Assessing Patient Safety Culture in Hospital Settings

The current knowledge about patient safety culture (PSC) in the healthcare industry, as well as the research tools that have been used to evaluate PSC in hospitals, is limited. Such a limitation may hamper current efforts to improve patient safety worldwide. This study provides a systematic review of published research on the perception of PSC in hospitals. The research methods used to survey and evaluate PSC in healthcare settings are also explored. A list of academic databases was searched from 2006 to 2020 to form a comprehensive view of PSC’s current applications. The following research instruments have been applied in the past to assess PSC: the Hospital Survey on Patient Safety Culture (HSPSC), the Safety Attitudes Questionnaire (SAQ), the Patient Safety Climate in Health Care Organizations (PSCHO), the Modified Stanford Instrument (MSI-2006), and the Scottish Hospital Safety Questionnaire (SHSQ). Some of the most critical factors that impact the PSC are teamwork and organizational and behavioral learning. Reporting errors and safety awareness, gender and demographics, work experience, and staffing levels have also been identified as essential factors. Therefore, these factors will need to be considered in future work to improve PSC. Finally, the results reveal strong evidence of growing interest among individuals in the healthcare industry to assess hospitals’ general patient safety culture.


Introduction
According to the World Health Organization, patient safety (PS) is about preventing medical errors and their adverse effects on patients during healthcare delivery [1][2][3]. Unsafe medical practices can lead to patient injury, death, or disability [4]. The proliferation of such incidents has led to the recognition of the need to improve patient safety culture (PSC) in the healthcare industry worldwide. Furthermore, patient safety has been considered as one of the strategic components of healthcare management [5]. Kohn et al. [6] argued that safety is a crucial and fundamental aspect of patient care research. Kohn et al. [6], in a landmark of PS publications, advocate for error prevention and mitigation using a systematic approach to PS management. Therefore, to ensure the highest level of safety culture in the healthcare industry, it is also essential to understand the beliefs, attitudes, norms, and values of PS and its thresholds [7].
The present study focuses on patient safety culture (PSC) in hospitals. This article's main objective is to discuss the research tools used to assess PSC and identify its essential components. The preferred reporting items for systematic reviews and meta-analyses (PRISMA) were used for this review to ensure reliable results. The PRISMA protocol contains 27 items that aim to analyze and report scientific evidence reliably [8].
This paper is structured as follows: the methodology section explains research questions and research strategy; the results section represents the primary outcomes; the discussion section answers research questions.
The eligibility criteria allowed us to narrow down the subject literature and to identify publications that were relevant to the stated research questions. The articles selected for this study met specific inclusion criteria; namely, these papers (a) were written in English; (b) had been peer-reviewed; (c) identified or described PSC; (d) applied to hospital settings; (e) utilized a survey tool to measure dimensions of PSC among acute care hospital personnel; and (f) applied to general, secondary, tertiary, teaching, or university hospitals. Exclusion criteria included (a) book chapters; (b) papers that, upon review, were found to not be related to the research questions; (c) opinions, viewpoints, anecdotes, letters, and editorials; (d) studies with small sample sizes; and (e) case studies that focused on only one specific hospital unit or sector. Paper titles and abstracts were analyzed based on the stated inclusion and exclusion criteria. Any discrepancies that arose during this phase were resolved through a process of discussion and consensus.
Hawker et al. [9] noted that the quality of any given paper must be assessed against a set of predefined criteria to determine whether it is appropriate for further study. They also proposed that such an appraisal should be performed through the use of appropriate appraising tools. The present study applied the Hawker Assessment Tool, which enables the user to score the quality of papers reviewed. This tool has a uniform assessment form for all types of papers, thereby providing consistency in the evaluation process. One of the assessment factors is the consideration of whether the abstract offers a description of the study. Other factors include the introduction of the paper under review, the paper's aims, background study, and findings. This tool also enables the user to analyze the study's implications concerning the topic under review and indicates how the findings can be converted into policies. A maximum score of 36 [9] was used to assess the quality of potential papers to be included in the present study. The range of the reviewed studies' quality score ranges from a minimum of 9 points to a maximum of 36 points. To create the overall quality grades, we used the following definitions: high quality (A), 30-36 points; medium quality (B), 24-29 points; and low quality (C), 9-24 points. A data extraction template from the Hawker Assessment Tool was used to collect data regarding the properties of the adopted studies. This template allows for a literature analysis with a minimal selection bias [10,11].
Through a search of all relevant databases, a total of 1339 publications were initially identified. The databases searched included CINAHL, MEDLINE, Embase, ProQuest, Google Scholar, PsycINFO, and PubMed. Further analysis was required to eliminate duplicate titles, which resulted in 601 duplicates being discarded. This step was followed by the application of exclusion criteria, as previously described. The abstracts for the remaining 261 titles were read, which led to the selection of 137 relevant articles whose entire texts were analyzed. It should be noted that no additional articles were added after the references from the initially selected papers were examined. Figure 1 provides a flowchart illustrating the article selection process. A total of 66 articles that met all eligibility criteria and that had been published between 2006 and 2020 were selected for the study. To identify research instruments used to study patient safety culture, two researchers (authors) independently read the selected articles' full texts to identify research instruments and their aspects. Subsequently, the two authors compared their findings to develop unified results. Disagreements between the two researchers concerning research instruments and their identified aspects were discussed and resolved in sessions with the third researcher.

