Safety attitudes in hospital emergency departments: a systematic review

Purpose The purpose of this paper is to perform and report a systematic review of published research on patient safety attitudes of health staff employed in hospital emergency departments (EDs). Design/methodology/approach An electronic search was conducted of PsychINFO, ProQuest, MEDLINE, EMBASE, PubMed and CINAHL databases. The review included all studies that focussed on the safety attitudes of professional hospital staff employed in EDs. Findings Overall, the review revealed that the safety attitudes of ED health staff are generally low, especially on teamwork and management support and among nurses when compared to doctors. Conversely, two intervention studies showed the effectiveness of team building interventions on improving the safety attitudes of health staff employed in EDs. Research limitations/implications Six studies met the inclusion criteria, however, most of the studies demonstrated low to moderate methodological quality. Originality/value Teamwork, communication and management support are central to positive safety attitudes. Teamwork training can improve safety attitudes. Given that EDs are the “front-line” of hospital care and patients within EDs are especially vulnerable to medical errors, future research should focus on the safety attitudes of medical staff employed in EDs and its relationship to medical errors.


Introduction
The effective delivery of hospital services and patient care is significantly tied to the safety attitudes and practices of hospital staff and management (Reason, 1993). Indeed, issues related to patient health and safety in hospitals throughout the world have resulted in-patient deaths, prolonged hospitalisations, irreversible disabilities and significant financial costs (Reason, 1993;Abdou and Saber, 2011;Alayed et al., 2014;Allen, 2009;Almutairi et al., 2013;Chaboyer et al., 2013;Duthie, 2006;Profit et al., 2012;Rodriguez-Paz and Dorman, 2008). To address these issues, recent research has focussed on the importance of a hospital safety climate to optimise the effective delivery of patient care. According to The Health Foundation (2011), safety climate focuses on staff perceptions about how safety is managed within their organisation in terms of measurable components. These measurable components include management behaviours, safety systems and employee's safety attitudes (The Health Foundation, 2011).
Measuring safety attitudes among hospital staff has been widely researched and reported in the literature to provide a lens through which to view and improve the patient safety culture in hospitals (Blegen et al., 2005;Bondevik et al., 2014;Carvalho et al., 2015;Sexton et al., 2011;Steyrer et al., 2013;Yaprak and Intepeler, 2015). Indeed, Sexton et al. (2006) maintain that attitudes gauged through surveys of the perceptions of frontline workers within hospitals provide a snapshot of hospital safety culture (Sexton et al., 2006). Safety attitudes have been investigated in a range of countries and different hospital departments. For example, Allen (Allen, 2009) employed the Safety Attitudes Questionnaire (SAQ; Sexton et al., 2006) to establish the safety culture in the maternity services of two Australian hospitals. He found the optimal safety culture was lacking across six safety domains, especially in the domain of management support and working conditions. Moreover, the safety culture was influenced by poor communication when the need for care escalated, lack of supervision of junior staff, issues with staffing, skill mix and low morale.
Along with the significant research focus on safety attitudes within hospital settings, there have been several systematic reviews of findings relating to patient safety attitudes. These have included systematic reviews relating to the safety attitudes of hospital staff in Arab countries (Elmontsri et al., 2017) and hospital in-patient settings (Weaver, 2013). Other systematic reviews have investigated research connecting patient safety attitudes and patient outcomes to determine nurse-sensitive patient outcomes in hospital settings (DiCuccio, 2015), studies on patient safety issues and practices in emergency medical services (Bigham et al., 2012) and studies on patient safety culture strategies to improve the hospital patient safety climate (Morello et al., 2013). Yet, there has been no systematic review of the state of research literature on safety attitudes of health staff employed in hospital emergency departments (Eds). This would appear to be an important issue to clarify, given that EDs are the "front-line" of hospital care (Rigobello et al., 2017) and patients within EDs are especially vulnerable to medical errors (Shaw et al., 2009). The primary objective of this study was to perform a systematic review of published research on the patient safety attitudes of health care professional staff employed in hospital EDs.

Data sources and search strategy
To meet the objective of this study, an electronic literature search was conducted in July 2018 using six different science, health and medicine focussed research databases: PsychINFO, ProQuest, MEDLINE, EMBASE, PubMed and the Cumulative Index to Nursing and Allied Health Literature (CINAHL). No limitations were set on the date of publications; however, search filters were used to limit search hits to publications published in English. The database search strategy entailed initial uses of a broad search term to capture a wide body of studies relevant to the review. Thus, the search process included combinations of the terms "Hospital Emergency department staff", "Patient Safety attitudes", "Safety Culture", "Safety Climate", "Medical Errors" and "Adverse Events" as well as combinations of MeSh terms "Safety Management", "Patient Care Team" and "Attitude of Health Personnel" (Appendix 1).

