“OK, So you Think It’s All About Work Engagement?” Exploring Precursors and Outcomes of Work Engagement Among Healthcare Professionals

Background: Health services organizations need to understand how they best can achieve important organizational goals such as lowering healthcare professionals’ turnover intentions, increase their job satisfaction and level of service quality provided to patients. This study aims to test whether work engagement of healthcare professionals is a core factor in the achievement of these preferred organizational goals. The study also aims to explore whether work engagement is manageable through the direct impact of organizational culture and climate and indirectly for the accomplishment of organizational goals. Methods: The proposed conceptual model was tested in a quantitative study where healthcare professionals, in this study represented by the group of hospital nurses, participated. The data were analyzed using conrmatory factor analysis and structural equation modeling in Stata. Results: The results can be divided into three sub-results. First, work engagement of health professionals was found to be positively related to service quality of care (b = 0.551) and job satisfaction (b = 0.883). Job satisfaction fully mediates the relationship between work engagement and turnover intention and in itself explains almost 60% (R 2 = 0.59) of turnover intentions. Second, health professionals’ perception of organizational culture (b = 0.278) and collaboration climate (b = 0.331) were both directly related to their work engagement. Third, work engagement fully mediates the relationship between organizational culture, organizational climate and service quality of care and job satisfaction. Moreover, work engagement particularly mediates the relationship between collaborative climate and job satisfaction. Conclusions: This study contributes to extending and deepening previous research on work engagement in health services research. Specically, it reveals the essential role that work engagement of healthcare professionals plays for the achievement of organizational goals. Consequently, leaders and managers of healthcare organizations should have a serious focus on health professionals’ work engagement and put it on their meeting agenda regularly because it is clearly a core driver to enabling multiple desirable outcomes for healthcare organizations.

