The Portable Clinical Obligation: a patient-held, condition-agnostic registry for closing diagnostic follow-up loops worldwide

Across every health system studied, a substantial fraction of correctly-made diagnoses and correctly-written follow-up recommendations are never acted on. The detection succeeded; the linkage failed. Reported failure rates range from roughly a quarter to a half of actionable incidental radiology findings in high-income settings, to 13 to 20% of bacteriologically-confirmed tuberculosis diagnoses never reaching treatment initiation in high-burden countries, where 17% of those lost patients died, most within 30 days. Existing solutions are institution-side, commercial, single-modality, single-country, and priced; they protect patients only while those patients remain inside one organisation. This work specifies a portable clinical obligation, a patient-owned, institution-independent, condition-agnostic record of the form {finding, recommended action, due date, source evidence, closure evidence}, together with a free, open-source registry that creates, tracks, escalates and closes such obligations from any source, including a photograph of a paper report, in any country, without institutional integration and without making any diagnostic claim. The system detects no disease; it carries detections that have already been made to the treatment that should have followed. Version 1.5 revises the specification to v0.6. Version 0.5 was not published separately, and its changes are included here. Both rest on measurement: the reference implementation was run against 500 hand-labelled radiology reports from MIMIC-IV-Note, and five of its rules were changed as a result. An obligation may now exist without a quoted finding where the report names none, as with a routine screening return; v0.4 refused every such duty. A candidate the extraction layer doubts is kept for a person rather than deleted. A later study found automatically only proposes closure, and closes an obligation on its own only when its modality, region and timing leave nothing in doubt, one proposal at a time. An obligation with no due date now escalates by asking for the missing date. Modalities are compared by a code aligned to DICOM rather than by matching words, which had discarded real follow-up studies and could accept the wrong ones. Section 12.1 records each change with the measurement behind it. The object format becomes cor.obligation/0.4, additive since 0.2, so stored obligations need no migration. The concept note is unchanged since version 1.2.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-10-07
DOI
https://doi.org/10.5281/zenodo.21706768
Primary Topic
Electronic Health Records Systems
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
preprint

The Portable Clinical Obligation: a patient-held, condition-agnostic registry for closing diagnostic follow-up loops worldwide

Mohamedsarfaraz Mohamadfiroz Dedrani
Zenodo (CERN European Organization for Nuclear Research)
Electronic Health Records Systems
preprint

The Portable Clinical Obligation: a patient-held, condition-agnostic registry for closing diagnostic follow-up loops worldwide

Mohamedsarfaraz Mohamadfiroz Dedrani
preprint en

Abstract

Across every health system studied, a substantial fraction of correctly-made diagnoses and correctly-written follow-up recommendations are never acted on. The detection succeeded; the linkage failed. Reported failure rates range from roughly a quarter to a half of actionable incidental radiology findings in high-income settings, to 13 to 20% of bacteriologically-confirmed tuberculosis diagnoses never reaching treatment initiation in high-burden countries, where 17% of those lost patients died, most within 30 days. Existing solutions are institution-side, commercial, single-modality, single-country, and priced; they protect patients only while those patients remain inside one organisation. This work specifies a portable clinical obligation, a patient-owned, institution-independent, condition-agnostic record of the form {finding, recommended action, due date, source evidence, closure evidence}, together with a free, open-source registry that creates, tracks, escalates and closes such obligations from any source, including a photograph of a paper report, in any country, without institutional integration and without making any diagnostic claim. The system detects no disease; it carries detections that have already been made to the treatment that should have followed. Version 1.5 revises the specification to v0.6. Version 0.5 was not published separately, and its changes are included here. Both rest on measurement: the reference implementation was run against 500 hand-labelled radiology reports from MIMIC-IV-Note, and five of its rules were changed as a result. An obligation may now exist without a quoted finding where the report names none, as with a routine screening return; v0.4 refused every such duty. A candidate the extraction layer doubts is kept for a person rather than deleted. A later study found automatically only proposes closure, and closes an obligation on its own only when its modality, region and timing leave nothing in doubt, one proposal at a time. An obligation with no due date now escalates by asking for the missing date. Modalities are compared by a code aligned to DICOM rather than by matching words, which had discarded real follow-up studies and could accept the wrong ones. Section 12.1 records each change with the measurement behind it. The object format becomes cor.obligation/0.4, additive since 0.2, so stored obligations need no migration. The concept note is unchanged since version 1.2.

Zenodo (CERN European Organization for Nuclear Research)
Good health and well-being
Electronic Health Records Systems
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.