Predictors and effects of hospital chief executive officer turnover: A systematic review

Abstract

Background

Chief executive officer (CEO) turnover is especially frequent in hospitals and represents a critical organizational event, yet its predictors and effects remain poorly understood.

Purpose

We conducted a systematic review of the empirical literature on the predictors and effects of hospital CEO turnover worldwide to synthesize and assess the multiple findings scattered across studies.

Methodology

In this systematic review, 30 empirical studies published between 1987 and 2024 were identified from three databases: Business Source Complete, MEDLINE, and APA PsycInfo. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

Results

We found that the methodological quality of the studies was highly variable. Furthermore, we identified 46 unique predictors of hospital CEO turnover, including organizational, environmental, and personal characteristics, as well as characteristics related to prior performance. The findings regarding the effects of CEO turnover suggest that it can result in a temporary decline in financial performance and an elevated risk of organizational failure for hospitals.

Practice Implications

Our results underscore that, in hospitals with a higher likelihood of CEO turnover, early and systematic succession planning is crucial to increase leadership stability, reduce recruitment costs, and ensure organizational resilience.

Chief executive officers (CEOs) hold the highest-ranking administrative position in organizations, representing them externally and bearing overall accountability for their economic success. The main res-ponsibilities of a CEO are to develop and implement business strategies, lead the management team, shape organizational culture, and oversee business operations and regulatory compliance (Wilson & Stranahan, 2000). Given the scope of a CEO’s decisions and actions, CEO turnover constitutes a critical event often marked by profound organizational change (Nyberg et al., 2021).

Organizations frequently initiate CEO turnover in response to poor financial performance (Ballantine et al., 2008), viewing it as an opportunity for organizational turnaround (Eldenburg et al., 2004). Indeed, a new CEO can bring fresh perspectives, ideas, and skills essential for addressing current and future challenges, potentially steering the organization toward economic success (Berns & Klarner, 2017; Haveman et al., 2001). However, CEO turnover also carries risks, including the possibility of radical shifts in strategy that may erode stakeholder trust, lead to conflict among employees, and increase turnover in other leadership positions, exacerbating the leadership vacuum (Khaliq et al., 2006; Mosadeghrad et al., 2013). Such dynamics can impair financial performance and amplify organizational instability, potentially leading to further CEO turnovers and creating a vicious cycle (Schepker et al., 2017).

Given the far-reaching implications of CEO turnover and its increasing rate worldwide in recent decades (PwC, 2019), it is unsurprising that the topic has attracted considerable research attention. Several literature reviews have examined related theories, causes, and effects, with their findings emphasizing the role of contingencies and developing comprehensive frameworks (Berns & Klarner, 2017; Nyberg et al., 2021; Schepker et al., 2017). None of these reviews, however, have focused on specific industries. This omission is particularly unfortunate in the case of the hospital industry, which represents a unique environment, making the findings from existing reviews less likely to be fully generalizable. Hospitals are complex, highly regulated organizations that provide a wide range of medical services, employ specialized staff across multiple disciplines, and operate under diverse ownership structures. They also face the challenge of maintaining high-quality care within economic constraints that are often exceptionally tight (Bertrand et al., 2005; Schneider et al., 2020).

These industry-specific differences are further evident in CEO turnover patterns, for which some evidence suggests that hospital CEOs may have shorter tenures compared to CEOs in other industries. For example, the average tenure of hospital CEOs in the United States is approximately 5 years (Khaliq et al., 2006), which is considerably lower than the U.S. average across all industries of 7 years (Korn Ferry, 2020). In Germany, a recent study showed an annual turnover rate of approximately 25% among hospital CEOs from 2011 to 2021, with large variations in tenure lengths (BDO & DKI, 2021).

Individual studies on hospital CEO turnover have attempted to take these unique characteristics into account, examining industry-specific factors such as community involvement, ownership type, and quality of care (Janke et al., 2019; Mathew et al., 2024a; Potter & Dowd, 2003). Other studies have adapted theoretical explanatory models to reflect industry-specific factors (e.g., Weil & Kimball, 1995), with their findings suggesting that CEO turnover patterns may vary based on industry dynamics.

Overall, studies that have investigated the predictors and effects of CEO turnover in the hospital industry have varied widely in terms of their methods, variables, designs, and methodological quality, raising questions about the consistency and validity of their findings. Additionally, the lack of an overview of this research area has left the findings scattered across individual studies, making it more difficult for decision-makers to develop evidence-based strategies for succession planning. Our systematic literature review therefore aims to fill this gap by providing a comprehensive review of the empirical literature on the predictors and effects of CEO turnover in hospitals. In doing so, it will also assess the methodological quality of the identified studies to determine the reliability and validity of their findings. Such an assessment is crucial because most previous studies offer only limited scope for causal inference—a key requirement for evidence-based decision-making in managing CEO turnover in practice and for furthering research in this area.

Theory and Concept

CEO turnover is a complex event with far-reaching effects on the organization and its environment. Accordingly, researchers from various disciplines, including strategic management, organizational behavior, and finance, have examined this topic and applied a variety of theoretical approaches to categorize and explain its causes and effects (for an overview, see Giambatista et al., 2005). For example, theories such as resource dependence theory (Pfeffer & Salancik, 1978), job match theory (Allgood & Farrell, 2003), and organizational life cycle theory (Cragun et al., 2016) conceptualize CEO turnover as an adaptive process that can help achieve a better fit between a CEO’s profile and new demands triggered by changes in the organization or its environment. Other theoretical perspectives include the theory of relative standing and agency theory, which explain CEO dismissals as a punitive response to poor financial performance (Bilgili et al., 2017). In contrast, scapegoat theory posits that financial performance is beyond CEOs’ control, and their dismissal primarily serves to appease shareholders (Rowe et al., 2005). Lastly, the circulation of power theory suggests that the influence a CEO wields within an organization diminishes with increasing tenure and age, thereby increasing the likelihood of turnover (Ocasio, 1994).