represent the group of health professionals in this study. Research and reports from several countries reveal that healthcare organizations experience an increased level of turnover among nurses [4] and characterize this challenge as an "ongoing problem" [4,1180]. Because of this high level of turnover, organizations experience several negative "domino effects" both externally and internally. Externally, turnover is associated with problems such as lower level of service quality of care provided to patients [5], lower patient satisfaction [6] and having a negative effect on patient outcomes [7]. Internally, turnover can harm areas such as productivity [8] and detrimentally in uence the work morale, which then in the next round leads to an additional increase in turnover [9].
When considering the abovementioned discussion about the critical role nurses, as healthcare professionals, play for healthcare organizations two conclusions can be drawn. First, bearing in mind that hospital nurses are working "face to face" with patients, hospitals should "especially prioritize focus on them" [2, p. 2]. Second, it becomes essential for hospital organizations to identify factors that have the potential to have a positive impact on nurses' perception of their work role and simultaneously are able to decrease problems such as nurses' turnover, improve job satisfaction and increase the service quality of care delivered to patients. This paper is centered on these aspects. It emphasizes whether the work engagement of health professionals (in this study represented by the group of hospital nurses) has a central role and consequently, something healthcare organizations should especially prioritize and put emphasis on. Keyko et al. support such an idea and focus by stating, "work engagement in professional nursing practice is critically important to consider when addressing key challenges of health systems" [14, p. 142]. In a similar way, Bargagliotti stated, "work engagement is the central issue for 21st century professionals and speci cally for registered nurses" [22, p. 1414].
In general, work engagement has for more than a decade been proposed as an important factor for organizations to focus on. For example, Lockwood noted, "Employee engagement has emerged as a critical driver of business success in today's competitive marketplace" [10, p. 2]. Research across different academic disciplines has linked work engagement to positive outcomes such as job performance [11], economic bene ts [12], employee commitment [13] and innovative behavior [15]. Also within healthcare research, with a special focus on healthcare professionals such as nurses, previous research has revealed that work engagement is linked to a variety of positive outcomes. Keyko et al. [14] reported an extensive systematic review of published studies on work engagement within the domain of professional nurses.
The ndings from this extensive review identi ed 77 in uencing factors and 17 outcomes linked to work engagement. In their paper, the authors developed a model to assist in pinpointing knowledge gaps in existing research to motivate and give direction to future research on work engagement. The extensive review by Keyko et al. inspired and motivated the speci c aims of the present study.
Four aims and their associated contributions regarding work engagement are considered in this study. First, Keyko et al. note a lack of "research on patient-related outcomes of professional nurses' work engagement" [14, p. 161]. This study follows this explicit recommendation by examining service quality of care provided to patients as one type of outcome of nurses' work engagement. By doing this, it contributes as Keyko et al. also noted to "provide further support for the importance and value of promoting work engagement in professional nursing practice" [14, p. 161]. Second, based on their review, regarding research identifying antecedents and outcomes related to nurse engagement, Keyko et al. also noted that "North American studies currently dominate the research in this eld." Consequently, they recommend performing more studies in other populations and settings. Based on these authors' suggestion, nurses' work engagement in this study is examined in a Norwegian hospital context and thus representing the Scandinavian population and setting. As such, it contributes to "analysis and discussion on potential geographical and cultural differences" [14, p. 161]. Third, although Keyko et al. identi ed 77 in uencing factors to nurses' work engagement, the ndings reveal that no study included in the review has examined how organizational culture is linked to nurses' work engagement. This is surprising because the "culture of an organization provides boundaries and guidelines that help members of the organization to know the correct way to behave towards each other, and how to perform their work tasks" [16, p. 2711]. According to Slåtten et al. [2,17], healthcare organizations need to foster a culture and climate that focuses on frontline employees. Recently, internal market-oriented culture (IMOC) has been suggested as one type of culture that healthcare organizations should develop regarding their frontline employees [2,17]. Based on IMOC as one type of culture orientation, Slåtten et al. suggest that "future research could examine whether and how IMOC is connected to other concepts such as … employees' individual and collective engagement" [17, p. 177]. This study follows this recommendation and accordingly contributes to revealing the linkage between IMOC and individual nurse's work engagement.
Fourth, in their review, Keyko et al. also requested that future research on nurses' work engagement should include "more complex statistical analyses … to improve the strength of study ndings" [14, p. 161]. This study aims to strive to follow this recommendation by examining both precursors and outcomes of nurses' work engagement in the same research model and thus offer an opportunity to perform more complex and complete statistical tests of both direct as well as mediating effects related to nurses' work engagement.
Considering the abovementioned four aims, this paper contributes to extending and deepening previous research on nurses' work engagement and thus contributes both theoretically as well as to suggest practical implications to "enable development of initiatives that enhance work engagement and its outcomes within the current health care context" [14, p. 143].
This paper is structured in the following way. First, it starts by presenting the conceptual model of the study. Then, the concepts of nurses' work engagement are described and de ned as well as other concepts proposed to be precursors and outcomes of work engagement. Then, the methodology and ndings from our empirical study are presented. The paper ends with a discussion. This latter part also includes suggestions for future research as well as conclusions of the study.
Conceptual Model Figure 1 visualizes the proposed conceptual model of this study. As seen at the foot of Fig. 1, the elements in the conceptual model are grouped into three different categories, Appraisal-Response-Outcome (ARO) that are connected in a speci c directional manner as shown by the arrows between them. On the top of the gure, each of the three categories is labeled with names re ecting the content and focus in each category. The category Response (R) is named "individual-level resource" and is represented as the most central concept in this study, that is, work engagement (WE). Furthermore, the category Appraisal (A) is named as "Organizational-level resources" and is represented by the concepts (i) internal market-oriented culture (IMOC) and (ii) collaboration climate (CC). It is notable to recognize that the term "resource" is used to label both organizational-and individual-level types of resources. The term "resource" signals something that is not static but dynamic and consequently varies, either positively or negatively, as time passes. Further, because a "resource" is dynamic, it also implies that a "resource" is potentially changeable or manageable. As seen in Fig. 1  The concept of WE originally comes from positive psychology, which focuses on the positive resources and strengths of humans, rather than their limits [18]. Speci cally, in this study WE is de ned as "positive ful lling, work-related state of mind" [19, p. 74]. Moreover, this positive work-related state of mind embraces three interconnected types of feelings, namely, absorption, vigor and dedication. Absorption is about employees' level of interest and engrossment or immersion of employees in their work so that the sense of time is lost [20]. Vigor re ects the workers' level of energy and mental resilience while performing their work. Dedication describes how involved employees are in the sense of signi cance, inspiration and challenge [20]. It is important to note that work engagement is a "… state of mind" [19, p. 74]. De ning the WE as a "state of mind" and not a "trait of a person" means that WE, compared with the relative stability of a trait of a person, is dynamic. This dynamic feature implies that all "ingredients" of the internal resource pool of individual employees' WE (referring to absorption, vigor and dedication) are potentially changeable and something that can uctuate, both positively and negatively within a certain timeframe.
As indicated in the previous discussion and visualized in Fig. 1, WE is a response (R) and an individuallevel resource. WE is heterogeneously distributed among individual employees in organizations. WE is a variable that can be described as a motivational-like variable with the potential to positively "improve … organizational outcomes" [21, p. 2]. In the next section, three outcomes of WE are proposed.
The outcomes of work engagement (WE) "It is pivotal to study which job resources can enhance … organizational outcomes" [21, p. 2]. WE is, as visualized in Fig. 1, an individual-level resource that is suggested to be linked to three different types of outcomes, namely, (i) turnover intentions (TI), (ii) job satisfaction (JS) and (iii) service quality of care (SQC). All outcomes of WE represent desirable organizational goals that healthcare organizations want to achieve.

