ImpossibleMA

Loading Python runtime (~6 MB)...

ImpossibleMA v0.1

k=1 - single-study borrowing

Target trial
Adjacent trials (>=3)
StudyEstimateSE

missing_se - SE reconstruction

Inputs

Route D - Figure extraction

Upload a horizontal forest plot (PNG or JPG). Click two x-axis reference points and type their values to calibrate. Then click once per study row; drag the proposed whisker-cap handles if needed; confirm to add the row to the extraction table.

Calibration (2 x-axis points)
#pixel xtyped value
Extracted rows
LabeleffectSEaudit

adversarial - worst/best defensible pool

or
Studies (>=3 needed; columns: study, estimate, se, rob, n, followup, language, pub_type)
studyestimateserob nfollowuplanguagepub_type