Development and professional validation of an App to support Oral Cancer Screening.

Abstract

The objective of this study was to develop and validate an App for identifying risk factors for oral cancer. To this end, we developed an App (OCS: Oral Cancer Screening) with predictors of Oral Cancer (OC) and algorithm assembly to estimate the risk of its development.

Methodology: Simulated clinical cases were designed so that 40 professionals with expertise in oral diagnostics could validate the algorithm and test its usability (SUS: System Usability Score) and acceptability (TAM: Technology Acceptance Model). Cronbach's alpha coefficient, Friedman/Dunn tests, and Spearman correlation evaluated the SUS and TAM scales. ROC curve was plotted to estimate the cutoff point of the algorithm in suggesting a high risk for OCS of the simulated cases. Chi-square and Fisher's exact tests were additionally used (p<0.05, SPSS v20.0).

Results: Professionals with expertise in oral diagnosis had usability of 84.63±10.66 and acceptability of 84.75±10.62, which correlated positively (p<0.001, r=0.647). Acting in clinical areas of dentistry (p=0.034) and history of performing OC risk factor orientation (p=0.048) increased acceptability while acting in higher education increased usability (p=0.011). The cutoff point suggested by the App after validation of the simulated clinical cases showed high sensitivity of 84.8% and lower specificity of 58.4%.

Conclusion: The OCS was effective and with adequate sensitivity, usability, and acceptability and may contribute to the detection of early oral lesions.