Introduction
Quality healthcare is dependent on effective organizational factors, including interdisciplinary teamwork, a supportive culture, and good leadership (Barr & Dowding, 2019; Marchionni & Ritchie, 2008; McAlearney, 2008). As the healthcare system has evolved to align business and medical imperatives (Murdock & Brammer, 2011), the tendency to separate leadership and administration from clinical care has given way to leadership development becoming a core approach in physician, nurse, and allied health training (Ackerman et al., 2019).
The development of leadership skills among healthcare professionals aims to improve performance, allow for succession planning, facilitate organizational change, and achieve organizational goals (Collins & Holton, 2004; Turner, 2019). Effective leadership grows an accountable culture that is aligned with these goals (Peters, 2019), which improves patient outcomes (Suhonen et al., 2019).
Research has shown the importance of developing leadership in health professionals, including specific clinical leadership skills (Cleary et al., 2005; Cutcliffe & Cleary, 2015; Daly et al., 2014). However, discussions of skills and capabilities in the literature often emphasize what should happen instead of evaluating successful strategies. Studying the comparative degrees of success of different leadership development programs allows future endeavors to benefit from previous experiences to better target strategies. For example, one study highlighted the importance of the context of the proposed change for successful leadership in complex healthcare environments (Kwamie et al., 2014). This study found that leadership change was not sufficiently institutionalized, directing future studies to consider context to enable a more reflexive organizational culture. Another study found that the development of leadership in physicians focused on individual skills rather than enhancing collaborative capacity (Frich et al., 2015).
Strategies to enhance leadership behavior have important effects on both public health workplaces and healthcare (Dellve et al., 2007; Jimenez et al., 2017). This systematic review seeks to evaluate the evidence of interventions aimed at improving leadership behavior in health professionals.
Methods
Study Design and Search Methods
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (Liberati et al., 2009) were used in this study to systematically review studies concerning the effectiveness of leadership development/intervention programs in improving leadership behavior in health professionals. A literature search was conducted in September 2018 of several databases, including PubMed, CINAHL, Embase, and Scopus. No time restriction was applied to these database searches. Boolean connectors combined Medical Subject Headings and the following search terms: leadership*, health personnel, allied health personnel, nursing, physicians, program development, intervention*, and program evaluation.
For example, the search strategy for CINAHL was ([MH nurses OR TI nursing OR AB nursing OR MH Allied Health Personnel OR MH Occupational Therapists OR MH Social Workers OR MH Physical Therapists OR TI physiotherapist OR AB physiotherapist OR MH Physicians] AND [MH Leadership OR TI "transformational leadership" OR AB "transformational leadership"] AND [MH Program Development OR TI intervention* OR AB intervention* OR MH Program Evaluation]) NOT student* NOT undergraduate NOT baccalaureate NOT supervision NOT mentor*. Across the four databases, the search was modified for variations in syntax and Medical Subject Headings terms.
Eligibility Criteria
The eligibility for inclusion encompassed qualitative or quantitative peer-reviewed articles published in English evaluating leadership development programs/interventions on leadership behavior among health professionals across health sectors. All prospective interventions that were designed to develop or enhance leadership were included. Kirkpatrick's evaluation model (Kirkpatrick & Kirkpatrick, 2009), which describes four evaluation levels of reaction (Level 1), knowledge (Level 2), behavioral change (Level 3), and system results (Level 4), was used to classify the outcome of leadership interventions. This review only included studies reporting behavior change (Level 3). Studies without interventions; studies not primarily addressing health professionals; studies on students, supervision, or mentorship; theoretical articles; commentaries; editorials; and review articles were excluded.
Search Outcomes
The search identified 660 articles, with an additional 34 articles identified in the reference lists of the identified studies and through hand searching. After the removal of 214 duplicates, the titles and abstracts of 480 articles were screened, resulting in the removal of an additional 361 articles. The remaining 119 articles were subject to full text review, resulting in the removal of an additional 39 articles. After removing one article (Blaney, 2012) because of the lack of reported findings, the remaining 79 articles were assessed for eligibility, of which 33 (from 31 studies) met the inclusion criteria (see Figure 1).
