Do people leave managers or organisations? An integrative review of employee turnover through the lenses of LMX and JD-R Open Access

People leave their jobs for many reasons. In a 2023 survey of 2,000 full-time employees in the UK, the top three reasons people quit their jobs are insufficient pay, burnout due to workload and poor relationships with line managers (Mercer, 2023). Based on their research of job switchers across industries over the course of 15 years, Bernstein et al. (2024) asserted that while compensation, career growth and work–life balance contribute to staff retention, employees often leave due to feeling undervalued or disrespected by their managers. Similarly, Kang et al. (2022) found that individuals are more likely to quit when they experience consecutive difficult tasks, which suggests that the way managers distribute workload can influence turnover.

These findings suggest the significance of the role of managers and their relationships with direct reports in affecting retention and attrition. These patterns point to both relationship dynamics with line managers and structural features of work design as plausible drivers of employee departure. To make sense of this indication, this study builds on two complementary lenses: leader–member exchange (LMX) theory and job demands–resources (JD-R) model. The intent is to specify boundary conditions between leadership and structure. The central claim advanced by this study is that leadership is influential, but not dominant, when job demands exceed the resource capacity of employees.

From an academic perspective, the widespread usage of the saying “people leave managers, not organisations” raises an interest to assess and validate how true the statement actually is. Existing research does acknowledge the role of leadership in employee retention (AbdELhay et al., 2025; Goens and Giannotti, 2024; Tyagi, 2021). However, there remains a need for further analysis to determine the extent of how leadership compares to organisational influences (Kanchana and Jayathilaka, 2023; Puhakka et al., 2021). This is particularly relevant in leadership research, where effective management practices are frequently linked to employee engagement and job satisfaction (Ozturk et al., 2021; Prentice, 2022; Van Tuin et al., 2021).

Similarly, in human resource (HR) strategy, understanding the factors that drive turnover is crucial for designing retention policies, training programmes and well-being initiatives (Dupe et al., 2022; Islam and Amin, 2022). Given the evolving nature of modern-day work, including inter-generational workforce and flexible working, the role of leadership in shaping employee experience has become increasingly complex. However, the literature remains unclear on the boundary conditions between leadership and structure in influencing turnover. Positioning LMX within JD-R provides a feasible way to examine these mechanisms.

Foundational turnover models view leaving as a withdrawal decision shaped by desirability and ease of movement (March et al., 1958; Price, 1977). Building on this, an LMX and JD-R framing specifies how manager–employee exchange quality and work allocation shape withdrawal processes and decisions (Hom et al., 2017). Existing turnover syntheses sit in parallel streams. Broad reviews summarise multiple predictors, but there remains an opportunity to integrate leadership and structural conditions in a single model (Griffeth et al., 2000; Hom et al., 2017).

Some relationship focused work links exchange quality to turnover intentions which paves the way to explore organisational contexts (Dulebohn et al., 2012), and some meta-analyses estimate overall associations and present some room to clarify how structural demands constrain leadership (Mackey et al., 2017). Similarly, existing JD-R reviews explain turnover through a demands–resources process in which leadership is one resource amongst many (Bakker and Demerouti, 2017). This review integrates LMX as the relational mechanism within a JD-R architecture to explain how leadership and structure jointly shape turnover decisions.

To examine this phenomenon, this study explores the following two questions:

An integrative review of 39 papers from leadership and HR management literature was conducted to analyse existing research from 2014 to 2025, using transparent search and selection procedures and theoretical synthesis of LMX and JD-R. This approach allows a comparison of leadership and organisational factors in a single framework, and clearer understanding of whether leadership is a dominant driver of turnover, or if broader organisational aspects play an equally significant role. The boundary conditions outlined in this study specify how and when leadership is insufficient to offset structural demands. It also proposes a research agenda and actionable implications for HR practice.

LMX and JD-R were chosen for this study because they link leadership processes to turnover through relational and demand–resource mechanisms.

The LMX theory (Dansereau et al., 1975) focuses on the dyadic or two-way relationships between leaders and employees. The emphasis is on how high-quality or low-quality exchanges can affect job satisfaction and morale. In this regard, leaders develop unique relationships with each subordinate, namely high-quality LMX relationships (i.e. trust, respect, support) and low-quality LMX relationships (i.e. limited interaction, lack of trust, minimal support). In the context of this study, LMX helps to analyse whether leadership quality is a dominant predictor of voluntary turnover (Harris et al., 2005). This can come in the form of either resource-related provisions (e.g. role clarity, access to information, autonomy) or factors related to work demands (e.g. uneven workload, role conflict, emotional labour).

