A mixed methods pilot study with a cluster randomized control trial to evaluate the impact of a leadership intervention on guideline implementation in home care nursing

Clinical practice guidelines synthesize and translate high quality research evidence into recommendations for practice, and provide an easy and accessible tool for bridging the evidence-practice gap [18‐21]. For practice change to occur however, guidelines must be utilized, and their timely and effective transfer into clinical practice remains fragmented and inconsistent [21‐24]. Implementation strategies directed at individuals, the environment and the organizational context are necessary for successful implementation and practice change to occur [20, 25‐27]. In recent Cochrane reviews, tailored interventions that focus on individual and organizational barriers to change showed promise for implementing change and improving patient care [28], and interactive workshops were found to have moderately large effects on changing professional practice [29].

The importance of top managers' involvement and commitment in implementing innovations such as guidelines and change have been emphasized outside [30‐39] and within healthcare settings [40‐45]. Descriptive and qualitative evidence has identified leadership and management behaviours as having an important impact on nurses' work environments [42, 46‐50] and their use of research evidence to inform practice [27, 51‐63]. Similarly, a systematic review of 30 studies identified the lack of support from managers, and 'other staff' to be one of the greatest barriers to nurses' use of research [60]. Management behaviours such as support and commitment [56, 58, 64‐69], policy revisions [66, 70] and monitoring of clinical outcomes [66, 71] have been described as enablers to nurses' use of research [72]. Limited experimental research exists however regarding which behaviours are most effective to facilitate guideline implementation in nursing. A recent mixed methods study of 37 organizations found leadership to be the only predictor of sustained use of clinical guideline recommendations two and three years post-implementation, accounting for 47 percent of the variance (p < .001) [73]. Using grounded theory to analyze 9 of the 37 organizations, Gifford et al. found patterns of leadership and managerial behaviours in organizations that sustained practice change based on guideline recommendations (n = 4) at 2 and 3 years differed when compared to organizations that did not sustain practice change (n = 5) [63]. A conceptual model was developed from the analysis that operationalizes leadership behaviours for implementing and sustaining practice change.

The aim of this pilot study is to evaluate the impact of a leadership intervention on implementing new recommendations from a clinical practice guideline on nursing assessments and management of foot ulcers for people with diabetes in community nursing practice. Specific objectives include:

We plan to test the following study hypotheses:

H ₁ : Nurses working in centers that receive the intervention will obtain significantly higher scores for practicing in accordance with guideline recommendations than control group.

H ₀ : No change in group means will occur following the intervention.

A two phase mixed method design is proposed (Figure 1). A pilot study is planned because there is little information regarding effective leadership behaviours for implementing practice change in nursing, and there is a need to test the intervention strategies prior to launching a larger multi-centered trial. Phase one involves descriptive qualitative methods to understand barriers to implementing the guideline recommendations and to refine the intervention strategy to be useful and appealing to leaders. A cluster randomized controlled trial, considered the optimal design when evaluating strategies to change professional behaviour [20, 74], will evaluate differences in outcomes between the two implementation strategies. Randomization will occur at the unit level to minimize threats of experimental contamination [20, 75, 76].

All centers (approximately 10) with the minimum number of clients being treated for diabetic foot ulcers to satisfy sample size calculations will be invited to participate in the study. Two centers will be randomly assigned to participate in phase one and four will be randomly assigned for phase two. The four sites in phase two will be randomly allocated to control (n = 2) or experimental (n = 2) groups.

Semi-structured interviews will be conducted at two centers with a sample of managers, supervisors, resource nurses and 2 'preceptor' staff nurses from each site (n = 10). Preceptor staff are experienced clinical nurses who volunteer to provide support to novice or newly hired nurses regarding clinical issues. The interview guide is based on previously published guides for assessing barriers and supports [80], and has been structured to understand components of an intervention strategy considered useful to managers and clinical leaders. Results of phase I will inform content and structure of the intervention strategy.

The theoretical underpinnings of the proposed intervention are based on mechanisms of planned change as described in the Ottawa Model of Research Use (OMRU^©) [52, 81], effective leadership behaviours described by Yukl [82], and leadership for guideline implementation described by Gifford et al [63].

The OMRU is a planned change framework for knowledge transfer in health care delivery [52]. Derived from evidence and theories of change, the OMRU recognizes that practice change is not a linear process, but involves simultaneous and interactive relationships between the nature of the innovation, the potential adopters, and the context within the practice environment. Three key processes involved are: 1) assessing barriers and supports; 2) developing and monitoring interventions tailored to barriers and supports; 3) evaluating outcomes. The underlying mechanism is that tailoring intervention strategies to address barriers and strengthen supports related to the innovation, potential adopters and practice environment will result in practice change.

