INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The hands of health care workers are the main vehicles of multidrug-resistant bacteria and nosocomial infections in hospital wards (1). Hand disinfection, generally achieved by mechanical washing, remains the cornerstone of prevention of such infection (4, 5). However, despite regular publication of guidelines and recommendations, compliance with handwashing by health care workers rarely exceeds 40% (6). Several factors may contribute to this unsatisfactory rate, including time limitation (9). Indeed, although conventional handwashing takes only about 10 s (7, 10), the entire procedure of leaving the room, moving to the sink, adjusting the tap, drying the hands, and returning to the patient takes more than a minute (11). More rapid and effective hand disinfection procedures have been proposed, such as rubbing with alcohol (12, 13). In this prospective study, we examined whether the availability of hand rubbing with an alcohol solution increased global compliance with hand disinfection by health care workers in a medical intensive care unit (MICU), and the acceptability of this procedure.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Design

The study was conducted in a 14-bed MICU in a 909-bed tertiary teaching hospital during two consecutive 5-wk periods. The unit contains 14 single bedrooms. Each bedroom is equipped with a manual sink. An electronic sink is located close to nurses' working surfaces (of which the unit has three). For the purpose of the study, opportunities for handwashing were clearly defined. Briefly, they were divided into three categories: (1) personal gestures (e.g., before and after duty hours and after eating, self-contact [other than hand-to-hand], visits to the lavatory); (2) patient care without exposure to body fluids; and (3) patient care with exposure to body fluids.

The medical and paramedical staff were regularly sensitized to the guidelines by means of regular meetings (once a week through a usual formal meeting with the paramedical and medical staffs) and wall posters (10 in the unit). This program began before the onset of the study, and was not reinforced between the two 5-wk study periods. Moreover, no performance feedback was reported during the study, and no encouragement was given to use any method. During the first 5-wk period (from January 1 to February 13, 1998), only conventional handwashing, using Savon Doux (imidazolin, glycerin) (Anios, Lille-Hellemmes, France) was available. During the second 5-wk period (from February 16 to March 20, 1998), both Savon Doux soap and alcohol rubbing with Sterillium (Bode Chemie, Hamburg, Germany, distributed by Rivadis, Thouars, France) were available. Sterillium contains 2-propanol [45%]; 1-propanol [30%]; and mecetronium [0.2 g/100 ml]). Each member of the staff chose the method he or she preferred. Alcohol rubbing was done as follows: the care worker rubbed his or her hands with Sterillium until they were dry. Sterillium was provided either in 100-ml individual capped flasks (easily carried in an ordinary pocket) or in wall distributors at each bedroom entrance and close to the nurses' working surfaces in each subunit (a total of 19 distributors). In our MICU, physicians and residents wear white coats, whereas nurses do not. Consequently, physicians and residents carried individual flasks in their coat pockets whereas nurses used wall distributors.

Compliance Study

During the first and second 5-wk study periods (P1 and P2), two members of the MICU medical staff (J.L.B. and H.N.) separately assessed the compliance rate with the previously defined handwashing opportunities during 20-min observation sessions selected at random. The sessions took place mainly during the duty hours of the two observers (8:30 A.M. to 7:30 P.M.), when their presence was not surprising; 15% of the observations were accomplished during the night, when the first observer (J.L.B.) was on duty and remained in the unit all night long. Both medical and nonmedical staff members were aware of the study, but did not know exactly when they were being observed. Three groups of health workers were defined as follows: paramedical staff (nurses and assistant nurses), physicians (senior physicians), and residents (residents and students).

Assessment of Acceptability

Each health care worker in the unit was given an anonymous questionnaire to assess his/her personal perception of alcohol rubbing and its potential unwanted effects (dry skin, pruritus, rash) on analogue scales (0 = manifestation absent, 10 = overwhelming reaction); such effects were considered significant when the score exceeded 6. The existence of a skin lesion with a score less than 6 on the application area of the alcohol solution, and not requiring cessation of alcohol rubbing, was defined as a mild skin lesion. The staff workers were also invited to report unlisted unwanted effects.

Complementary Study

To assess whether the availability of alcohol rubbing induced a lasting improvement in hand disinfection compliance, we conducted an additional study, 4 mo after the completion of the first one. This complementary study was conducted between July 18 and August 23, 1998, by the first observer (J.L.B.), using the same methodology as in the first study, except that during this period health care workers did not know they were being observed.

