ClarusC64/ABX-CS-001_sepsis_treatment_response_decay-v0.1

ABX-CS-001 Sepsis Treatment Response Decay

Purpose

Detect early failure of bloodstream infection clearing when clinical response stops tracking antibiotic exposure.

Core pattern

• —severity_index high
• —abxexposureindex high
• —sepsiscoherenceindex drops
• —cfupersistencevs_expected stays high
• —lactatenonresponsevs_expected stays high
• —latersepticshockorpersistentbacteremiaflag appears

Files

• —data/train.csv
• —data/test.csv
• —scorer.py

Schema

Each row is one timepoint in a within series sepsis course.

Required columns

• —row_id
• —series_id
• —timepoint_h
• —host_model
• —abx_drug
• —abxconcmg_L
• —expectedabxconcmgL
• —abxconcdeviation
• —abxexposureindex
• —severity_index
• —bloodculturecfu_mL
• —expectedcfumL
• —cfupersistencevs_expected
• —lactatemmolL
• —expectedlactatemmol_L
• —lactatenonresponsevs_expected
• —sepsiscoherenceindex
• —latersepticshockorpersistentbacteremiaflag
• —assay_method
• —source_type
• —sepsisresponsedecay_signal
• —earliestsepsisresponse_decay

Labels

• —sepsisresponsedecay_signal
• —1 for rows at or after first confirmed decay onset

• —earliestsepsisresponse_decay
• —1 only for the first onset row in that series

Scorer logic in v1

• —candidate onset point
• —abxexposureindex at least 0.80
• —severity_index at least 0.80
• —sepsiscoherenceindex at most 0.40
• —cfupersistencevs_expected at least 15000
• —lactatenonresponsevs_expected at least 0.50
• —for two consecutive points
• —ignore one point culture spike then recovery artifacts
• —confirmation
• —latersepticshockorpersistentbacteremiaflag equals 1 later in series

Evaluation

Run

• —python scorer.py --path data/test.csv