Right-sizing expectations for hand hygiene observation collection

Background

Hand Hygiene (HH) monitoring is essential for patient and staff safety, but the optimal number of observations remains elusive. This project aimed to determine a statistically comparable number below the current standard of 100 to 200 per month per unit.

Methods

HH observations from various hospitals were grouped in strata by facility, unit, and month, then resampled into sets of 25, 50, 100, and 150 for comparison with 200. Significant comparisons and confidence interval width differences were calculated. A power analysis, accounting for adherence rates (50%-90%), identified sample sizes needed for significant differences from 200 at the 0.05 level.

Results

A total of 873,618 observations across 68 facilities, compared sample sizes (25, 50, 100, 150) to 200, showed 2.6% to 4.3% significant differences at P = .05, with median confidence interval width differences from 0.05% to 0.68%. The power analysis showed required percentage differences for significance at P = .05 ranged from 7.8% (150 vs 200 at 90% HH adherence) to 30% (25 vs 200 at 50% adherence).

Conclusions

The findings suggest hospitals could reduce HH observations to 50 per unit per month without affecting data quality even at lower adherence. It is recommended that standards shift focus from monitoring (with fewer observations) to training, education, culture, infrastructure, and feedback.