A balanced approach involving hard and soft factors for internalizing Lean Management and Six Sigma in hospitals

Purpose – This studyexaminesthe theoretical and empiricalrelationships between LM&SS, humanresource management(HRM),climateforLM&SSandoutcomes(employeewell-beingandperformance)inhospitals.Aspartofthisresearch,theauthorsexaminetheinterplaybetween “ hard ” and “ soft ” practices for LM&SS and “ soft ” HR practices. Design/methodology/approach – A cross-sectional, multisite survey study covering all internal service units at all eight Dutch university hospitals was conducted (42 units, N 5 218 supervisors, N 5 1,668 employees), and multivariate multilevel regression analyses were performed. Findings – A systems approach involving “ soft ” LM&SS practices that are specifically HR-related has a positiveeffect ( β is0.46) ona climatefor LM&SS.A climatefor LM&SSisnot relatedto perceived performance or employeehealth. It is, however,positivelyrelated to employee happiness and trustingrelationships (both β s are 0.33). We did not find that a climate for LM&SS had a mediating effect. Research limitations/implications – This study shows that a balanced approach involving both “ hard ” and “ soft ” factorsiscrucialtoachieving thedesiredbreadthanddepthofLM&SSadoptionatthe macro,meso, and micro levels. The authors found that a climate for LM&SS positively affects employee well-being in hospitals. Practical implications – In their attempt to create mutual gains for both their organization and their employees,hospitalsthatadoptLM&SSshouldfosteraclimateforLM&SSbyembracingabalancedapproachconsistingofboth “ hard ” and “ soft ” practices, thereby internalizing LM&SS at the macro, meso, and micro levels. Originality/value – Thisisoneofthefirststudiestoexaminein-depththeimpactof “ hard ” and “ soft ” LM&SS on both employee well-being (subdivided into different components) and performance in healthcare, as well as the role of “ soft ” HRM in this relationship. Linking LM&SS, HRM and outcomes to a climate for LM&SS is relatively a new approach and has led to a deeper understanding of the mechanisms underpinning the internalization of LM&SS in healthcare.


