Authors: Kathy B. Chappell, Phd, RN; Kathy C. Richards, Phd, RN, FAAN; Scott D. Barnett, Phd

1

SDC#1

Title: New Graduate Nurse Transition Programs and Clinical Leadership Skills in Novice Registered Nurses

Authors: Kathy B. Chappell, PhD, RN; Kathy C. Richards, PhD, RN, FAAN; Scott D. Barnett, PhD

Authors’ Affiliations: Director, American Nurses Credentialing Center; University Professor and Assistant Dean Doctoral Programs and Research Development, School of Nursing, College of Health and Human Services, George Mason University; Statistician, James A Haley Veterans Hospital

Corresponding Author: Kathy B. Chappell, MSN, RN,

Conflicts of Interest and Sources of Funding: Ms. Chappell received grants from George Mason University and from the Association for Nursing Professional Development. Dr. Richards has received support from the National Institutes of Health. Dr. Barnett has received support from the Department of Veterans Affairs and University of South Florida.

Abstract

Objective: Determine predictors of clinical leadership skill (CLS) for registered nurses (RN) with ≤ 24 months of clinical experience.

Background: New graduate nurse transition programs (NGNTP) have been proposed as a strategy to increase CLS. CLS is associated with positive patient outcomes.

Methods: Hierarchical regression modeling to evaluate predictors of CLS among individual characteristics of RNs and characteristics of NGNTPs.

Results: Perceived overall quality of a NGNTP was the strongest predictor of CLS (R2= .041, p < .01). Clinical experience and NGNTP characteristics accounted for 6.9% of the variance in CLS; 12.6% of the variance among RNs with assigned mentors (p < .01). RNs participating in NGNTPs > 24 weeks were 21 times more likely to remain employed within the organization when compared to NGNTPs ≤ 12 weeks, a significant cost-benefit to the organization.

Conclusions: Although perceived overall quality of a NGNTP was the strongest predictor of CLS, much of the variance in CLS remains unexplained.

The majority of literature on leadership and nursing focuses on nurses in formal leadership positions. Evaluating leadership skill at the clinical level has been a recent focus. Bedside clinical nurses must utilize clinical leadership skill (CLS) to ensure safe, high-quality patient care. Research supports the relationship between actions of bedside clinical nurses and positive patient outcomes (1 – 3). CLS is defined as “staff nurse behaviours that provide direction and support to clients and the healthcare team in the delivery of patient care.” (4; p. 450). Clinical leadership skill has 5 defining characteristics: clinical expertise; effective communication; collaboration; coordination; and interpersonal understanding (4). IncreasingCLS in staff nurses improves their ability to identify clinical problems, implement innovative change and evaluate outcomes (5). Increased CLS has been associated with leadership roles, initiation of research studies, and publications/presentations at local and national levels (5 – 7).

New graduate nurse transition programs (NGNTP) arecited as 1 strategy to increase CLS, however much of the outcome research on NGNTPs has focused on turnover and vacancy (8). Minimal research has explored the relationship between individual characteristics of registered nurses (RNs) and CLS; or the relationship between NGNTPs and CLS. The purpose of this study was to determine what individual characteristicsof RNs and NGNTPssubsequently predictincreased CLS in RNswith ≤ 24 months of clinical experience. Characteristics of RNsevaluated were: age; primary nursing degree; previous leadership or healthcare experience;and months of clinical experience as an RN. Characteristics of NGNTPsevaluated were: NGNTP length; assigned mentor/quality of mentor support; participation in/perceived improvement inprofessional development skill through participation in supplemental courses to promote critical thinking ability, leadership, and/or delegation skills; and overall NGNTP quality as perceived by RN participants(Figure 1).

Review of the Literature

We conducted a systematic review of the literature following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines to evaluate the relationship between NGNs and CLS, and NGNTPs and CLS (9).Electronic databases (CINAHL, Medline and Cochrane Library) were searched from January 2000 to January 2013. Search terms included “new nurse”, “new graduate”, “new graduate nurse”, “residency”, “internship”, “orientation”, “transition” and “leadership.” A total of 93 articles were identified from the search, and 3 additional articles were identified through bibliography review. Outcomes from 17 studies meeting inclusion requirements with data from more than 4000 NGNswere analyzed (9, 10 - 26). Studies were conducted in the United States (n = 15), in Scotland (n = 1), and in New Zealand (n = 1). Study designs included repeated-measures with (n = 3) and without (n = 2) a comparison group, pre/posttest with (n = 2) and without (n = 6) a comparison group, case study (n = 1) or were classified as program evaluation (n = 3). The most common comparison group was NGNs completing orientation prior to implementation of a NGNTP.

