Introduction

Health care-associated infections (HCAIs) are commonplace. The term encompasses a range of infections that may be contracted by a patient during a stay in the hospital for a pathology other than that infection [1]. Such healthcare-associated infections may, similarly, be contracted following contact with health services in clinics, long term care facilities, etc. Collectively, these infections represent a serious public health challenge linked with increased patient morbidity and mortality, and an economic burden for healthcare systems [2]. Unsurprisingly, infection prevention and control practices in developed countries have targeted reduction of HCAI impact, resulting in approximately 10% of hospitalised patients in higher-income countries affected. However, developing countries report considerably higher incidence (in some cases, greater than 25% of all hospitalised infections) [3]. In Europe, prevalence rates of HCAIs fluctuate between 4.6% and 9.3% [4].

The financial costs related to HCAIs are well understood, estimated at between $28 and $45 billion [5]. In the UK, public hospitals registered more than 650,000 HCAIs in 2016/17 amongst 13.8 million inpatients, including 22,800 fatal cases [6]. According to Stewart et al., the HCAI may create a further 7.8 days to the patient's hospital length of stay (LOS) with a median LOS of 30 days for HCAI patients and 3 days for non-HCAI patients. The paper concluded that a 10% decline in HCAI incidence could probably free up to 5800 bed-days [7]. It has been estimated that the NHS spent almost £2.1 billion on HCAIs in 2016/17 [6].

Causes of HCAIs are also reasonably well understood and relate mainly to patient and health care worker (HSW) transmission of potentially pathogenic microbes, often involving contamination of surrounding or nearby surfaces (e.g., equipment, clothing, sanitary ware) [8,9,10]. Efforts to mitigate risk of such transmission has focused, to a great extent, on hand hygiene (HH) [11], epitomised by the 2009 World Health Organisation “5 Moments for Hand Hygiene” [12]. In that context, Chen et al. estimated that US$25 can be saved for each US dollar spent on HH awareness [13].

Notably, HH compliance (HHC) in the emergency department (ED) has remained recalcitrant to significant improvement [14,15,16,17]. This is particularly challenging as, in most developed countries, the ED is the principal point of access for life-saving assistance by critically ill or injured patients [18]. Distinct from inpatient units, the ED environment and practices reflect elevated complexity associated with critically ill or polytrauma patients and overcrowding; contributing factors to the relatively low compliance rates observed. More specifically, Muller et al. [16] reported that ED crowding contributes to lower ED hand hygiene compliance highlighting typical restricted physical spaces within EDs, relatively close proximity of patients and staff, and unpredictable acuity level of patients as some of the more important HHC challenges in the ED. They further refer to ED staff often consulting more patients per shift than colleague sin other specialties, and the fact that such consults commonly occur in corridors and exposed spaces due to overcrowding and attempts to expedite patient care; often without ready access to alcohol-based hand rub [16]. Venkatesh et al. [19] reported complementary observations. Paradoxically, the high frequency of invasive procedures performed in the ED represent an opportunity for improvement in HH and reduction in HCAIs [16].

It is unfortunate then that systematic reviews and meta-analyses of HHC in EDs are scarce. Seo et al. [20] completed an appraisal of interventions to improve HHC in Eds [20] (from 1948 to 2018); reporting that 80% of eligible papers applied multimodal or dual interventions to improve HHC albeit that only half of the included studies observed a HHC rate above 50%. The review, however, utilised narrow inclusion criteria and included only studies reported in English and Korean. Cartel et al. [2] conducted a review of common infection control practices in the ED setting, between 2002 to 2012, (finding HHC rates ranging between 7.7% and 89.7%). Given such disparity, it is compelling to conduct a more inclusive systematic review that focuses on handwashing in the ED for the prevention of nosocomial infections associated with admitted infectious patients.