Regarding the effects of CEO turnover, common sense theory suggests that new CEOs bring fresh perspectives and abilities to the organization, leading to a positive impact on performance (Giambatista et al., 2005). Conversely, CEO turnover can also cause major disruptions in the organization, which, according to the theory of the vicious cycle, can trigger a downward spiral of organizational instability, poor performance, and further CEO turnover (Grusky, 1960; Schepker et al., 2017).

Beyond the general direction of the determinants and effects of CEO turnover, theories have also been applied to make more differentiated explanations. For example, upper echelons theory has been used to explain differences in the strategic actions and outcomes of CEO turnover based on the individual characteristics of the CEO (Hambrick, 2007). In turn, contingency theory has been drawn upon to posit that there are a variety of situational factors that influence both the causes and consequences of CEO turnover (Giambatista et al., 2005).

Overall, CEO turnover theories can be described as fragmented and ambiguous. Many are used to explain different types of relationships, and some even contradict each other regarding cause-and-effect mechanisms. However, the prevailing view in recent years is that most of the theories are valid in their respective contexts (Giambatista et al., 2005; Schepker et al., 2017). As a conceptual framework for this study, we followed Berns and Klarner (2017) and posit broadly that:

1. there are various predictors of hospital CEO turnover, which can be subsumed under the categories of environment, organization, and personal characteristics of the CEOs;

2. the outcomes of hospital CEO turnover are the result of strategic choices (reflecting, in particular, changes of strategy) and hospital performance; and

3. the links between predictors and CEO turnover and between CEO turnover and outcomes are subject to contingencies, which again comprise factors related to the environment, the organization, and the personal characteristics of the CEOs.

Methods

Design

Our systematic review followed the guidelines set out in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. The PRISMA statement includes a 27-item checklist, which is designed to ensure transparent, complete, and accurate reporting in systematic reviews (Page et al., 2021).

Search Strategy

We developed the following algorithm-based search strategy, which included several key words and synonyms derived from existing reviews of the literature on executive turnover and succession (Berns & Klarner, 2017; Nyberg et al., 2021), but with additional minor adjustments for the hospital sector: (“ceo” OR “chief executive” OR “executive director”) AND (turnover OR succession OR change OR dismiss* OR replac* OR transition OR appoint* OR terminat* OR quit* OR leav* OR retention OR tenure) AND (hospital OR “medical center” OR “medical facilit*” OR “health* system” OR “health* facilit*” OR “health* center” OR “secondary care” OR “tertiary care” OR “care facilit*” OR “care center”). We subsequently used the algorithm to search titles, abstracts, key words, and related concepts via EBSCOhost in the following databases: Business Source Premier, APA PsychINFO, and MEDLINE. The search was conducted on September 13, 2024. Lastly, we conducted backward-looking and forward-looking citation searches of included publications using Google Scholar on September 16, 2024.

Eligibility Criteria

Our review only included publications if they met all of the following inclusion criteria:

• (a) Publication type: Publications had to be research articles published in peer-reviewed journals in order to ensure a minimum level of methodological quality.
• (b) Language and year of publication: The articles had to be published in English between 1980 and 2023. We chose this period because the early 1980s marked the beginning of research on hospital CEO turnover (Dwore & Murray, 1996).
• (c) Thematic relevance: Articles had to explore the predictors or effects of CEO turnover in hospitals.
• (d) Study design: Articles had to report the results of quantitative empirical studies or qualitative empirical studies. We excluded commentaries, reviews, and theoretical analyses in order to focus exclusively on empirical results.
• (e) Study population: To ensure that our analysis focused on the relevant population, articles had to report the results of studies whose sample consisted of secondary data from hospitals (of any type and ownership) or data from surveys or qualitative interviews of hospital CEOs. These could also include articles reporting the results of intersectoral studies that analyzed hospitals or hospital CEOs in one or more subsamples.

Study Selection

Study selection followed a three-stage process as seen in Figure 1. First, the bibliographic information from the records identified in the algorithm-based search was screened to determine whether the records met the inclusion criteria related to publication type, language, and year of publication. Second, the remaining titles and abstracts were evaluated for thematic relevance, study design, and study population. Third, full-text reports of the remaining records were sought for retrieval and assessed for eligibility. The three stages were performed independently by two of the authors (MH and VW). At regular intervals during each stage, disagreements were discussed and resolved bilaterally through discussion.

Figure 1.

PRISMA flow chart of study selection.

In total, the systematic literature search yielded 4,711 records. Out of these, 3,069 were excluded in the first stage and an additional 1,541 records were excluded in the subsequent title and abstract screening. The remaining 101 records were retrieved and screened as full-text reports. Among these, 74 were eliminated because they lacked an empirical approach (n = 29), lacked thematic relevance (n = 40), focused on other executive levels (e.g., ward management, chief nursing officer; n = 3), or were redundant (n = 2) because they analyzed a subsample of a sample in other identified articles and did not generate additional findings relevant to this review (see Table 1, Supplemental Digital Content, http://links.lww.com/HCMR/A170).

In addition, the backward and forward citation searches identified two peer-reviewed articles that met our inclusion criteria. These searches also identified one recent working paper that met all of our inclusion criteria apart from the requirement to be peer-reviewed. After thoroughly reviewing the methodological quality of the paper using the Quality Assessment for Diverse Studies (QuADS) tool developed by Harrison et al. (2021) and determining it to be of high quality, the two authors conducting the search mutually agreed to include it. Therefore, a total of 30 articles were ultimately included in the review, each of which reported the results of one study.

