Hand Hygiene among Physicians: Performance, Beliefs, and Perceptions

Context Why do physicians fail to practice good hand hygiene? Contribution This observational study of 163 physicians in a university hospital found that overall adherence to hand hygiene guidelines was 57%. Factors associated with poor adherence included having busy workloads, performing activities with high risks for cross-transmission, and being in technical specialties (such as surgery and anesthesiology). Adherence was higher when hand-rub solutions were easily accessible and when physicians valued hand hygiene and considered themselves role models. Implications Providing easy access to cleansing materials and improving attitudes toward hand hygiene, particularly among physicians working in technical specialties, merit emphasis. The Editors Hand hygiene is recognized as the leading measure to prevent cross-transmission of microorganisms and to reduce the incidence of health careassociated infections (1, 2). Despite the relative simplicity of this procedure, adherence to hand hygiene recommendations is unacceptably low, usually well below 50% (1-4). Risk factors for nonadherence have been extensively studied (1, 4-7), and physicians have been repeatedly observed as being poor compliers (1, 3, 4, 8, 9). At our hospital, physician behavior did not improve substantially despite a hospital-wide hand hygiene promotion campaign that had a positive and marked effect on adherence among all other health care workers (1). That study highlighted the need for improved knowledge of behavior determinants among physicians. Promotion of hand hygiene behavior is a complex issue (7, 10-12). Adherence to hand hygiene recommendations is influenced by knowledge; awareness of personal and group performance; workload; and type, tolerance, and accessibility of hand hygiene agents (2, 4, 12). Over the past 50 years in particular, the assumption that an individual's perceptions have a strong effect on his or her behavior gave birth to social cognitive models of human behavior (13). Some of these models have been applied to individual factors (that is, knowledge, attitude, intentions, beliefs, and perceptions) to help build strategies that improve specific health behaviors (14). To date, individual cognitive factors related to hand hygiene have not been adequately studied among physicians. Our study aimed to investigate risk factors for nonadherence among physicians and to identify beliefs and perceptions associated with hand hygiene in this population. Methods Setting The University of Geneva Hospital is an acute care center that provides primary and tertiary medical care for residents of Geneva, Switzerland, and the surrounding area. Hand-washing facilities are conveniently located throughout the center; each patient room and all lavatories have 1 to 3 sinks, unmedicated soap, and paper towels (4). Individual bottles of an alcohol-based liquid hand disinfectant (Hopirub, B. Braun Medical AG, Sempach, Switzerland) are available in all areas, and pocket carriage of these bottles by each health care worker is strongly encouraged to facilitate bedside hand antisepsis (1). Infection-control structures and activities have been described elsewhere (15-17). Study Design We performed a cross-sectional study of physician hand hygiene practices and of physician beliefs and attitudes toward hand hygiene. Individual physicians were directly observed during routine patient care, and each physician completed a self-report questionnaire administered immediately after patient contact. All physicians at the University of Geneva Hospital were informed by mail about the upcoming study and were eligible for inclusion. At the time of the study, 1266 physicians were practicing at the University of Geneva Hospital: 440 staff physicians, 767 fellows and residents, and 59 medical students. Anonymity was guaranteed. The institutional review board approved the protocol as a Continuous Quality Improvement project. A hospital epidemiologist recorded all potential opportunities for hand hygiene among selected physicians. Observations were distributed throughout the hospital over a 6-month period in such a way that the observer would obtain a balanced distribution of observation periods throughout the entire institution. On entry to the predetermined ward, the investigator observed the first-encountered physician, provided he or she was involved in patient care activities. Each physician was included only once in the study. Although no physician declined to participate, 10 did not return the questionnaire. Multiple opportunities for hand hygiene (4) were observed during a single observation that varied in length, according to the physician's activity. As described elsewhere (4), the observer was trained and validated before the study during 40 monitoring sessions in which 2 observers worked simultaneously; at the time of the study, interrater agreement was high for all variables ( = 0.94 [range, 0.83 to 1.0]). Instruments