Data Extraction and Synthesis
A data extraction sheet was developed in consensus with all of the authors and finalized using an iterative process. Two of the authors extracted the information from the included articles in five domains: (a) study characteristics, namely, author, publication year, country, objective(s), and study design; (b) study settings and participants; (c) intervention characteristics, namely, duration, content, and training methods; (d) behavioral outcome and assessment, namely, assessment measures and follow-up; and (e) significant findings (see Table 1). The other authors assessed the data, with discrepancies resolved by consensus. In line with the previous reviews of leadership interventions (Collins & Holton, 2004; Frich et al., 2015), the reported behavioral outcomes were differentiated into subjective and objective assessments of the behaviors. The results of this review are presented in narrative form.
Results
Characteristics of the Studies
The characteristics of the included studies are presented in Table 1. Of the 31 included studies, 19 used quantitative methods, five used qualitative methods, and seven used mixed methods to measure the intervention outcomes. Seventeen of the studies used a pretest-posttest design, and the remaining 14 used only a posttest (with or without follow-up) design. Only five of the included studies used a control group(s) for comparison. Ten of the studies were conducted in the United States; seven, in Canada; six, in the United Kingdom; two each, in Australia and Ireland; and three, in other European countries. One study was conducted in both the United States and Canada. The included studies were published between 2002 and 2016.
Study Settings and Participants
Most of the included studies evaluated different types of leadership development programs in clinical settings. Study participants included nurses in 12 studies, physicians in six studies, healthcare educators in three studies, and mixed-group participants (including physicians, nurses, clinical managers, ward managers, allied health professionals, and administrators) in 10 studies. The number of trainees who participated in the interventions ranged from eight to 48 in qualitative and seven to 550 in quantitative and mixed-methods studies. Female participants outnumbered male participants in many of the studies. Some studies did not exclusively report the number of participants who provided the evaluation data used in analyses.
In addition to the participants, key stakeholder informants were used in nine studies to provide information using observer evaluations of leadership behavior (objective behavior) of the participants (Study nos. 3, 4, 7, 10, 18, 23, 26, 30, 33). These informants included participants' immediate supervisors, peers and colleagues, dyad partners, and patients.
Intervention Characteristics
The interventions differed in terms of duration, content, and mode of delivery. The duration of the interventions ranged from a half-day workshop (Gilfoyle et al., 2007) to a 4-year program that was integrated into standard clinical training (Agius et al., 2015). Eleven studies reported on interventions lasting for 12 months or longer, whereas five studies had intervention lasting for less than 1 week. Five studies (Study nos. 5, 10, 17, 19, 32) did not specify the duration of the intervention. The intervention dose also varied across the studies, and the time devoted to the intervention was difficult to ascertain in many studies.
Despite the diversity across interventions in terms of training content, most interventions addressed leadership, group dynamics, teamwork, communication, personal development, change management, conflict resolution, time management, and supervision skills. Some studies also included clinical skills such as perioperative issues in geriatric care (delirium, functional assessment, polypharmacy, and discharge planning; Levine et al., 2008) and pediatric resuscitation (resuscitation skills and avoidance of fixation errors; Gilfoyle et al., 2007).
All of the included studies used face-to-face training, with the exception of Brown et al. (2003), which used an online leadership course. Maddalena and Fleet (2015) employed both face-to-face and online modules. The interventions used different types of teaching/learning methods, among which group-based workshops were the most common, with many studies using more than one type of training method. Other methods included lectures, coaching, 360[degrees] feedback, mentorship, learning reflection, group discussion, team activities, plenary sessions, role play, presentation, and simulation exercises. Some of the studies also included work-based action learning in which participants developed their own action plan and implemented this plan over the intervention period. Five studies (Study nos. 4, 9, 10, 19, 32) did not clearly specify the mode of training delivery.