Data Extraction

We designed a data extraction sheet in Microsoft Excel, which two authors (MH and VW) used to extract data independently from the included studies. Extracted data were compared at regular intervals to ensure accuracy, with any discrepancies being resolved bilaterally through discussion. The following data were extracted: author names, publication year, data (type and sources), study design and analyses (type and methods), setting (country, observation period, number of hospitals or CEOs), theories used to explain hypothesized or identified relationships between predictors, contingencies and effects and CEO turnover, and the main results regarding the predictors and effects of hospital CEO turnover. The type of data extracted for the main results varied by study type: In the case of quantitative inferential studies examining secondary data from hospitals, the direction of coefficients and statistical significance were extracted. For quantitative descriptive studies involving surveys of hospital CEOs, information on the frequency of the three most frequent responses was extracted. For qualitative studies involving interviews of hospital CEOs, information on the frequency of responses was extracted or, if no aggregated information was available, a condensed summary of the main findings was created and recorded in the extraction sheet. In instances where data were not reported, we made estimates where feasible. For example, in some studies, we estimated the number of CEO turnovers using the number of observed hospitals and information on turnover rates (e.g., Weil, 1990). See Table 1 in the Supplemental Digital Content, http://links.lww.com/HCMR/A170, for a comprehensive summary of the extracted data.

Data Synthesis

After data extraction was completed, two of the authors (MH and VW) jointly categorized the predictors of CEO turnover reported in the quantitative inferential studies through a three-step process. First, they deductively derived four thematic blocks of factors that have been recognized as predictors of CEO turnover in previous research (Berns & Klarner, 2017). These comprise organizational, environmental, and personal characteristics of the CEO, as well as characteristics related to the hospital’s prior performance. Prior performance was extracted as a distinct dimension from the organizational characteristics due to its dynamic nature and its particular relevance to CEO turnover, as reported in the literature (e.g., Eldenburg et al., 2004; Mathew et al., 2024b). In the second step, all identified predictors were assigned to one of the thematic blocks. In the third step, the authors developed subcategories for the predictors in each thematic block based on thematic similarity using an inductive, iterative process. This final step was performed by each of the authors individually. Subsequently, the subcategories and assigned predictors were compared, with any discrepancies being resolved bilaterally through discussion.

The two authors also categorized the effects of CEO turnover that had been reported in the included studies. Almost all of the quantitative inferential studies focused on the effects of CEO turnover on hospital performance. Following the same procedure as that used to categorize the predictors of CEO turnover, the following categories were formed: financial performance, quality of care, and other organizational performance measures such as the number of day cases and staff satisfaction. Furthermore, based on the results of the quantitative descriptive and the qualitative studies, the two categories “organizational activities” and “competitor behavior” were added.

Quality Assessment

We assessed the methodological and reporting quality of the included studies using a modified version of the QuADS tool originally developed by Harrison et al. (2021). Because a large proportion of the studies used secondary data, we made some modifications to the assessment criteria. For example, we did not assess the extent to which the studies reported stakeholder involvement in the development and conduct of the studies, as this would be unusual for studies based on secondary data. Our adapted scale comprised 11 criteria grouped into four dimensions, as follows: concept and design of the study (Items 1–4), transparency of methodological procedures (Items 5–7), methodological appropriateness (Items 8–10), and limitations (Item 11). The final assessment template can be found in Table 2 in the Supplemental Digital Content, http://links.lww.com/HCMR/A170. For each criterion, studies were assigned a score between 0 and 3, with 3 indicating the highest quality ranking.

The methodological and reporting quality of each study was assessed independently by two of the authors (MH and VW). Disagreements in the assessment were discussed and resolved bilaterally through discussion. Based on its total score, each study was grouped into one of the following three categories to provide a basic indication of overall quality: low (score <60%), moderate (score ≥60% but ≤80%), and high (score >80%). When reporting the results of our systematic review, we emphasize the results of studies with high quality ratings but do not exclude studies with low quality ratings because these also provide important insights and excluding them would contradict the concept of QuADS (Harrison et al., 2021).

Results

Characteristics of the Included Studies

Table 1 summarizes the characteristics of the 30 studies included in our systematic review. The majority of studies were published before 2000 (n = 12) or in the early 2000s (n = 11), whereas only seven studies were published between 2011 and September 2024. Thirteen of the studies were published in journals related to health care management and health care administration. The remaining studies were published in journals with a different focus, such as hospital and medicine (n = 6), general management (n = 7), and other (n = 3).

TABLE 1.

Characteristics of hospital CEO turnover literature

Literature characteristics	No. of articles	% of articles
Publication period
1987–1999	12	40
2000–2010	11	37
2011–2024	7	23
Journal typea
Health care management and health care administration	13	45
Hospital or medical	6	21
General management (including finance, administration, organizational behavior)	7	24
Socioecological, law, and labor economics	3	10
Data source
Mixed	2	7
Secondary data	18	60
Primary data	10	33
• Survey	6
• Interview	3
• Mixed	1
Country
United States	24	80
Other countries	6	20
Research design
Longitudinal	18	60
Cross-sectional	12	40
Research type and method of analysis
Descriptive	11	37
• Quantitative descriptive (e.g., number and distribution of responses)	8
• Qualitative descriptive (e.g., responses of interviewees)	3
Analytical	19	63
• Analysis of variance and hypothesis tests (e.g., Pearson chi-square)	2
• Basic (multiple) regression models (e.g., linear or logistic or probit regression)	10
• Elaborate regression models (e.g., DiD estimator, survival estimators, fixed or random effects, matching)	7
Content
Predictors	16	53
Effects	8	27
Both	6	20

^a The working paper was not included in this element.