Introduction
The COVID-19 pandemic has highlighted strengths, weaknesses, opportunities and challenges within healthcare systems across the globe (Bhat et al., 2022). Changed circumstances lead to different value trade-offs. For example, during the coronavirus crisis we started to look differently at lean approaches to organizing care, using as few supplies as possible (also called just-in-time management) (De Graaff et al., 2021;Kuiper et al., 2021). Other existing issues were put on edge by the crisis. Consider the challenges of access to healthcare services and of enhancing the quality of care and patient safety while reducing costs (Hundal et al., 2021;McDermott et al., 2021).
These challenges, amplified during the COVID-19 pandemic, make healthcare an attractive domain for operations management (Barjis, 2011). The Lean Management and Six Sigma (LM&SS) approach has taken a central role in healthcare quality management (Antony et al., 2019;Niñerola et al., 2020;Bhat et al., 2022) within the larger framework of TQM (Klefsj€ o et al., 2001). There is, however, a need for more empirical research on the application of LM&SS in healthcare (Bertolaccini et al., 2015;Ko et al., 2016;Da Silva et al., 2018;Wilson et al., 2018;Anthony et al., 2018;Bhat et al., 2022). One of the main gaps in studies on LM&SS in healthcare is that they fail to address the interplay between its "hard" and "soft" factors (Lau and Idris, 2001;Ershadi et al., 2019;Khalili et al., 2019). For example, although research shows that "soft" practices, i.e. those concerning people and relationships (Mamata et al., 2015), are crucial for achieving superior performance, LM&SS is often perceived as a set of "hard" practices, i.e. practices involving tools and techniques for improving processes (e.g. Poksinska, 2010;Stamatis, 2011), especially in healthcare. This study distinguishes itself from earlier research by examining a systems approach that encompasses interrelated "hard" and "soft" LM&SS practices (Bortolotti et al., 2015;De Koeijer et al., 2022) adapted from manufacturing.
Another gap in the literature on LM&SS is the inconsistent evidence regarding the effect of these "hard" and "soft" LM&SS factors on outcomes (Abdullah and Tar ı, 2017). Some studies suggest a positive effect between "hard" practices and performance (Kaynak, 2003;Rahman and Bullock, 2005), while other studies state the opposite (Ho et al., 2001;Parast et al., 2011). In addition, few studies have attempted to assess how implementing LM&SS in hospitals impacts outcomes (Antony et al., 2019). This study is one of the first to examine in detail the impact of a systems approach encompassing interrelated "hard" and "soft" LM&SS practices on both employee well-being and performance in hospitals.
Finally, there is a dearth of research investigating the organizational patterns (routines) that LM&SS implementation may enable (Antony et al., 2019). Adopting LM&SS in such a way that it becomes a permanent part of the organization's routines can be described as internalization (Kostova and Roth, 2002). Although growing research on LM&SS internalization underlines the importance of "soft" human resource (HR) practices (e.g. Schneider, 1975;Ostroff and Bowen, 2000;Veld et al., 2010;Thirkell and Ashman, 2014), empirical evidence is lacking. Whereas most studies consider HR practices as part of the LM&SS systems approach, we constructed a separate HRM systems approach for "soft" LM&SS practices that are specifically HR-related, such as teamwork, participation and training. This is relevant in light of researchers' disagreement about the interplay between different "soft" practices in establishing a commitment to organizational excellence (Durairatnam et al., 2021). Such considerations call for additional research that clarifies the importance of "soft" HR practices in enacting an organizational climate supportive of a quality orientation (Cavallone and Palumbo, 2021). The present study is unique in that it focuses on both the role of "soft" HR practices and climate as related to "hard" and "soft" LM&SS factors and outcomes in healthcare. The main research question is therefore: Is a balanced approach involving "hard" and "soft" LM&SS practices combined with "soft" HR practices positively related to a climate for LM&SS, and is a climate for LM&SS positively related to outcomes in hospitals?
organizations, the narrow focus on "hard" LM&SS practices disregards issues concerning people and relationships Mamata et al., 2015). For example, studies by Chung and Kwon (2016), Begen et al. (2016) and Cardon and Bribiescas (2015) on LM&SS in healthcare emphasize supply chain principles and cost reduction while largely ignoring the human side. A recent study by De Koeijer et al. (2022) shows that a combination of LM&SS and HRM positively impacts both performance and employee well-being in hospitals. Although the evidence suggests that HRM is important for a successful implementation of LM&SS (e.g. Jørgensen et al., 2007;Zacharatos et al., 2007), little research has been done on this topic in the context of healthcare (Hasle et al., 2012;Cullinane et al., 2014), or the outcomes of research are contradictory (e.g. Sepp€ al€ a and Klemola, 2004;Bonavia and Marin-Garcia, 2011).
Unlike previous research in which HR practices are often regarded as part of the LM&SS systems approach, we constructed a separate HRM systems approach for those "soft" LM&SS practices that are part of the systems approach Description (Cua et al., 2001;McKone et al., 2001;Zu et al., 2008) Special aspects in a healthcare setting There are many areas of knowledge and practice. In general, each specialty has preference for certain suppliers and assortments Table 1. LM&SS systems approach LM&SS practices that are specifically HR-related. For example, cross-functional teams could help to generate ideas for science-based, systematic quality initiatives (N€ aslund, 2013). Performance appraisal and rewards could also function as morale boosters and encourage employee engagement (Sreedharan et al., 2018;Sony et al., 2020;Clegg et al., 2010;Manville et al., 2012). In addition, training and development are crucial to getting skilled and motivated people to work on LM&SS projects (Sanders and Karr, 2015;Anthony et al., 2018). Employee participation and engagement in decision-making and problem-solving can also help inspire commitment to organizational excellence (Cavallone and Palumbo, 2021).
The rationale behind constructing an HRM systems approach is threefold. First, the growing critique of the negative effect of LM&SS on employees argued for the HR side to be viewed separately (Holden, 2011;Moraros et al., 2016;Goodridge et al., 2015). Second, LM&SS practices such as process management and the focus on metrics appear to be of a different order than, for example, LM&SS practices such as training and teamwork. Where the first two practices are usually related directly to the adoption of LM&SS, the latter two are likely to have been standard in healthcare organizations for quite some time. More specifically, while LM&SS is often programmatic and temporary in nature, HRM is often a consistent component in a hospital's business operations. Third, research on the interplay between "soft" HR practices and "soft" and "hard" LM&SS practices is scarce. To gain a thorough understanding of the relationship between "soft" LM&SS practices that are specifically HRrelated and the internalization of LM&SS, we follow Wright and Boswell (2002) in including an HRM systems approach in our research (see Table 2). Radnor et al. (2012) argue that the narrow focus on "hard" LM&SS practices, such as "focus on metrics" and "process management", has led healthcare to neglect activities that encourage employees to develop shared perceptions of LM&SS. These shared perceptions are important for the internalization of LM&SS interventions (Taylor et al., 2013). For the effects of LM&SS to become visible and measurable, a process of routinization must take place in which professionals adopt these new work practices and adapt their existing organizational routines accordingly. Adopting LM&SS in such a way that it becomes a permanent part of the organization's daily routine can be described as internalization (Kostova and Roth, 2002). New routines cannot be sustained in a setting that does not support and enable their performance, however. For example, unless the LM&SS climate reflects employees' belief in the real value of LM&SS for their organization, there is a significant risk that LM&SS will never be internalized (Tolbert and Zucker, 1996). This risk is particularly acute in healthcare because healthcare professionals fear that adopting LM&SS will lead to over-standardization (Holden, 2011) and that LM&SS redirects clinical practice away from patient care towards more administrative and management tasks (e.g. Radnor, 2011). Shared perceptions support employees in their drive to sustain quality improvement initiatives (Tan et al., 2014) and in their commitment to accomplishing organizational excellence (Beraldin et al., 2020;Cavallone and Palumbo, 2021). Creating a climate for LM&SS that reflects positive shared perceptions of employees about LM&SS practices and their commitment to them is therefore crucial to the internalization of LM&SS (Kostova and Roth, 2002;Ostroff et al., 2003;Patterson et al., 2005).