Results
All included records were categorized according to objective, strength, limitation, finding and quality score as it is represented in Table 2. Only four hospitals considered for data collection.
• Healthcare professionals have a positive perception of patient safety.

•
Organizational learning was the strongest area in PSC.

•
Professionals with a greater number of employment years were more willing to communicate. • Among respondents, 63.53% stated that they had never reported a case of patient safety.

•
The low rate of reported cases was attributed to fear of the cases being recorded in the respondent's file.

HSPSC, 2012, Saudi Arabia, [13]
Identify general strengths and recognize areas of patient safety improvements                  • TFL has a role in creating a PSC through the actual PSI execution. • TFL has an indirect relationship with the implementation of initiatives, and ultimately improved PSO.

•
The characteristics of inspirational leaders are linked with the creation and promotion of a safety culture, making safety a priority and investing resources to PSI to realize maximal improvements in PSO.

•
There was a noteworthy positive correlation between lower safety climate and higher rates of readmission among AMI (acute myocardial infarction) and HF (heart failure) (p 0.05 for both models).

•
Frontline workers perceptions of safety climate were linked to readmission rates (p 0.01), however, the management's perceptions contradicted this.

•
The results demonstrate that hospital patient safety climate has a connection with readmission outcomes patients with AMI and HF. The associations were specific to management and leadership. Discipline had an impact on the differences at management level: senior managers had less differences than frontline workers. Additionally, the differences were more pronounced for nurses than physicians and other disciplines.
A total of 1,690,225 participants took part in the reviewed studies. The response rate ranged from 17% [53,58] to 100% [39]. However, some studies did not report the response rate [15,28,34,45,46,67]. The study participants included nurses, doctors, and administrators. Figure 2 shows the distribution of participants. Seventeen papers focused on nurses, 38 studies included clinical and non-clinical staff, and 11 studies included clinical staff only.
A total of 1,690,225 participants took part in the reviewed studies. The response rate ranged from 17% [53,58] to 100% [39]. However, some studies did not report the response rate [15,28,34,45,46,67]. The study participants included nurses, doctors, and administrators. Figure 2 shows the distribution of participants. Seventeen papers focused on nurses, 38 studies included clinical and non-clinical staff, and 11 studies included clinical staff only. The reviewed articles reported several limitations concerning the applied methodology and results. First, articles mainly used quantitative approaches to measure PSC, where these methods are not efficient for measuring complex and dynamic attributes such as culture. Second, cross-sectional designs were commonly used among included articles with data collected at one point at a time. Therefore, it is not possible to determine the causal relationships between PSC and the explanatory variables. Third, self-reported questionnaires were applied to collect data, which introduced social desirability biases to the reported research results. Fourth, seven articles did not report their participants' response rate, and 26 articles reported a relatively low response rate (less than 60%). The majority of the reviewed papers concluded that their results could not be generalized because their studies represented unique cultures, the large variations of the applied research instruments, variation in sample sizes, differences in the type of healthcare facilities, and the diversity of study participants.
