Background
Hand hygiene is referred to as either hand washing with soap and water or hand disinfection. Important benefits of proper hand hygiene include reduction of nosocomial infections [
1], reduced transmission of multi-drug resistant pathogens [
2,
3], and cost effectiveness [
4,
5]. Alcoholic handrub is regarded to be superior to washing hands with soap and water. It has greater activity against microorganisms, less time constraints, and better skin tolerability [
5‐
7]. Furthermore, alcoholic handrub is better accessible in most settings as it can be provided in pocket bottles and may thus be available at any time at the point of care. The World Health Organization (WHO) has identified formulations for the local preparation of alcohol-based handrubs with substantially lower costs compared to commercial products [
8].
Compliance with hand hygiene varies greatly between countries and settings but is globally low [
5]. Several factors have been shown to be related to low compliance with hand hygiene in developed countries [
9]. In low-income countries the major reason for non-compliance with hand hygiene may be lack of adequate facilities [
10]. The burden of health-care associated infections (HAIs) is high in developing countries [
11]. WHO has established a multimodal implementation strategy to improve compliance with hand hygiene [
12]. Furthermore, the concept “my five moments for hand hygiene“was developed to perform hand hygiene in key moments [
13]. Allegranzi et al. found that the implementation of WHO’s hand-hygiene strategy is feasible and sustainable in different settings and countries and leads to significant compliance and knowledge improvement in health-care workers (HCWs) [
14]. There are, however, few data on the implementation of the WHO multimodal hand hygiene improvement strategy in Ethiopia, a country with high rates of nosocomial infections [
15].
The main objective for this study was to assess compliance with hand hygiene in selected wards of the Asella Teaching Hospital (ATH) before and after the implementation of the hand hygiene campaign. The secondary objectives were to assess compliance with hand hygiene for the different professional categories and the different wards and to assess perception and knowledge for the different professional categories before and after the implementation of the hand hygiene campaign.
Discussion
In our study, we found a very low compliance with hand hygiene at baseline. Compliance at baseline was similar to two studies that had been undertaken in Ethiopia [
19,
20] but was lower compared to a study from Bamako, Mali [
21]. The main reason for the low baseline compliance appears obvious: hand hygiene products and facilities were not available on the wards. Alcoholic disinfectants were only used for disinfection of patients’ skin prior to aseptic procedures. For this purpose, usually one bottle of gentian violet-stained alcoholic solution was provided on each ward. The accessibility of soap and water was similarly difficult. The majority of sinks were non-functional for different reasons. Furthermore, the water supply of ATH was limited. This was especially true during the dry season when water supply was completely cut for several days in a row on various occasions.
In addition to the lack of alcoholic handrub and water – although presumably less important – compliance with hand hygiene at baseline may have been low for social reasons. One senior physician mentioned that he was reluctant to use his own pocket bottle hand disinfectant in order not to create envy and shame among other HCWs.
Observation at follow up showed a significant increase of compliance with hand hygiene. This increase was consistent across both predefined professional categories and in all four wards and persisted in the second follow up after responsibility for hand hygiene had been transferred to ATH. Improvement was associated with the intervention and this association remained significant after adjustment for potential confounders. However, compliance with hand hygiene remained low compared to data from developed countries. In a landmark study, which was conducted in the University of Geneva Hospitals, compliance with hand hygiene rose from 47.6% at baseline to 66.2% over a four-year period [
3]. Particularly high rates of compliance with hand hygiene were observed in selected sites with vulnerable patients like intensive-care units after the implementation of a hand hygiene campaign [
22,
23]. It must be considered, however, that in these studies only specific indications for hand hygiene were assessed (e.g. after completion of patient contact, on entrance into the unit). The concept “my five moments in hand hygiene” was not applied. One important study assessed the implementation of WHO’s hand hygiene improvement strategy in five countries with different socio-economic background [
14]. Overall compliance before the intervention was significantly lower in low-income and middle-income countries than in high-income countries.
Compliance with hand hygiene improved across all indications except for “before aseptic procedures” in the first follow up. To our knowledge, this finding was not reported from previous studies. After having noticed preferential use of gloves before aseptic procedures at baseline and at first follow up, the use of gloves was systematically assessed during the second follow up. Observation revealed a high percentage of use of gloves (>90%) solely in the indication ‘before aseptic procedures’. The failure to change or remove contaminated gloves has been identified as major component of inadequate infection control practices [
24]. We had emphasized the need for hand hygiene regardless of the use of gloves in our trainings. However, data show that this imperative was not understood and must be stressed further. Compliance with hand hygiene was highest after body fluid exposure and after patient contact. This possibly reflects the HCWs’ priority for self-protection rather than for protection of the patients. Self-protection has been shown to be the engine for hand hygiene adherence in several studies [
5,
14,
25].
Interestingly, compliance with hand hygiene was similar in HCWs from category I and category II. This is in contrast to many studies, which found lower compliance in doctors than in nurses [
5,
9], although Allegranzi et al. observed better compliance with hand hygiene in doctors than in nurses in Mali [
21]. They hypothesized that the better compliance with hand hygiene in doctors could be due to a higher level of education and a stronger perception of their professional role [
14]. The compliance with hand hygiene was higher in the neonatology ward when compared to the paediatrics ward as a reference. One possible reason for this – apart from the presumptions that hygiene is of critical importance in neonatology and many HCWs in neonatology may be particularly dedicated to their work - may have been the presence of a professor who emphasized a lot the importance of hand hygiene in routine patient care. It has been shown that role models may influence compliance with hand hygiene [
26]. In our case the professor was a person that many HCWs looked up to and thus at least some HCWs may have tried to copy his behavior.