Turnover intentions (TI)
The concept of TI is "particularly useful in nursing studies because of their effectiveness in predicting actual turnover" [23, p. 191]. In this study, TI refers to nurses' psychological response to conditions in organizations [24] and consequently their re ection about whether they should stay with the organization by which they are employed. As such, TI re ects an important "cognitive step in the decision-making process" [25, p. 23] about leaving the organization.
Previous research has suggested that WE is negatively related to TI. When employees experience a positive state of mind characterized of vigor, dedication and absorption, referring to those elements embraced by the concept WE, this should lead to a more favorable evaluation and thoughts of their workplace and thus be associated with decisions or intention to not leave the organization in which they are employed. Previous research on nurses supports the existence of a negative relationship between WE and turnover intention. For example, in their study, Moloney et al. [26], including 2876 nurses working in New Zealand as respondents, the ndings reveal that nurses' WE is directly negatively linked to the intention to leave the organization. Rodwell et al. [23] also found in their study of 459 nurses across Australia a directly negative linkage between nurses' WE and nurses' intention to leave the organization. Similar to the two aforementioned research studies, Wan et al. [8] found, when performing a study among 778 nurses employed in a hospital in China, a signi cant negative relationship between nurses' WE and TI. In line with the aforementioned research ndings, it is expected to nd a similar negative relationship between nurses' WE and TI among those nurses included in this study (referring to nurses from Norway).
Therefore, the following hypothesis is proposed: Work engagement (WE) is negatively related to turnover intentions (TI)

Job satisfaction (JS)
According to Tomietto et al. [21], job satisfaction (JS) requires "strategies … to enhance job satisfaction … one of the most challenging topics in nursing research" [21, p. 2]. In this study, WE is suggested to be one such strategy to achieve JS. Therefore, as seen in Fig. 1, JS is proposed to be one outcome of nurses' WE.
In this study JS is about the "extent to which employees [nurses] like their jobs" [28, p. 13]. JS is a cognitive concept regarding nurses' job satisfaction. Speci cally, JS refers to nurses' global assessment of their job [27].
Although JS has generally been included in several previous studies on nurses, it seems that relatively few of these have examined the speci c link between WE and JS. For example, in their extensive review of antecedents and effects of work engagement by Keyko et al. [14], the authors only reported four previous studies examining this linkage among professional nurses. Of these four studies, two studies identi ed a signi cant relationship between WE and JS. One study found a nonsigni cant relationship between WE and JS. One study was qualitative. Consequently, there is a need for more research on this topic that speci cally focuses on the domain of nursing. In general, it seems that previous research most often has identi ed a positive relationship between WE and JS [29]. It is reasonable to assume a similar pattern and relationship between these concepts among nurses. Both WE and JS are work-related concepts. As noted in the previous discussion, WE is de ned as a motivational concept characterized by feelings of vigor, dedication and absorption in employees' work role. Consequently, when positively experienced, these three "ingredients" of WE should lead employees to have more favorable thoughts and feelings about their job. It is di cult to imagine a situation where an employee (e.g., a nurse) has a high level of WE and at the same time dislikes/hates his/her job and thus has low JS. Clearly, there is probably variation in the level of both WE and JS among nurses. However, in line with previous research [29] on this topic, this study assumes that professional nurses' WE is positively related to their JS. This reasoning leads to the second hypothesis in this study: Hypothesis a: Work engagement (WE) is positively related to job satisfaction (JS).
Previous research has found that the JS of nurses is directly linked to TI among nurses [21]. However, in this study, it is also suggested that JS functions as a mediator between WE and TI. Consequently, this represents a complementary "route" to TI compared with what was proposed in hypothesis one. In addition, to have a direct impact on TI, it is also assumed that there is some inner mechanism in humans that explains why there exists a relationship between WE and TI. In this study, this inner mechanism in humans refers to JS. JS does not arise by itself or occur in isolation. JS is always triggered or created by someone or something. In this study, "something" refers to nurses' WE. As discussed in the second hypothesis, previous research on WE suggested its impact on JS [14,29]. Based on this, when nurses experience a positive WE, their JS will be positively affected. Further, when JS increases because of an increase in WE, it is reasonable to assume that this next also will lower the level of TI among nurses. To the authors' knowledge, no previous research has examined JS as a mediator between WE and TI. However, it is interesting to note that in the study by Tomietto et al. [21], the authors actually conceptualize nurses' JS as one potential mediator between nurses' WE and nurses' TI. Nevertheless, the authors did not make any empirical test of whether JS was operating as a mediator. Consequently, this study is the very rst study to test whether JS is operating as a mediator between WE and TI. The aforementioned reasoning can be expressed by this formal hypothesis: Hypothesis b: The relationship between work engagement (WE) and turnover intention (TI) is mediated by job satisfaction (JS).