and Variables Observation We assessed opportunities for hand hygiene and adherence as described elsewhere (1, 4, 18, 19) and according to published guidelines (2). Opportunities were stratified into 3 categories (1, 4, 18, 19): high risk for cross-transmission (before direct patient contact; between care of a dirty and a clean body site; before intravenous or arterial care; before urinary, respiratory, or wound care); medium risk for cross-transmission (after direct patient contact; after intravenous or arterial care; after urinary, respiratory, or wound care; and after contact with biological body fluid); and low risk for cross-transmission (other conditions). Of note, we considered both hand hygiene after patient contact to prevent contamination among patients and hand hygiene between a dirty and a clean body site in the same patient to prevent cross-transmission as opportunities for hand hygiene (1, 2, 4, 18, 19). Failure to remove gloves after patient contact or between a dirty and a clean body site on the same patient was considered nonadherence (1, 4, 18, 19). Hand hygiene action, whether by hand-washing or alcohol-based hand-rubbing, was the main outcome variable. Hand-washing refers to washing hands with plain soap and water or water alone, and hand-rubbing refers to the application of an alcohol-based solution on hands (2). Study variables included sex, medical specialty, type of hand hygiene opportunity (high, medium, or low risk for cross-transmission), availability of the hand-rub solution at the bedside or in an individual bottle for pocket carriage, glove use, activity index, and duration of the observation period. The activity index (1, 4, 18, 19) was estimated by the number of observed opportunities for hand hygiene per hour of patient care for each physician observation. Since the activity index represents the hand hygiene workload (1, 4, 18, 19), it is called the workload throughout the paper. At the end of the observation period, the observer asked the physician whether he or she was aware of being observed. Self-Report Questionnaire Immediately after physicians were observed, we gave them a self-report questionnaire to collect data on cognitive factors related to hand hygiene. We followed guidelines from social cognitive theories applied to health-related behaviors (14, 20-23) in the construction of the questionnaire. By using single items for measures and a 7-point scale for answers, we assessed cognitive factors, that is, intention to adhere to hand hygiene, perception of knowledge of hand hygiene indications, attitude toward hand hygiene, perception of social norms concerning hand hygiene (both behavioral and subjective norms) (24), perception of difficulty of adhering to hand hygiene, and perception of the risk for cross-transmission linked to nonadherence (Table 1) (25). The last 2 points of the scale closest to the positive perceptive evaluation were considered positive answers; all other points were considered negative answers (25). Motivation to improve hand hygiene was assessed by using a 3-point scale, and only the answer yes was considered a positive answer (Table 1). We measured knowledge of hand hygiene indications for 4 types of contact, according to standard definitions (2, 26), with structured questions. Good knowledge was defined as a correct answer to all 4 questions. Age, sex, professional status (medical student, resident, fellow, attending physician, or professor) and type and duration of medical practice were also recorded. Table 1. Assessment of Physicians' Individual Cognitive Factors Related to Hand Hygiene Statistical Analysis We investigated factors associated with adherence to hand hygiene (1, 4). Variables included were those collected during the observation periods and related to patient care activities, as well as cognitive factors gathered by the questionnaire. We performed all group comparisons by using logistic regression, with generalized estimating equations (27) to account for interdependence of clustered observations; each observed physician was included as a cluster. We built 2 forced-entry models: The first included only variables gathered during the observations, and the second included all variables (collected through the observations and the questionnaire). For both multivariate models, all observations related to 1 physician were excluded when any of the variables included in the model had missing values. The magnitude of the association between adherence and explanatory variables was measured by odds ratios and corresponding 95% CIs. All tests were 2-tailed, and a P value less than 0.05 was defined as statistically significant. We conducted all analyses with Stata software, version 7 (Stata Corp., College Station, Texas). Role of the Funding Source The funding source had no role in the collection, analysis, or interpretation of the data or in the decision to submit the manuscript for publication. Results Study Group We observed 163 physicians during 573 patient-care episodes, whi