs reviewed?
Output review can form part of the approved governance design, particularly where disclosure or re-identification risk requires it.
What happens when approval ends?
Access is designed to end with the approved purpose or duration, subject to the governing protocol and applicable retention obligations.
How do we begin a feasibility discussion?
Bring the research question, target population, required variables, study design, governance basis and expected outputs.
Next step