Service quality of care (SQC)
The best level of SQC to patients is a highly desirable organizational goal for hospital organizations to achieve. Frontline nurses make a considerable contribution to the total "service package" regarding SQC that patients receive or experience during their stay at a hospital. According to Chen et al., this group of frontline workers "tend to have the longest and closest contact with patients" [30, p. 1]. In this study, the concept of SQC refers to nurses' perceptions of the overall quality of services provided. SQC is studied from an employee perspective and not from a patients' or "customer" perspective. Although this is a subjective in contrast to an objective way to capture the content of SQC, it follows from how previous research sometimes has studied service quality both within and outside healthcare and other contexts [2,31]. Moreover, previous research also suggests that there exists a "psychosocial closeness" between providers' and receivers' perception of service quality [32]. Consequently, parallel to how SQC is studied and de ned in previous research, it is assumed in this study that frontline workers are capable of considering whether their level of quality of services is of high or low standard or lies within a zone that "customers generally perceive as acceptable" [33, p. 208].
There are many examples in the literature that employees' WE is positively associated with a variety of types of job performances such as innovative behavior [15], work effectiveness [34], productivity [35], patient satisfaction [35], innovative capability [36], creative strategy generation [36] and market orientation [37] among several other types of performance concepts. In the extensive systematic review of studies on WE by Keyko et al. [14], the authors identi ed seven different types of performance and care outcomes associated with WE. Among this group, two outcomes of WE, referred to in the paper as perceived care quality [38] and quality care [35], are relatively similar to the concept of SQC in this study. In both cases, WE was positively associated with service quality. Similar to previous research, it is expected to nd the internal resource pool embraced in WE (referring to vigor, dedication and absorption) of an individual as something that positively contributes to frontline employees to work more intensively and goal-oriented to provide excellent service quality to hospital patients. In contrast, a negative WE would have a negative impact on SQC. However, in this study, a positive perspective of the association between WE and SQC is considered. Consequently, it is expected that the more WE increases the more positively it drives SQC. The assumption about the association is formulated by this hypothesis:

Hypothesis 3
Work engagement (WE) is positively related to service quality of care (SQC).

Precursors to work engagement (WE)
In Fig. 1, precursors to WE are suggested to be (1) collaboration climate (CC) and (2) internal marketoriented culture (IMOC). Consequently, the discussion in focus is whether the appraisal of CC and IMOC, representing two organizational-level resources, promotes or causes a response in the individual-level resources of employees referred to as WE in Fig. 1. This is discussed in the following.