LM&SS, HRM and climate
Climate is consistently conceptualized as employees' shared perceptions about the nature of their organization in terms of events, policies, practices, and procedures (Ostroff et al., 2003;Patterson et al., 2005). Internally, climate is often considered actionable, i.e. management can try to shape climate to pursue organizational goals and influence performance (Denison, 1996;Haakonsson et al., 2008). Like climate in the general sense, the climate for LM&SS can be conceptualized in two different ways. First, there is the individual level or psychological climate, which refers to the individual's perceptions of organizational practices and procedures (Patterson et al., 2004). Second, these individual perceptions can be aggregated to the unit level or organizational climate if there is a certain degree of consensus among organizational members (Joyce and Slocum, 1984;Bergmann et al., 2018). Disparities between the work environment of organizational units can lead to different climate manifestations within the same organization (Bergmann et al., 2018). We therefore focus on the climate in work units rather than the whole organization as the appropriate level of analysis (Zohar and Luria, 2005). Schneider and Reichers (1983) argue that, for the concept of climate to be meaningful, it must have a specific reference. In the past twenty years, a growing body of research has therefore focused on a climate "for something" (e.g. Patterson et al., 2005;Schulte et al., 2009). HR practices that are part of the systems approach Generic description (Boon et al., 2011) Specific description  (Hines et al., 2018). Bhat et al. (2022) argue that an integrated LSS strategy ensures a climate of continual improvement in the healthcare setting. Goodridge et al. (2015) state that LM&SS seeks to create an environment in which mistakes are opportunities for learning, with consistent application of no-blame approaches to mistakes and errors. Ambekar and Hudnurkar (2017) claim that people with a positive attitude and criticalthinking capability innovate and ideate solutions. While researchers agree that a successful LM&SS implementation will aim to achieve climate change and succeed, they fail to agree on the specific characteristics of such a climate for LM&SS. In this study, we focus on a climate for LM&SS that reflects employees' perceptions regarding the extent to which the organization emphasizes specific LM&SS values, goals, expected behaviors and contributions at work, related to quality, innovation and efficiency (Schneider, 1975;Veld and Alfes, 2017) (see Table 3). . Drawing from the above-mentioned studies, we focus in our research on perceived organizational performance reflecting a wide range of improvements, including internal processes, customer satisfaction and finances (see Table 4). Perceived organizational performance can be defined as employees' individual perceptions of organizational efficiency (Giauque et al., 2013) and refers specifically to employees' subjective evaluation of an organization's overall performance compared to its rivals in the same sector (Allen and Helms, 2002;Berberoglu, 2018). Research shows positive relationships between organizational climate and employee wellbeing (Parker et al., 2003;Veld et al., 2010). Employee well-being is a multidimensional construct (e.g. Van de Voorde et al., 2012). Where the traditional view held that well-being was mainly about affect, several broader conceptualizations have more recently been proposed that include behavior and motivation (Ryff and Keyes, 1995;Warr, 2007). In the context of Description (Patterson et al., 2005) A climate for LM&SS A reflection of employees' perceptions of the extent to which the organization emphasizes specific LM&SS values, goals, expected behaviors and contributions at work (Schneider, 1975;Veld and Alfes, 2017) Description (e.g. Holden, 2011;Shah and Ward, 2003;Wiklund and Wiklund, 2002;Habidin et al., 2012) Organizational performance Measures related to organizational performance, that reflects a wide range of improvements such as internal process-, customer-, innovation, and financial performance Table 3. Climate Table 4. Performance organizations, well-being can be broadly defined as the overall quality of an employee's experience and functioning at work (Peccei et al., 2013). Following current HRM literature (e.g. Grant et al., 2007;Van de Voorde and Boxall, 2014), we identify three core components of wellbeing: health, happiness and trusting relationships (see Table 5). Subdividing well-being into these components is important for several reasons. First, the dominant models within both HRM and LM&SS theory and research continue to focus largely on ways to improve performance, with employee concerns mainly as a secondary consideration (Fotopoulos and Psomas, 2009;Guest, 2017). Second, scholars do not agree about the effectpositive or negativeof LM&SS on employee well-being (e.g. Conti et al., 2006). With the evidence being inconsistent, there is a need for more in-depth research that focuses on both the positive and negative effects of LM&SS on employee well-being.

Research hypotheses
Over the past ten years, a growing body of research has examined the concepts of psychological and organizational climate among healthcare employees (e.g. Veld et al., 2010;Purohit and Ashok, 2012). Systematic research on the relationship between a LM&SS systems approach and a climate for LM&SS is lacking in healthcare, however. We expect that the more an organization adopts LM&SS practices, the more LM&SS is internalized. Internalization takes place when employees develop shared perceptions concerning the value of LM&SS practices, referred to as climate (Patterson et al., 2005). We therefore expect that: H1. The adoption of a LM&SS systems approach is positively related to a climate for LM&SS in hospitals.
Previous studies have confirmed that HRM plays a vital role in shaping organizational climate (Gelade and Ivery, 2003;Ali et al., 2018). Thirkell and Ashman (2014) claim that it is essential to combine LM&SS and HRM to achieve the desired breadth and depth of LM&SS adoption at the macro, meso and micro levels. In this context, HRM can be seen as a signaling system that constantly sends messages to employees stressing the attitudes and behaviors desired within the organization (Bowen and Ostroff, 2004). In other words, HR practices can be used to strengthen goal alignment and foster specific work behaviors (Veld and Alfes, 2017), thereby creating a desired climate. For example, hospital management can use HRM practices to create a desired climate where LM&SS initiatives take root by communicating to employees that quality improvement is important and that improvement initiatives and innovative behavior are expected and rewarded (Bowen and Ostroff, 2004;Veld and Alfes, 2017 The relationship dimension of employee well-being focuses on the quality of trusting relationships between employees and their employer and colleagues The hierarchical structure impacts the relations between employees and their employer and colleagues Table 5. Employee well-being TQM on both the role of HRM and organizational climate relating specifically to LM&SS are scarce. Recent research shows that a system of HRM practices can be used to create climate perceptions (e.g. Veld and Alfes, 2017), leading us to expect that: H2. An HRM systems approach is positively related to a climate for LM&SS in hospitals.
Researchers suggest that a positive climate, for example a climate for LM&SS, influences quality and quantity of work done in organizations as well as employee productivity (Mullins, 2010;Permarupan et al., 2013). If employees perceive a positive LM&SS climate in their organization, they tend to rate organizational performance higher than that of competitors (Berberoglu, 2018). Although there is evidence showing that organizational climate is an important determinant of organizational performance (Burton et al., 2004), there has been little systematic research on the relationship between a climate for LM&SS and performance. Recent studies identify a climate for LM&SS as the missing link in achieving performance improvements that are sustained over time (e.g. Bortolotti et al., 2015;D'Andreamatteo et al., 2015). We therefore expect that more shared perceptions among employees concerning LM&SS will lead to higher levels of performance:

