Pulmonary Function Test Study - Automated Interpretation

Completed

• Conditions: Respiratory Disease

• Registration Number: NCT03264417
• Lead Sponsor: KU Leuven

Brief Summary

A multicentric study to explore variability when clinicians are intrepreting lung function tests. Comparison with results of in-house built software for automatic interpretation.

Detailed Description

This is a multi-center non-interventional retrospective study that will use data obtained from clinical routine. In the hospital, subjects with or without respiratory complaints undergo subset of (or complete) pulmonary function tests if clinically indicated. Results of these tests, expressed as numerical values are send to the clinical workstation (KWS) for immediate clinical use. These results are subsequently being interpreted by a clinician.

A random sample of 50 lung function printouts of subjects followed at the Leuven pulmonary service with a final diagnosis of a respiratory disease have been randomised. These anonymized printouts will be given to different pulmonologists for examination in 18 different centers and compared with the interpretation of in-house developed software.

Study Timeline

• Study Start: 2017-08-01
• Study First Submitted: 2017-08-01
• Study First Submitted QC: 2017-08-25
• Study First Posted: 2017-08-29
• Primary Completion: 2017-12-30
• Study Completion: 2018-04-01
• Status Verified: 2018-09
• Last Update Submitted: 2018-09-03
• Last Update Posted: 2018-09-05

Recruitment & Eligibility

• Status: COMPLETED
• Sex: All
• Target Recruitment: 50

Inclusion Criteria

1. Age more than 18 years
2. Completed one or more pulmonary function tests

Exclusion Criteria

1. Younger than 18
2. Lung transplantation, lung cancer.

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Primary Outcome Measures

Name	Time	Method
Interclass correlation coëfficiënt for whole group	6 months	Variation in preferred primary diagnosis of respiratory disease given by every physician for every case (selection of one diagnosis out of a fixed panel of 9 possible diagnoses).

Secondary Outcome Measures

Name	Time	Method
Accuracy of software expressed as a percentage of correct diagnoses in total group, per hospital and per case	6 months	Comparing labels of in-house developed software for the clinical interpretation and diagnostic labeling of complete pulmonary function tests with GOLD standard based on all additional tests
Interclass correlation coëfficiënt for every case	6 months	Variation in preferred primary diagnosis of respiratory disease given by every physician for an individual case (selection of one diagnosis out of a fixed panel of 9 possible diagnoses).
Accuracy of physicians expressed as a percentage of correct diagnoses per case (%)	6 months	Comparison of diagnostic labels per case provided by the physicians with gold standard diagnosis based on clinical characteristics, complete pulmonary function test data and all other investigations (CT scan, biochmestry, functional test, biopsy, etc...) needed to come to a gold standard diagnosis.
Accuracy of physicans expressed as a percentage of correct diagnoses in total group (%)	6 months	Comparison of diagnostic labels per case provided by the physicians with gold standard diagnosis based on clinical characteristics, complete pulmonary function test data and all other investigations (CT scan, biochmestry, functional test, biopsy, etc...) needed to come to a gold standard diagnosis.
Accuracy of physicians expressed as a percentage of correct diagnoses per hospital (group) (%)	6 months	Comparison of diagnostic labels per case provided by the group of physicians per hospital with gold standard diagnosis based on clinical characteristics, complete pulmonary function test data and all other investigations (CT scan, biochmestry, functional test, biopsy, etc...) needed to come to a gold standard diagnosis.
Comparison of accuracy (% of correct diagnoses) of physicians with software	6 months	Comparing the diagnostic labels provided by in-house developed software for the clinical interpretation and diagnostic labeling of complete pulmonary function tests with those of physicans (in total group, per hospital and per case)

Trial Locations

Locations (1)

UZ Leuven; 🇧🇪 Leuven, Belgium