Skip to content

Finding-type subgroup analysis of imaging contagion #214

Description

@sebasmos

Lane: NIH imaging | API cost: none (pure re-analysis of committed transcripts)

From an exhaustive MedQA + NIH ChestX-ray14 Gemini-only experiment design pass (six lenses: cue/stimulus, committee structure, referee/oversight, shortcut mechanism, pure re-analysis, robustness).

The finding field is on every imaging row but the paper only reports pooled contagion. Cells are small (n=35) so this is explicitly exploratory. It is pure re-analysis.

Conditions / arms

shared_adopt rate per finding (cardiomegaly 9, hernia 8, infiltration 6, effusion 3, ...) with Wilson CIs; cochran_q across findings for cue-invariance within a finding; reads imaging_cascade.jsonl + the three per-cue files.

Expected output artifact

Per-finding contagion table with CIs (exploratory).

Why this is net-new

The finding field is logged but the paper reports only pooled contagion. Not tracked.

Strengthens

Exploratory subgroup texture on claim 4.

Metadata

Metadata

Assignees

Labels

analysisMetrics, stats, visualization of resultsdifficulty: beginnerSelf-contained, no deep context neededpriority: lowNice to have

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions