# Assessment of the impact of power business intelligence on adenoma detection rate: a prospective observational trial

## TL;DR

This study found that using Power BI for real-time feedback did not significantly improve adenoma detection rates in colonoscopies, though it helped one low-performing endoscopist.

## Contribution

The study evaluates the impact of Power BI on adenoma detection rates in a high-performing hospital setting.

## Key findings

- Power BI use was not associated with higher overall adenoma detection rates.
- One low-performing endoscopist improved ADR by 20% with Power BI.
- Technical factors like withdrawal time and preparation method significantly influenced ADR.

## Abstract

Adenoma detection rate (ADR) is a key quality indicator in colonoscopy, with low ADRs linked to higher risks of post-colonoscopy colorectal cancers. Feedback to endoscopists has been shown to improve ADRs; however, many feedback systems lack automation and real-time interactivity. This study evaluates the effectiveness of Power Business Intelligence (Power BI) on ADR enhancement.

This prospective observational study compared ADRs before (2021) and after (2022) the implementation of Power BI at Dalin Tzu Chi Hospital, Taiwan. Power BI automatically processed pathology data to calculate ADRs and provided real-time visual feedback on endoscopy quality indicators. A total of 4,306 colonoscopies performed by 10 endoscopists were analyzed. Logistic regression was employed to identify factors associated with ADR.

The overall ADR was high and comparable between the periods without and with Power BI (50.1% vs. 47.9%, P = 0.152). Individual performance was stable, though one low-performing endoscopist improved ADR by 20.0%. Adjusted multivariate analysis found no association between Power BI and ADR. Higher ADRs correlated with male gender (odds ratio [OR], 1.638; 95% confidence interval [CI], 1.438–1.864; p < 0.001), advanced age (OR, 1.642; CI, 1.439–1.875; p < 0.001), elevated BMI (OR, 1.642; CI, 1.439–1.875; p < 0.001), and positive stool occult blood (OR, 1.829; CI, 1.545–2.167; p < 0.001). Effective technical practices for improving ADRs included polyethylene glycol preparation (OR, 1.246; CI, 1.063–1.462; p = 0.007), water-method colonoscopy (OR, 1.321; CI, 1.134–1.538; p < 0.001), and withdrawal times ≥ 6 min (OR, 6.370; CI, 5.179–7.837; p < 0.001).

The use of Power BI was not associated with a higher ADR at a high-performing institution but may benefit low-performing endoscopists. Efforts should target behavioral changes in modifiable technical factors to drive meaningful ADR improvements.