Statistical Analysis

Compliance rates are expressed as percentages. Comparisons are based on the chi-square test or Fisher's exact test as appropriate. Statistical significance is defined as a value of p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The observers assessed 68 sessions during P1 and 79 sessions during P2, totaling 621 and 905 opportunities, respectively, for hand disinfection. During each period, 10 sessions were assessed from 7:30 P.M. to 8:30 A.M. During P1 and P2, the breakdown according to health care worker categories was, respectively, paramedical: 256 and 359; physicians: 250 and 326; and residents: 115 and 220. The breakdown according to handwashing categories was, respectively, personal gestures: 46 and 132; care without body fluid contact: 445 and 587; and care with body fluid contact: 130 and 186. The first observer observed 339 and 636 situations during P1 and P2, respectively, and the second observer observed 282 and 269 indications, respectively. The mean volume of alcohol solution dispensed by the personal flasks and wall distributors was 1.3 ml. The mean duration of the entire alcohol rubbing procedure was 25 s. Conventional handwashing (excluding the time required to move to and from the sink) took a mean of 35 s.

Compliance Study

The general compliance with hand disinfection was 42.4% during P1 and 60.9% during P2 (p = 0.001). The increase in compliance from P1 to P2 was also observed in each professional category as follows: paramedical staff, from 45.3% to 66.9%; physicians, from 37.2% to 55.5%; residents, from 46.9% to 59.1% (Table 1).

The compliance rate with indications linked to care procedures (with and without body fluid contact) also increased significantly from P1 to P2, whereas there was a slight but nonsignificant fall in compliance with indications relating to personal gestures (Table 1). Both observers reported similar improvement rates.

The proportion of use of conventional handwashing and alcohol rubbing during P2 among the three health care worker categories was not homogeneous (Table 2). Alcohol rubbing was clearly preferred by physicians and residents, whereas the members of the paramedical staff used the two methods with a quite similar frequency. The rates of use of the two methods were not different among physicians and residents (p = 0.86), whereas the rates differed significantly as compared with that for the paramedical staff (p < 0.001).

The frequency of use of the two methods differed according to handwashing categories (Table 2). Alcohol rubbing was significantly preferred for personal gestures and care procedures without body fluid exposure as compared with care procedures with body fluid exposure (p < 0.001). The frequency of use of the two methods was not different for personal gestures and care procedures without exposure to body fluids (p = 0.56).

Acceptance

Forty-eight of the 53 questionnaires distributed were completed. The responses showed that alcohol rubbing was easy to use (100%). None of the respondents reported having to cease using alcohol rubbing because of unwanted effects. Pruritus occurred in 4%, mild skin lesion in 10%, and dry skin in 4% of respondents. All respondents reported the need for a conventional handwashing after three or four alcohol rubbings because of the sensation of a deposit. Conventional handwashing induced mild skin lesions (pruritus, dry skin) in 6% of users, but none of the respondents reported having to cease using the handwashing soap as a result.

Complementary Study

Compliance with hand disinfection during the complementary study period remained better than during P1 (51.3% versus 42.4%, respectively, p = 0.007), but was lower than the rate during P2 (51.3% versus 60.9%, respectively, p = 0.002).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main result of this study was that the availability of an alcohol solution increased general compliance with hand disinfection in a MICU. Since Semmelweis, hand disinfection has been seen as an important part of patient care (1). Until now, few randomized controlled studies have suggested that handwashing could decrease nosocomial infections (14, 15); however, such trials would now be deemed unethical (16).

Compliance with handwashing remains low mainly because it is a time-consuming procedure (9, 11). The overall compliance rate with conventional handwashing in our study (P1) was 42.4%. Alcohol hand rubbing, which in our study took only about 25 s, improved compliance with hand disinfection.