Outcome Measures
This review only included studies reporting the behavioral outcomes of the leadership interventions. Most of the included studies (n = 21) reported only subjective behavioral outcomes, consisting of intervention-related, self-reported alterations in behavior. Gilfoyle et al. (2007) reported objective behavior in which participants were asked to participate in a mock leadership scenario. The remaining nine studies reported on both subjective and objective behaviors. Nine studies used standard scales to measure leadership behavior, of which four used the Leadership Practices Inventory (LPI; Table 2).
Eight studies (Study nos. 3, 5, 10, 11, 13, 14, 22, 28) used only one measurement during the postintervention period. Nine studies (Study nos. 1, 2, 7, 17, 19, 24, 30, 32, 33) used two assessments during the preintervention and postintervention periods, and seven studies collected follow-up data (Study nos. 4, 12, 18, 21, 23, 26, 27). Dannels et al. (2008) used baseline assessments and a 4- to 5-year follow-up period. The remaining six studies (Study nos. 6, 15, 16, 20, 25, 29) used measurements during the postintervention and follow-up periods. Follow-up periods ranged from 6 weeks (Leeson & Millar, 2013) to 8 years (Margolis et al., 2013).
Effectiveness of the Interventions
Key findings related to the effects of the interventions on subjective and objective leadership behaviors are summarized in Table 2.
Subjective behavior
Subjective behavior outcomes included participants' self-reported change in behavior as a result of the leadership intervention. Of the nine studies that used standard scales to measure these outcomes, six reported a significant postprogram improvement. Among the four studies that used the LPI, three showed significant improvements on different subscales of the LPI. Significant improvements in "inspiring a shared vision" (Krugman & Smith, 2003; Lee et al., 2010; Martin et al., 2012), "challenging the process" (Krugman & Smith, 2003; Martin et al., 2012), and "modeling the way" (Krugman & Smith, 2003) were observed. In the case of Tourangeau (2003), the increase in LPI scores was not significant.
Cunningham and Kitson (2000b), using the Multifactor Leadership Questionnaire, reported significant changes in the dimensions of inspiration, active management by exception, effectiveness, and satisfaction among the nurses participating in the Clinical Nurse Leadership program. Werrett et al. (2002) found significant pre-post improvements in the Importance-Performance Scale, particularly in aspects of team and management issues, staff development, and assertiveness. Boyle and Kochinda (2004) found a significant increase in postintervention scores (56.7 in pretest to 75.3 in posttest, p = .021) on the Collaborative Communication Simulation Vignette (maximum score = 100). Moreover, the follow-up results showed significant improvements in self-perceived leadership and leadership and communication skills satisfaction (measured using an ICU Nurse-Physician Questionnaire) among participants. Fennimore and Wolf (2011), using a modified Nurse Manager Inventory Tool, reported an average improvement of 0.68 in scores across 15 different competency areas among nurses who participated in the Leadership Development for Nursing Middle Managers program. Competency showed major improvements, particularly in the areas of "the science of managing people," "the art of leading people," and "creating the leader within." In Malling et al. (2009), the improvement in pre-post Multisource Feedback (MSF) was not significant for the intervention group (who received both the leadership intervention and MSF) and between the intervention and control groups, the latter of which received MSF only.
Maddalena and Fleet (2015) found that more than 95% of the participants were better prepared for leadership responsibilities, with participants reporting planned changes in the workplace because of program participation. Weston et al. (2008) reported that participants applied learned skills such as conflict, time management and communication skills. Chief residents who participated in a 2-day leadership intervention reported better leadership and conflict resolution skills, more collaboration between disciplines (Levine et al., 2008) as well as the belief that the training would affect their future work as physicians. Physicians who participated in a 20-month leadership program reported increased effectiveness in their leadership role (mean = 4.2 on a 5-point scale), teamwork abilities (mean = 4.0), ability to lead teams (mean = 4.3), and experience of new and expanded leadership roles (mean = 4.0; McAlearney et al., 2005).
