Application of Electronic Algorithms to Improve Diagnostic Evaluation for Bladder Cancer

Abstract

Background: Strategies to ensure timely diagnostic evaluation of hematuria are needed to reduce delays in bladder cancer diagnosis.

Objective: To evaluate the performance of electronic trigger algorithms to detect delays in hematuria follow-up.

Methods: We developed a computerized trigger to detect delayed follow-up action on a urinalysis result with high-grade hematuria (>50 red blood cells/high powered field). The trigger scanned clinical data within a Department of Veterans Affairs (VA) national data repository to identify all patient records with hematuria, then excluded those where follow-up was unnecessary (e.g., terminal illness) or where typical follow-up action was detected (e.g., cystoscopy). We manually reviewed a randomly-selected sample of flagged records to confirm delays. We performed a similar analysis of records with hematuria that were marked as not delayed (non-triggered). We used review findings to calculate trigger performance.

Results: Of 310,331 patients seen between 1/1/2012-12/31/2014, the trigger identified 5,857 patients who experienced high-grade hematuria, of which 495 experienced a delay. On manual review of 400 randomly-selected triggered records and 100 non-triggered records, the trigger achieved positive and negative predictive values of 58% and 97%, respectively.

Conclusions: Triggers offer a promising method to detect delays in care of patients with high-grade hematuria and warrant further evaluation in clinical practice as a means to reduce delays in bladder cancer diagnosis.

Keywords: Electronic health records; data mining; delayed diagnosis; hematuria; medical informatics; monitoring and surveillance; triggers; urologic neoplasms.

Conflict of interest statement

Conflict of Interest

The authors declare that they have no conflicts of interest in the research.

Figures

Fig. 1

Study Flow