Methods

Study design

Registered with PROSPERO (https://www.crd.york.ac.uk/prospero) as CRD42021234488 [21], we conducted searches in the following databases: Web of Science, PubMed, CINHAL & Medline (EBSCO Host), Cochrane and Embase, on any article related to ED hand hygiene compliance published between 1st January 2010 and 31st December 2020. The following search terms were used alone or in combination: PubMed—(hand hygiene) OR (handwashing) AND (emergency room OR emergency department OR ER OR ED) [Title / Abstract] ) AND compliance [Title / Abstract]; CINHAL and MEDLINE and EMBASE and Cochrane—(hand hygiene) OR (handwashing) AND compliance AND (emergency room OR emergency department OR ER OR ED); Web of Science—(hand hygiene) OR (handwashing) AND compliance AND (emergency room OR emergency department OR ER OR ED) AND article, a summary of published article [type of document].

Articles were deemed eligible if they described any intervention, peer-reviewed or grey literature, seeking to affect HHC by ED staff. In summary, this included observational studies, randomised and non-randomised controlled studies, before and after studies, and cross-sectional studies that investigated effectiveness of interventions on HHC in EDs determined either by observation or the use of the camera surveillance, with no limitation to English language or geographical location. Unlike previous reviews on the topic, our criteria encompassed all healthcare providers working in emergency departments: physicians, residents, nurses including assistant nurses and student nurses, healthcare assistants, technicians, physiotherapists, and any other ED personnel.

Specifically excluded were studies conducted outside the ED settings (ICU, prehospital setting, any ward, military bases, prisons, schools, etc.), editorial opinions, posters presentations (not full article), unpublished conference proceedings (not full article), letters to the Editor, opinion-based publications, doctoral theses, and quality improvement projects.

Data extraction

Two reviewers (MI and CPD) hand-reviewed search results independently to identify articles meeting the inclusion criteria. Wherever inconsistencies arose, disagreement was discussed and resolved by consensus. Data extraction was completed by the reviewers separately and disagreement again resolved by consensus. A data extraction form was utilised that captured: the study design; year of publication; database source; country of the publication including World Bank regions and income classification; sample size; types and characteristics of the participants; recorded number of HH opportunities; HHC rate; study intervention; details of the intervention; strengths and weaknesses of the study; and outcomes. Meta-analysis could not be conducted due to abundant heterogeneity of the studies. Therefore, analysis was only possible by compiling results manually in tables.

Assessment of bias risk in eligible studies

The reviewers (MI and CPD) analysed the methodological quality of the included studies. For observational studies, the Risk of Bias Assessment for Nonrandomized Studies (RoBANS) tool [22] was utilised. For cross-sectional study designs, the Appraisal tool for Cross-Sectional Studies (AXIS) was used to assess five elements of each study (Introduction, Methods, Results, Discussion and other) by answering 20 stipulated questions [23].

Results

Following the search, 15,484 publications were retrieved of which 15,355 papers were excluded as duplicates or evidently ineligible (Fig. 1). Of the remaining 129 papers, a further 78 were excluded subsequent to abstract screening (60%), as conducted outside ED settings (69 papers, 53%) and additional duplicates were found (9 papers, 7%). Therefore, full-text appraisal was performed for 51 studies. Of those, 31 studies failed to fully meet the inclusion criteria, with 20 studies (16%) of HHC eventually included in this review (Table 1).

Fig. 1
figure 1

PRISMA flow diagram for the selection of studies

Full size image
Table 1 Included articles
Full size table

Details of the included studies

Of the 20 included articles (Table 1), four papers were cross-sectional studies, two studies were retrospective, and 14 publications were observational studies (Table 2). Notably, 18 papers (90%) were in English, one study (5%) in Portuguese, and one publication (5%) in Spanish (Table 1). Six of the papers (30%) were retrieved from CINAHL & Medline, six (30%) from Embase, three (15%) from PubMed, four (20%) from Web of Science, and one (5%) from Cochrane (Table 1).

Table 2 Study design, setting, and location of included articles
Full size table

Geographic location

Using World Bank country classification [24], seven studies were found to have been conducted in North America (USA n = 5, Canada n = 2). Three were performed in East Asia & Pacific (Australia n = 1, Japan n = 1, Republic of Korea n = 1), five articles originated from Europe (Italy n = 2, France n = 1, Germany n = 1, Netherlands n = 1). Two studies originated in Latin America & the Caribbean (Brazil n = 1, Ecuador n = 1). The Middle East & North Africa generated two papers (Saudi Arabia n = 1, Iran n = 1) and one study was conducted in Ethiopia (Table 2).