H3.
A climate for LM&SS is positively related to organizational performance in hospitals.
Research by Gouldner (1960) suggests that employees are expected to reciprocate the organization for its encouragement, benefits and support for, in this case, LM&SS with positive employee outcomes. We could therefore argue that a climate for LM&SS is positively related to employee well-being. Others authors, however, point out that emphasizing efficiency and productivity puts employees under greater pressure and intensifies their workload (Holden, 2011). The nature of the relationshippositive or negativebetween climate and employee well-being can vary between the three components of well-being (Peccei et al., 2013). Internalization is linked to employee commitment and employee perceptions of trust (Kostova and Roth, 2002). Following Blau's (1964) social exchange theory, we expect that employees interpret a climate for LM&SS as indicative of organizational support and care, and reciprocate accordingly with commitment, satisfaction and trust (Whitener, 2001;Van de Voorde et al., 2012). Based on research suggesting that LM&SS has a negative effect on employee health (Hasle et al., 2012), for example that LM&SS leads to higher levels of stress, we furthermore expect that a climate for LM&SS negatively impacts employee health.

H4.
A climate for LM&SS is positively related to the happiness (H4a) and trusting relationships (H4b) components of well-being, while it is negatively related to the health (H4c) component of well-being in hospitals. Bowen and Ostroff's (2004) proposition that climate can be seen as a mediating factor between HRM and outcomes is confirmed by studies in healthcare (e.g. Veld et al., 2010;McCaughey et al., 2013). For example, Veld et al. (2010) report that climate in hospitals mediates the effect of perceived HRM systems and unit commitment. Unlike HRM literature, research on LM&SS has produced scant evidence as to what role climate plays between LM&SS and outcomes. Our study builds on growing evidence in the field of HRM of climate's mediating role, and we therefore expect that climate mediates the relationship between LM&SS and HRM on the one hand and performance and employee well-being on the other. Although we expect the relationships between the described variables to differ, the nature of this relationshippositive or negativedepends on our findings for hypothesis 4 (relationship between a climate for LM&SS and employee well-being). We have therefore formulated a neutral hypothesis regarding the relationships between a climate for LM&SS, HRM and outcomes:

H5.
A climate for LM&SS mediates the relationships between LM&SS (H5a) and HRM (H5b) on the one hand and organizational performance and employee well-being in hospitals on the other. Figure 1 summarizes the proposed relationships in this study.

Research methodology 3.1 Population and sampling
This paper describes a cross-sectional, multisite study that uses quantitative research methods and nested data from internal service units in Dutch university hospitals. In healthcare, LM&SS is often introduced, first of all, in high-volume processes such as cleaning, logistics and food service (Stamatis, 2011;Goodridge et al., 2015). Hospital service processes differ fundamentally from processes at a fast-food restaurants or cleaning companies. The employees of internal service units are usually assigned permanently to a hospital ward and, therefore, perceive nurses and physicians as their direct colleagues, have direct contact with patients, and see their work as a link in the value chain for delivering top-notch care. While most of the above-mentioned studies focused on a single hospital unit or department, our study covers more than 40 internal service units at hospitals in the Netherlands (A to H). These hospitals deliver highly specialized patient care, combined with specialist diagnosis and treatment, and are closely linked to medical research and education. The internal service units differ in size and structure (see Table 6). The more homogeneous a population, the easier (i.e. more likely) it is to generate a representative sample (Jager et al., 2017). To ensure a homogeneous sample and internal and external validity and reliability, we applied four criteria for participation in our research: (1) Services had to be similar in nature and present at four or more university hospitals.
(2) At least ten employees and three supervisors per unit had to assess the theoretical concepts at unit level for reliability.
(3) Employees and supervisors (including temporary workers) had to have been employed by the internal service unit for at least one year.
Applying these criteria gave us a sample of 1,668 employees and 218 supervisors in 42 units (response rate of 55%, varying from 20% to 96% per unit). The average group size per unit was 40 employees and five supervisors. Table 6 shows the response rates at unit level for each of the eight hospitals. Following Cohen (1992), we categorize effect sizes into small (0.10), medium (0.30), and large (0.50).   At the time of the research, according to the Dutch law approval from the 'Central Committee on Research Involving Human Subjects', ethical approval is only necessary when research concerns medical/scientific research and participants are subject to procedures or are required to follow rules of behavior. In this study, the study population involved employees of internal services only. Therefore, these criteria were not met and ethical approval of the Central Committee on Research Involving Human Subjects was not found applicable.