Overall, most studies analyzed secondary data (n = 18) and had a longitudinal design (n = 18). Most of the studies used data from the United States (n = 24). Of the remaining six studies, two used data from the United Kingdom, two used data from Iran, and two used data from Australia. In terms of their research focus, 16 of the studies focused on predictors, eight focused on effects, and six focused on both. With regard to study type, the included studies were dominated by quantitative inferential approaches (n = 19), followed by quantitative descriptive (n = 8) and qualitative approaches (n = 3). Among these, 19 studies used inferential analyses. The remaining 11 studies relied solely on descriptive methods, reporting the distribution of responses.

Regarding methods of analysis, two of the quantitative inferential studies used basic methods, such as Pearson chi-square or analysis of variance, to analyze variation in their data. The remaining 17 studies estimated the predictors and effects with regression models, comprising studies that used basic regression models (n = 10), including ordinary least squares, logit or probit regressions with or without lagged independent variables, and studies that used sophisticated analytic methods (n = 7), such as difference-in-differences estimators, survival estimators, or regression models with fixed or random effects following a matching approach.

The distribution of analytical methods differed by research topic. For instance, with the exception of one survival analysis and one random effects regression, studies on predictors of CEO turnover exclusively used analysis of variance and basic regressions, which were dominated by logistic regressions with lagged independent variables. In contrast, studies dealing with the effects of CEO turnover (n = 8) or with both predictors and effects (n = 6) comprised seven of the 11 noninferential studies. They also accounted for five of the seven studies that used sophisticated analytic methods, including two survival analyses, two quasi-experiments with matching, and one multiple regression on pooled time series data.

Six of the quantitative descriptive studies involved surveys of hospital CEOs, whereas three of the qualitative studies involved semistructured or structured interviews of hospital CEOs. These nine studies were dominated by frequency tables presenting data from responses to closed-ended or open-ended questions (with grouped similar responses).

Quality Assessment

An overview of the quality assessment for each study with respect to each criterion and the overall rating is presented in the Table 2 in the Supplemental Digital Content, http://links.lww.com/HCMR/A170. In terms of overall methodological and reporting quality, we rated 10 of the included studies as low, 14 as moderate, and six as high. The average total quality score was 21.2 out of a maximum of 33, which was in the lower range of a moderate rating. The studies received an average of 1.9 points per item across all dimensions. Regarding the concept and design of the studies, the studies performed well in reporting the theoretical or conceptual underpinning of the research and the statement of research aims, as well as in choosing an appropriate study design, with an average rating of 2.2 in this dimension. Quantitative descriptive and qualitative studies were rated below average because the research aims were often only briefly stated without providing a theoretical grounding or adequately explaining the current state of research (e.g., Dwore & Murray, 1996; Wilson & Meadors, 1990). With an average score of 2.1, the studies scored well in terms of the transparency of their methodological procedures. Within this dimension, the studies scored lowest with regard to describing the data collection procedure, receiving an average of 1.8 points. Regarding the appropriateness of the methodology, the included studies achieved a modest average score of 1.8. In studies using secondary data, the representativeness of the sample was, in most cases, insufficiently addressed or not discussed at all. In general, the quality of reporting increased with the complexity of the analysis methods used. Finally, the included studies performed worst in critically addressing their own limitations, with an average score of 0.8.

Theoretical and Conceptional Frameworks of Hospital CEO Turnover

The included studies varied substantially in the extent to which they drew upon theoretical and conceptual frameworks, as well as in the specific theoretical and conceptual frameworks they applied (if any). Thirteen of the included studies did not draw on any theory. Two studies that examined both the predictors and effects of CEO turnover considered theoretical frameworks only for predictor or only for effects. Seven studies based their rationale on key terms closely linked to theoretical concepts (e.g., agency costs) without explicitly referring to the corresponding theories. Eight studies provided explicit theoretical foundations for the relationships examined.

The diversity of theories used and their content is comparable to that of studies on the causes and effects of CEO turnover in the general management literature. However, institutional theories were considered somewhat more frequently in the studies included in our systematic review, probably due to the unique set of ownership types in the hospital sector. For an overview of the relationships examined and the theories applied to them, see Table 1 in the Supplemental Digital Content, http://links.lww.com/HCMR/A170.

Theoretical Explanations for the Reasons of CEO Turnover

Of the 22 studies investigating predictors of CEO turnover, 10 provided theoretical explanations for the observed relationships. Institutional theories were applied in four studies, and agency theory and organizational life cycle theory were each applied in two studies. The remaining two studies focused their theoretical considerations on a model of voluntary turnover and the resource-based view. As some studies examined multiple predictors, the punctuated equilibrium model, person–organization fit theory and leadership instability theory were each used once to explain additional predictors of CEO turnover. Interestingly, the studies also varied in terms of whether they reported hypotheses or expectations regarding the impact of the investigated predictors on CEO turnover. Eight studies (36%) did not make any predictions; rather, they analyzed or derived a set of variables without an a priori assumption about whether these variables would increase or decrease CEO turnover. Twelve studies (55%) predicted the direction of effects for all of their predictors, whereas two further studies predicted the direction of effects but only for a subset of variables (e.g., Eldenburg et al., 2004, formulated predictions for their focal variables but not for their control variables). In general, studies that drew on specific theories more frequently provided directional predictions for relationships.