The global distribution of the included articles is represented in Figure 3. Several studies targeted more than one country.  The reviewed articles reported several limitations concerning the applied methodology and results. First, articles mainly used quantitative approaches to measure PSC, where these methods are not efficient for measuring complex and dynamic attributes such as culture. Second, cross-sectional designs were commonly used among included articles with data collected at one point at a time. Therefore, it is not possible to determine the causal relationships between PSC and the explanatory variables. Third, self-reported questionnaires were applied to collect data, which introduced social desirability biases to the reported research results. Fourth, seven articles did not report their participants' response rate, and 26 articles reported a relatively low response rate (less than 60%). The majority of the reviewed papers concluded that their results could not be generalized because their studies represented unique cultures, the large variations of the applied research instruments, variation in sample sizes, differences in the type of healthcare facilities, and the diversity of study participants.
The global distribution of the included articles is represented in Figure 3. Several studies targeted more than one country.
A total of 1,690,225 participants took part in the reviewed studies. The response rate ranged from 17% [53,58] to 100% [39]. However, some studies did not report the response rate [15,28,34,45,46,67]. The study participants included nurses, doctors, and administrators. Figure 2 shows the distribution of participants. Seventeen papers focused on nurses, 38 studies included clinical and non-clinical staff, and 11 studies included clinical staff only. The reviewed articles reported several limitations concerning the applied methodology and results. First, articles mainly used quantitative approaches to measure PSC, where these methods are not efficient for measuring complex and dynamic attributes such as culture. Second, cross-sectional designs were commonly used among included articles with data collected at one point at a time. Therefore, it is not possible to determine the causal relationships between PSC and the explanatory variables. Third, self-reported questionnaires were applied to collect data, which introduced social desirability biases to the reported research results. Fourth, seven articles did not report their participants' response rate, and 26 articles reported a relatively low response rate (less than 60%). The majority of the reviewed papers concluded that their results could not be generalized because their studies represented unique cultures, the large variations of the applied research instruments, variation in sample sizes, differences in the type of healthcare facilities, and the diversity of study participants.
The global distribution of the included articles is represented in Figure 3. Several studies targeted more than one country.  The map of the co-occurrence of terms in included papers is depicted in Figure 4. The nodes represent specific terms, their sizes indicate their frequency, and links show the co-occurrence of the terms. In the title and abstract of included papers, frequently co-occurring terms created a cluster that appeared with the same color (green, blue, and red color). The three core nodes of these clusters are safety climate, safety culture, and survey. Furthermore, the relationship between the core node of "safety culture" and other high-frequency terms is shown in Figure 5. The thickness of links between nodes represents the strength of the co-occurrence relationships.
Int. J. Environ. Res. Public Health 2021, 18, x 20 of 31 The map of the co-occurrence of terms in included papers is depicted in Figure 4. The nodes represent specific terms, their sizes indicate their frequency, and links show the cooccurrence of the terms. In the title and abstract of included papers, frequently cooccurring terms created a cluster that appeared with the same color (green, blue, and red color). The three core nodes of these clusters are safety climate, safety culture, and survey. Furthermore, the relationship between the core node of "safety culture" and other highfrequency terms is shown in Figure 5. The thickness of links between nodes represents the strength of the co-occurrence relationships.   The map of the co-occurrence of terms in included papers is depicted in Figure 4. The nodes represent specific terms, their sizes indicate their frequency, and links show the cooccurrence of the terms. In the title and abstract of included papers, frequently cooccurring terms created a cluster that appeared with the same color (green, blue, and red color). The three core nodes of these clusters are safety climate, safety culture, and survey. Furthermore, the relationship between the core node of "safety culture" and other highfrequency terms is shown in Figure 5. The thickness of links between nodes represents the strength of the co-occurrence relationships.

