Map a research question to biobank schema fields

Why use this

1. A researcher translating a question into biobank variables may find field names that sound relevant but measure a different concept, time point or proxy. Discovering a schema field also does not establish permission to use participant-level records.

2. A metadata-based candidate list could explain why each field might matter and which definitions need expert checking. It would support data-access planning while keeping public schema review separate from controlled-data access and study feasibility.

Why an agent / simpler alternative

Useful judgment

Compare the question with field descriptions and identify direct measures, proxies and ambiguous matches.

Simpler alternative

Use the official schema search directly when the required field identifier is already known.

Unsuitable when

The task requires participant data, clinical inference or a claim that the available fields answer the research question.

Proposed inputs

1. Name

   Research question and concepts

   Owner

   Researcher

   Minimum scope

   Target concepts, population scope and required measurement context

   Quality check

   Separate preferred variables from acceptable proxies.

2. Name

   Public schema snapshot

   Owner

   Researcher

   Minimum scope

   Field identifiers, descriptions and supplied coding/time-context metadata

   Quality check

   Confirm source version and retain full definitions for review.

Access and setup — proposed or unknown

Proposed minimum

1. Read-only public metadata and an offline field-ranking worksheet.

Compatibility unknowns

1. Unknown: exact runtime, schema-index freshness, search tooling and controlled-data permissions.

External actions requiring approval

1. Applying for data access, retrieving participant records or sharing restricted research information.

Expected deliverable and review checks

Proposed artifact

Ranked field candidates with definition pointers, relevance rationale and data-access questions.

Critical checks

1. Relevance scores are editorial aids, not validated scientific measures.

2. Public metadata access does not imply participant-data availability or research authorization.

Missing output evidence

No new workflow run or output inspection was performed for this overlay.

Failure signals and proposed responses

1. Signal

   A proxy variable is labelled as the direct measurement sought.

   Why it matters

   The study plan could rest on a false data assumption.

   Proposed response

   Label the proxy relationship and request domain-expert evaluation.

2. Signal

   A matching field lacks the required measurement timing.

   Why it matters

   The variable may not support the proposed analysis.

   Proposed response

   Flag timing uncertainty before treating the candidate as usable.

Effort and cost — unknown

Setup work

1. Define concepts and preserve schema-version context.

Ongoing review

1. Verify field meaning and plan access through the proper research process.

Unknowns

1. Unknown: field suitability, data availability and downstream analysis effort.

First check — proposed, unexecuted

Status

proposed_unexecuted

Synthetic fixture

Invented metadata includes a similarly named proxy and a direct measure with unspecified timing.

Acceptance checks

1. Neither becomes an unconditional research-ready selection.

2. No participant access is implied.

Stop conditions

1. Controlled data or scientific suitability judgments are required.

Limits

1. Editorial schema discovery; a synthetic relevance check is not scientific or participant-data validation.