Collaboration climate (CC)
An organizational climate is about employees' shared perceptions of their organization. This shared perception of climate varies from strongly negative to strongly positive. The literature emphasizes the signi cance and value for organizations to focus on organizational climate. For example, in the research undertaken by Kieft et al., the authors found that "it is important to develop and maintain collaborative working relationships with professionals, including those in their own eld" [39, p. 5]. Consequently, there are good reasons to include collaborative climate (CC) and how frontline employees perceive and appraise this organizational-level resource. However, in an organization, there are a variety of alternative aspects to focus on when studying a collaborative climate. Therefore, it is necessary to identify aspects that are both relevant and speci c for the phenomenon in focus. Consequently, "climate is best regarded as a speci c construct having a referent" [40, p. 1278]. The literature suggests several ways to study organizational climate. For example, D'Amour et al. suggest four ways to analyze cooperation in healthcare organizations [41]. In this study, CC is about two work-related concerns regarding interdepartmental collaboration in the health organization. Speci cally, CC refers to frontline employees' perception of interdepartmental (i) con ict and (ii) connectedness, which both are suggested in previous research [42,43]. The rst climatic aspect of "con ict" of CC is about whether there exist tensions, caused by inconsistency regarding actual and desired responses between departments in the organization. The second climatic aspect, "connectedness," focuses on whether there are formal and informal contacts across departments in the organization. This latter aspect intends to reveal frontline employees' perceptions of whether departments are operating dependently or independently from each other. The two aspects of CC represent and re ect frontline employees' perceptions of the supportive work environment or what alternatively can also be labeled as the internal service climate in an organization [40].
According to the job demands-resources (JD-R) model, WE is fostered by job resources [44]. In this study, CC represents this organizational level of resource in the JD-R model that promotes or fosters WE among frontline employees. According to Wan et al., "a supportive work environment … offers various resources to foster employees' willingness to dedicate their efforts and abilities to job tasks" [8, p. 1334]. Although it seems that no study has examined the link between the (work environmental factor) CC and WE speci cally, previous research has found that the work environment, in general, is positively associated with WE. For example, the work environment has been found to be associated with WE among nurses [45]. Furthermore, research has found that perceptions of organizational climate are linked to employees' attitudes [46]. Consequently, based on the JD-R model and ndings in previous research there are good reasons to expect CC to be positively associated with WE. Therefore, the following hypothesis is proposed: Collaboration climate (CC) is positively related to work engagement (WE).
Internal market-oriented culture (IMOC) In Fig. 1, internal market-oriented culture (IMOC) is considered as an organizational-level resource.
Organizational culture is, according to Banaszak-Holl, said to "pervade all aspects of organizational life" [47, p. 462]. Organizational culture is a "stable element deeply rooted in employees' mentality" [48, p. 585]. Moreover, organizational culture embraces norms that "provide the rules for behavior" [49, p. 2]. Of the different components an organizational culture consists of, norms and behavior are the two most observable components [50]. In this study, organizational culture refers to frontline employees' perception of norm-based behavior regarding the IMOC in the organization. As such, as also noted in Slåtten et al., IMOC "focus on more tangible or visible aspects of organizational culture that frontline employees of hospitals experience or observe daily" [17, p. 160]. The basic idea with the concept of IMOC is to treat employees in organizations as customers. Parallel to external customers, it is important to treat these internal customers (referring to employees) in the best possible way. Consequently, IMOC focuses on employees' perception of whether managers' norm-based behavior in the organization is oriented toward satisfying the needs and wants relevant to employees' working conditions [17]. Three systems constitute the norm-based behavior concept of IMOC: (i) internal-market intelligence generation, (ii) internalintelligence dissemination and (iii) response to internal intelligence [51]. The three systems are closely connected and represent a logical ow of information from system one to system three. The rst system is about management-related activities to collect information regarding the needs and wants of employees. The second system is about management interpretation and understanding of employees' needs and wants. Finally, the third system is about the willingness and capability of the management in an organization to take steps to perform real actions and actively do what is necessary to satisfy the needs and wants of their employees. All systems of IMOC are interconnected. However, system three (referring to "response to internal intelligence") is probably the most visible part employees experience or observe in their day-to-day work and thus most prominently brings IMOC to "life" in the organization.
Organizational culture "strongly in uences employee behaviors" [50, p. 1] and thus is an employeeimpacting instrument to create the desirable and necessary behavior in an organization. Previous research within healthcare organizations has found organization culture associated with such areas as job satisfaction, leadership behavior, turnover intention, organizational attractiveness ( [2], [17], [49]). Regarding this and speci cally referring to IMOC, Slåtten et al. state, "IMOC is a type of organizational culture affecting frontline employees" [17, p. 161]. The core of IMOC is about management ability to satisfy needs and wants with a speci c focus on employees' work role. With this in mind, it is reasonable to assume that when IMOC is perceived by employees as something good, it should positively in uence employees' vigor, absorption and dedication, which all are embraced in the concept of employees' WE. The association between IMOC and WE is also supported in the JD-R model. The JD-R model [44] highlights that different resources in the work environment can promote or act as motivational factors for employees' dedication and efforts (or what this study refers to as WE) to perform work tasks [44]. In this study, IMOC is termed an organizational-level resource. Therefore, based on the JD-R model and previous research, IMOC should have an impact on employees' WE. This reasoning leads to the following hypothesis: Hypothesis a: Internal market-oriented culture (IMOC) is positively related to work engagement (WE).
Although IMOC is expected to be directly associated with WE, it is also assumed that this relationship is mediated through the concept organizational climate, in this study represented as CC. Accordingly, it is proposed that two "routes" exist from IMOC to WE. In the literature, the concept of climate and culture are often suggested to be closely related. Regarding conceptual closeness Carlucci noted, "… culture and climate are similar concepts" [48, p. 585]. Although they at rst hand seem to be rather similar, they diverge. Culture is about relatively stable and deeply rooted norm-based behavior while climate is more "super cial elements such as employees' reactions, opinions and tendencies" [48, p. 585]. Therefore, the climate is a surface manifestation of culture [48]. Based on this it is expected that when employees perceive IMOC in a positive manner this would have a positive impact on the organizational climate of CC. Moreover, when CC increases, because of IMOC, this next should have a positive impact on employees' WE. To the authors' knowledge, no previous research has examined this speci c linkage. However, two arguments support this idea. First, IMOC as de ned in this study focuses on managers' norm-based behavior and their orientation toward satisfying the needs and wants in employees' work role. Consequently, the norm-based behavior of management, manifested in IMOC, models the "correct standard" of organizational climate for all employees such as how CC should be in the organization. As such, and based on social learning theory [52], employees learn appropriate behavior from their managers as signi cant role models. When IMOC and CC are perceived as positive, this should signi cantly increase employees' WE. The second argument supporting a linkage between the three variables can be found in the JD-R model. IMOC and CC are both based on the JD-R model, suggested to be resources. In Fig. 1, they are both labeled as organizational-level resources. However, as indicated in the aforementioned discussion, IMOC is critical for CC in the organization. Thus, the resources of IMOC serve as necessary input or "ingredients" to positively build and increase the CC resources. Consequently, when the (resources in) CC increase because of a positive impact of (the resources in) IMOC, this should next lead to a positive increase in employees' absorption, vigor and dedication which all are embraced in the concept of WE. Thus, IMOC and CC are expected to work in tandem to strengthen employees' WE. The assumption about WE as a mediator between IMOC and CC can be formulated by this formal hypothesis: Hypothesis b: The relationship between internal market-oriented culture (IMOC) and work engagement (WE) is mediated by collaboration climate (CC).