Discussion
In this section, two research questions are answered in two subsections of PSC instruments and PSC dimensions.

PSC Instruments
This review identified five primary instruments that have been used to assess PSC in hospital settings. The first instrument, the Hospital Survey on Patient Safety Culture (HSPSC), was used in 54 studies. By contrast, the Safety Attitudes Questionnaire (SAQ) tool was used in five studies, and the Patient Safety Climate in Health Care Organizations (PSCHO) was used in five studies. The Scottish Hospital Safety Questionnaire (SHSQ) and the Modified Stanford Instrument-2006 (MSI-2006) were used by one study each as shown in Table 3.

Hospital Survey on Patient Safety Culture (HSPSC)
In 2004, the AHRQ developed the HSPSC within the United States (U.S.) Department of Health and Human Services, which became a widely used survey. This survey allows for an assessment of staff opinions concerning medical errors, adverse event reporting, and other issues relevant to PS [12,13]. Although the original survey was primarily intended for use by hospitals, it has been enhanced with various versions. This survey currently measures the safety culture of patients in ambulatory settings, outpatient health offices (such as primary care), nursing homes, and public pharmacies. The HSPSC is available in different languages, including Arabic, Spanish, French, and Dutch. The hospital questionnaire version contains 42 items and assesses 12 composites that are treated as subscales.

Safety Attitudes Questionnaire (SAQ)
The SAQ was developed by Sexton and colleagues at the University of Texas in the U.S. The SAQ comprises six main components ( Table 3). The primary advantage of the SAQ is that it can be applied to different healthcare settings. The complete version of the SAQ uses a total of 60 components or items, with 30 items considered as standard across all environments. The survey utilizes a five-point Likert scale ranging from strongly agree to strongly disagree. In addition to the 30 standard items, this survey can incorporate another 6 items, with 3 additional items that focus on demographic studies. The statements utilized by the short-form SAQ can also be addressed using the five-point Likert scale. The short form is easily accessible and available in different languages, including English, Swedish, Dutch, Norwegian, German, Arabic, and Chinese [73].

Patient Safety Climate in Health Care Organizations (PSCHO)
According to Singer et al. [74], PSCHO was designed with the aid of the Stanford Safety Instrument. The PSCHO tool includes 38 items that are used to assess work units, interpersonal factors, and inter-related organizational topics [74]. Using a Likert scale, items are rated via a two-page form. PSCHO is considered to be the first tool that analyzed safety constituents and provided information by measuring the safety climate in corporations outside hospitals. Information from this survey regarding management and clinical personnel can be applied to a wide range of healthcare organizations. PSCHO has undergone psychometrical tests and can be used to compare the performance of several types of hospital units. The earlier form of this tool has been modified with respect to its length [75] and has been adapted for use in multiple languages [61]. The SHSQ was designed for the Scottish NHS clinical staff, with the main aim of gauging the safety outcomes and climate for both patients and staff. The SHSQ includes 4 primary components: 44 items related to the hospital survey (HSPSC), 10 worker safety behavior aspects, 2 items concerned with self-reported patient and worker injuries (see Table 3), and 7 items that focus on demographics [54].

Modified Stanford Instrument-2006
The MSI-2006 Patient Safety Culture in Healthcare Organizations Survey [53] was designed to evaluate 32 unique items encompassing five aspects. These aspects include, but are not limited to, issues associated with seeking help, shame, and self-awareness ( Table 3).
Modification of the MSI-2006 tool has facilitated the assessment of perceptions of a wide range of hospital staff, including direct care workers, technicians, health practitioners, managers, and nurses. This tool also includes assessments of other aspects, such as support service personnel, as these workers are an essential part of the hospital and healthcare setting. MSI-2006 was developed for a wide range of hospital settings with the aim of generating relevant and accurate data over the long term.