Theoretical Explanations for the Effects of CEO Turnover

Of the 14 studies investigating effects of CEO turnover in hospitals, eight referred to theories to explain the observed relationships. Two of these studies were based on the theory of leadership instability and two on the resource-based view. The remaining studies drew upon organizational ecology theory, the manager socialization model, the strategic collaborative quality management model, and the new public management theory in their assumptions. As a second theory, one of the studies referred to the person–organization fit theory. Ten studies (71%) reported expectations about the effects of CEO turnover. Among these, three assumed that CEO turnover would have beneficial effects, whereas six expected effects to be detrimental and one hypothesized that the effect would be either negative or positive depending on the time frame considered. Interestingly, six of the studies that only investigated predictors also elaborated on the potential effects of CEO turnover; of these studies, two assumed CEO turnover would have detrimental consequences, and four provided arguments for why CEO turnover might have detrimental or beneficial effects.

Predictors of Hospital CEO Turnover

The included quantitative inferential studies examined a total of 78 unique predictors of hospital CEO turnover, of which 46 were statistically significant. We assigned each of the 78 unique predictors to one of four thematic blocks and then created 22 subcategories (see the Data Synthesis section). These subcategories were distributed among the thematic blocks as follows: organizational characteristics (10), prior hospital performance (3), personal characteristics (5), and environmental characteristics (4). Only the most frequently investigated and relevant predictors are discussed here because predictors examined in only a single study lack sufficient evidence of consistency across studies to be considered robust. For the sake of transparency, a complete list of the predictors investigated, including their categorization, references and supplemental information, is provided in the Table 3 in the Supplemental Digital Content, http://links.lww.com/HCMR/A170, for quantitative inferential studies and in Table 4 in the Supplemental Digital Content, http://links.lww.com/HCMR/A170, for quantitative descriptive and qualitative studies. We summarize the results for each of the thematic blocks in the following subsections. When doing so, we first report the result of the quantitative studies that used statistical inference to analyze the relationships and subsequently complement each result, where feasible, with findings from the quantitative descriptive and qualitative studies.

Organizational Characteristics

Organizational characteristics were the most frequently examined predictors of CEO turnover, with 10 such characteristics being analyzed in 14 quantitative inferential studies. Starting with hospital size, a consistent picture emerged. CEO turnover occurred significantly less often in larger hospitals, measured in terms of the number of beds, total assets, or the number of admissions. Hospitals that were members of a hospital system or hospital network had higher CEO turnover rates than free-standing hospitals. A higher probability of CEO turnover was also found for hospitals that had either recently joined a hospital system or had been members of a system for a long time. Concerning hospital ownership, a range of studies found that the frequency of hospital CEO turnover varied according to ownership type. The studies consistently showed that CEO turnover occurred most frequently in for-profit hospitals compared to nonprofit or public hospitals. With regard to the other forms of ownership, a higher frequency of CEO turnover in public hospitals than in nonprofit hospitals was not consistently identified. Furthermore, it appears that changes in ownership increased the likelihood of CEO turnover.

The available evidence suggests that being a teaching hospital does not have a significant effect on the likelihood of CEO turnover. Two out of four studies showed lower CEO turnover, but the results were only significant at an alpha level of .10 and varied by the type of teaching hospital. As for other organization-specific predictors, one study found a positive relationship between both the number of services offered and the service usage intensity and CEO turnover, whereas other studies found no significant relationship for either. Furthermore, it was found that older hospitals had a higher probability of CEO turnover, while hospital accreditation and the circumstances surrounding the previous CEO turnover did not serve as predictors of CEO turnover. Additionally, there were conflicting results regarding the impact of time since the last turnover. Alexander et al. (1993) found that some characteristics of the board of directors or the board–CEO relationship—including CEO power and influence over the board, and board size and heterogeneity—significantly related to CEO turnover. This relationship is supported by a number of survey and interview studies, which suggest that a poor working atmosphere and existing conflicts between the board and the CEO are major reasons for the change of CEO.

Prior Hospital Performance

In terms of prior performance, six studies found that weak financial performance was associated with increased hospital CEO turnover. This was also stated by CEOs and board members in various survey and interview studies as one of the most frequent reasons for CEO turnover. However, the relevance of finance-related predictors seemed to vary depending on hospital type. For example, the sensitivity of CEO turnover to performance was found to be at least as strong in nonprofit as in for-profit hospitals (Brickley & van Horn, 2002) but was lowest in public hospitals. Additional differences between hospital types were found by Eldenburg et al. (2004). For instance, they showed a lack of significance for the negative relationship between excess profit margin and CEO turnover in church-owned hospitals, as well as for the positive relationship between high administrative costs and CEO turnover in nonprofit and government hospitals. In terms of organizational performance, the likelihood of CEO turnover was higher when the hospitals were in an unstable or declining stage of their life cycle, had a poor multidimensional performance rating or a low occupancy rate, as investigated in five studies. Furthermore, two studies looked at altruistic performance measures, such as the nurse-to-patient ratio and the level of uncompensated care provided, as predictors of CEO turnover, but found no statistical correlation.

Personal Characteristics

Personal characteristics as potential predictors of CEO turnover were examined in five of the included quantitative inferential studies and can be summarized as follows. First, three studies investigated CEO age and tenure, pointing to an inverse U-shaped relationship between age and the probability of CEO turnover. In addition to age, some evidence suggests that gender plays a role. Wilson and Stranahan (2000) found that CEO turnover was significantly more common among female than male CEOs, albeit only in large hospitals. Three other personal characteristics—job satisfaction, education, and life situation—were analyzed, each in one study only. Other personal characteristics related to work that were frequently cited in surveys and interviews included a lack of career opportunities and feelings of isolation or lack of support, both of which are probably related to job satisfaction.