LMX differentiation between leader and team members also leads to comparisons between employees which can amplify these effects. For substantial evidence linking poor leadership to turnover, LMX theory can be a foundational explanation as low-quality exchanges limit resource provisions and increase demand factors. Conversely, when organisational factors emerge as stronger predictors, this suggests that LMX quality alone is insufficient to explain turnover as a whole.

The JD-R model (Demerouti et al., 2001) offers a broader organisational perspective by identifying two key factors that shape employee well-being and retention, namely job demands (i.e. work-related stressors) and job resources (i.e. managerial support that helps employee engagement). The model has parallels to LMX and posits that when job demands exceed available resources, employees experience burnout and reduced engagement (Bakker and Demerouti, 2007). In the context of this study, leadership mitigates these effects by supporting career development and workload management (Schaufeli, 2017).

Leaders add resources (e.g. clarity, support) and shape demands (e.g. task assignment, time allocation) which would place LMX within the JD-R construct. The assertion that employees are more likely to quit after experiencing consecutive difficult tasks, for example, aligns with JD-R (Kang et al., 2022). When burnout and excessive workload act as primary drivers of turnover, instead of poor leadership, it suggests that organisational factors play a stronger role in turnover decisions.

In this study, LMX and JD-R are integrated to explain turnover as a demand–resource process in which leadership operates as a relational mechanism. LMX clarifies how relationship quality shapes key job resources that strengthen commitment, whereas JD-R captures how job demands can drive exhaustion and withdrawal even when relationships are positive. LMX shapes how demands are experienced, buffering strain when support is consistent or compounding strain when exchanges are low quality or uneven across team members. However, leadership is not unconstrained. Organisational structures set baseline demands and resources, which limits how much leadership can influence the overall balance. Leadership can support retention when resource gains offset demands, but its effect weakens when structural hindrances remain persistently high.

Based on this theoretical framework, the guiding proposition for this study is outlined across three segments. Firstly, the resource pathway whereby high-quality LMX reduces turnover intentions by increasing key resources. Secondly, the demand pathway whereby leadership behaviours that concentrate hindrances increase turnover intentions independent of baseline LMX. Finally, the boundary condition whereby retention effect of LMX weakens as demands rise or structural resources fall.

In conducting the integrative literature review, three key databases were referred to, namely Scopus, Web of Science and Directory of Open Access Journals. Searches combined turnover terms (e.g. turnover, quit, resign, attrition, retention) with leadership terms (e.g. leader, supervisor, manager, LMX, leadership) and organisational terms (e.g. salary, compensation, work design, workload, JD-R, career progression). As the search was conducted iteratively, the exact database-specific search strings were not archived.  Appendix 1 reports the keyword blocks used and screening logic applied.

The review timeline for this study is between 2014-2025 to identify developments before COVID-19 and over the past five years since it has changed the workplace environment. The findings were also compared descriptively across two periods, i.e. 2014–2019 and 2020–2025, noting shifts in constructs and mechanisms rather than effect sizes, given the diversity of measures across studies. Inclusion criteria include peer-reviewed journal articles in English conducted in organisational settings that examined leadership behaviours in relation to turnover. Exclusion criteria include non-organisational settings and non-peer-reviewed sources.

Based on these criteria, records were screened in two stages, beginning with title and abstract screening to remove irrelevant items. This was followed by a full-text assessment for fit with the research questions  Appendix 1. The final set comprised 39 studies, and formal quality scoring tool was not applied due to diversity of research designs adopted across the literature. Instead, methodological credibility was considered during extraction by noting clarity of design, transparency of analysis, and whether conclusions were supported by reported findings, as outlined in  Appendix 2. The PRISMA diagram for this study is presented in  Appendix 3.

