JSTOR

Methods

Our study was conducted in a 500‐bed university‐affiliated teaching hospital that has had a hand hygiene promotion program since 2001, with alcohol‐based hand rub dispensers located throughout the hospital, except for the psychiatric wards. The 6‐month trial was conducted during the period from September 2007 through February 2008. Prototypes of an electronic device designed to record alcohol‐based hand rub use (iSIGNOL; GOJO Industries) were placed in existing alcohol‐based hand rub dispensers located on a general medical ward and in the surgical intensive care unit (SICU). The general medical ward has 22 beds (in 14 rooms) and 23 alcohol‐based hand rub dispensers; 13 of these dispensers are located inside patient rooms, and 10 of these dispensers are located immediately outside patient rooms and at several other locations on the ward. The SICU has 15 single‐bed rooms and 31 alcohol‐based hand rub dispensers; 14 of these dispensers are located inside patient rooms, and 17 of these dispensers are located in corridors immediately outside patient rooms and in several other locations in the unit. Pocket bottles of alcohol‐based hand rub were not available on the general medical ward or in the SICU. The electronic devices were not placed in soap dispensers, which are from a different manufacturer. The alcohol‐based hand rub dispensers deliver an average of 1.5 mL of alcohol‐based hand gel with 1 press of the lever. The electronic devices record each time an alcohol‐based hand rub dispenser is used (defined as a hand hygiene event for the purposes of our study). When a dispenser lever is depressed twice within a 2‐second period, it is recorded as a single event. The time and date of each hand hygiene event are recorded in an electronic database located inside the device. Each electronic device has a unique identification number. The data records of hand hygiene events stored in each electronic device’s database were downloaded to a wireless, portable hand‐held “data logger” by moving the data logger in front of the alcohol‐based hand rub dispenser. A dispenser did not need to be opened to download information to the data logger. The prototype devices were designed to meet standards of electrical emission from Underwriters Laboratories, to avoid interference with other electronic equipment on the general medical ward and in the SICU. Records on the data logger were transferred to a computer in the infection control department. Data were subsequently uploaded to a password‐protected, dedicated Internet Web site for data analysis. For each device, the dedicated Web site provides analysis of hand hygiene events by month, by day of the week, by hour of the day, and by location. It also provides an analysis of the 5 most frequently used dispensers and the 5 least frequently used dispensers. For analysis of use by hour, the day shift was defined as being from 7:00 am to 6:00 pm, and the evening shift was defined as being from 7:00 pm to 6:00 am.

The prototype electronic devices, the data logger and accompanying software, and the access to the dedicated Web site were provided to the hospital without expense by the manufacturer for the purposes of the trial. Floor plans of the general medical ward and the SICU were obtained from the hospital’s engineering department, and the locations of the alcohol‐based hand rub dispensers equipped with the electronic devices were plotted on diagrams of the general medical ward and the SICU. The number of patient‐days occurring each month on the general medical ward and in the SICU was obtained from the hospital’s information services department, and the frequency of hand hygiene events was expressed as the number of hand hygiene events per patient‐day. Statistical analysis was performed using the χ² test and the Mann‐Whitney U test.

Results

During the 6‐month trial period, a total of 150,307 hand hygiene events were recorded on the general medical ward and in the SICU. In the SICU alone, 105,462 hand hygiene events using alcohol‐based hand rub dispensers were recorded, for an average of 579 hand hygiene events per day in the SICU. The average number of hand hygiene events per patient‐day performed with alcohol‐based hand rub was 48.7. On the general medical ward, 44,845 hand hygiene events using alcohol‐based hand rub dispensers were recorded, for an average of 246 hand hygiene events per day on the ward. The average number of hand hygiene events per patient‐day performed with alcohol‐based hand rub was 12.2. Figure 1 shows the monthly variation in the number of hand hygiene events per patient‐day for both the general medical ward and the SICU.

(19 KB)

Figure 1.  Hand hygiene (HH) events per patient‐day per month on the general medical ward (GMW) and in the surgical intensive care unit (SICU) of our university‐affiliated teaching hospital, from September 2007 through February 2008.