Three of the four studies with control groups showed increased leadership competencies and changed behavior among the intervention groups compared with controls. Dannels et al. (2008) found significant improvements on 12 of the 16 leadership indicators for participants in the intervention group. Follow-up results also showed a significantly higher number of participants reporting attainment of a higher administrative position. Bergman et al. (2009) compared long-term support groups with a 1-week leadership intervention for healthcare managers and found that both interventions strengthened leadership roles. The 1-week course was better for inexperienced managers, whereas the long-term group benefited experienced managers. Interdisciplinary Leadership Development Program participants improved their interdisciplinary practices more than nonparticipants (Margolis et al., 2013). Malling et al. (2009) did not find significant differences in pre-post MSF between intervention and control groups.
Castillo and James (2013) found that 70% of the training participants reported improved leadership behavior because of training, including being a role model, encouraging problem solving, and critical thinking. Moreover, Debono et al. (2016) found that 70% of participants reported improvements in job performance after the leadership program. In Gagliano et al. (2010), 79% of the physicians participating in the Leadership Development Program reported a change in their approach to specific projects and greater positivity in addressing problems. In addition, more than 65% of the nurses who participated in a web-based leadership course reported improved confidence in applying skills such as change management, conflict resolution, and ethical decision making (Brown et al., 2003).
Other qualitative studies also had similar results, with participants reporting changed leadership behavior and performance because of the intervention (Boomer & McCormack, 2010; Ford et al., 2008; Gifford et al., 2011; Macphee et al., 2012). Participants reported consciously seeking positive behavioral change such as having a positive attitude, encouraging others to speak up, facilitating communication, and seeking input from colleagues (Singer et al., 2011). Other behavioral changes reported included becoming more reflexive (Boomer & McCormack, 2010; Fennimore & Wolf, 2011), facilitating teamwork (Study nos. 3, 6, 20, 23, 28, 32), effective conflict resolution (Brown et al., 2003; Levine et al., 2008; Maddalena & Fleet, 2015; Weston et al., 2008), and time management (Steinert et al., 2003; Weston et al., 2008).
Graham and Jack (2008) found no significant pre-post change in the quantitative results related to leadership behavior among leadership program participants. However, participants in qualitative interviews described being better able to be action focused, democratic, and participative. Steinert et al. (2003), at 1-year follow-up, found the participants in their study less satisfied with the changes they could make in leadership style and skills.
Interventions that incorporated the elements of personal development planning, self-directed learning, and reflection achieved relatively better results. Developing a leadership action plan and a practice-based action learning project (Study nos. 3, 6, 7, 13, 15, 17, 21, 26) was particularly effective in supporting the participants to operationalize leadership strategies and become more engaged with unit staff.
Objective behavior
Nine of the included studies reported objective behavior outcomes, including behavioral change exhibited by the participants as observed and rated by supervisors, colleagues, peers, and/or staff from the working unit of the participants (Table 2). Gilfoyle et al. (2007) used a mock leadership scenario to assess the effect of a leadership intervention, and hence it was considered an objective outcome in this review.
Three studies administered the LPI to the observers of the training participants. Martin et al. (2012) found that observers reported significant improvements in the "inspiring a shared vision" and "challenging the process" subscales of the LPI. Tourangeau et al. (2003) distinguished the observers into supervisors, dyad partners, and peers, with supervisors and dyad partners both reporting significant improvements in "challenging the process" and "inspiring a shared vision" and peers reporting significant improvements in all of the five subscales of the LPI. Conversely, in Krugman and Smith (2003), observers reported a significant decrease in the "modeling the way," "enabling others to act," and "encouraging the heart" subscales of the LPI. Observers, including colleagues of the participants, reported significant improvements in the dimensions of charisma, inspiration, idealized contribution, extra effort, and effectiveness on the Multifactor Leadership Questionnaire (Cunningham & Kitson, 2000a, 2000b). The unit staff in Boyle and Kochinda (2004) reported significant improvements in collaborative communication, problem solving, and nursing leadership, as measured using a modified ICU Nurse-Physician Questionnaire at 6 months of follow-up.