From each study, the parameters recorded include author, publication year, sampling, design, key findings and limitations. For synthesis, a deductive and inductive approach was applied. Findings were organised into two areas, resources and demands. The synthesis involved an initial deductive process using JD-R and LMX as organising constructs. This involves categorising extracted findings as demand- or resource-related and noting whether leadership operated through LMX quality, support or demand increase. Inductive themes were then identified through manual coding of extracted findings. This involved clustering recurring patterns of predictors that appeared across multiple studies. Labels such as “leadership behaviours” emerged when multiple studies described supervisory actions (e.g. micromanagement, abuse of power, lack of support) as factors of turnover intentions. Similarly, “organisational constraints” emerged when studies pointed to structural conditions (e.g. compensation, workload, organisational change) as dominant drivers that constrained leadership effects. Emergent themes are discussed in the next section.

From the integrative review of conducted, the extent to which leadership behaviours influence employee turnover was analysed in comparison to organisational factors. Several themes were identified to understand the link between leadership and turnover. In this section, three key themes are discussed, namely the role of leadership in employee turnover, managerial behaviours influencing turnover intentions and the relative impact of organisational and leadership factors. The discussions were interpreted through the lenses of LMX and JD-R.

A minor descriptive split was noted across the reviewed papers from 2014-2019 and 2020-2025. This indicates a marginal shift in emphasis within the literature. Studies in the 2014–2019 subset primarily model leadership as a predictor of turnover intentions via attitudinal mechanisms (e.g. commitment, job satisfaction). These findings treat organisational context as background condition rather than central factor (Lim et al., 2017; Waldman et al., 2015). Studies from 2020-2025 foreground how leadership effects operate through (or constrained by) demand and strain conditions. These include workload pressures, burnout, emotional labour, uncertainty and work design (Buss and Kearney, 2024; Oh and Chhinzer, 2021; Puhakka et al., 2021). This pattern aligns with the overarching JD-R framing where leadership contributes to resource and demand balance.

Across the reviewed studies, leadership constructs are linked to turnover in two ways: increasing or eroding job resources and shaping job demands. Among the variety of leadership approaches, transformational leadership is often linked to reduced turnover intentions, whereas toxic or abusive leadership is associated with higher attrition.

Transformational leadership is an approach whereby managers seek to inspire and mentor team members in enhancing autonomy and competence which can raise job satisfaction and commitment (Cao et al., 2025; Yusoff, 2025). These traits can reduce turnover intentions as employees who experience this approach exhibit strong loyalty from increased trust, improved capability and better perceived work–life balance (AbdELhay et al., 2025; Goens and Giannotti, 2024). The literature points to transformational leadership as a moderator to turnover intentions given the positive impact that it generates in the leader–follower relationship (Chandra et al., 2023; Kaymakcı et al., 2022; Oh and Chhinzer, 2021; Waldman et al., 2015). In contrast, transactional leadership which keeps engagement only at the surface level is associated with higher turnover especially when hindrances such as corporate bureaucracy or lengthy processes are high (Lim et al., 2017; Prentice, 2022; Waldman et al., 2015). This approach is commonly demonstrated by managers who focus solely on task execution with minimal effort to foster constructive relationships with employees.

Another leadership style linked to positive work environments is servant leadership where leaders prioritise the development of staff contributing to stronger work engagement (Ozturk et al., 2021). The literature indicates that servant leadership is imperative to influencing positive outcomes both in terms of operational productivity and staff well-being (De Clerck et al., 2024; Khanam and Tarab, 2024; Wibowo and Paramita, 2022). A similar approach is ethical leadership whereby fairness and integrity are strongly fostered, consequently lowering turnover intention and actual decisions (Choi, 2025; Ustun et al., 2024). Several findings from the literature highlight how ethical leadership reduces turnover while moderating the effects of resignation (Choi, 2025; Ustun et al., 2024). It fosters a climate that strengthens retention, although its impact depends on follower moral identity and strengthened by behaviours of leaders (Marquardt et al., 2022; Yasin et al., 2021). A summary of the leadership styles discussed in this section is outlined in Table 1.

It needs to be highlighted, however, that the literature indicates how the moderating impact of leadership on turnover decision can be industry specific. Healthcare and education are two industries that stand out in terms of correlation between leadership and turnover in view of job demand (Schildhouse and Kim, 2022; Zhang and Zhang, 2024). In these roles, the baseline demands on workload and emotional labour are high. Similarly, cultural expectations can also diversify the way leadership affects turnovers. In collectivist cultures, the tolerance of poor leadership is stronger when compared with individualistic cultures where career growth is prioritised more significantly (Waldman et al., 2015; Wibowo and Paramita, 2022). These are specificities that can inform future research in contextualising any empirical study to industrial or cultural aspects.