Inclusion criteria and study selection
Two reviewers performed an assessment of the eligibility of potential studies for inclusion in the review of research on safety attitudes in EDs. All identified records from the aforementioned database searches (total of n ¼ 617) were imported into EndNote citation software where duplicates were first identified and then removed. The 503 remaining titles and abstracts were screened against the predetermined inclusion and exclusion criteria such that studies where attitudes of hospital staff towards patient safety had been assessed and/ or measured were included at this point of the review. Based on this set of criteria, an additional 443 studies were further excluded from the review, leaving a total of 60 eligible research articles. A final inclusion/exclusion criterion was then applied by removing articles where the study setting did not include a hospital ED. From this investigation, a total of 48 papers were excluded from the final review leaving 12 full-text research papers for in-depth analysis and review. Full-text articles were retrieved from an electronic library and examined in detail for the study design, sample, measures and findings. The study selection process is summarised in Figure 1 using the PRISMA flow diagram (Moher et al., 2009).
Records identified through database searching n = 617 Records after duplicates removed n = 503 Records screened n = 503 Records excluded n = 443 Full-text articles assessed for eligibility n = 60 Full-text articles excluded n = 48 Studies included n =12 Figure 1.

PRISMA flow diagram
Data extraction and quality appraisal Data were extracted included study sample and setting, type and number of participants, study design, variables and measurement tools and study findings. The quality of the reviewed articles was assessed through the NIH Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (NIH, 2017) which gives a score out of 14 to indicate the quality of research studies.

Results
In total, 12 studies met the inclusion criteria. The methodological characteristics, measures and findings from these articles are summarised in Table I. The studies covered a wide variety of settings including three studies in the USA, two studies in Sweden, and one study each in Australia, Brazil, Cyprus, Denmark, Iran, China and the Netherlands. Four studies were conducted in a single ED site, two studies were conducted in two sites, and six studies included participants from multiple ED sites (from 5 to 62). Whereas ten of the studies were quantitative cross-sectional designs with survey methods, two studies entailed the use of a qualitative phenomenological methodology with semi-structured interviews. Of the ten cross-sectional studies, two used a repeated measures design whereby participants completed a survey prior to and after a safety quality improvement intervention.

Participants and measures
Across the 12 studies there were a total of 7,645 participants. Most participants were either nurses or physicians working in an ED. In the ten cross-sectional studies, participants completed a validated measure of patient safety culture attitudes, whereas participants in the qualitative studies answered open-ended questions about patient safety attitudes. Three cross-sectional studies also measured the number of adverse patient events to compare against safety attitudes.

Findings
The two studies that used a quantitative repeated measures design (Burstrom et al., 2014;Lisbon et al., 2016) entailed the use of a team building intervention to test the effects of the intervention on the safety attitudes of participants. Together, the interventions had some success because, post-intervention, the safety culture attitudes demonstrated improved teamwork and communication. Nevertheless, the safety attitudes of physicians and nurses from EDs were generally less than positive in both study settings even after the intervention.
A further eight studies were survey-based using cross-sectional designs where ED staff completed different measures of safety attitudes on one occasion (Rigobello et al., 2017;Shaw et al., 2009;Camargo et al., 2012;Lambrou et al., 2015;Rasmussen et al., 2014;Tourani et al., 2015;Verbeek-Van Nord et al., 2014;Wang et al., 2014). In two of these studies, physicians' safety attitudes were reported as more positive than nurses (Shaw et al., 2009;Verbeek-Van Nord et al., 2014) although overall safety attitudes reported in six of the eight cross-sectional studies were generally low, especially on teamwork, in-patient coordination and management support. In contrast, job satisfaction was comparatively high in one study (Rigobello et al., 2017). Moreover, the findings from three studies showed more positive safety attitudes were associated with teamwork, communication and management support (Verbeek-Van Nord et al., 2014), improved management of EDs and the presence of an ED safety committee (Shaw et al., 2009), and leadership and autonomy, control over practice, and cultural sensitivity (Lambrou et al., 2015). Importantly, three studies compared safety attitudes to patient adverse event data (Camargo et al., 2012;Rasmussen et al., 2014;Wang et al., 2014)  The overall rating of safety culture on most dimensions by doctors and nurses at both hospitals and at both measurement points was low (below the midpoint). However, a higher score was measured post-intervention on two dimensions with participants from the country hospital: teamwork within hospital and communication openness.
At the university hospital, a higher score was measured at follow-up for the two dimensions: teamwork across hospital units and teamwork within hospital safety and team climate, poor inter-departmental working relationships, and increased cognitive demands (Tourani et al., 2015). Of the reviewed studies, only two employed a qualitative research design (Grover et al., 2017;Källberg et al., 2017). These studies reported similar findings in that teamwork and team support, workload, and communication and organisational failures were found to be critical to enhanced patient safety.