WE as a mediator between IMOC, CC and SQC, TI and JS
As visualized in Fig. 1, WE is suggested to play a role as a mediator between the two organizational-level resources (IMOC and CC) and the three organizational goals (SQC, TI and JS). WE, labeled as an individual-level resource, is expected to play a central key role. Consequently, WE is assumed to function as the mediating mechanism that binds or ties the suggested variables together to a whole.
As noted in the previous discussion, IMOC and CC are closely related concepts. Because of their close relationship and association, it is reasonable to assume that the two concepts should work in the same direction depending on how employees appraise or perceive them. As Trus et al. noted, "culture and climate represent a social context … that constrains and promotes certain behaviors and interactions" [53, p. 55]. This study takes a positive perspective when studying the social context of IMOC and CC. Consequently, rooted in the assumption of the close relationship between IMOC and CC, it is expected to nd both as positive promoters to employees' WE. Moreover, when WE increases, because of employees' positive perception or appraisals of IMOC and CC, this should next also lead to several positive job-related outcomes and thus an achievement of organizational goals such as increased SQC, JS and reduced TI of employees in the organization.
The assumption of the mediating role of WE is summarized in these two nal hypotheses:

Hypothesis 6
Work engagement (WE) functions as a mediator between internal market-oriented culture (IMOC) and a) service quality of care (SQC), b) turnover intention (TI) and job satisfaction (JS).

Hypothesis 7
Work engagement (WE) functions as a mediator between collaboration climate (CC) and a) service quality of care (SQC), b) turnover intention (TI) and job satisfaction (JS).

Aim, design and setting of the study
This study aimed to test whether WE among frontline healthcare professionals is a key to the achievement of the organizational goals JS, TI and SQC. To reveal further whether WE is a key element, the two organizational-level resources IMOC and CC were included to test their direct effect on WE and whether the relationship with JS, TI and SQC is mediated by WE.
Frontline healthcare professionals in this study are represented by the group hospital nurses. Before the study was executed, six public hospitals located in southeast Norway were contacted. Two team members of the research group contacted the director of each hospital. In their meeting with the directors, the two researchers explained in detail the overall aim and focus of the study. At the end of the meeting, the director received an open invitation for the hospital to be included in the research project. Of the six invited hospitals, four hospitals agreed to take part in the research project.
A survey was designed and developed, and the nal questionnaire was then distributed by e-mail to the nursing staff. All invited participants were informed about the background, aim and focus of the study.
Furthermore, the shared information clearly indicated that all participation was voluntary and all information collected in this research project would be handled strictly con dentially. Further, the distributed e-mail provided invitees with information about the time to complete the questionnaire and a telephone number if they had any questions about the study. To collect the data, the software Checkbox was used. One hundred sixty-four questionnaires were received. All returned questionnaires had satisfactory quality and, consequently, all were included as the basis for testing the proposed hypotheses. Table 1 shows the personal characteristics of the study sample. Most responses were received from those frontline employees who worked as nurses or specialist nurses. A large portion of frontline employees had signi cant experience. Of the total sample, 64% had worked in the investigated hospitals for more than 10 years. About 50% worked full-time. Some 35% were under 40 years of age, about 30% were between 41 and 50, and about 35% were older than 50.