PSC Dimensions
To understand the effect of PS on healthcare organizations and their staff, the process and structure of each system needs to be broken into subsystems. The type of instruments and their varying dimensions, as well as the groups targeted in each study, were among the most interesting points to be considered when attempting to understand PS.
Five instruments were used in the reviewed studies to measure PSC within the healthcare facilities examined. As indicated in Table 2, teamwork, organizational and behavioral learning, reporting of errors and safety awareness, gender and demographics, work experience, and staffing levels were perceived as factors that significantly impacted patient safety. Personal variables, such as the age and experience of medical professionals, were also related to PS perceptions. By examining results from individual hospitals or groups of hospitals, we identified the aspects of safety culture that need improvement, including considerations of working conditions and management support.
The reviewed studies differ in their focus on relevant PS variables across different hospitals in various geographical regions. However, many standard components of safety culture indicators and risk factors have been identified [14,15,18].

Teamwork
Teamwork and mutual help provided by team members in task performance within specific hospital units were the factors that represented PS through the use of different instruments [77]. A high score of positive teamwork within units indicates the existence of healthy work relationships and respect among members within a unit [67]. Moreover, vertical hierarchy, horizontal hierarchy, and years of working within a unit influenced the level of teamwork within units. The level of skill competency also affected teamwork within units [57]. However, teamwork across units was reported to have low positive scores [15,21]. Besides, attitudes towards colleagues from different units and managers' or supervisors' actions and expectations towards PS affected teamwork performance across units [18]. According to Hamdan and Saleem [19], skills and organizational learning were significantly related to knowledge teamwork across units. However, supportive managers or supervisors increased the level of teamwork across units. Moreover, colleagues who worked closely together and supported each other in their work duties often resulted in mutual respect [19]. Therefore, while it could be concluded that teamwork is one of the important factors that impact PS, there are always opportunities for improvement.
After reviewing the studies, the HSPSC and SAQ instruments are the only two that are focused on the teamwork dimension. Among the studies that used the SAQ, the pronounced difference in PSC was notable among the front-line healthcare staff, supervisors, and managers [65]. Furthermore, a great variance in PS perception was observed within specific hospital units compared with differences between units. Chakravarty et al. [39] reported low variations in scores between hospitals based on the PS index. However, their study also revealed significant differences in individual measures of PS, including the perception of management, teamwork, and stress recognition, when using the PS index score [39].
The HSPSC provides more details about teamwork performance within and between units of hospitals. Additionally, teamwork is the most factor that has a relationship with the other characteristics of PS. Among studies using the HSPSC, high scores were obtained for teamwork within units, especially in different developing countries [18,30,35,43,45,47,56,67]. These results confirm that the healthcare industry greatly relies on interdisciplinary teams of specialists with the skill sets needed to perform specialized tasks. Such teams also collaborate to achieve common safety goals [40]. Different teams use shared resources and rely on communication to adapt to ever-changing healthcare environments. The behavior of these teams was analyzed through observational studies. The results indicated that the teams' clinical performance was influenced by how they communicated, coordinated, and practiced effective leadership [40].