Measurement instruments
To operationalize the theoretical concepts of LM&SS, HRM, climate, performance and employee well-being, we searched the literature for existing validated measurement instruments. In consultation with experts, we selected empirical studies that applied validated measurement instruments to healthcare. A translator translated our original surveys into English, with an independent bilingual native speaker of Dutch and English performing the back translation. Under the supervision of a research assistant made available for a week at each research site, the cross-sectional survey was distributed to the supervisors and employees of eight university hospitals to collect survey data on LM&SS, HRM, climate, performance, and employee well-being. Our LM&SS systems approach incorporates the following practices: top management support, customer relationship, quality information, process management, structured improvement procedure, focus on metrics, and supplier relationships. We adapted the original manufacturing-oriented items (e.g. error rates, defect rates, scrap, defects, cost of quality) to reflect a healthcare perspective (e.g. mistakes, throughput time, productivity). We excluded elements of the survey that focus specifically on the industrial context of plants (for example: "We design for manufacturability"). With the exception of the LM&SS customer relationship practice measured at supervisor level (Cronbach's α 5 0.66), the reliability of all scales exceeded 0.70.
We studied a wide range of HR practices: training and development, performance appraisal and rewards, team working and autonomy, participation and job design, employment security, and work/life balance. We included 27 items on HR practices, measured using the scale proposed by Boon et al. (2011) (for example: "My unit offers me work that gives me the opportunity to express myself"). Response choices were on a five-point Likert-type scale ranging from "completely disagree" (1) to "totally agree" (5). We constructed the HRM and LM&SS system approaches by calculating the sums of the mean scores for the separate practices and separating them into two bundle variables. With the exception of the work/life balance HR practice (α 5 0.69), the reliability of all scales exceeded 0.70.
We included seven items on organizational performance (Zu et al., 2008) (for example: "The quality of our units' products and services has been improved over the past 3 years."). After consultation with the author of the original scales, we changed the scale from a seven-point to a five-point Likert scale, ranging from "completely disagree" (1) to "totally agree" (5), because this is more in line with other parts of the survey.
As employee well-being is an individual characteristic, we measured it at the individual employee level. Regarding the health component of employee well-being, we used subscales of the Dutch standardized survey on the experience of work (VBBA) (Van Veldhoven et al., 2002) to measure workload and strain. The scale for strain captures small deficits in employee functioning at the end of, or just after, a workday (Van Veldhoven, 2005). Sample items include "Do you have too much work to do?" and "It is an effort for me to stay focused in my free time after work". Response choices were on the original four-point Likert-type scale ranging from "never" (1) to "always" (4). There are several measures of intra-organizational trust available that differ depending on who is being trusted (Dietz and Den Hartog, 2006). We focused on trust between an employee and his or her direct supervisor, using the seven-item scale devised by Robinson (1996). Sample items include "I can expect my supervisor to treat TQM me in a consistent and predictable fashion". Response choices were on a five-point Likert-type scale ranging from "completely disagree" (1) to "totally agree" (5). The reliability of all scales was 0.84 or higher (see Table 7). To measure the happiness component of employee wellbeing, we included items on satisfaction and commitment. In contrast to the health and trusting relationships components, we measured the happiness component of well-being at the group level. Mason and Griffin (2005) show that assessing the satisfaction of the group directly, rather than simply aggregating the individual job satisfaction ratings of group members, explains additional variance in outcomes. We therefore transposed the items on commitment and satisfaction from the individual to the unit level. To measure employee satisfaction, we used one other VVBA item: "All things considered, my colleagues are satisfied with their job". Organizational commitment was measured using four items from Allen and Meyer's (1990) Affective Commitment Scale (for example, "My colleagues feel like 'part of the family' at their unit"). Response choices were on a five-point Likert-type scale ranging from "completely disagree" (1) to "totally agree" (5). The reliability of all scales exceeded 0.70.
To measure climate for LM&SS, we used 14 items on important LM&SS aims, namely quality, innovation and efficiency climate, as proposed by Patterson et al. (2005). We reformulated the original items, which reflected an organizational-level perspective (e.g. "People in this organization are always searching for new ways of looking at problems"), to represent a unit-level perspective (e.g. "People in my unit are always searching for new ways of looking at problems"). This reformulation was necessary because each climate item should focus on the specific collective unit corresponding to the climate being studied (in this case, the unit). By specifying a clear frame of reference, we precluded the risk that respondents would describe the perceptions of different parts of the organization (Patterson et al., 2005). Response choices were on the original four-point Likert-type scale ranging from "absolutely not true" (1) to "absolutely true" (4). The reliability of all scales was 0.71 or higher. Table 7 shows the psychometric characteristics of the measurement instruments as well as the respondents (employees or supervisors) for each measurement instrument.
As potential control variables, we included the general characteristics of respondents (age, gender, educational level), general characteristics of the job (work unit, number of years working for the organization, number of years working in the specific work unit and job, type of employment contract) and general characteristics of the work unit (size). We dummy coded categorical variables and added familiarity with LM&SS and experience participating in LM&SS projects to our control variables.

Data analysis and results
We used descriptive statistics to describe our research population at the unit level. As our HRM, climate and employee well-being data was collected from a single source, the employees, we used a split sample for our analysis: we randomly split the units in half and obtained the values for the HRM and climate perceptions from one-half and the employee well-being variables from the other. We followed the same procedure regarding the values for the LM&SS perceptions and the performance variables. As these split-sample results were robust compared to the whole-sample results, we concluded that the common method bias was unlikely to be a serious problem in our data. To identify which control variables to include in regression analyses, we examined the degree of correlation between these variables and the dependent variables. Criterion for inclusion in the regression was an effect size of 0.30 or higher (reflecting medium to strong relationships) (Cohen, 1992). As no control variable exceeded this minimum level, none of the control variables were entered in the multilevel regression analysis. To test our hypotheses, we performed multivariate regression analyses. We employed multi-level analysis techniques to allow for the hierarchical structure of the  Table 7. Psychometric characteristics measures TQM data, in which employees and supervisors are nested within units. We conducted the analyses with performance as a dependent variable at the unit level. The HRM and climate variables then had to be aggregated. To support the aggregation of individual scores to unit-level scores, we calculated ICC1 and ICC2 values (intra-class correlations; to measure inter-rater reliability) and tested whether the average scores differed significantly across units. The ICC1 value was 0.05 and 0.03 for HRM and climate respectively, and the ICC2 value 0.66 and 0.57 respectively, exceeding the minimum value of 0.50 (Klein and Kozlowski, 2000) and supporting the aggregation to unit level. The ICC1 values of the three components of employee well-being implied that 6-13% of the variance in these components can be attributed to the unit level (see Table 7). Since we expected mediating effects, we used the mediation framework developed by Zhao et al. (2010) to test hypothesis 5.