Quality rating
The quality rating of each study was assessed through the NIH Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (NIH, 2017). According to the rating system, the quantitative intervention studies by Burstrom et al. (2014) and Lisbon et al. (2016) were the highest quality research with a score of 8/14 and 6/14, respectively (Burstrom et al., 2014;Lisbon et al., 2016). The fact that both studies employed an intervention to test the direct effect of an independent variable (IV ) on a dependant variable (DV ) distinguished the quality of these studies from the other studies in the review. Nevertheless, the study by Lisbon et al. (2016) had lower quality because the study population was not clearly defined and over 20 per cent of the participants were lost to follow-up. The quality rating of the eight other quantitative studies in the review (Rigobello et al., 2017;Shaw et al., 2009) was quite low (between 3/14 and 6/14) and reflected the fact that each study employed a cross-sectional survey design with little control over extraneous or intervening variables where only the relationship between the IV and DV could be established. Similarly, the qualitative studies by Grover et al. (2017) and Källberg et al. (2017) were rated low (2/14) because each study did not employ a systematic sampling procedure or use valid and reliable measures (Grover et al., 2017;Källberg et al., 2017). Overall, each of the 12 studies failed to justify the sample size through appropriate use of power analysis and estimates of effect size. Furthermore, only one study (Wang et al., 2014) made an adjustment in analysis to take into account key potential confounding variables such as the gender, profession and years of practice of participants.

Discussion
The primary objective of this study was to perform a systematic review of published research on the patient safety attitudes of health care professional staff employed in hospital EDs. This systematic review of the current literature identified 12 studies, including 10 quantitative studies and 2 qualitative studies that met the inclusion criteria of studies where the safety attitudes of health care professionals from hospital EDs was ascertained. Given the number of studies to have investigated the safety attitudes of the front-line emergency staff of hospitals is comparatively few and patients within hospital EDs are especially vulnerable to medical errors (Shaw et al., 2009), there is justification for addressing the lack of research on the safety attitudes of emergency hospital staff in future studies.
Furthermore, additional research into the safety attitudes of hospital staff is justified because the current systematic review revealed the overall methodological quality of the reviewed studies was comparatively low. Despite some of the reviewed studies having large participant numbers which contribute to the validity of the findings, all the quantitative studies employed cross-sectional research designs which undermines the internal validity of the findings such that it is not possible to observe the direct effects of an IV on a DV. Nevertheless, two higher quality studies employed team building interventions that showed safety culture attitudes improved teamwork and communication post-intervention (Burstrom et al., 2014;Lisbon et al., 2016).
The importance of teamwork and communication to safety attitudes in hospital EDs was also evident in three of the other reviewed quantitative studies. In two of these studies (Shaw et al., 2009;Camargo et al., 2012), more positive safety attitudes were associated with teamwork, communication and management support as well as improved management of EDs and the presence of an ED safety committee. Similarly, one reviewed qualitative research design reported teamwork and team support as critical to enhanced patient safety (Grover et al., 2017). Nevertheless, teamwork and management support are often rated comparatively low on multidimensional safety attitude scales (Chaboyer et al., 2013;Profit et al., 2012;Alzahrani, 2015) such as the studies reviewed here show (Shaw et al., 2009;Verbeek-Van Nord et al., 2014). It would appear from the literature and the review of research reported here that human resource issues like teamwork and management support are related to lower safety attitudes of hospital staff and that interventions to improve these factors in the EDs of hospitals are likely to impact positively on safety attitudes.
The findings from this review that ED physicians' safety attitudes were reported as more positive than nurses (Shaw et al., 2009;Verbeek-Van Nord et al., 2014) is consistent with previous research in other hospital departments. For example, Thomas et al. (2003) reported nurses rated the quality of collaboration and communication with physicians to be lower than the ratings of doctors. As surmised by Thomas, the findings are likely to be associated with differences in status/authority between nurses and physicians, differential responsibilities and training, gender issues, and nursing and physician cultures. Nevertheless, the findings of this review suggest the safety issues associated with the human resource components of a hospital ED are a particular focus for nurses.
Altogether, the findings contribute to the literature by being one of the first studies to systematically review the safety attitudes of health professionals in hospital EDs. Although the numbers of studies on this topic are limited, they do show that teamwork, communication and management support are central to positive safety attitudes, and that teamwork training can improve safety attitudes. Nevertheless, a strength of three of the reviewed studies was an investigation of the relationship between safety attitudes and adverse patient events (Camargo et al., 2012;Rasmussen et al., 2014;Wang et al., 2014) with one study showing the number of adverse events was related to a poor safety and team climate, poor inter-departmental working relationships, and increased cognitive demands (Rasmussen et al., 2014). Yet, the assumed relationship between safety attitudes and hospital error rates has not been clearly and unequivocally shown in the research literature on hospital safety (Steyrer et al., 2013;Ausserhofer et al., 2012). Given that EDs are the "frontline" of hospital care (Rigobello et al., 2017) and ED patients are especially vulnerable to medical errors (Shaw et al., 2009), future research on the safety attitudes of medical staff employed in hospital EDs and how they relate to medical errors is warranted.