The supervisors of nurses participating in the leadership program reported a positive impact on the leadership skills (79%) and job performance (83%) of the nurses who participated in the program (Debono et al., 2016). In other studies, the observers rated postintervention improvements in leadership behaviors such as conflict management (Maddalena & Fleet, 2015; Weston et al., 2008), problem solving (Maddalena & Fleet, 2015), dealing with difficult people (Weston et al., 2008), and better teamwork and communication (Boomer & McCormack, 2010). Gilfoyle et al. (2007) found significantly improved performance during a 6-month follow-up of the mock leadership scenario among trainees who participated in a half-day workshop compared with both baseline and the control group.
Discussion
The primary aim of this study was to review the behavioral outcomes of leadership interventions that were conducted on health professionals. This review shows the beneficial effects on the leadership behavior of participants across a range of leadership interventions. Specific behavioral improvements as perceived by participants as well as observers were observed postprogram for most courses. The participants were found to implement more efficient processes and to engage more frequently with staff (Debono et al., 2016). Six of the nine studies that used standard leadership behavior scales (Study nos. 4, 7, 18, 19, 26, 32) showed significant postprogram improvement. Castillo and James (2013) identified change in organizational culture, with constructive use of feedback and improved communication. Both participants and observers reported the effective use of learned skills such as conflict resolution, communication, time management, teamwork, problem solving, critical thinking, and being reflective.
This review identified that the methods and processes used to implement leadership interventions are important in improving leadership behavior among participants. In this review, programs that targeted personal development, self-direction, and reflection were more likely to produce behavioral outcomes. Moreover, activities that were designed to assist participants to operationalize leadership strategies were particularly effective. Leadership action plans and practice-based action projects assisted participants to implement practical leadership strategies and promote staff engagement. Moreover, these action plans provided a tangible framework for showing leadership behavior and tracking the progress of leadership development (Gifford et al., 2011).
Among the included studies that used observer (e.g., supervisor, peer, or unit staff) ratings, most found that observers rated the participants' leadership behavior higher than participant self-ratings (Martin et al., 2012; Tourangeau, 2003; Tourangeau et al., 2003; Weston et al., 2008). Thus, objective observations are more likely to show the benefits of the leadership interventions. In general, objective behavior, as reported by the observers, improved from preintervention to postintervention. Unit staff perceived changes in participant behavior as the participants became more approachable, encouraging, and supporting (Maddalena & Fleet, 2015).
There is evidence across the studies showing improved leadership development where interventions use face-to-face delivery and showing that participants valued the opportunities provided by this mode of delivery. Participants from the included studies that used online course also preferred face-to-face delivery and reported that the absence of nonverbal cues and body language impeded their learning progress (Brown et al., 2003). Furthermore, participants in the included studies that used face-to-face courses valued the networking opportunities (Debono et al., 2016; Levine et al., 2008) and feedback from peers (Gagliano et al., 2010). Participants perceived these events as an opportunity to collaborate with other like-minded professionals (Lee et al., 2010).