Description of the study sample
Within the units, female employees account for 13% of the workforce and the average age of the respondents is 45 (see Table 6). The relatively small percentage of female employees can be explained by the technical nature of the internal service units, for example maintenance, logistics and security. More than 80% of the respondents have a permanent employment contract and only 17% attended higher education. Respondents have worked an average of ten years for their internal service unit, and eight years in their job (see Table 6).

Testing the hypotheses
The results of the regression analyses (see Table 8) show that the LM&SS systems approach has a significant, but very small, effect on a climate for LM&SS (β is 0.07). Hypothesis 1 is therefore not supported. In addition, HRM systems approach has an almost strong positive effect on a climate for LM&SS (β is 0.46). In total, 39% of the variance in climate for LM&SS is explained. Hypothesis 2 is supported. The results of the regression analysis (see Table 9) indicate that a climate for LM&SS is not related to perceived performance (β is À0.05). Hypothesis 3 is therefore not supported.
With respect to the three components of employee well-being, the results of the regression analysis (see Table 10) indicate that a climate for LM&SS is positively related to the happiness and trusting relationships components (medium effects, both βs are 0.33) and negatively related to the health component (small effect, β is À0.13). The total explained variance in happiness is 22% and in trusting relationships 25%. Hypothesis 4 is supported.
Hypothesis 5a concerns the mediating effects on employee well-being. The analysis shows that a climate for LM&SS has small mediating effects on the three components of well-being (βs varied from À0.07 to 0.17) (see Table 11). There was evidence of complementary mediation for the three components, with both the mediated effects and the direct effect apparent and pointing in one and the same direction (Zhao et al., 2010). The direct effect of LM&SS and HRM on employee well-being decreases, however, and the mediating effects are small (Cohen, 1992). Hypothesis 5b concerns the mediating effects on performance. The analysis shows that a climate for LM&SS has no effect on perceived performance (see Table 11). We conclude that a strategic climate does not mediate the relationship between LM&SS and HRM on the one hand and performance and employee well-being on the other.