Open New Window

In the SICU, 49,567 (47%) of the 105,462 hand hygiene events using alcohol‐based hand rub were performed using the 14 dispensers located inside patient rooms, and the remaining 55,895 hand hygiene events (53%) were performed using the 17 dispensers located in corridors outside patient rooms. On the general medical ward, 16,144 (36%) of the 44,845 hand hygiene events using alcohol‐based hand rub were performed using the 13 dispensers located inside patient rooms, and 28,701 hand hygiene events (64%) were performed using the 10 dispensers located in corridors outside patient rooms. The proportion of hand hygiene events performed inside patient rooms in the SICU was significantly higher than that on the general medical ward ( ).

Analysis of the number of hand hygiene events recorded by each electronic device revealed that there were 4 dispensers on the general medical ward that were used much less frequently than other dispensers on the ward. These 4 dispensers were located inside patient rooms that have anterooms equipped with a sink and that are used most of the time to care for patients placed under contact precautions (Figure 2). During the 6‐month trial period, the median number of hand hygiene events for each of these 4 dispensers was 391 (range, 210–670), compared with a median number of hand hygiene events per dispenser of 2,000 (range, 809–2,272) for the other dispensers located on the ward ( ). A review of the number of hand hygiene events for all the other dispensers (on the ward and in the SICU) revealed that the 4 most frequently used dispensers in the SICU (range, 5,493–4,757 hand hygiene events) and on the general medical ward (range, 5,370–3099 hand hygiene events) were located in the corridors outside patient rooms. Analysis of the average number of hand hygiene events by day of the week on the general medical ward and in the SICU revealed that hand hygiene events occurred most frequently on Tuesdays (896 events), Wednesday (903 events), and Thursdays (882 events) and less frequently on Saturdays (742 events) and Sundays (735 events). Analysis of alcohol‐based hand rub use by hour of the day on the general medical ward and in the SICU revealed that hand hygiene events occurred least frequently from midnight to 2:00 am. The median number of hand hygiene events per hour on the general medical ward and in the SICU was significantly greater during the day shift than during the evening shift (43.0 vs 23.3 hours; ).

(68 KB)

Figure 2.  Location of the 23 alcohol‐based hand rub dispensers on the general medical ward of our university‐affiliated teaching hospital. Dispensers with the lowest rates of use are indicated by an “X.”

Open New Window

We did not witness any HCWs using dispensers very frequently within a few minutes with the intent of artificially increasing the number of hand hygiene events recorded. No electrical interference attributable to the prototype devices or the data logger were noted during the trial period. Six prototype devices that malfunctioned early during the trial period were replaced, and no further malfunctions occurred during the trial period. None of the devices required a battery change during the trial period.

Discussion

The prototype electronic devices used in our study provided specific data on how frequently HCWs used alcohol‐based hand rub dispensers when they performed hand hygiene (ie, the time of day and the day of the week), revealing the locations of dispensers with the highest and the lowest rate of use. Another unique finding of our study was that 36%–47% of all hand hygiene events using alcohol‐based hand rub were performed using dispensers located inside patient rooms, even though dispensers were located both inside patient rooms and immediately outside patient rooms. The electronic devices used in our study, unlike those used in earlier studies by Larson et al,^8,9 recorded the time and date that each dispenser was used and permitted the data to be downloaded wirelessly without having to open the dispenser. The electronic devices provided substantially more detailed information about patterns of alcohol‐based hand rub use than did studies in which use was expressed as the number of liters of hand hygiene product used per 1,000 patient‐days on a hospital‐wide basis or unit‐wide basis.^6,10‐12 The devices also provided a more accurate record of the number of times alcohol‐based hand rub was used than do systems that estimate the frequency of hand hygiene events by dividing the volume of product used by a proposed average amount utilized per hand hygiene event.⁷

The less frequent use of dispensers located in rooms used primarily for patients placed under contact precautions may be due to several factors. When leaving this type of room, HCWs often discard gloves and/or gowns in the anteroom, and they may have used the anteroom sink or dispenser located outside the room for cleaning their hands. Also, several previous studies have shown that HCWs enter these rooms less frequently because of the inconvenience of donning gloves and gowns for room entry.^13,14 The reason(s) for the greater use of patient‐room dispensers in the SICU than on the general medical ward could not be determined for our study. It may have been due in part to the fact that a greater number of hand hygiene opportunities during an episode of patient care occur in ICUs than in other patient‐care areas.^3,9