This review identified the difficulties of and barriers to translating learned knowledge and skills into action. Steinert et al. (2003) argued that one of the reasons for failing to implement behavioral changes was the absence of dedicated workshop time during the sessions to apply their newly learned skills. In addition, leadership behavioral change may require time and experience. Other training areas such as resilience have shown context and experience as important to effective training outcomes (Cleary et al., 2018). Boomer and McCormack (2010) identified that lack of support from organization leaders adversely affected the articulation of learned skills into practice. Lee et al. (2010) further observed that, although leadership development programs often promote and encourage leadership practice, recurrent organizational barriers make participants reluctant to initiate and continue efforts to change. The increasing incongruities between workplace reality and leadership idealism promoted in training programs may result in a sense of isolation, skepticism, and frustration among program participants (Spiers et al., 2010). Furthermore, participants who are involved in a leadership development intervention may hold the preconception that senior executives are not in favor of change and that organizational culture inhibits change (Lee et al., 2010). Iles and Preece (2006) posited that "leadership development" differs from "the development of leaders." Interventions aiming to develop leadership, in addition to focusing on individual skill development, should consider the social, cultural, and political contexts of the organization in which behavioral change is expected. Leadership should not be considered as an individual activity but rather as a collective cultural activity with collective identity, interdependence, and collaborative accountability.
Program Implications
Efforts to improve leadership behavior among health professionals should involve the active collaboration and support of senior executives. As leadership is a team process, the entire organization should facilitate sustainable change in leadership. Face-to-face courses are recommended, as they support collegial feedback and provide networking opportunities to positively impact learning, behavioral change, and professional development. Training should incorporate elements of work-based and experiential learning. Interventions in which participants determine behavioral change needs, develop action plans, and have opportunities to apply these in real scenarios are more likely to succeed. Assessments performed with participants before training help identify the appropriate content and focuses for training interventions (Collins & Holton, 2004). Changing leadership behavior is a long-term process that requires continuous and sustained interventions and long-term follow-up.
Research Implications
Few studies have assessed the leadership behavior of health professionals using a robust evaluation design. Studies evaluating leadership interventions with a focus on leadership behavior should use specific leadership behavioral outcome measures, a strong evaluation design, and multiple outcome assessments. Studies should consider confounders that may have significantly affect leadership behavior. In evaluating leadership interventions, Saleh et al. (2004) recommended considering factors such as the support of senior executives, organizational culture (openness to change), organizational resources, and the opportunity to apply learning. Randomized control designs with wait-list controls and studies with an adequate sample size and follow-up periods are required to identify the long-term effects of the interventions.
Limitations
The considerable heterogeneity among the studies precluded a quantitative synthesis of the results. In addition, this review found that most interventions focus on individuals and thus ignore teamwork and collaboration. Most of the included studies on the behavioral impact of leadership interventions adopted a narrow set of outcome measures that primarily addressed self-reported behavioral change. Furthermore, as only a few of the included studies used control groups and none used a randomized control trial design, it is possible that other, unaccounted-for factors may have contributed to the outcomes. This review may also experience publication bias, as negative and nonsignificant findings may not have been published. The use of small sample sizes and self-selected convenience sampling in most of the included studies is another limitation that may affect the outcomes of this review. Finally, in many of the included studies, the researchers were also responsible for implementing the change in training programs and/or developing the training resources.
Conclusions
The culture of healthcare is defined by constant development, where effective leadership is central to addressing and managing change. Leadership development programs produce positive results, including implementation of efficient processes, staff engagement, and improved satisfaction for patients and staffs. This review identified many supporting factors for successful leadership development programs. Interventions that are designed to promote leadership development require preplanning, leadership needs assessments, considerations of organizational context, strong focuses on self-awareness and collaboration, experiential work-based learning to apply skills within the desired context, and proper evaluation and incentivization. The successful development of leadership in healthcare professionals promotes evidence-based best practice, positive organizational culture, and improved patient outcomes and keeps healthcare responsive to its ever-changing environment.
Acknowledgments
This work was supported by University of Tasmania funding awarded under the UTAS Research Theme: Better Health Research Development Grant Scheme, supported by the Office of the Deputy Vice-Chancellor and the Faculty of Health (C0025653).
Author Contributions
Study conception and design: MC, DV, RK, DKT
Data collection: MC, DV, RK, DKT
Data analysis and interpretation: All authors
Drafting of the article: All authors
Critical revision of the article: All authors
References