Instruments
The process to develop an instrument for this study included several workshops. Participants from both academia and the target group were represented in the workshops. Combining capabilities from these two groups contributed to developing and designing a more appropriate tailor-made survey. Most instruments used were based on previous research. However, the instruments had not been previously used in a healthcare context. As a consequence of this, through the research process to make sure to develop a tailor-made questionnaire, several changes had to be made in the best possible way to adapt and modify instruments to the selected population of participants. The nal questionnaire was pretested on the target group. The pretest resulted in some minor changes. No questionnaire used in the pretest was included in the nal analysis or test of the hypotheses.
In this study, JS refers to frontline employees' global assessment of their job. Items for JS are based on and adapted from Anaza and Rutherford [27]. The concept of IMOC refers to frontline employees' perception of managers' norm-based behavior regarding the IMOC re ected in three aspects: (i) internalmarket intelligence generation, (ii) internal-intelligence dissemination and (iii) response to internal intelligence. The items included in IMOC are based on and adapted from Gounaris [51]. The concept of WE in this study was de ned as "positive ful lling, work-related state of mind" [19, p. 74] and represented three facets of state of mind: (i) absorption, (ii) vigor and (iii) dedication. Items included in WE are based on Schaufeli et al. [54] and modi ed for this study. In this study, the concept of SQC is about frontline employees' perceptions about their overall service quality provided. Items included in SQC are based on Slåtten [55]. The concept of CC refers to frontline employees' perception of two aspects of interdepartmental collaboration: (i) level of con ict and (ii) level of connectedness. These two aspects are suggested in previous research. Items for CC are based on the work of Kohli and Jaworski [42,43] and modi ed for this study. The concept of TI re ects frontline employees' cognitive re ection about whether they should stay with the organization by which they are employed. Items included in TI are based on Boshoff and Allen [56]. All items for each construct were measured using a Likert scale from (1) strongly disagree to (7) strongly agree. Table 2 lists all items included in each construct.

Data analysis
Structural equation modeling was applied to explore the relationships among the constructs. The rst step was to assess the measurement model (consisting of re ective latent constructs); step two tested the structural model. The estimation was done with the "sem" package in Stata [57]. Testing of the mediation hypotheses was conducted using the "medsem" package in Stata [58], following the approach proposed by Zhao et al. [59].
The measurement model was assessed by examining several criteria. Goodness-of-t indices: SRMR, RMSEA, CFI, TFI; indicator reliability (item loading); latent construct reliability (Raykov's reliability coe cient); convergent validity (average variance extracted (AVE, all AVE values should be larger than the squared correlations among the latent constructs)); and discriminant validity. Convergent and discriminant validity make up the construct validity. The "rules of thumb" used in Tables 2 and 3 are based on Mehmetoglu and Jacobsen [57].
The structural model was assessed with the same goodness-of-t measures as the measurement model.
The structural model's path coe cients reported are standardized values, which ranged between − 1 and 1. The closer a path coe cient is to ± 1, the stronger is the relationship. Table 2 shows that the reliability and validity measures were all within the rules of thumb, and Table 3 shows that the goodness-of-t indices were also within the commonly accepted thresholds, indicating that a sound measurement model was established.   Table 3 indicates that the goodness-of-t indices for the structural model were also within the commonly accepted thresholds. Figure 2 presents the standardized path coe cients and their signi cance level. We  Standardized coe cients (*** p < 0.01, ** p < 0.05)   [8,23,26] and JS [14,29]. In contrast to previous research, this study undertakes a more complex test of how WE is potentially associated with TI, suggesting JS as a mediator. Previous research on this is limited because it has only theoretically conceptualized JS as a mediator between WE and TI (e.g., [21]). To the authors' knowledge, no study in health services research has empirically examined whether JS of healthcare professionals mediates the relationship between WE and TI. Consequently, this study contributes to revealing an alternative, nuanced understanding and a more complex explanation, of how WE is linked to TI. Although WE and JS are conceptually different, they share a commonality because both focus on healthcare professionals' perception of aspects of their work role. In contrast, TI is about the healthcare professionals' perception of their employer and organization and whether healthcare professionals evaluate this as a great place to work. Consequently, TI is related to a "decision … process" [25, p. 23]. Based on these distinctions, it is reasonable to assume that JS is closer to and signi cantly more an integral part of healthcare professionals' decision process embraced in TI compared with WE. Therefore, it is natural and more logical, as this study also has revealed, that JS is functioning as a mediator between WE and TI.