Four cross-sectional studies (20%) were included in the review, each emanating from a different region (Canada, Ecuador, Iran, and Italy). Among the sixteen observational articles, four studies (20%) were before-and-after designs, conducted in Ethiopia, Netherlands, France, and Japan. Nine studies (45%) were descriptive observational papers from Australia, Brazil, Canada, Italy, Saudi Arabia, Korea, and the USA. Two of the twenty articles (10%) were retrospective observational studies performed in the USA, and one article (5%) described tri-phase observational research conducted in Germany (Table 2).

Quality assessment results

Using the RoBANS tool [22] (Table 3), 14 studies had low risk of bias for selective outcome reporting, 16 had low risk of bias for incomplete outcome data, one paper had low risk of bias of blinding for outcome assessment, eight had low risk of bias for intervention (exposure) measurement, 11 had low risk of bias for confounding variables. All papers had high risk of bias for selection of participants. Four cross-sectional study designs were appraised using the AXIS tool (Table 4); all four associated with low risk of bias, albeit that two did not discuss the limitations of their studies.

Table 3 Risk of bias assessment tool for nonrandomized studies (RoBANS) [22]
Full size table
Table 4 The appraisal tool for cross-sectional studies (AXIS tool): Yes, No, Don’t know [23]
Full size table

Study objectives

Of the 20 articles, 55% (N = 11) intended to evaluate HHC among ED HCWs based on the WHO-recommended 5-moments of hand hygiene. Fifteen percent (N = 3) assessed the effect of ED crowding on ability of HCWs to correctly perform HH. Similarly, fifteen percent (N = 3) estimated the bacterial loads of HCWs hands (N = 2) or equipment (N = 1) within EDs., while the same number (N = 3) assessed compliance using ABHR solutions (Table 5).

Table 5 Study objectives of included articles
Full size table

Study design and settings

Of the included studies, only four (20%) were multicentre studies, with the remainder performed in discrete EDs in single hospitals (Table 2).

Healthcare worker category and time and place of intervention

Although all studies stipulated HCW categories involved, 12 of 20 studies (60%) did not provide specific detail of included participants differentiated by profession. Irrespective, HCWs involved were mostly nurses and nurse's assistants, physicians of all grades, healthcare assistants, technicians, and ‘other’ staff (Table 6). More specifically, 16 of the 20 articles listed nurses, 13 studies associated physicians, seven papers included residents, seven studies cited technicians, and six studies involved nursing assistants among their population sample.

Table 6 Health care workers category and time/place of interventions
Full size table

With respect to timing and location at which observations were conducted, most studies ran across days, evening and nights shifts throughout both weekdays and weekends. Exceptions were Muller et al. [16] who did not include weekends shifts in their investigation, and Sakihama et al. [37] and Schmitz et al. [39] who conducted their study only during day shifts. Notably, some studies focused on the influence of location within EDS on HHC, including Cartel et al. [26] who included in their observation the location of patients such as hallways, semiprivate and private ED areas to estimate the effect of crowding on HHC. Use of technological surveillance was detailed by Strauch et al. [41] who used a retrospective reading of an electronic device to track the HH activity data on each personalised badge when returned to a charging station.

Study interventions

Twelve studies (60%) implemented a specific strategy to enhance HHC within the ED either by multimodal or single intervention strategy. Education and training were employed in 9 of the 20 studies (45%) [25, 27, 29,30,31, 34, 37,38,39]. Reminders applied in 4 of the 20 articles (20%) [25, 37,38,39]. The use of feedbacks was highlighted in three publications (15%) [25, 37, 38]. Emails to alert participants were sent by three authors (15%) [26, 40, 41]. Posters were utilised by two authors (10%) [25, 38]. Evaluation and Questionnaires were reported in three articles (15%) [30, 34, 37]. Feedback approaches were associated with three publications (15%) [25, 37, 38]. Two authors incorporated the method of role models in their study (10%) [27, 39]. Video training was associated with Ghazali et al. [30], while two authors included presentations to improve the HCWs HH practice [30, 38]. Simulation and live demonstration were incorporated in Fouad et al.'s article [29]. Eight studies (40%) were unclear in terms of their intervention. Amongst them, seven papers (35%) used direct observations techniques [15, 16, 19, 28, 32, 33, 36] and one study employing a video surveillance approach [31] (Table 7).

