MétaCan
Menu
Back to cohort

Accuracy of Administrative Hospital Data to Identify Use of Life Support Modalities. A Canadian Study

2020· article· en· W3003349073 on OpenAlexaffabout
Allan Garland, Ruth Ann Marrie, Hannah Wunsch, Marina Yogendran, Daniel Chateau

Bibliographic record

VenueAnnals of the American Thoracic Society · 2020
Typearticle
Languageen
FieldMedicine
TopicSepsis Diagnosis and Treatment
Canadian institutionsUniversity of ManitobaManitoba HealthUniversity of Toronto
Fundersnot available
KeywordsMedicineIntensive care unitDiagnosis codeMechanical ventilationEmergency medicineIntensive care medicineVasoactiveConfidence intervalRenal replacement therapyDatabaseAcute careHealth careInternal medicine

Abstract

fetched live from OpenAlex

Abstract Rationale Accurately identifying use of life support in hospital administrative data enhances the data’s value for quality improvement and research in critical illness. Objectives To assess the accuracy of administrative hospital data for identifying invasive mechanical ventilation (IMV), acute renal replacement therapy (RRT), and intravenous vasoactive drugs in unselected adult intensive care unit (ICU) patients. Methods We employed the administrative dataset of the Discharge Abstract Database from the Province of Manitoba during 2007–2012, using nationally standardized diagnosis and procedure codes to identify the three types of life support. The criterion standard was the Winnipeg ICU Database, which contains daily clinical information about all admissions to all 11 adult ICUs within the Winnipeg Regional Health Authority. For all individuals aged 40 years or older at ICU admission, we calculated sensitivity, specificity, positive predictive value (PPV), and negative predictive value of the administrative data for identifying life support. We also assessed the ability of the administrative data to identify overlapping use of the forms of life support. Results Over the study period, there were 20,764 eligible ICU admissions; 52.6% (10,914) involved IMV, 46.8% (9,724) involved vasoactive agents, and 4.4% (907) involved acute RRT. Identification of IMV from administrative data procedure codes was good, with all four parameters exceeding 90%. The procedure code for use of selected vasoactive drugs had a sensitivity of zero; addition of diagnosis codes for shock raised the sensitivity to only 23% (95% confidence interval [CI], 22–24%). Both the sensitivity and specificity for acute RRT procedure codes exceeded 92%, but owing to low prevalence of RRT, the PPV was only 55% (95% CI, 53–58%). Addition of diagnosis codes for acute renal failure did not appreciably improve performance. Overlapping use of the three types of life support was substantial. Among those receiving any one of the types of life support, 68–76% received at least one of the two other types assessed. Considering use of any one or more of the three forms of life support, the administrative data had a PPV of 97% (95% CI, 96–97%) and a negative predictive value of 69% (95% CI, 68–70%). Conclusions Administrative data accurately identify IMV but not use of vasoactive drugs or acute RRT.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame machine prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.

metaresearch head score (Codex)0.015
metaresearch head score (Gemma)0.065
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Observational · Consensus signal: Observational
GenreCandidate signal: Empirical · Consensus signal: Empirical
Teacher disagreement score0.031
Threshold uncertainty score0.228

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0150.065
Meta-epidemiology (narrow)0.0010.001
Meta-epidemiology (broad)0.0010.001
Bibliometrics0.0070.015
Science and technology studies0.0030.002
Scholarly communication0.0030.001
Open science0.0040.002
Research integrity0.0010.001
Insufficient payload (model declined to judge)0.0010.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.597
GPT teacher head0.531
Teacher spread0.066 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.

The models applied no category: nothing in the taxonomy fit this work.
Study designObservational
Domainnot available
GenreEmpirical

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations11
Published2020
Admission routes2
Has abstractyes

Explore more

Same venueAnnals of the American Thoracic SocietySame topicSepsis Diagnosis and TreatmentFrench-language works237,207