Comparison of Two Glove-Sampling Methods to Discriminate Between Study Arms of a Hand Hygiene and Glove-Use Study.

Abstract

BackgroundOur group will be assessing in a randomized trial whether the use of alcohol-based hand rub (ABHR) to clean non-sterile gloved hands (instead of performing hand hygiene (HH) and changing gloves) for the WHO 5 Moments of HH during a single patient encounter is an efficient and effective practice. In preparation for this future study and in the absence of a well-established gold standard, this pilot study aimed to identify an effective glove-sampling method with the ability to detect a difference between our two study arms.MethodsHealthcare workers (HCWs) entering Contact Precaution rooms were recruited and randomized into the intervention arm (HCWs directed to clean gloves with ABHR at each opportunity) or the usual care arm (HCWs behavior is silently recorded at each opportunity). In both arms, HCWs individual gloved hands were randomized then sampled, using both sponge stick and direct imprint methodology (Fig 1–2). The outcome of interest was total bacterial colony forming units (CFUs). A Wilcoxon rank-sum test was performed to compare the median distribution of CFUs recovered by each sampling method between study arms. Sampled gloved hands were also categorized by level of bacterial detection (0 CFUs vs. more than 0 CFUs). For each sampling method, the proportional difference between categories was analyzed using a Fisher exact test.Results42 unique HCWs participated in the study (25 intervention, 17 usual care). For each HCW interaction an average of 3 HH moments were recorded. Using the direct imprint technique, the median CFUs in the intervention arm compared with the usual care arm were 2 vs. 31, P < 0.01. Using the sponge stick technique, the median CFUs were 1 vs. 6, P = 0.25. (Fig 3). Using the agar method, bacteria was detected on 16/25 (64%) of gloved hands in the intervention arm compared with 17/17 (100%) in the usual care arm (P < 0.05). Using the sponge-stick method, bacteria was detected 16/25 (64%) of gloved hands in the intervention arm and on 15/17 (88%) in the usual care arm (P = 0.15).ConclusionThe direct imprint method is an effective method to discriminate between our intervention arms. In addition, this method is more affordable and less work-intensive in the lab and as a sampling method than the sponge-stick method.Figure 1.Figure 2.Figure 3.Disclosures A. D. Harris, CDC Epicenter Program: Grant Investigator, Grant recipient; K. A. Thom, AHRQ: Grant Investigator, Grant recipient