Environmental Characteristics

Environmental characteristics include the overall socioeconomic context in which a hospital operates. A number of studies examined differences in the incidence of CEO turnover by geographic location, the degree of urbanization and the intensity of competition, yielding mixed results. Four studies examined the relationship between CEO turnover and regulatory changes. Although certain regulatory changes in the United States, such as the New Charity Period or the Affordable Care Act, led to periods of higher CEO turnover rates, Haveman et al. (2001) could not find a systematic relationship between regulatory changes and CEO turnover. In terms of the regional socioeconomic environment, hospital CEO turnover was found to be more likely to occur in areas with shortages of health professionals or a high share of minority residents. Additionally, dissatisfaction with cultural offerings and lifestyle in a particular area or conflicts with the local community were frequently stated as reasons for CEO turnover.

Effects of Hospital CEO Turnover

In total, 15 studies examined the effects of CEO turnover. We present the results of the quantitative inferential studies in Table 3 in the Supplemental Digital Content, http://links.lww.com/HCMR/A170, and the results of studies using descriptive methodologies in Table 4 in the Supplemental Digital Content, http://links.lww.com/HCMR/A170. In the case of the quantitative inferential studies, we classified the effects into four groups: financial performance, quality of care, other organizational performance measures, and organizational activities. In the case of the qualitative and quantitative descriptive studies, we added the category of competitor behavior. For each potential effect, we first report the results of the quantitative inferential studies and then complement these with insights from the qualitative and purely descriptive studies. Where applicable, we additionally highlight the results of the studies with the most sophisticated analytic approaches because these come closest to providing causal evidence on the effects of CEO turnover.

Most studies considered the financial impact of CEO turnover, yielding mixed results. The studies with the highest quality scores reported negative effects on financial performance. For example, Ford et al. (2018) showed in their quasi-experimental study that hospital CEO turnover led to short-term productivity losses (cost efficiency). In addition, they found that outsider CEOs were able to close the performance gap on average 1 year after taking office, whereas this was not demonstrated for insider successors. Two other longitudinal studies showed in their survival analysis a relationship between hospital CEO turnover and hospital closure. In contrast, Janke et al. (2019) found no association between CEO turnover and the so-called retained surplus of National Health Service hospitals in England, and Haveman et al. (2001) showed that CEO succession following regulatory change had an effect on the operating margin that was positive but diminished over time. In addition, the majority of CEOs stated in surveys that their hospitals’ profitability and revenues had improved since their arrival. Yet, in the study of Leibert and Leaming (2010), financial performance was among the top three most frequently mentioned negative impacts of CEO turnover, with 35% agreement.

Apart from a slight deterioration in the length of stay, which was inconsistent, and one study that concluded that patient satisfaction decreased as a result of several CEO turnovers, no effect of CEO turnover on the quality of care was found in the included quantitative inferential studies. However, some CEOs reported negative effects on quality of care in an interview study (Mathew et al., 2024a).

Regarding other organizational performance aspects, Janke et al. (2019) detected a positive effect of CEO turnover on the number of day cases and a negative effect on staff job satisfaction. However, while the study is of high methodological quality, its results stem from different models (nonparametric/parametric), and the authors advise caution in interpreting the results. Yet, various qualitative and quantitative descriptive studies found comparable results.

With respect to organizational activities as a result of CEO turnover, Mick et al. (1993) found a positive effect on strategic changes, such as entry into hospital systems, networks, sale, or hospital restructuring. Consistent with this, Goodstein and Boeker (1991) found that CEO turnover leads to changes in the range of services provided, including the addition and divestiture of services. The qualitative and quantitative descriptive studies pointed to increases in the number of cases treated (Dworkin et al., 2010), while statements about strategic activities were contradictory. For instance, some studies suggested that strategic planning and action were initiated after CEO turnover, whereas other studies suggested that these activities were halted or postponed. In the descriptive studies, additional organizational and operational changes triggered by CEO turnover were mentioned. These included changes in organizational culture, the leadership style, as well as restructuring and the turnover of key personnel, such as chief nurse officers (Freund, 1987). Furthermore, Khaliq et al. (2006) investigated whether CEO turnovers affected competitor behavior and detected increases in marketing activities, the opening of new clinics, and the recruitment of key employees.

An overview of the empirical results regarding the direction of the coefficients of all considered predictors and effects of CEO turnover can be found in Figure 2.

Figure 2.

Comprehensive framework for hospital CEO transition literature.

Discussion

We conducted a systematic literature review with the aim of providing a comprehensive overview of the predictors and effects of hospital CEO turnover and assessing the methodological quality of existing research to determine the reliability and validity of its findings. In the following section, we discuss and interpret our main findings, starting with the predictors.

Predictors of Hospital CEO Turnover

We compiled a comprehensive list of 46 unique and statistically significant predictors of CEO turnover, as identified in quantitative inferential studies included in our systematic review. These predictors fell into four broad categories: organizational, environmental, personal characteristics, and prior performance. For the sake of parsimony and robustness, we focus in our discussion on the most frequently examined predictors.

The evidence from our review suggests that organizational predictors play an important role in CEO turnover. For example, smaller hospitals are more likely to have higher CEO turnover rates, a finding that might be attributable to a common career trajectory of hospital CEOs, who often start in smaller facilities and progress to larger ones (Hearld et al., 2019). Furthermore, CEO turnover is more likely to occur in hospitals after major organizational changes like joining a hospital system, ownership changes, or shifts in the mix of services provided. Such situational changes may lead to a reassessment of CEO profile requirements, presenting an opportunity to identify a CEO whose skills align more closely with the organization’s needs (Allgood & Farrell, 2003; Haveman et al., 2001). Additionally, CEO turnover rates are higher in for-profit hospitals compared to nonprofit or public hospitals. One reason for this might be the clearer, more quantifiable goals of profitability in for-profit hospitals, contrasting with the multiple, less specific goals in nonprofit and public institutions. This clarity makes it easier to evaluate CEO performance in for-profit hospitals (Potter & Dowd, 2003). In addition, the focus on short-term performance targets in for-profit hospitals, driven by shareholder interests, may cause managers to prioritize immediate value creation at the expense of sustainability and long-term value creation strategies (Haessler, 2020). This approach could influence CEO turnover in these hospitals by incentivizing CEOs to leave voluntarily while the financials are still favorable to avoid reputational damage later when the lack of sustainability becomes apparent (Garay et al., 2007).