Organizational and Behavioral Learning
Organizational learning is also a critical factor that affects the PS. In most of the survey studies examined, positive responses were given for organizational learning/continuous improvement as a composite for PS [12,29,31,34]. Continuous improvement can be gained from daily work routines and incidents. PS can also improve by enhancing relevant personnel's skills and knowledge based on incident analysis. Additionally, the junior staff can learn from more experienced staff as they worked together [74].
Although organizational and behavioral learning had positive responses, the outcome dimension, frequency of events reported, did not have positive responses in all the studies included in this review. Therefore, the learning process in PSC should be enhanced by establishing formal methods instead of informal practices to avoid harming patients. In the U.S., as a result of the IOM's report, the U.S. Congress passed the Patient Safety and Quality Improvement Act in 2005, which aimed to improve quality and safety via the collection and analysis of data on patient events. This shows that PS has to be enhanced by the participation of healthcare providers and patients.
In 28 of the studies examined, 55% of the participants agreed that these factors were important components of organizational learning and continuous improvement processes at the examined healthcare facilities. These processes are also responsible for communicating and conveying information that is essential for PS and healthcare. Such processes occur in both formal and informal learning environments within healthcare systems that perform complex and interconnected operations, which should be considered to enhance the PSC.

Reporting of Errors and Safety Awareness
Two of the dimensions that received low positive scores were non-punitive responses to errors and the frequency of event reporting [32]. That is because a large percentage of respondents indicated that they do not report incidents to their managers or supervisors. The reason behind this could be that staff members fear being reprimanded for an error and the lack of safety awareness. Such a culture might cause the staff to hide issues that could later influence the efficacy of PS. A culture that includes non-punitive responses to errors could arise from managers, supervisors, and colleagues [46]. Another reason behind this finding could be the risks of patients complaining; patient demands for compensation might have also reduced the frequency of event reporting [52].
Moreover, another study that was conducted in Saudi Arabia illustrated that one of the dimensions that indicated a high positive response was feedback and communication about errors [24]. The factors requiring improvement included non-punitive responses to error reporting and adequate personnel staffing [24]. The survey showed that the overall perception of PS was 59.9%, while the reporting frequency was 68.8% [24]. Another study that was conducted in Scotland by Agnew et al. [54] found that the overall perception of PS was judged at 56%; the reported frequency of incident reporting was also 56%. Another study in Saudi Arabia showed that the frequency of reporting adverse safety events was 57% [23]. Additionally, A study conducted by Khater et al. [26] among senior nurses in Jordan showed a positive correlation between non-punitive responses to medical errors and the frequency of medical error reporting. The result was a reduction in adverse events regarding PS and risks of complaints from patients. The overall perception of senior nurses was 51.5% before education and 60.6% after educational sessions. The frequency of event reporting increased from 54.2% to 64.3% after implementing suitable educational training [26].
In a related study, Hellings et al. [49] described a PSC improvement approach implemented in five Belgian hospitals. The results showed that management support for PS increased along with supervisor expectations and actions that promoted safety practices. Medical personnel from Dutch-speaking hospitals had a higher positive perception of PS compared with French-speaking hospitals [49]. The survey also showed that respondents working in pediatrics, rehabilitation, and psychiatry departments (units) provided more positive feedback about perceived PSC. By contrast, medical professionals working in emergency departments (units) provided lower positive feedback [49]. These differences in the hospitals' departments and languages are some of the reasons for reporting low scores in the non-punitive responses to errors [49].
A positive perception of PS was observed among medical personnel in China and U.S. managers. In both countries, these individuals expressed a higher level of perceived PS compared with front-line personnel. However, Chinese staff had higher scores for work-related fear of shame and blame compared with their American counterparts [61]. The U.S. hospitals have fewer cases of "fear of blame" compared to Chinese hospitals [61].
As noted earlier, a reduction in avoidable incidents with potential or actual medical harm is a key objective in developing a robust PSC [31,34,36]. Harm can be measured by the frequency of reported events. Effective reporting of safety incidents is essential for identifying the causes of failures in a healthcare work environment. The present analysis indicates a need to implement more effective reporting systems. Reporting provides relevant information about the frequency of events that can adversely affect PS.
A culture of blame was evident in 22 studies, representing 43% of those examined. In these studies, punitive responses to medical errors were prevalent and created a culture that discouraged personnel from reporting safety incidents and occurrences [42]. Such a culture impeded the hospitals' ability to determine the causes of errors and, consequently, to learn from previous mistakes [13,15,17]. In instances in which an influential safety culture exists, workers can highlight potential risk factors and also identify failures when they occur with a focus on PS [38]. Additionally, adverse events arise from multiple unintentional causes. Blame was judged to be appropriate when addressing individuals who consistently commit frequent and careless errors or who ignore established safety standards and policies. Competent institutions should maintain a culture of accountability to ensure that patient care is maintained at the highest levels.
A study conducted in Canada by Zaheer et al. [53] focused on supervisory and senior leadership support for PS. The survey noted that ease in reporting provided the hospital with a platform for learning and improving through reported incidents. Among the supervisory and senior leadership, ease in reporting was recorded at 11% and 12%, respectively. These findings suggest that hospitals should ensure that front-line staff are aware of and contribute to their organization's reporting systems. Ease in reporting should provide organizations with an opportunity to improve strategy, commitment, and the overall efficacy of PSC in sample facilities [53].

