MétaCan
Menu
Back to cohort
Record W2944436640 · doi:10.2196/13151

Evaluation of a Health Information Technology–Enabled Collective Intelligence Platform to Improve Diagnosis in Primary Care and Urgent Care Settings: Protocol for a Pragmatic Randomized Controlled Trial

2019· article· en· W2944436640 on OpenAlexvenueno aff
Valy Fontil, Elaine C. Khoong, Mekhala Hoskote, Kate Radcliffe, Neda Ratanawongsa, Courtney R. Lyles, Urmimala Sarkar

Bibliographic record

VenueJMIR Research Protocols · 2019
Typearticle
Languageen
FieldMedicine
TopicClinical Reasoning and Diagnostic Skills
Canadian institutionsnot available
FundersAgency for Healthcare Research and QualityNational Cancer InstituteGordon and Betty Moore Foundation
KeywordsRandomized controlled trialMedicineIntervention (counseling)Collective intelligenceHealth careUsabilityNursingFamily medicineArtificial intelligenceComputer sciencePathology

Abstract

fetched live from OpenAlex

BACKGROUND: Diagnostic error in ambulatory care, a frequent cause of preventable harm, may be mitigated using the collective intelligence of multiple clinicians. The National Academy of Medicine has identified enhanced clinician collaboration and digital tools as a means to improve the diagnostic process. OBJECTIVE: This study aims to assess the efficacy of a collective intelligence output to improve diagnostic confidence and accuracy in ambulatory care cases (from primary care and urgent care clinic visits) with diagnostic uncertainty. METHODS: This is a pragmatic randomized controlled trial of using collective intelligence in cases with diagnostic uncertainty from clinicians at primary care and urgent care clinics in 2 health care systems in San Francisco. Real-life cases, identified for having an element of diagnostic uncertainty, will be entered into a collective intelligence digital platform to acquire collective intelligence from at least 5 clinician contributors on the platform. Cases will be randomized to an intervention group (where clinicians will view the collective intelligence output) or control (where clinicians will not view the collective intelligence output). Clinicians will complete a postvisit questionnaire that assesses their diagnostic confidence for each case; in the intervention cases, clinicians will complete the questionnaire after reviewing the collective intelligence output for the case. Using logistic regression accounting for clinician clustering, we will compare the primary outcome of diagnostic confidence and the secondary outcome of time with diagnosis (the time it takes for a clinician to reach a diagnosis), for intervention versus control cases. We will also assess the usability and satisfaction with the digital tool using measures adapted from the Technology Acceptance Model and Net Promoter Score. RESULTS: We have recruited 32 out of our recruitment goal of 33 participants. This study is funded until May 2020 and is approved by the University of California San Francisco Institutional Review Board until January 2020. We have completed data collection as of June 2019 and will complete our proposed analysis by December 2019. CONCLUSIONS: This study will determine if the use of a digital platform for collective intelligence is acceptable, useful, and efficacious in improving diagnostic confidence and accuracy in outpatient cases with diagnostic uncertainty. If shown to be valuable in improving clinicians' diagnostic process, this type of digital tool may be one of the first innovations used for reducing diagnostic errors in outpatient care. The findings of this study may provide a path forward for improving the diagnostic process. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/13151.

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)0.050
metaresearch head score (Gemma)0.059
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Randomized trial · Consensus signal: Randomized trial
GenreCandidate signal: Protocol · Consensus signal: Protocol
Teacher disagreement score0.061
Threshold uncertainty score0.266

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.0500.059
Meta-epidemiology (narrow)0.0080.003
Meta-epidemiology (broad)0.0140.009
Bibliometrics0.0030.004
Science and technology studies0.0040.005
Scholarly communication0.0060.005
Open science0.0040.003
Research integrity0.0100.009
Insufficient payload (model declined to judge)0.0610.010

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 head0.099
GPT teacher head0.526
Teacher spread0.427 · 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 designRandomized trial
Domainnot available
GenreProtocol

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

Citations14
Published2019
Admission routes1
Has abstractyes

Explore more

Same venueJMIR Research ProtocolsSame topicClinical Reasoning and Diagnostic SkillsFrench-language works237,207