Measurement model
Accordingly, JS of healthcare professionals, stemming from the level of WE, is linked to individual healthcare professionals' decisions about whether they should either continue or stop working for an organization they are employed in at present.
As noted previously in this paper, WE as an individual-level resource is considered and described as something that is not static but dynamic. This dynamic aspect implies that a WE of healthcare professionals is potentially changeable or manageable. This study shows that this assumption is correct.
Both IMOC and CC, representing two types of organizational-level resources can positively change or "manage" WE. The ndings also reveal a close association between organizational culture (refer to IMOC) and organizational climate (refer to CC). Simultaneously, it also reveals how organizational culture (IMOC) has an impact on WE through organizational climate (CC professionals on their meeting agenda because it seems to be a core variable and key tool to enable multiple desirable organizational outcomes. In summary, this study has important implications of how to handle one of the biggest challenges today's organizations are facing, mentioned in Kaye and Jordan-Evans, stating, "The challenge today is not just retaining talented people, but fully engaging them, capturing their minds and hearts at each stage of their work lives" [61].

Limitations And Future Research
Although this study contributes to health service research by including several interesting and important precursors and outcomes associated with WE, there is clearly more research that should be undertaken on the WE topic in future research.
First, this study limited its focus on WE of health professionals from an individual-level perspective. While this has been the most conventional way to study WE, other levels could be included. One such approach would be to focus on the WE from an organizational-level perspective and what the literature refers to as collective work engagement. This implies a shift in focus from "I" or individual engagement to "We" and an organizational climate engagement. In general, there is limited research undertaken focusing on collective engagement [62]. To the authors' knowledge, no study has been undertaken within health service research that speci cally has focused aspects associated with health professionals' collective engagement. Research outside health service research, but within knowledge-based rms, has revealed that collective engagement is positively associated with such areas as relationship learning, rms' innovative capability and employee commitment [36]. Given these positive outcomes, it is strongly recommended that future research in health services also includes a focus on collective engagement. One concrete suggestion of study would be, as also suggested in Slåtten et al. [17], to "examine whether and how IMOC is connected to … collective engagement" [17, p. 177]. Such a focus, among several other potential factors, would clearly contribute to broadening our understanding of the concept of WE.
Second, according to Mes n et al., "… a good hospital culture leads to better individual performance" [60, p. 3]. This study limits its focus to study organizational culture represented by the concept of IMOC.
Future research could include other potential types of organizational culture. One concrete suggestion would be to include one or more of the four organizational culture typologies (that is, 1) Clan, 2) Adhocracy, 3) Market and 4) Hierarchy) mentioned in Cameron and Quinn, known as "the competing value framework" [63]. Studying these organizational culture typologies and their association with potential antecedents or outcome variables contributes to extending our understanding about what type of culture should be put in place to develop and manage employee WE for successful achievement of organizational goals in healthcare organizations.
Third, this study only includes the concepts of culture and climate as drivers of WE. Consequently, examining the impact of leadership is left out and therefore represents a limitation of the study. Although one could argue that both IMOC and CC are affected by leadership, future research should clearly include leadership as a precursor to WE. The literature provides a range of potential leadership perspectives and styles that could be recommended to be included such as transformational leadership, transactional leadership, empowering leadership and leadership autonomy support. However, one concrete suggestion would be to include the four leadership styles mentioned in the Path-Goal Theory Conceptual Framework (that is, 1) Directive, 2) Achievement-oriented, 3) Participative and 4) Supportive) [64]. Similar to the suggestion including other types of organizational culture, examining a range of leadership styles would extend our insight of "what to do" in the best possible way to manage health professionals' WE for the achievement of organizational goals.

Conclusions
This study examines whether work engagement of healthcare professionals is a core factor in the achievement of the preferred organizational goals in healthcare organizations. Thus, it contributes to extending and deepening previous research on work engagement in health services research. The study demonstrates strongly that healthcare organizations should have a clear focus on WE of their health professionals and continuously put WE on their agenda. The "goodness" of cultivating WE of health professionals is having more satis ed employees, a higher level of service quality provided to patients and not least it signi cantly lowers health professionals' turnover intentions.

Consent for publication
Not applicable

Competing interests
The authors declare that they have no competing interests.

Figure 2
Results of the structural model to analyze the precursors and outcomes of work engagement.