Furthermore, we found that hospitals with poor performance metrics, such as low occupancy rates or poor financials, were more likely to undergo CEO turnover in the following years. Poor performance and the failure to meet organizational goals are presumably two of the most frequent reasons for forced CEO turnover in any type of hospital (Brickley & van Horn, 2002; Eldenburg et al., 2004). This could be due to the pursuit of a better-fitting CEO who can provide new impulses and commit to major changes (Allgood & Farrell, 2003; Haveman et al., 2001; Mick et al., 1993).

In terms of environmental characteristics, we found mixed evidence of a relationship between regulatory change and CEO turnover, suggesting that the impact of regulatory changes on the likelihood of CEO turnover is highly dependent on the specific regulation and cannot be generalized. There is also mixed evidence for the relationship between competition or the degree of urbanization in a hospital’s region and CEO turnover. It appears that competition may play a secondary role compared to internal challenges that can pose an acute threat to a hospital’s survival, such as the pressure to reduce costs and improve efficiency due to tight refinancing (Cantor & Poh, 2017) or difficulties in recruiting and retaining qualified staff (Chan et al., 2013). With regard to the degree of urbanization, in contrast to studies from the 1990s (e.g., Weil & Timmerberg, 1990), the most recent studies suggest that there is a higher probability of CEO turnover in rural areas. This may be due to the structural decline observed in many rural areas in high-income countries and the progressive closure of rural hospitals (Kaufman et al., 2016). Structural decline, in turn, might lead to an increasing number of young professionals and managers moving to urban areas, where they have better career prospects and a more diverse range of leisure activities (Hart et al., 1993; Manshadi et al., 2022).

Regarding CEOs’ personal characteristics, the studies identified a U-shaped relationship between age and CEO turnover. This phenomenon could be due to the frequent job changes common in the early stages of a professional’s career, which decrease as experience and professional development increase (Ng & Feldman, 2009). However, turnover tends to rise again in later stages of life, often due to retirement or health issues (BDO & DKI, 2021).

As for the relationship between gender and hospital CEO turnover, the included studies found no statistically significant differences in general. However, there is evidence in the general management literature that the determinants of CEO turnover and their importance may differ by gender. For example, V. K. Gupta et al. (2020) found that in well-performing organizations, female CEOs face a higher risk of dismissal compared to their male counterparts. Moreover, Silvera and Clark (2021) detected that female hospital CEOs are more likely to improve the interpersonal care experience, especially in large hospitals and those located in metropolitan areas. These findings suggest the need for a nuanced approach in future research when addressing the role of gender in CEO turnover in hospitals.

Other personal characteristics contributing to CEO turnover that were frequently stated by CEOs in surveys and interview studies were a loss of confidence and a lack of peer support. This issue could be connected to the limited practice of succession planning in hospitals. For example, one study found that only 21% of freestanding hospitals in the United States routinely engage in succession planning (Garman & Tyler, 2004). Yet, various studies show that new hospital CEOs need support, particularly in the transition phase, such as when familiarizing themselves with core tasks, getting to know key personnel, and building a professional network (Dworkin & Goldstein, 2004; Khaliq et al., 2007). The absence of such support can have a negative impact on CEOs’ confidence in their abilities and increase the likelihood of turnover (P. D. Gupta et al., 2018). This not only risks losing capable CEOs but also incurs additional time and costs in finding replacements (Nyberg et al., 2021).

Effects of Hospital CEO Turnover

The studies examining the impact of CEO turnover on hospital performance present a mixed and complex picture that is difficult to interpret due to differences in outcome variables and research methods. The quantitative inferential studies included in our review generally concluded that CEO turnover has a negative impact on financial performance. In more extreme cases, the results of survival analyses have linked CEO turnover and short tenures with an increased risk of hospital closure (Alexander & Lee, 1996; Lee & Alexander, 1999). Especially in scenarios of frequent CEO turnover, this is not surprising given evidence suggesting that CEO turnover is associated with high costs and strategic shifts (Nyberg et al., 2021), which can result in organizational instability (Schepker et al., 2017). Supporting this, studies like Ford et al. (2018) have found that CEO turnover can adversely affect hospital productivity in terms of cost efficiency in the short term. However, it should be noted that other characteristics such as inadequate financial performance and hospital size, which are common predictors of CEO turnover, were also associated with the risk of closure in the included studies (Alexander & Lee, 1996; Lee & Alexander, 1999). Therefore, CEO turnover may just be a symptom of other contextual characteristics rather than a direct cause of hospital closures. Furthermore, Ford et al.’s (2018) study of postturnover productivity performance did not account for additional CEO turnover during their study period, making it difficult to attribute the effects unambiguously to a specific CEO turnover event.

On the other hand, while Janke et al. (2019) observed a negative relationship between CEO turnover and hospital performance, they did not statistically demonstrate a decline in performance as measured by retained surplus. Moreover, Haveman et al. (2001) found that CEO turnover following regulatory change actually led to an increase in return on patient services revenue. However, the validity of these studies was limited by their sample sizes, with a maximum of 185 hospitals included. Additionally, the use of retained surplus as a financial measure to assess performance effects within 1 year after CEO turnover seems inadequate, as investments or restructuring measures might reduce surplus while improving operational efficiency. Thus, when assessing the financial impact of CEO turnover, it would be more appropriate to use performance indicators that more accurately reflect efficient service delivery. These indicators can provide a clearer picture of the short-term change in a hospital’s financial health.