Although leadership styles can shape the overall culture of a team, direct managerial behaviours are often cited as a leading cause of voluntary resignations. Poor managerial conducts ranging from micromanagement to abusive supervision significantly influence employee turnover (Griep et al., 2025). Several studies highlight the detrimental effects of toxic leadership, including authoritarianism, lack of empathy and favouritism. These experiences get exacerbated when uncertainties come into the picture (Buss and Kearney, 2024; Lopes et al., 2025; Potipiroon and Chumphong, 2024). As a result, employees who experience such leadership often report increased burnout, emotional labour and intentions to leave (Nunes and Palma-Moreira, 2024; Zhang and Zhang, 2024). Furthermore, as employees expect mentorship from managers as part of their role, the absence of guidance and career support would lead to disengagement (Kanchana and Jayathilaka, 2023; Prentice, 2022).

Studies within the literature also suggest that when employees feel ignored or unappreciated by direct supervisors, which can affect their overall sense of belonging, they are likely to seek employment elsewhere (Islam and Amin, 2022). A summary of the managerial behaviours discussed in this section is outlined in Table 2. These indications might sound elementary. Nevertheless, the evidence from the literature signals how an understanding of why these long-standing issues continue to occur is needed, especially as the workplace becomes more diverse and dynamic.

Although much of the literature models turnover as an individual decision, several studies indicate that it is socially embedded within teams. In particular, turnover can become contagious, where resignations by colleagues increase the perceived feasibility of leaving. This can amplify exit decisions beyond individual job attitudes alone (Oh and Chhinzer, 2021), and aligns with the LMX lens as leaders do not develop uniform relationships with all team members. LMX differentiation can trigger social comparisons and perceptions of unfairness, which intensifies withdrawal (Dansereau et al., 1975). This suggests how the “people leave managers” stance operates through dyadic relationship quality and the wider team climate that shapes exit norms.

Several mitigating strategies are discussed in this literature in addressing the challenges of adverse leadership. Leaders who adopt coaching behaviour can build stronger empathy in understanding their direct reports, consequently leading to stronger commitment (Lim et al., 2017; Wesemann, 2024). In JD-R terms, coaching-oriented behaviours described in these studies operate by strengthening job resources to buffer the strain effects of high workload. Coaching can also reduce perceived demands by enabling earlier clarification of priorities reduce stress driven by uncertainty.

Furthermore, coaching trains leaders to adopt active listening which is a skill that is not inherent to many but can be acquired with practical applications. Another potential action is the empowerment of employees which can be done via participative decision-making and inclusive policies (Alhajaj and Ahmad, 2024; Ustun et al., 2024). Empowerment, however, requires strong trust levels between leaders and followers. Autonomy granted to employees requires clear and robust guidelines so that governance could be upheld while innovation and process improvement could still be encouraged. Organisational readiness is critical in enabling these intentions.

Although leadership is a critical factor in influencing turnover decisions, organisational factors such as compensation, career growth and work–life balance retain significant roles in inducing turnover. Studies indicate that compensation structure remains a primary cause of turnover as financial incentives remain a key driver of retention (Croes et al., 2025; Wesemann, 2024). The literature suggests that compensation is a factor that influences a sense of empowerment among employees, which also suggests how both psychological and financial resources play a critical role in mitigating turnover and enhancing retention (Hong et al., 2025; Li et al., 2025). Given the priorities of modern-day employees, the mechanisms for rewards in organisations have also changed. An example is the development of compensation to grow from salary remuneration to also include other aspects of benefits that are conventional (e.g. medical insurance, leave allowances) and modern (e.g. flexible working, career-break programmes).

In terms of professional development, employees are more likely to leave if they perceive limited career growth opportunities. This aspect can be a function of the performance management system in place as well as the sense of competence in the work that they do as employees (Puhakka et al., 2021; Qureshi et al., 2024). At the organisational level, a collaborative performance management framework can also empower staff and foster career growth, while doing the opposite could encourage them to look for better stability in other companies (Alhajaj and Ahmad, 2024; Qureshi et al., 2024). These findings suggest that development opportunities shape retention through motivation and job autonomy. This relates to the purpose and meaning of their respective roles. Furthermore, organisational changes, either by structure or by hierarchy, can also cause uncertainties among employees. This may lead to turnover, especially when coupled with increasing workload and limited support (Bolt et al., 2025; Van Der Baan et al., 2025). These organisational factors show how structural elements can influence turnover decisions, even if leadership styles are favourable. A summary of the organisational influences discussed in this section is outlined in Table 3.