The OMRU provides a template to assess barriers and supports for implementing change and will facilitate the selection of intervention strategies with the best probability of success. The relevance and pragmatic utility of the OMRU for guiding implementation of innovations (including nursing guidelines) has been demonstrated in previous research [83‐87].

Leadership is "the process of influencing others to understand and agree about what needs to be done and how to do it, and the process of facilitating individual and collective efforts to accomplish shared objectives" [[82], p.8]. Three meta-categories of effective leadership behaviours described by Yukl and supported by decades of research [82, 82, 88, 89], provide the foundation for this study: 1) relations-orientated, 2) change-orientated and 3) task-orientated. Relations-oriented behaviours include supporting, developing personal skills and job adjustments, and recognizing others and their contributions. Relations-oriented behaviours increase mutual trust, cooperation among members, and commitment to a unit and organization. Change-oriented behaviours are concerned with integrating a vision, developing strategies and building coalitions to support change, creating a sense of need and demonstrating commitment to change. Task-oriented behaviours include clarifying roles, monitoring operations and performance, and the efficient use of resources [82].

Three leadership themes emerged as central to implementing guidelines in the grounded theory study by Gifford et al., and these align closely with Yukl's [82] metacategories of effective leadership behaviours. Leaders were found to have: 1) facilitated staff through relations-oriented behaviours (e.g.: support, encouragement and recognition); 2) created a positive milieu within the clinical practice environment through change-related behaviours (e.g.: reinforced goals and philosophies of care); and 3) influenced organizational structures and processes through task-oriented behaviours (e.g.: providing resources, policies and monitoring). Together these behaviours influenced individuals, practice environments and infrastructures to enable nurses to practice based on guideline recommendations.

Drawing on the work of Van de Ven et al. (1999), effective leadership at different hierarchical levels is necessary for the adoption of new innovations in organizations [90]. Successful implementation in healthcare is dependent on strong effective leadership to create a context which is receptive to change [26, 27, 51, 63, 82, 90‐96]. The organizational context exerts a particularly powerful set of influences on nurses' adoption of new innovations [81, 97, 98]. Extensive managerial involvement, commitment and attitude toward change, role clarity, and leadership styles are significantly associated with maintaining the momentum of innovation adoption in organizations [32, 33, 90, 99, 100]. A 'road map" that explains what leaders do is not however possible due to the inherent unpredictability and nonlinear processes of innovation adoption [90]. "Management cannot ensure innovation success but can influence its odds" (p.11, 88). Leadership is an integral part of managerial roles, and is necessary for managers to influence change [34, 82, 96, 101‐104]. Individuals and organizational context must be influenced for practice change to occur based on new innovations [20]. The proposed intervention aims to influence individuals, the practice environment and organizational context through leadership processes and behaviours that manage barriers and enable practice change to occur. (Figure 2)

Chart audits will be conducted on all patients being treated for diabetic foot ulcers up to 12 weeks following the intervention. To understand the leadership and management behaviours that influenced nursing practice, semi structured qualitative interviews will be conducted with managers, supervisors and resource nurses and staff nurses at control and experimental sites (n = 20). The experimental group interview guide will also ask for participants' opinions regarding the usefulness of the intervention. The interview guides are based on previously published guides for assessing barriers and supports [80], and previous research on implementing guidelines [105]. To evaluate satisfaction and perceived utility of the one day workshop, an evaluation form, based on previously evaluations from RNAO guideline implementations, [106] will be administered at the end of the workshop.

Pre/post univariate descriptive data will be computed for demographics of patients and staff.

An inherent limitation of collecting data through chart audit is the documented data obtained may potentially underestimate actual care [112]. Other methods of data collection, such as direct observations are not feasible for this pilot study due to geographical distances and associated costs of observing home-care nurses provide care in patients' homes throughout the province. A second limitation of collecting data through chart audits involves reviewers accuracy, impartiality, attentiveness and consistency in extracting data [112]. Having an experienced research manager overseeing the process, and pilot testing for interrater and test-retest reliability will assist with addressing this limitation. Additionally, this is a pilot study and not sufficiently powered to account for the effect of clustering.

Prior to commencement, ethical approval will be obtained from University of Ottawa Research Ethics Board which follows Tri-council guidelines [113]. Details of ethical considerations, including informed consent, anonymity and confidentiality are found in ethics submission form. Briefly, a numerical coding system will be used to track individual participant and chart audit data. Names of interview participants will be kept separated from data collection forms and locked at the University of Ottawa Nursing Best Practice Research Unit. Names from chart audits will be kept by the research manager at the participating organization in a secured place; only numerically coded data will be sent to investigators. Only aggregated data will be reported. Information consent forms will be available in English and French. Data will be securely stored for 5 years after study conclusion (e.g. December, 2014).