Table 7 Study interventions and compliance outcome
Full size table

Arntz et al. [25] and Schmitz et al. [39] utilised additional education, feedbacks, and reminders, while Sakihama et al. [37] recruited a new infection control specialist in 2 of their 3 centres to assist with improvements. Di Martino et al. [27] provided formal training and employed role models wearing reminder buttons to promote HH. Fouad et al. [29] mounted live demonstrations and posters as nudges towards change. Of interest, Stackelroth et al. [40] installed a video surveillance system but without clear HH training for staff. In contrast, Sakihama et al. [37] supplemented available training and then instigated a competition to reward best performing groups while, in a questionable aspect to their study, Zottele et al. [15] organised seven sessions with their study observers to increase their familiarity with HH, but no formal training was provided to the subject ED staff.

Measurement methods and main outcomes

Fifteen studies (75%) assessed primary outcomes according to the WHO-recommended 5-moments [34]. One paper assessed the microbial loads on the hands of the ED HCWs [28], one focused on the adherence rate to preventive measures (PM) against blood-borne disease. Martel et al. [34] centred on respiratory hygiene compliance, with HH as only one of its components. One article has as the primary endpoint the use of alcohol-based hand cleanser in the ED with non-sterile glove use, and Strauch et al. [41] emphasised decline in the rate of sick calls for nurses and technicians using an automated HHC system (HHCS). Almost all studies mentioned a certain degree of training of the involved observers, of which only three studies (15%) used qualified infection control experts [19, 37, 38]. Three studies (15%) used trained medical students as observers [25, 29, 34]. Seven studies (35%) used a single observer [15, 16, 32, 34, 37,38,39], three studies (15%) involved two observers [25, 27, 30]. Cartel et al. [26] included four observers, Martel et al. [34] worked with five medical students, and Venkatesh et al. [19] conducted their research with five trained observers and one qualified infection control specialist (Table 7). Coming to statistical analysis, fifteen studies (75%) used frequency percentage (%). Eight papers (40%) used the 95% Confidence interval (95% CI) in their studies. The P-value was described in twelve of twenty studies (60%). Mean and standard deviation for parametric data was used by five authors (25%) [15, 30, 33, 36, 40] while the median and interquartile range (IQR) was associated with four papers (20%) [28, 34, 35, 40]. Odd ratio was applied in four publications [15, 16, 34, 35]. Venkatesh et al. [19] brought in an adjusted risk ratio (aRR) (Table 8).

Table 8 Results
Full size table

Hand hygiene opportunities

Fourteen studies described 25,192 HHOs and 1664 hand rubs (HR) across all HCWs observations. The highest opportunities were described by Scheithauer et al. [38] and Venkatesh et al. [19], respectively, with 7338 (5674 HHOs and 1664 HR) and 5865 HHOs. Zottele et al. [15] reported the lowest number with 116 HHOs. Other studies recorded considerably less participants and HHOs; specifically: Ghazali et al. [30] 22 participants, < 220 HHOs; di Martino et al. [27] conducted 3 intervention phases referring to > 420 HHOs in each; Cartel et al. [26] appraised nurse HHO rate (55%) comparing to that of physicians (32%) or Nursing assistants (10%) (Table 7); Fouad et al. [29] reported similar HHOs among nurses (83.55%) and physicians (14.41%); as did Scheithauer et al. [38]. Muller et al. [16] found that nurses had 68% HHO, physicians totalling 18%, and others (nursing assistant, housekeeping, transport, IV team, radiology technicians, and nursing students) scoring 10% (Table 8).