Discussion of findings
In this study, we focused on Lean Management and Six Sigma (LM&SS) as a widely adopted operations management approach in healthcare (Antony et al., 2019;Niñerola et al., 2020; Table 9. Hierarchical multilevel analysis climate and performance Table 10. Hierarchical multilevel analysis climate and employee well-being Table 11. Hierarchical multilevel analysis mediating role of climate, performance, and well-being TQM Bhat et al., 2022) within the larger framework of TQM (Klefsj€ o et al., 2001). Specifically, we concentrated on several gaps in existing research, namely a failure to address the interplay between "hard" and "soft" factors of LM&SS (Ershadi et al., 2019;Khalili et al., 2019;De Koeijer et al., 2022;Durairatnam et al., 2021), inconsistent evidence for the effect of these "hard" and "soft" LM&SS factors on outcomes (Abdullah and Tar ı, 2017;Wilson et al., 2018;Anthony et al., 2018;Bhat et al., 2022) and the potential of climate and Human Resource Management (HRM) for internalizing LM&SS (Thirkell and Ashman, 2014;Antony et al., 2019;Cavallone and Palumbo, 2021). We conceptualized "hard" and "soft" practices for LM&SS and defined a separate bundle of "soft" LM&SS practices that are specifically HRrelated. Our research shows that a balanced approach to internalizing LM&SS at the macro, meso and micro levels encompassing both "hard" and "soft" factors is indeed crucial and sheds light on the interplay between these practices, the internalization of LM&SS, and outcomes in several ways.
First, there appears to be a logical relationship between a climate for LM&SS and LM&SS adoption (Ambekar and Hudnurkar, 2017;Bhat et al., 2022). We found, however, that the HRM systems approach, consisting of "soft" LM&SS practices that are specifically HRrelatedand not the "soft" and "hard" practices that are part of the LM&SS systems approachis crucial for creating shared perceptions among employees and, consequently, a climate for LM&SS. Our results indicate that to create a desired climate where LM&SS initiatives take root, management can use HRM practices by communicating to employees what is valued and considered important in the organization (Bowen and Ostroff, 2004;Veld and Alfes, 2017) and by encouraging them to attain organizational excellence (Beraldin et al., 2020;Cavallone and Palumbo, 2021). Our findings are in line with research by Bowen and Ostroff (2004), Knies and Leisink (2014), and Ali et al. (2018) showing that HRM can influence climate by signaling what strategic goals are most relevant and what kind of employee behaviors are expected, supported and rewarded in connection with these goals. Combining LM&SS and HRM is therefore essential to achieving the desired breadth and depth of LM&SS adoption at macro, meso and micro levels.
An explanation for the absence of a relationship between LM&SS and a climate for LM&SS could be that employees of internal service units find it hard to grasp the concept of LM&SS and struggle to adapt the approach to their daily practice. For example, process management and a focus on metrics are LM&SS practices that require analytical skills from those who apply them. Employees of internal service units are usually lower educated: only 17% of our respondents attended higher education. When employees perceive LM&SS practices as distal and abstract, this may weaken the development of shared employee perceptions about the real value of such practices. In contrast, employees of internal service units are likely to have a much better understanding of HRM than LM&SS because HR practices are developed specifically for employees. For example, quality management training can be tailored to specific employee groups and their educational background. Following this line of thinking, it can be said that HRM boosts employee engagement and involvement in continuous quality improvement (Keng Boon et al., 2007;Assarlind and Gremyr, 2014;Cavallone and Palumbo, 2021), which may explain the relationship between HRM and a climate for LM&SS. Also, an explanation for the relationship between HRM and a climate for LM&SS may be that the growing (internal and external) focus on efficiency in healthcare induces employees to perceive HR practices in light of these efficiency goals, perhaps fostering a climate for LM&SS (Nishii et al., 2008).
Second, our results show that a climate for LM&SS leads to higher levels of happiness and trust among employees and that it has no effect on the health of employees. In addition, we found a weak relationship between a climate for LM&SS and performance. This suggests that internalizing LM&SS is important for employee well-being but not for performance. One explanation for this outcome may be that LM&SS is implemented in healthcare mainly with the aim of improving short-term efficiency and quality (Drotz and Poksinska, 2014), but not to sustain it in the longer run. Given the ambition of hospitals to maintain higher standards of both organizational performance and employee well-being, however (Kowalski et al., 2015), we argue that hospitals that adopt LM&SS should also foster a climate for LM&SS by combining LM&SS and HRM, thereby internalizing LM&SS. Blau's (1964) social exchange theory states that employees interpret management activities as indicative of organizational support and care and reciprocate accordingly with commitment, satisfaction and trust (Whitener, 2001;Van de Voorde et al., 2012). In that sense, employees of internal service units may experience HRM as a form of recognition and concern, creating a climate for LM&SS and affecting their well-being. We did not find that a climate for LM&SS has a strong mediating effect, but it seems likely that there would be a positive effect that spirals upwards: the more an organization adopts LM&SS in combination with HRM, the more LM&SS is internalized and the more both overall performance and employee well-being improve, and vice versa. In other words, it could be argued that a coherent set of "soft" and "hard" LM&SS and HR practices is essential for achieving both internalization and outcomes at the micro, meso and macro levels.
Another line of thinking takes into account that the average interval between adopting LM&SS at the participating hospitals and our data collection was 18 months, and that most hospitals started out with a top-down LM&SS program. We did include a time lag for LM&SS adoption to gain a better understanding of the relationship between adoption, internalization and outcomes, but without any conclusive results. It is possible that at the time of data collection, there was a gap between supervisors and employees in terms of the degree of LM&SS internalization. The first groups of hospital employees to be impacted by strategic goals are usually the managers and supervisors. It is they who decidewhen simple cost-cutting measures prove to be inadequateto adopt LM&SS as a programmed approach to efficiency. In that sense, supervisors have a head start when it comes to creating shared perceptions about LM&SS and we can imagine that the climate for LM&SS among supervisors would have a stronger impact on outcomes than the climate for LM&SS among employees. There is therefore some likelihood that, over time, as employees increasingly internalize a systems approach encompassing both "soft" and "hard" LM&SS practices, the mediating effects of a climate for LM&SS on the relationship between LM&SS, HRM and outcomes will become stronger.

Conclusion
Our research shows that an integrated approach to implementing "soft" and "hard" LM&SS and HRM shapes a climate for LM&SS conducive to the pursuit of organizational goals. We found that the HRM systems approach, consisting of "soft" LM&SS practices that are specifically HR-related (e.g. teamwork and autonomy, appraisal and rewards, participation and job design, training and development)and not the "soft" and "hard" practices that are part of the LM&SS systems approachare crucial for creating a climate for LM&SS. In addition, we found that a climate for LM&SS is not related to perceived performance or employee health, and that it is positively related to the employee happiness and trusting relationships. This means that the combination of "soft" and "hard" LM&SS and "soft" HR practices allows healthcare organizations to internalize LM&SS through a climate for LM&SS, supporting happy and trusting employees. In their attempt to create mutual gains for organization and employees, hospitals that adopt LM&SS should foster a climate for LM&SS by embracing a balanced approach that consists of "hard" and "soft" LM&SS and HR practices, thereby internalizing LM&SS at the macro, meso and micro levels.