We found that a majority of hand hygiene events using alcohol‐based hand rub dispensers were performed by accessing dispensers located in corridors just outside patient rooms. It is difficult to know whether this is an unusual finding, because we are not aware of other studies that have determined the relative frequency of use of alcohol‐based hand rub dispensers located in corridors outside patient rooms, compared with those located inside patient rooms. The HCWs in our facility are accustomed to using dispensers in corridors because during the early years of our hand hygiene promotion campaign, all alcohol‐based hand rub dispensers were located in corridors rather than in patient rooms. Of note, a study by Broughall¹⁵ of an automatic monitoring system for measuring the frequency of hand washing found that the sinks used most frequently were located in a corridor near a nursing station. Our findings suggest that having alcohol‐based hand rub dispensers in corridors as well as in patient rooms may be useful. The fact that hand hygiene events occurred more frequently during the day shift is not surprising, given the greater number of hand hygiene opportunities that occur during the day.^3,16

The prototype devices that were used proved to be quite reliable, without the need for battery changes during the 6‐month trial or during the 10‐month period after the trial. Two devices disappeared from the dispensers in which they were located during the early part of the trial period. We have no way of knowing whether they accidentally became dislodged when the dispensers were being refilled or whether they were purposely taken by personnel. Some HCWs undoubtedly observed the infection control practitioner waving the data logger in front of the dispensers on one or more occasions, which could conceivably have led some individuals to “game” the system. However, we never observed HCWs depressing dispenser levers repeatedly within a few minutes with the intent of inflating the number of hand hygiene events recorded.

The ability to upload data to a password‐protected dedicated Internet Web site with data‐analysis capabilities substantially reduced the amount of time infection control practitioners had to spend on data analysis. In addition, the Web site provides a wider range of analyses than do other Web sites (eg, McGuckin Methods International)¹⁷ that collate data on hand hygiene product use from healthcare facilities.

Our study has a number of limitations. Prototype electronic devices were only used on the general medical ward and in the SICU, and the patterns of use for those locations may not be representative of the patterns of use for all types of hospital wards and units. Our study underestimated the total number of times that HCWs on the general medical ward and in the SICU cleaned their hands, because electronic devices were not present in one room of the general medical ward and in one room of the SICU, and because devices were not placed in soap dispensers, which were from a different manufacturer. As a result, there are few studies to which we can compare our findings. Larson et al⁹ found that manual alcohol‐based hand rub dispensers were used an average of 25.6 times per dispenser per day in 2 locations (a pediatric emergency department and a pediatric ICU). This average number of events was higher than the average number of hand hygiene events using alcohol‐based hand rub per dispenser per day on our general medical ward (10.3 events) or in our SICU (18.7 events). McGuckin et al⁷ reported hand hygiene rates of 5–9.7 events per patient‐day in an inpatient rehabilitation unit, but these rates were estimated on the basis of the volume of product used, rather than on the actual number of times alcohol‐based hand rub dispensers were used. The electronic devices used in our study recorded all hand hygiene events using alcohol‐based hand rub, including those by visitors and HCWs who may not have entered patient rooms. This factor, which is true for nearly all methods that monitor product use, may have led to an overestimation of the number of hand hygiene events performed by HCWs directly involved in patient care. Because the devices that we used were completely hidden inside dispensers, it is unlikely that their use resulted in a Hawthorne effect, something that often occurs with overt observation of hand hygiene or when electronic monitoring equipment is readily visible to HCWs. We did not record the volume of alcohol‐based hand rub used on the general medical ward or in the SICU, so we were unable to correlate the number of recorded hand hygiene events with the number of liters of alcohol‐based hand rub used per 1,000 patient‐days. Finally, we did not conduct concurrent observational surveys of hand hygiene compliance on the general medical ward or in the SICU, so we were unable to compare data on alcohol‐based hand rub use to actual compliance rates.

In conclusion, electronic devices provided an efficient and accurate method of determining the number of hand hygiene events performed with alcohol‐based hand rub at 2 study locations (ie, the general medical ward and the SICU). The large number of hand hygiene events recorded in the present study (ie, more than 150,000 events) was feasible because of the length of the study and limited time necessary to collect data from dispensers by using wireless technology. Unlike other methods for determining product use, the electronic devices that were used provided specific data on the number of hand hygiene events per month, per week, per day, and per hour. The devices allowed us to study the impact of the location of a dispenser on the frequency of hand hygiene and to establish which dispensers had the highest and lowest rates of use. Such devices are useful for monitoring hand hygiene rates before and after various types of interventions designed to improve hand hygiene performance among HCWs.