Distinctly, Espinoza Diaz et al. [28] conducted a microbiological evaluation of HHC stating high HH compliance amongst nursing staff, higher than interns, residents, physicians, and nurse's assistants respectively. Hong et al. [32] swabbed ED electronic devices (56 keyboards and 56 mice), while Martel et al. [34] targeted respiratory compliance among ED HCWs. Similarly, Parmeggiani et al. [35] surveyed ED staff with a 55.8% response rate out of the 550 questionnaires distributed. Reardon et al. [36] evaluated the use of ABHR and Mahfoozpour et al. [33] assessed preventive measures against blood-borne disease, but without providing details of HHO observed.

Hand hygiene compliance

Overall, six studies [25, 27, 35, 37,38,39]  documented baseline HHC rates, with an overall median of 14% (range 5.1–21%). Typically, nurses displayed higher averages (median 17%, range 3.5–31.4%) than physicians (median 11.5%, range 2.9–26%). Of note, post-intervention, all studies reported an improvement in HHC rate to an overall median of 45%, ranging between 8% and 89.7% (Table 9 in the Appendix); again nurses improved HHC most (median 46%, range 8–100%) than physicians (median 40%, range 5.8–91.9%). Of these studies, three adopted multimodal approaches successfully [25, 37, 39]. However, sustainability of improvements is questioned by di Martino et al. [27], who appraised compliance in the ED a year following intervention, and noted disempowerment from initial high levels achieved.

From a microbiology perspective, it was notable that some groups investigated microbial burden on hands as surrogate markers of HHC [28, 29]. Specifically, Espinoza et al. [28] found higher bacterial loads on nursing assistants’ hands (545 CFU/g, IQR 30–2300) than on interns’ hands (nurses or medical) (335 CFU/g, IQR 60–785), while nurses, physicians, and residents yielded significantly lower levels (Table 8). Alternative approaches assessed computer accessories, with Hong et al. [32] reporting contamination of 92% of equipment culture with significant levels of culturable species, including potentially pathogenic isolates from 50% of samples.

Impact of overcrowding and workload

Overcrowding and high numbers of transiting patients are features of EDs. In that context, studies focused on influence of overcrowding on HHC reported direct relationship between overcrowding and poor HHC [16, 19, 26[. Of these, Cartel et al. [26] observed poor HHC elevated by 63% during overcrowding (OR = 0.63, 95% CI: 0.46–0.86) (Table 8), while Muller et al. [16] ascertained a direct correlation between overcrowding and poor soap and water HHC (33%) or ABHR (66%). (Mean = 29%).

Mahfoozpour et al. [33], when assessing the rate of adherence to preventive measures (PM) amongst 80 EM residents (EMR), determined poor behaviour as being related to workload and the need for speed when performing ED clinical duties [33] (Table 8). Interestingly then, Ghazali et al. [30] employed simulation-based training (SBT) to estimate the duration and the quality of HH before and after the SBT, reporting an increase of the duration of HH from 31.2 s (± 13.6 s) at baseline to 35.8 s (± 16.6 s, P = 0.04) post-SBT, probably reflecting rushed practices in real-world scenarios settings. Indeed, Reardon et al. [36] quantified the time burden of ABHR linked with the use of non-sterile glove by ED HCWs, determining that only 3% (95% CI, 0–9%) of the participants adhered to WHO recommendations of 20 s of hand rubbing.

Evaluating with technological aids

Assessment of HHC using technological tools Is not new but remains uncommon. In the ED setting, Haac et al. [31] used video surveillance in resuscitation bays to appraise HHC, while Stackelroth et al. [40] used 24 h-period video surveillance of ED HCWs staff and non-ED staff (EMS personnel) (Table 8). Perhaps most usefully, Strauch et al. [41] introduced electronic badges to remind staff of maintaining high standards of HHC. Participants achieved a 94% HHC rate.