Based on the three themes discussed, leadership is a meaningful influence on turnover, but its relative weight varies and is often constrained by structural features. In addition to leadership style, managerial behaviours also significantly affect staff retention. These aspects co-exist with prevailing factors that are driven by the organisation. The coexistence of these elements increases the complexity of how the root causes of turnover could be tracked down. The subjectivity of these factors, and the minimal control that management has over the daily impact of these aspects, further complicates the ability for organisations to fully grapple with the challenge. In the next section, the findings from the literature are discussed in relation to leadership and HR strategies.

In response to the research questions of this study, the findings from this integrative review reveal several insights.

Firstly, transformational and ethical leadership styles are associated with lower turnover intentions and behaviour. This reinforces the notion that leadership quality has a meaningful impact on employees’ decisions to resign or remain, as asserted by LMX theory. Secondly, although poor leadership traits contribute to turnover decisions, they rarely operate in isolation. Organisational factors such as work–life balance, compensation structure and career growth opportunities also shape employee decisions to stay or leave. When this occurs, leadership effects become weaker or more indirect if these structural conditions are unfavourable (Croes et al., 2025; Hong et al., 2025; Puhakka et al., 2021). A smaller subset of studies also indicates that turnover can be shaped by collective dynamics, which reinforces that turnover can also be amplified by the positioning taken at the team level.

In JD-R terms, leadership primarily works by adding resources and shaping demands. However, when organisational factors are significantly high or structural resources are scarce, the effects from leadership weaken. Therefore, leadership behaviours may not always mitigate turnover. Under demanding work situations or toxic organisational cultures, even supportive leadership may fail to prevent employee attrition. Collectively, the evidence challenges the simplified view that “people leave managers, not organisations”. Consequently, the evidence supports a more precise claim whereby people leave when work demands consistently exceed resources. This assertion aligns with the core principle of JD-R and specifies how LMX quality operates as a mechanism that shapes resources and demands. It also clarifies when leadership effects are likely to weaken. The causes for this situation could vary from the leadership style of managers, issues with the existing organisational systems, processes and structures. From this study, it is indicative that the intention and decision for turnover is an amalgamation of both leadership and organisation factors.

Sector context should also be treated as a boundary condition in interpreting leadership effects on turnover. Across the reviewed literature, leadership operates as a resource mechanism, but its retention value weakens when the demand–resource balance shifts. This is especially salient in higher demand sectors (e.g. healthcare, hospitality, education) where workload intensity and emotional labour can dominate turnover decisions. Cultural context is also a boundary condition. Norms relating to authority and mobility can shape how employees interpret LMX quality. Therefore, the suggested framework is most generalisable at the level where demand–resource imbalance is shaped by LMX, which can vary by sector and culture.

The findings contribute to leadership and turnover research by positioning LMX within the JD-R construct rather than treating leadership as a standalone variable. This review advances theory in three ways. Firstly, it positions LMX within the JD-R architecture. Secondly, it synthesises the reviewed evidence into two explanatory pathways (i.e. building resource, intensifying demand). Thirdly, it specifies a boundary condition that helps reconcile findings in turnover literature, namely how retention value of leadership weakens when structural demands or resources change.

Although LMX theory suggests that high-quality exchanges improve retention, the review indicates that leadership quality alone is necessary but insufficient. Leadership can moderate and mediate the effects of workplace features. This is done by buffering the impact of workload strain (e.g. better clarity, moral support) or intensifying demands (e.g. unfair allocation, limiting resources). This finding resurfaces the need to reframe leadership in a contextualised manner in its role of influencing turnover .

The findings also align with the broader JD-R model whereby leadership can protect employees against excessive stress. That said, leadership effectiveness is constrained when structural demands remain unaddressed. In such cases, JD-R predicts limited marginal gains from leader behaviour as it gets overpowered by the wider organisational ecosystem. This reinforces the premise that retention strategies must balance leader-focused interventions with structural redesign.