We produced alcohol-based handrub locally. Costs of local production were less than one fifth compared to commercially bought products. The skin tolerability of the handrub was perceived to be very good. After the intervention, hand hygiene actions were almost exclusively performed with handrub indicating high acceptance. The consumption of alcohol-based handrub increased steadily from the intervention until the end of the first follow up. The amount of alcohol-based handrub used was selected as indirect marker for compliance with hand hygiene in many settings although assessment based on product consumption cannot determine whether hand hygiene actions are performed in the right indications [
5].
There were several challenges in our setting, which may have hampered achievement of better compliance with hand hygiene. First, we faced intense staff rotation on all levels within ATH. Since the beginning of the planning process until the end of the second follow up assessment, the position of medical director of ATH changed three times. Staff rotation on the wards resulted in observation of entirely different teams at baseline and follow up. Second, during the follow up assessment we were frequently shown empty wall-fixed alcoholic handrub dispensers and empty private pocket bottle hand disinfectants. Refill of alcoholic hand-rub had occasionally been fetched at HITM but then no further distribution among HCWs of the respective ward had been undertaken. On one ward, alcoholic handrub had been locked and was only accessible for one HCW. Third, although we did not measure the consumption of alcoholic handrub from wall-fixed dispensers and from pocket bottles independently, we felt that wall-fixed dispensers were used preferentially. According to our ward infrastructure, wall-fixed dispensers were mounted in selected sites with intense patient care. It seems obvious that the provision of wall-fixed dispensers to every room where patient care is performed would have been preferable. The concentration solely on the provision of alcohol-based handrub was regarded as a limitation by many HCWs and parts of the management of ATH. Whitby and McLaws demonstrated, however, that improved accessibility to sinks does not lead to improvement in compliance with hand hygiene [
27]. In addition, WHO recommends the combined provision of pocket bottles and wall-fixed dispensers filled with alcohol-based handrub without focusing on water supply [
5]. Just before the end of this study, management of ATH implemented a hygiene committee. The committee took over responsibilities like identification of further structural necessities. It has been shown that designated staff is one major critical component of an effective infection control program [
28].
Results of the hand hygiene knowledge questionnaire were significantly better after the training than before the training. The improvement was seen in both professional categories and was similar to the improvement detected in a study from Bamako, Mali [
21]. It may seem surprising that even immediately after the training the median scores reached were still far below the possible maximum score. We found that the way several questions should have been answered was not understood by many examinees (e.g. in some questions it is stated “tick one answer only” whereas in others there is no such statement. Many examinees wrongly only ticked one answer in these questions, too and therefore lost the chance to reach higher scores). We had already adapted the questionnaire to the local situation in accordance with WHO recommendations. However, some structural modifications may be necessary especially for HCWs who are not used to multiple choice exams and may be confused by the changing design of the questions. In further studies, current tools could be compared with adapted tools to confirm or refute our concern.
In contrast to Allegranzi et al., who detected significant increases in median perception scores for all five components of the multimodal WHO hand hygiene improvement strategy, we only found increased median scores in the component “reminders at the workplace” [
21]. This may be explained by the high baseline scores in our study. HCWs estimated their compliance with hand hygiene and the compliance of their co-workers to be high. Estimates differed greatly from our observation findings. This is in line with various studies, which reported that the correlation between self-assessment and observation findings is low [
29,
30]. The acceptance of the different elements of the hand hygiene campaign and the perceived impact of hand hygiene were very high as indicated by maximum scores after the training. This finding supports the multimodal approach recommended by WHO. We cannot know, however, which element was most important for the observed outcomes and how the outcomes would have been if one or several elements had been omitted.
Our study has several limitations. First, although English is the official language of medical education in Ethiopia, not all HCWs have good English language skills. HCWs insisted on English presentations and WHO could not provide working materials in Amharic. We managed to establish question and answer sessions in Amharic after each training but we cannot exclude that outcomes would have been better if Amharic had been used preferentially.
Second, it seems logical that HCWs are reminded of performing hand hygiene actions in the presence of an observer. Observation at second follow up was entirely performed by local staff (TN), whereas international staff (NS) did most of the observation at baseline and at first follow up. Assessments of different observers may vary. However, all observers were well trained in WHO hand hygiene observation methods, and criteria defined by WHO are straightforward to minimize inter-observer differences [
17].
Third, the hospital management addressed the wish of extension of our hygiene activities to the entire hospital arguing that all HCWs and patients should benefit from the positive effects of proper hand hygiene. We felt that our approach was adequate as pilot project to demonstrate feasibility and efficacy. Hospital management was in charge of ensuring sustainability of the project and of extending activities to the ward that had not yet been covered.
Fourth, we performed two follow up assessments in relatively short time intervals after the intervention. It would have been preferable to perform a time series analysis with several follow ups to longitudinally assess compliance with hand hygiene.
Last but not least, the clinical relevance of our intervention remains unclear in the light of compliance rates that were still low at follow up. To assess the rates of HAIs in ATH surveillance activities would have to be implemented. Surveillance is essential to record the burden of infectious diseases and the effect of interventions. Moreover, by itself, surveillance can lead to reduction in HAIs [
31]. The most widely used surveillance definitions for HAIs come from the Centers for Disease Control and Prevention (CDC) and the National Healthcare Safety Network [
32]. They are rarely applied in low-income countries as strict criteria have to be used including bacterial culture in most settings. Future research may help to develop criteria, which are adapted to the settings in resource-constrained countries. Ideally, prospective investigations should assess both compliance with hand hygiene and rates of HAIs.
Acknowledgements
Not applicable.