WHO 5 moments

Despite ubiquitous information regarding the WHO 5 moments, Zottele et al. [15] analysed ED HHC based on the WHO 5-moments of HH and noticed a profoundly low rate (54.2%), with nurses demonstrating a higher compliance rate than physicians (66.6% vs 41.3%, OR = 2.83, 95% CI, 1.09–7.34). Other studies [16, 25, 29,30,31] detailed HHC rates for the distinct moments, with moment-1 (before touching a patient) ranging between 3% [31] and 36% [30], moment-2 (before clean/aseptic procedures) between 0% [31] and 25.5% [29], moment-3 (after body fluid exposure) between 2% [30] and 26% [16], moment-4 (after touching a patient) between 15% [31] and 31.6% [25] and moment-5 between 2% and 18.5% [29].

Discussion

While there have been previous reviews of hand hygiene practices, and associated interventions for improvement, across healthcare settings generally, there has been comparatively little emphasis placed on emergency facilities specifically. In that context, this review focuses on emergency departments, with particular discussion of the hand hygiene challenges that may be encountered due to large numbers of patients transiting services, often requiring urgent care that is frequently delivered in overcrowded spaces. This review provides a comprehensive systematic appraisal of published studies of hand hygiene and interventions relevant to this setting in particular.

Hand hygiene compliance

Twelve of the twenty included studies reported participant HHC. In the nine observational studies in which HHC was documented, we calculated a median post-intervention HHC rate of 45% (range 8–89.7%) (Table 9 in the Appendix). This rate correlated with those detailed by Cartel et al. (7.7 to 89.7% in a review of studies published between 2002 and 2012) [2] and Seo et al. (7 to 89% in 12 cross-sectional studies identified in a search of literature published between 1948 and 2018) [19]. Encouragingly, we determined a slightly higher HHC rate than 40% HHC described by Erasmus et al. [43] in 2010. It is, perhaps, reasonable to suggest that this modest increase in HHC rate can be attributed to awareness of the WHO recommended 5-moments of hand hygiene in the last decade promoted through education and training campaigns.

Critical appraisal of the observational studies eligible for our review highlighted complexity in both data extraction and comparison across publications. Specifically, while eight of the nine studies defined HHC according to the WHO recommended 5-moments, Muller et al. [16] combined moments 4 and 5 as a single moment. Further variation in methodology related to choice of time points when HHC was to be determined. For instance, some studies observed HHC immediately after intervention [16, 25, 26, 29, 37], while others delayed observation from six weeks [27, 38] to a year post-intervention [27]. There was similar diversity across with regard to whether HHC was monitored before and after a patient contact [33, 34] or before and after gloving [36] or before and after entering the patient room [41]. However, the greatest disparities related to HH techniques employed, participant educational attainment, cohort sample sizes, and ED level of activity during period of observation (especially staffing level and workload, and overcrowding). To exemplify the impact of such variation, in an African study published in 2014, Schmitz et al. [39] described a post-interventional average HHC rate of 24.8%, representing a HHC rate of almost twice the 14.9% average detailed five years later by Engdaw et al. [44] in the same country. However, the two studies are not comparable directly as, unlike the participant cohort studied by Schmitz et al. based in academic institutions in Addis Ababa [39], that of Engdaw et al. worked at a district level in primary health care centres [44].

Despite a paucity of analogous studies, there is an evident association of ED overcrowding with poor HHC [15, 16, 26].

Hand hygiene interventions

Like others, we observed the most effective interventions ED HCWs to be education and training either alone or complemented by posters, feedbacks, presentations, live demonstration, simulation, or video surveillance. Exemplifying this, Ward et al. [45] found that interventions such as reminder sounds, practical simulations, videos, and audio-visual media significantly improved handwashing compliance. Seo et al. [20] affirmed that the use of multimodal tactics was an effective way to promote HHC among ED HCWs, while Sakihama et al. [37] commented favourably on addition of contests successfully improved HHC. The latter approach was similarly highlighted by Luangasanatip et al. [46] in their study of comparative efficacity of HH interventions where they introduced reward incentives and accountability metrics.

