Data-driven analysis of influence between radiologists for diagnosis of breast lesions

Breast lesions are the most common threat to the health of women. The accumulation of historical examination reports for diagnosing breast lesions in clinical practice provides the necessary foundations for analyzing the diagnostic preferences of radiologists and the mutual influence between radiolo...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Annals of operations research Ročník 328; číslo 1; s. 419 - 449
Hlavní autoři: Fu, Chao, Wang, Dongyue, Chang, Wenjun
Médium: Journal Article
Jazyk:angličtina
Vydáno: New York Springer US 01.09.2023
Springer Nature B.V
Témata:
Algorithms
Business and Management
Combinatorics
Data analysis
Diagnostic systems
Hospitals
Lesions
Machine learning
Operations research
Operations Research/Decision Theory
Original Research
Qualitative analysis
Regression models
Statistical analysis
Theory of Computation
Womens health
Machine learning regression algorithm
Diagnosis of breast lesion
Statistical comparison
Influence analysis
Organizational development
ISSN:0254-5330, 1572-9338
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Breast lesions are the most common threat to the health of women. The accumulation of historical examination reports for diagnosing breast lesions in clinical practice provides the necessary foundations for analyzing the diagnostic preferences of radiologists and the mutual influence between radiologists in a hospital. This mutual influence is important for indicating the development of an ultrasonic department in which radiologists work. To conduct a data-driven analysis of the influence between the two radiologists, the influence of the diagnostic preferences of one radiologist on the other was qualitatively defined using regression models. Following the qualitative definition, the process of analyzing the influence between two radiologists was designed, in which ten machine learning regression algorithms were included to make a reliable analysis. A statistical comparison method was developed using each machine learning regression algorithm to generate the indicator pair. The indicator pairs generated by ten machine learning regression algorithms were integrated using absolute majority voting to derive the overall indicator pair, from which the influence between two radiologists was determined, namely the unclear influence, clear influence, or significant influence. Experiments were conducted based on historical examination reports collected from two hospitals in Hefei, Anhui, China. The experimental results indicate that the trend in the influence between two radiologists in one hospital is different from that in the other hospital, which is associated with the management pattern, innovation incentive, and reward pattern of the two hospitals. A general conclusion on managerial insights was drawn to generalize the findings of this study.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0254-5330
1572-9338
DOI:10.1007/s10479-022-05086-4