User perception data were also collected in Kehewin First Nation and Cold Lake First Nations. Study Site 1: We observed zero errors with barcode scanning, compared to seven errors in six immunization records (1.7%) in the manual arm (p = 0.04) ( see more Table 3). The latter included one instance of the nurse recording the wrong vaccine name, and three instances each of incorrectly recorded lot numbers and expiry dates. Study Site 2: We observed zero errors for the barcode arm and 26 errors in 19 immunization records (5.6%) for the non-barcode arm (p < 0.001) ( Table 3). Eight errors were from choosing

the wrong vaccine name from the drop-down menu, and 18 were from typing lot numbers incorrectly. Study Site

1: Mean time per vial to enter vaccine data did not differ between scanning and manual methods (27.6 s vs. 26.3 s; p = 0.39) ( Table 4). The mean scan time was 8.8 s/vial (range = 0.1–94.5 s). Study Site 2: Barcode scanning was significantly faster than entering data using the manual method (30.3 s vs. 41.3 s; p < 0.001) ( Table 4). For scanning alone, the Selumetinib mean time was 4.4 s/vial (range = 0.29–58 s). Study Site 1: Immunizers reverted to the manual method for data entry for 15 vials (5.3%). The mean scanning time before the nurse switched to manual entry was 32.9 s (range = 1.6–87.2 s). Study Site 2: Immunizers switched to the manual method for four (0.98%) barcoded vials. The mean scanning time before switching to manual entry was 5.1 s/vial (range = 1.2–15.3 s). Study Site 1: We conducted interviews with eight immunization nurses (the remaining

two were trainees who only administered non-barcoded vaccines during the study). All reported that the training was adequate and appreciated the opportunity to practice with dummy vials. They also noted that the designated resident “barcode scanning expert” (nurse who learned the process early on) was valuable in supporting the adoption of the technology, helping to resolve issues that arose. All noted the benefits of scanning for recording accurate and complete information. Nearly all interviewees mentioned early difficulties with scanning, leading to the discovery that the pattern on the countertop heptaminol surface was creating interference. A blank white sheet placed under the scanner improved the scanning success rate. Many nurses felt that the barcode readability was not consistent; using a particular technique to scan one vial successfully did not always translate into success with subsequent vials, and multiple attempts were often needed. “I would like it [barcode scanner] to be more sensitive because […] our site was doing it yesterday and there were some [scanners] that you have to, turn and turn and up and down, and it takes… I could’ve typed it in ten times by the time it actually scanned it.