Irrespective of the potential of such approaches, ED HHC rates remain relatively poor. Interventions may be hampered by ED overcrowding [16, 26] or the amount of time it takes to follow the WHO steps while dealing with a real emergency that may be perceived as barriers. It appears that further robust studies, preferably randomized controlled trials or interrupted time series, are required to draw definitive conclusions as to which interventions may be appropriate and effective in influencing HHC in ED settings.

ED healthcare workers

Across the twenty studies eligible for our review, 19 involved ED nurses, with 16 studies focused on physicians (80%). Of note, nurses demonstrated higher HHC compared to physicians and other personnel with an overall mean rate of approximately 46% (8–100%), exceeding that of physicians at around 40%. Proportionally, this is similar to the findings of Erasmus et al. [43] who observed physician HHC to be lower (32%) than nurses (48%).

These observations, however, are challenging as, unfortunately, the definition of physician and nurse have not been consistent across studies. For example, the term “physicians” may represent residents, consultants, general Doctors, or trainees; yet the denomination "physician" was applied in multiple studies without clarity regarding actual clinical grade. Five studies [25, 27, 30, 33, 36] provided data relating to differentiated ED HCWs, while two studies presented data attributable to all ED HCWs collectively [35, 39]. It seems reasonable to suggest that studies ought to refine their design to better understand the impact of experience and clinical responsibility across physician and nursing staff. Similar attention to detail should be applied to investigation of all other staff, clinical, allied health professionals or support staff (porters, hygiene or catering staff) and their contribution to ED HHC and HCAI rates.

Geographical region

Employing World Bank region and income classification [24], we found that of the twenty eligible papers sixteen studies (80%) were conducted in higher-income countries (Europe, North America, East Asia and Pacific and Middle East Asia), while one originated from Ethiopia [39] and three (15%) from upper-middle-income countries (Latin America and the Caribbean and Middle East Asia) [15, 27, 33].

While Africa is home to 17% of the world's population [47], Sub-Saharan EDs reported only one study. Latin America and the Caribbean were represented twice (10%). Asia, with close to 60% of the world population, contributed 25% of the papers. Europe, with 9%, and North America, with 5% of the world population [47] provided 60% of eligible studies. This aligns with a large review conducted by Clancy et al. [3] who commented that 79% of HH studies originated from Asia, Europe, and North America combined.

These evident disparities are, of course, influenced considerably by availability of resources to either perform studies or publish them. However, taken at face value, there are many determinants that may explain the poor HH practices in low- or middle-income countries. Engdaw et al. [44] stated that half of participants had no access to sinks and ABHR, therefore a major factor in poor HHC. Other influencers include lack of basic infrastructure and equipment; limited financial support; inadequate healthcare systems; scarcity of satisfactory HH training; deficiencies in infection control and prevention programs; and poorly developed hand hygiene awareness and attitudes [44].

Training and level of knowledge of ED HCWs

It is noteworthy that a small number of studies attempted to quantify the impact of knowledge and training on ED HHWs’ HHC. Eighteen of the studies referred to training of the involved observers. Amongst them, Parmeggiani et al. [35] utilised a self-reported survey to understand the level of knowledge by ED HCWs and determined that those who use gloves when at direct contact with a patient are 8 times more aware that HH following glove removal is a HCAI control measure; displaying higher compliance rates. In Kuwait, Al-Wazzan et al. [48] concluded that those with a good understanding of HH demonstrated almost seven times more adherence than those with lesser knowledge. Clearly, however, self-reporting is less than accurate and ethical studies utilising anonymous video surveillance are probably more accurate and reflective of actual practice [31, 39, 40, 45, 49], albeit those considerable resources are required to perform such research.

In conclusion, our results indicate that studies conducted in ED on HHC are essentially observational, weak, and prone to multiple biases. Hence, outcomes lack external validity and are difficult to be generalise. ED overcrowding and associated stressors are major barriers to HCW HHC in the ED. An unpredictable environment may require an adjusted strategy to improve HHC rates in the context of critically ill or polytrauma patients. A prudent recommendation may be to simplify the WHO advice and to emphasise the value of ABHR with gloves in improving HHC. Stating the obvious, there is an urgent need for robust and well-designed research to better approach ED poor HHC rates.