Managerial and practical implications
From a management perspective, our results emphasize that fragmentation into "hard" and "soft" LM&SS and HR practices may lead to suboptimal results, which will not be conducive TQM to establishing a fully-fledged quality philosophy (Cavallone and Palumbo, 2021). Both "soft" and "hard" LM&SS and HR practices should be incorporated into a systems approach, allowing organizations to capitalize on their synergies for internalizing LM&SS and employee well-being. Our research shows that management can use HRM to shape a climate for LM&SS conducive to the pursuit of organizational goals and the happiness and trust of employees. Hospitals should therefore involve their HR departments right from the start when introducing LM&SS programs to ensure that a HRM systems approach is in place. In the Netherlands, many HR practices are predetermined in national Collective Bargaining Agreements (CBA) for hospitals, making HRMunlike LM&SSa consistent component of healthcare organizations and covering all employees. These HR practices are practical and can be tailored to specific employee groups and their educational background. For example, HR practices such as teamwork, participation and training involve employees at different levels in continuous quality improvement. Management can use these HRM practices to create a desired climate in which LM&SS initiatives can take root. It is important that managers are consistent in communicating to employees what is valued and considered important in the organization and the kind of behaviors and attitudes that are expected and rewarded (Bowen and Ostroff, 2004;Veld and Alfes, 2017). For example, they should emphasize the importance of continuous improvement and of achieving quality outcomes, and discuss with employees how they can contribute in practical terms.
Where many studies so far have argued for the inclusion of HR practices in an LM&SS systems approach, our results show that LM&SS and HRM should be viewed as two different things that require constant alignment and that should be managed integrally. In practice, this could mean that when hospital leaders share the "why" of LM&SS within the organization, they should emphasize both performance improvements as well as higher levels of employee well-being. Another recommendation is to monitor progress in LM&SS integrally by focusing not only on the number of LM&SS initiatives and their progress but also on the happiness, health and trusting relationships of employees, and by explicitly including performance indicators in the "LM&SS dashboard". In addition, since direct supervisors play a prominent role in transmitting values and climate (Kuenzi and Schminke, 2009), they should actively support their employees with a balanced approach that incorporates both "hard" and "soft" factors into the improvement process (Poksinska, 2010). For example, appraisal interviews should not only focus on "hard" key performance indicators, but also on improvement efforts and more narrative input. This may also mean that employee productivity would temporarily decline to allow time for improvement projects or quality training.

Limitations and future research
This study has a number of limitations. First, it focuses on cross-sectional relationships and is therefore not suitable for establishing cause-and-effect relationships. We included a time lag for LM&SS implementation to gain a better understanding of the relationship between intervention and outcome, but without any conclusive results. The findingsalthough based on a thorough literature studyshould therefore be interpreted with some caution. We recommend that future research apply a longitudinal and intervention design (including control settings) to acquire a deeper understanding of the causal relationships between LM&SS, HRM, climate for LM&SS, performance and well-being. Such research could, for example, examine a potential spiraling positive or negative effect, i.e. that the more LM&SS in combination with HRM is adopted, the more LM&SS is internalized and the more performance and employee well-being improve, and vice versa. Longitudinal research could also verify whether the relationships that we found, for example, between LM&SS and performance, HRM, climate and well-being, are cause-and-effect relationships. That is plausible for the relationship between HRM and climate based on earlier extensive research that confirms our findings (e.g. Gelade and Ivery, 2003;Ali et al., 2018).
Second, efforts to improve quality and efficiency are nothing new for healthcare professionals; they have been part of their jobs for many years. Following this line of thinking, it could be argued that a climate for LM&SS is not necessarily the result of adopting LM&SS, but more a prerequisite for the successful adoption and internalization of LM&SS. This could explain why we did not find that a climate for LM&SS had mediating effects on outcomes (neither performance nor well-being). It would be interesting to examine this idea in subsequent research.
Third, we dealt only with internal service units at university hospitals. For most employees, including physicians, LM&SS is something new, in large part because medical school and residency training do not emphasize them (Blumenthal et al., 2012). We consequently argue that our findings are valid for every hospital ward, but it would be interesting to investigate the interaction between hospital wards and internal service units in future research. Including direct care processes and healthcare professionals could lead to a more in-depth understanding of the mechanisms underpinning the internalization of LM&SS in hospitals, for example, how professionals from internal service units and hospital wards develop and adapt their existing organizational routines to LM&SS practices. Differences between the characteristics of hospitals and other types of healthcare organizations (e.g. elderly and disabled care) may imply that our findings will not be easy to generalize to other types of healthcare organizations (De Koeijer et al., 2019) and should be tested in future research.
Fourth, the measures we included regarding the health component of employee well-being focus on the quantitative burden of work, i.e. workload and recovery time after a workday. It would be interesting to include other health-related measures in future research, for example (early) physical, mental and behavioral symptoms of burnout (Maslach and Schaufeli, 1993), especially with the health of healthcare employees being a pressing issue at present (Taris et al., 2013;Drenth, 2016).
Fifth, we need a broader definition of performance in relation to LM&SS, as well as a more comprehensive set of performance measures. Recent debates have focused on how performance in healthcare should be defined and measured (Willems and Ingerfurth, 2018). For example, is performance about costs, efficiency (e.g. shorter waiting times and improved utilization) or about customer satisfaction, quality and health-related outcomes? Or is performance about all of the foregoing (e.g. Porter, 2010;Arora et al., 2016)? In light of these recent debates, we argue that the definition of performance in relation to LM&SS should be updated and clarified specifically in the context of healthcare. In addition to our research setting out a wide range of perceived improvements (e.g. internal processes, customer satisfaction and financial results), we propose incorporating objective outcome measures into any future research.
It is also worth mentioning the strengths of this research. Our study is one of the first to examine in-depth the impact of "hard" and "soft" LM&SS practices on both employee wellbeing (subdivided into different components) and performance in healthcare, as well as the role of "soft" HRM in this relationship. Linking LM&SS, HRM and outcomes to a climate for LM&SS is a relatively new approach in operations management research and has led to a deeper understanding of the mechanisms underpinning the internalization of LM&SS in healthcare. By studying employee-level as well as unit-level data, we gained a better understanding of the relationships between concepts at both levels. In addition, our sample is unique in several ways. To begin with, all Dutch university hospitals participated, which is remarkable given the increased competition between hospitals in the Netherlands. Our sample also consists of 42 units with an acceptable response rate of 55% (Baruch and Holtom, 2008), while earlier studies focused on a single hospital ward or department.