Gender and Demographics
PSC is a multidimensional concept that requires a strict analysis to identify its vital elements. The perception of PSC is always measured through the dimensions of the tools used. However, gender and demographic characteristics can be used to analyze participants' responses to a survey [16]. Many of the studies analyzed herein demonstrate the correlation between PSC perception with gender and demographics.
Numerous differences in nurses' perceptions of PSC arose due to demographic characteristics, including gender, age, level of education, years of experience, the language used, and length of work shift [27]. In general, female nurses had a more positive view of the prevalent PSC than did their male counterparts. Moreover, nurses between the ages of 40 and 60 years had a more positive view of the PSC than nurses between 20 and 40 years of age [53]. As 85.4% of the nurses had a Bachelor of Science in nursing, it is plausible that their education levels did not affect their perception of PS [16]. However, as Hamdan et al. [19] observed, education is generally one of the most critical aspects of healthcare delivery to patients worldwide.
Elsous et al. [27] evaluated nurses' perception regarding PSC and investigated the influence of age, hierarchal position, working hours, and experience. Job satisfaction and perception by management concerning PS had a strong influence on these variables. Frontline clinicians had a less positive attitude toward PS than did nurse managers. Moreover, positive attitudes increased with years of experience. Work shift hours and ages of the nurses had a direct effect on the perception of PS. Nurses working within the normal hours allocated per week and aged 35 years or older showed a better PS perception [27]. The study also reported no differences in safety attitude scores between nurses and doctors due to gender, age, and work experience [27]. The studies of the potential effects of gender and demographics on the perception of PSC should be expanded in the future.

Work Experience
Relevant work experience was strongly related to the perception of the PSC. Work experience was also associated with the perceived quality of care among nurses [19]. Furthermore, more experienced healthcare providers had a better understanding of patient care needs than did less experienced nurses [53]. A study conducted in the U.S. by Hansen et al. [76] investigated the relationship between hospital PSC and rehospitalization rates within 30 days of discharge. A survey done in 67 hospitals discovered that higher readmission rates of acute myocardial infarction and heart failure patients were directly related to a lower safety climate [76]. Additionally, frontline staff workers reported a lower level of perceived safety climate with the readmissions, which were the management's responsibility [76]. In another study, a survey was conducted in 97 hospitals in the U.S. that revealed that frontline workers perceived a climate of safety more frequently than did the managers and the supervisors [77]. Furthermore, among the clinicians, nurses perceived a safety climate more than physicians [77]. Based on that, it could be concluded that the work environment plays a key role in perceiving the PSC.
Moreover, another study illustrated that language also has effects on perceiving the PSC [16]. Non-Arabic-speaking nurses had more positive views of PSC than did Arabicspeaking nurses [16]. This finding was unanticipated as the Arabic-speaking nurses and their patients spoke the same language. The low PSC scores might have been due to disparities in educational systems affecting PS perceptions. Furthermore, nurses working on day shifts had more positive PSC perceptions than nurses working night shifts or alternating shifts [16]. It was noted that day-shift nurses were more time engaging with and involved in their patients' progress, which resulted in a positive PSC [16]. Day-shift nurses also interacted with their managers and became more familiar with relevant aspects of the PSC [16]. Therefore, it could be concluded that work experience and the possibility of knowledge exchange had a measurable impact on perceptions related to the PSC.