Taking a broader perspective, similar patterns can also be interpreted through Conservation of Resources theory, where turnover risk increases when employees experience resource loss (or gain) in the face of sustained demands (Chen et al., 2024; Kim et al., 2025). Similarly, job embeddedness theory further explains why some employees stay despite demands, particularly when links, fit and perceived sacrifice are high (Hassan et al., 2024; Sender and Eberly, 2025). These lenses are consistent with the LMX and JD-R framing of this study.

In terms of practice, the value of this review lies in its sequenced guidance that links HR levers and leader behaviours to the underlying demand–resource imbalance. Retention efforts should begin by diagnosing and addressing structural demands and resources where feasible (e.g. workload, scheduling, remuneration, growth). Where structural change is slow or constrained, leadership interventions can run in parallel as near-term levers by developing leaders as reliable resource providers. This includes coaching to strengthen support and role clarity. It can also avoid unproductive practices such as arbitrary tasking or last-minute changes. Organisations can also monitor and intervene proactively by tracking attrition to detect emerging toxic patterns and pairing these signals with targeted corrective actions.

These insights should be applied with contextualisation to sectoral needs, particularly in high-demand environments. In these scenarios, workload, staffing and scheduling pressures can dominate turnover decisions. Furthermore, implementation is often constrained by budgets and longer timelines, creating trade-offs between immediate delivery and redesign. A pragmatic approach is simultaneously pursuing targeted structural improvements while developing leaders to reduce demand intensification. This avoids the expectation for leaders to compensate structural problems they cannot control.

Although this review synthesises evidence across multiple industries and settings, sector and culture remain as boundary conditions that shape how leadership translates into retention. The findings are best interpreted as mechanisms that show how LMX quality shapes resources and demands. Thus, future research should test these mechanisms across sectors with high demands and across cultural contexts where leadership norms and expectations differ.

This integrative review challenges the conventional wisdom that “people leave managers, not organisations” by showing that while leadership significantly influences turnover intentions, its effects are conditioned by organisational demands and resources. Leadership quality is important, but it is not a silver bullet as leader effects become less significant when structural demands are high or resources are scarce.

From a theoretical standpoint, the review clarifies when leadership acts as a primary mechanism and when structural conditions dominate. This reframes the leadership factor as context-sensitive rather than a standalone cause. This contextualised view moves beyond the either–or accounts of turnover decisions and specifies the boundary conditions under which leadership and organisational levers interact.

From a practical standpoint, the findings support retention strategies that address structural constraints alongside developing leaders who avoid demand intensifying practices. In this regard, transparent and fair policies set the foundation on which effective leadership can work. Ultimately, organisations should treat leadership and organisational interventions as complementary instead of competing forces, given their counteracting influences. The goal is a well-designed system where good leadership thrives within strong organisational practices.

This research did not receive any specific grant from funding agencies in the public, commercial, or non-profit sectors.

Microsoft Copilot was used in editing this manuscript.

Databases searched: Scopus; Web of Science; Directory of Open Access Journals.

Time window and limits: 2014–2025 (inclusive), English language, peer-reviewed journal articles.

Keyword blocks used (iterative search)

Searches combined keyword blocks, iteratively refined during screening for relevance to research questions:

• Turnover terms (action words): turnover, quit, resign, attrition, retention.

• Leadership terms (leader nouns): leader, supervisor, manager, leader–member exchange, leadership.

• Organisational terms: salary, compensation, work design, workload, job demands–resources, career progression.

Screening process

• Stage 1: Title and abstract screening to remove clearly irrelevant items.

• Stage 2: Full-text assessment for fit with the two research questions (leadership influences on turnover and comparison with organisational factors).

Inclusion criteria

• Peer-reviewed journal articles, English language

• Organisational employment settings

• Examined leadership or managerial behaviours and/or organisational factors in relation to turnover.

Exclusion criteria

• Non-organisational settings.

• Non-peer-reviewed sources.

A formal quality scoring tool was not applied due to heterogeneity in research designs across the included studies. Instead, methodological credibility was considered during extraction using the following criteria:

• Design clarity: the study aim, setting, sample and design were described clearly enough to interpret the findings.

• Transparency of measures and analysis: key constructs (turnover outcome and predictors) and analytic approach were reported with adequate transparency (including reliability/validity information where relevant).

• Conclusions supported by results: conclusions were clearly grounded in the reported findings and limitations were acknowledged where appropriate.

Note: These criteria were used to guide interpretation and weighting of evidence in the narrative synthesis, rather than to exclude studies.