Several authors have investigated health care workers' motivation for hand hygiene (13, 14, 16, 17). A recent study done in the restrooms of a western hospital revealed than no more than two-thirds of persons washed their hands after using the toilet facilities, pointing to a real problem of education about sanitation (18). Previous studies in ICUs expressed this same view (1). It is noteworthy that nurses are more compliant than physicians with handwashing in ICUs (1). In our study, compliance with hand disinfection was significantly improved when alcohol rubbing was available, both among the entire MICU team and also in each of the three professional categories. Alcohol rubbing was made readily available by means of personal flasks and numerous wall-mounted distributors, whereas conventional handwashing required leaving the room, moving to the sink, and returning to the room. The availability of alcohol rubbing significantly improved compliance with hand disinfection in situations involving patient care (with or without body fluid contact).

Alcohol rubbing was used to a greater extent by physicians and residents than by the members of the paramedical staff. This discrepancy could have been related to a different acceptance of the alcohol solution in the three health care worker categories in the MICU. This hypothesis seems unlikely to have occurred, since availability of the alcohol solution for handwashing increased compliance with hand disinfection in each worker category. Moreover, the alcohol solution was equally evaluated by each health care worker. The difference could have been due to several other reasons. The most relevant of them is that, despite its lack of simplicity (when compared with alcohol rubbing), conventional handwashing with mechanical drying remains the most efficient method of eliminating contaminating particles, and is warranted during the most soiling instances of care (12). The lesser utilization of alcohol rubbing could have been related to the greater percentage of care procedures with exposure to body fluids performed by the paramedical staff (58% of care procedures with exposure to body fluid contact were performed by the paramedical group), and to the general involvement of nurses and nurses' aides in the most soiling care procedures (contact with feces, tracheostomy care). Moreover the nonavailability of personal flasks to the paramedical staff could in part explain the lesser use by this group of the alcohol solution. However, when opportunities for hand hygiene following personal gestures and patient care without exposure to patient fluids were considered as a separate category, nurses used alcohol rubbing in a similar fashion to that of physicians and residents.

All members of the MICU staff knew that they were being observed, but during P2, when the study subjects knew that they were definitely under observation, compliance with hand disinfection improved, suggesting that this improvement was related to the introduction of alcohol rubbing rather than to the fact that workers knew they were observed. Furthermore, the persistent effect observed during the complementary study strongly suggests that modification of hand hygiene was probably caused by the introduction of the new method of alcohol rubbing. Although this work was prospective, its design cannot exclude a relative learning effect occurring when alcohol rubbing was made available. However, we consider this hypothesis unlikely, because no additional educative activity toward hand hygiene was achieved when rubbing was introduced. We offered the MICU staff a choice between handwashing and alcohol rubbing. As we have observed, alcohol rubbing cannot completely replace conventional handwashing when the hands are soiled or covered with proteineous materials (19). Furthermore, all respondents to the study questionnaire stated the need for classical handwashing after three or four alcohol rubs because of the sensation of a deposit.

The goal of hand disinfection is to decrease the rate of nosocomial infection. Unfortunately, the rate of nosocomial infection also depends on the severity of patients' illnesses, patients' own flora, antibiotic use, invasive procedures, and other factors. These points probably explain why handwashing alone is often insufficient for infection control (2, 6, 8, 9, 19, 20) although it has been previously shown that an improvement in compliance with hand sterilization could decrease the nosocomial infection rate (15). For these reasons, we wanted to assess the efficacy of alcohol hand rubbing for increasing hand hygiene in our unit before assessing the effect of this method on the incidence of nosocomial infections.

We limited the use of the alcohol hand rubbing to conventional hand disinfection, even though this method has been proposed for thorough antiseptic or surgical washing (19). The availability of alcohol rubbing for staff personnel caring for high risk patients (e.g., with human immunodeficiency or hepatitis B virus infection) would probably not have markedly improved hand disinfection compliance, which is generally higher with high-risk patients than with low-risk patients (9). The "effect" of alcohol rubbing may be related in part to its novelty. In fact, as recently emphasized (21), we consider it essential that ICU staff personnel be regularly sensitized to the guidelines for infection control in order to maintain an acceptable rate of compliance. At 3 mo after its introduction in our MICU, alcohol rubbing maintained its beneficial effect on hand hygiene, with handwashing compliance being significantly better than that observed during P1.

When health care workers carry their own flasks of alcohol solution, hand disinfection can be achieved anywhere, even outside the ICU, when caring for patients in other units. The excellent availability of this method, combined with its simplicity (no need for a sink, soap, or paper towel distributor) should translate into an improvement in ICU infection control, as recently suggested (22).