Staffing
The availability of human resources also impacts the perceptions of the PSC. A study conducted in Scotland by Agnew et al. [54] aimed to analyze the relationship between the medical personnel safety behavior and reported injury measures for patients and healthcare providers. At the hospital level, the authors found a strong correlation between overall PS scores and patient and personnel injury measures and behavior [54]. Therefore, the level of hospital staffing, coupled with management support for PS, also influenced the perception of PS within the studied facilities [54]. Generic safety climate factors and patient-specific items showed a strong correlation with perceived safety outcomes [54]. To summarize, a total of 24 studies reported on the issue of healthcare personnel understaffing. The staff reported feelings of being overburdened and overloaded with their daily responsibilities in approximately half of the hospitals [18,37,47,48,60,64,66]. Consequently, this issue had a negative impact on the quality of care provided by the staff [45]. Therefore, the availability of adequate staffing plays a critical role in perceiving the PSC because employees' focuses might be harmed due to overload.

Study Limitations
The present study has some important limitations. This systematic review focused only on articles written in English; moreover, a meta-analysis was not performed. The results of the reviewed studies are difficult to generalize due to the application of a diverse set of PSC measures with different dimensions. Furthermore, the reviewed studies also varied in the type of participants included (doctors, nurses, and administrators), the periods over which the measurements were conducted, the sampling strategies used, and the cultural settings. For example, the results that focused primarily on results from nurses were obtained from convenience samples of participants and as such cannot be generalized to the entire nursing staff. Finally, this study did not account for language and cultural disparities prevalent in the specific countries in which the reported studies were conducted.

Conclusions
Enhancing the perception of the PSC in health sectors plays a key role in improving their overall quality, efficiency, and productivity. This paper contributes to the body of knowledge related to PSCs by identifying important critical factors and illustrating the instruments that have been developed and used to generate data. A comprehensive review of perceiving the PSC in hospital settings was provided. A systematic literature review was conducted using the PRISMA protocol for the period of 2006 to 2020. The paper reviewed 66 studies that were identified based on carefully selected keywords. The Hawker Assessment Tool was also implemented in this paper to enable the researcher to score the quality of the papers reviewed. The paper analyzed PSC perception in the hospital setting, determined available instruments, and identified the most critical factors that have an impact on the PSC. Our findings revealed that teamwork and organizational and behavioral learning are some of the factors that have a significant impact on the PSC. This paper also illustrated that reporting errors and safety awareness, gender and demographics, work experience, and staffing are additional critical factors that need to be considered further to improve perceptions of PSCs.
In the future, the impact of culture on PS might be analyzed in greater depth. PS, particularly in hospitals, is a dynamic and complex phenomenon. Therefore, it is recommended that research and surveys be performed every two to three years to ensure the best practices for PS. Such an approach could also enhance the quality of healthcare delivery. A large number of hospitals in many different countries have been studied and the specific characteristics of the healthcare management systems in these countries greatly vary. Consequently, for future studies, a broader study population crossing the national boundaries would help to ensure that the findings can have an impact on the development of high-quality, affordable healthcare worldwide.
Finally, it should be pointed out that although the reported survey questionnaires described in the reviewed studies were anonymous, some respondents might not have been candid in providing their answers. Some of the questionnaires were long and some of the respondents may have become distracted during the process, lost interest, or answered some questions inaccurately. Additionally, some inconsistencies in using different survey tools due to cultural and language diversities were noted. For future, investigations including qualitative evaluations of these relationships should be conducted. Finally, the long-term effects of safety incidents on patients' health and their long-term impact on families have not been investigated. Future studies should evaluate the effects of such experiences in hospital settings.