|
?:abstract
|
-
BACKGROUND Transition to digital pathology (DP) usually takes months or years to be completed. We were familiarizing with DP solutions, when COVID19 epidemic forced us to embark in an abrupt transition. OBJECTIVE The aim of this paper is to describe quantitatively how the abrupt transition to DP might affect the quality of diagnosis, to model the possible causes via probabilistic modeling and to gauge qualitatively the perception of this abrupt transition. METHODS Pathologists involved received 25 additional test cases from the archive and a final psychologic survey. For each case, several different diagnostic tasks were recorded and answers were compared to the original diagnosis. We performed a Bayesian data analysis with probabilistic modeling. RESULTS The overall analysis, 1345 different items, showed a 9% error rate with digital slides. The task of differentiating a neoplastic disease from a non-neoplastic one accounted for about 10% (42/392) of error; whereas the distinction of a malignant process from a benign one accounted for less than 5% (11/258). Apart from residents, senior pathologists generated the majorities of discrepancies (8%; 13/164). Our model showed that after adjusting for other factors, these differences in career levels remained. CONCLUSIONS Our diagnostic performance is in line with literature, emphasizing that the duration of transition (lengthy or abrupt) might not influence it. Moreover, we highlight that digital gap of senior pathologists negatively affects the performance with DP. These results can be of help during the process of digital transition. CLINICALTRIAL
|
![[This page as RDF]](https://research.tib.eu/covid-19/static/rdf-icon.gif){kind=icon}