In contrast, the results from the quantitative descriptive and qualitative descriptive studies included in our systematic review suggest a mostly positive picture. For instance, CEOs often reported that they had been able to improve their hospital’s profitability and revenue since their appointment. However, these studies must be interpreted with caution due to the methodological limitations identified in our quality assessment, including small sample sizes and potentially inadequate sampling methods, which could lead to selection bias. In addition, subjective statements by survey respondents can introduce a self-serving bias, as seen in studies where self-report and third-party ratings differed significantly. For example, Hart et al. (1993) found that CEOs tended to rate their own performance much more positively compared to evaluations by their respective boards in various performance categories. Similarly, Khaliq et al. (2006) observed a pronounced disparity in the assessments of the outcomes of CEO turnover, with CEOs tending to overstate their own performance while underestimating that of their successors.

Limitations and Further Research

When interpreting the results of our systematic review, it is important to consider its various limitations, some of which also provide avenues for future research:

Magnitude and Causality of Associations

Our review focused on the direction and statistical significance or frequency of agreement of the identified associations. Due to the heterogeneity of methods and indicators used across the studies, however, we cannot draw definitive conclusions about the magnitude of effects or the relative importance of different predictors. Our review can thus only provide an overview of which predictors seem to affect the probability of CEO turnover in hospitals and which outcomes appear to be affected by it; however, we cannot ascertain the practical relevance of these relationships. Furthermore, despite the predominance of longitudinal data in the included studies, many relied solely on descriptive data analysis, correlations, and simple regression models. There is a remarkable lack of studies, which enable causal inference. Further quantitative research using longitudinal data and more sophisticated methods to investigate the predictors and effects of CEO turnover is clearly needed. These could include sophisticated survival analysis to investigate predictors, as well as quasi-experimental designs or other causal methods to assess the effects of CEO turnover.

Heterogeneous Treatment Effects

It is also important to acknowledge that most of the included studies did not adequately account for potential heterogeneity in treatment effects. There is a strong likelihood, with some evidence already supporting this notion, that the impact of CEO turnover depends on the circumstances. These include prior organizational performance, the nature of and reason for the turnover (voluntary or involuntary; internal vs. external), and environmental characteristics like competition (Berns & Klarner, 2017; Ford et al., 2018; Nyberg et al., 2021). Recognizing and incorporating these diverse characteristics in future analyses could greatly enrich our understanding of the effects of CEO turnover in hospitals.

Underexplored Predictors and Outcomes

Many of the predictors and effects identified in our review were examined only once or twice. This lack of replication presents challenges in making precise statements about many of the relationships because of the possibility that the results may be due to random, systematic, or other errors. Consequently, there is a need for further research on less frequently examined predictors, including hospital size, ownership type, and degree of urbanization. Moreover, future studies should explore additional outcomes beyond financial performance metrics. These could include a range of competitive factors for hospitals, such as technical efficiency and patient satisfaction.

Measurement Bias

An additional concern arises from the reliance of several studies on self-reported or peer assessments by current or former CEOs. The discrepancies in the statements and results of these studies, such as those highlighted by Khaliq et al. (2006), suggest the presence of self-serving bias, which affects the reliability of their findings. Future research, especially when using qualitative interview or survey methods, should strive for a multiperspective approach or use data triangulation to mitigate respondent bias.

Limitations of Review Methodology

In addition to the limitations inherent to the data or methodologies of the included studies, our systematic review itself has limitations that should be considered. One of these is the potential exclusion of relevant literature, primarily due to our focus on articles published in peer-reviewed journals, which could introduce publication bias. Furthermore, our categorization of data was conducted inductively based on the perceived thematic context in the studies. Thus, these categories should be viewed as fluid rather than fixed, allowing for variation in categorization depending on the context. For example, the number of hospital admissions, while highly correlated with hospital size, can be categorized as a predictor within the hospital size category. At the same time, the number of admissions can also serve as a performance indicator for evaluating hospital CEOs, leading to a different categorization. This highlights the importance of not only considering aggregate results but also paying attention to specific indicators and individual findings.

Practical Implications and Conclusion

The findings from our review suggest that CEO turnover in hospitals is associated with short-term financial performance losses and that frequent CEO turnover can pose a threat to organizational stability and financial viability. Hospitals with characteristics associated with a higher likelihood of CEO turnover should consider implementing early and systematic succession planning to mitigate the substantial costs of CEO recruitment and fill leadership gaps more efficiently. A well-structured succession plan can also ease transitions for the CEOs by ensuring a smoother adjustment to a hospital’s corporate culture and job requirements (Khaliq et al., 2007).

Despite the potential for financial setbacks, CEO turnover can serve as a strategic catalyst for reorienting and potentially improving hospital performance. Often, CEO turnover is accompanied by strategic shifts that can yield substantial long-term benefits for hospitals, offsetting any short-term financial losses. However, the impact of these strategic changes frequently outlasts the tenure of the CEOs, who are typically assessed based on immediate financial results and may be dismissed for poor performance. The delayed effects of a CEO’s strategic decisions are rarely included in performance evaluations, and current problems may stem from decisions made by predecessors. Hence, decision-makers should carefully consider the potential negative consequences of frequent CEO replacements.

That being said, it appears that, overall, little is known about the subsequent processes and effects of CEO turnover in hospitals. Future research should explore these topics using causal methods to provide more meaningful insights into the effects of CEO turnover. This research should take into account the different contextual conditions and situational challenges that hospitals face. A deeper understanding of these dynamics will aid in developing tailored strategies to facilitate smooth transitions and the successful integration of new CEOs.