In the inclusion studies, NGNs were predominately female and 35 years of age (8, 10 – 13, 15 - 18, 21, 24 - 26). Evidence demonstrated that NGNs participating in NGNTPs ≥ 24 weeks had higher levels of CLS when compared to NGNs not participating in a NGNTP (18 - 20). There was insufficient evidence to evaluate the effect of NGNTPs 24 weeks. There was some evidence that NGNTP curriculum impacted the magnitude of change in CLS from baseline to post-program. There werelargewithin-subjects effects on CLSfor NGNs (n = 1334) participating in NGNTPs using the University HealthSystem Consortium/American Association of Colleges of Nursing (UHC/AACN) Nurse Residency™program curriculum (Cohen’s d = 1.29) (9). Moderate to large within-subjects effectswere also found onCLS for NGNs (n = 75) participating in NGNTPs using the Versant New Graduate RN Residency™curriculum (Cohen’s d = .58) and on CLS for NGNs (n = 33) participating in NGNTPs developed by 1 organization (Cohen’s d = .57) (9). The majority of studies included only baccalaureate-prepared nurses or failed to differentiate outcomes by academic preparation (9).

Evaluating impact of NGNTPs on CLSwas limited by single group and historical control group designs. The majority of studies used pre/post measures with no comparison group. Four of the studies used NGNs transitioning into the clinical setting prior to implementation of the NGNTP as the comparison group, which limited ability to control for confounding variables. There was lack of data on individual characteristics of NGNs such as age, nursing degree, and previous leadership or healthcare experience and the outcomes of NGNTPs.

Conceptual Model

The study variables and conceptual model were based on Bandura’s Self-Efficacy Theory andBenner’s Novice to Expert framework (27 – 28). According to Bandura’s Self-Efficacy Theory, individuals who have had successful experiences in 1 setting are more confident and may translate those experiences to different settings (27). The study variables of previous leadership or healthcare experience associated with individual characteristics of RNs were based on self-efficacy theory hypothesizing that there would be a relationship between previous experience and CLS (27). Benner’s Novice to Expert framework describes the process by which new graduate nurses progress through stages using experiential learning and draw from previous experiences as a reference for decision-making (28). Study variables associated with NGNTP characteristics were based on Benner’s framework with the hypothesis that NGNTPs are designed to help facilitate NGNs’ growth and developmental needs by providing a gradual increase in clinical responsibility in a supportive environment. It was hypothesized that there would be a relationship between components of a NGNTP and CLS. Therefore,NGNTP length, assigned mentor/quality of mentor support, and participation in professional development classes/perceived improvement inprofessional development skill were included in the study conceptual model. The number of months of experience as an RN was chosen as a control variable andbased on self-efficacy theory and Benner’s framework (27 – 28). Overall NGNTP quality as perceived by RN participantswas included as a variable(Figure 1).

Methods

Design, Sample, Setting

This research utilized a non-experimental, retrospective design to determinepredictors ofCLS in RNs with ≤24 months of clinical experience. Twenty-three United States hospitals representing Colorado, Hawaii, Kansas, Illinois, Nevada, New York, Pennsylvania, Texas, Virginia, and Washington, DC participated. All but 2 hospitals were part of a larger healthcare system. Criteria for participant selection included: currently employed in an acute care hospital, ≤ 24 months of clinical experience as an RN, consent to participate, and fluent in English. For hierarchical regression analyses with 9 predictors, Mertler and Vannatta (29) recommend using 15 subjects per independent variable to provide a reliable regression equationwhich required a sample size of 135 participants for this study (29). A total of 306participants were recruited which provided sufficient power for subsample analyses. To test the hypothesis that individual characteristics of RNs and characteristics of NGNTPs were associated with CLS, assuming alpha = .05, beta = .20, there was approximately 80% power to detect a statistically significant difference using hierarchical regression modeling.

Data Collection

Instruments

Individual characteristics of RNs and NGNTPs were collected using a researcher-developed, self-report, on-line questionnaire. CLS was measured using the Clinical Leadership (CL) Survey,aninstrument with 5 subscales based on the 5 leadership practices in Posner and Kouzes’ leadership model (4, 30 – 31) (SDC #1). The 5 leadership practices of nurses in formal leadership positions as described in Posner and Kouzes’ model of leadership include: challenging the process, inspiring a shared vision, enabling others to act, modeling the way, and encouraging the heart (30, 31). The instrument subscales of the CLSurvey that mapped to the leadership practices of Posner and Kouzes’ model were: clinical expertise (challenging the process), effective communication (inspiring a shared vision, enabling others to act, encouraging the heart), collaboration (inspiring a shared vision, enabling others to act, modeling the way, encouraging the heart), coordination (modeling the way), and interpersonal understanding (challenging the process, enabling others to act, modeling the way, encouraging the heart) (4). The original instrument was modified following confirmatory factor analysis resulting in a final instrument of 15 items with 3 items measuring each clinical leadership construct. Responses were rated using a 5-point Likert scale ranging from “almost never” to “almost always.” Mean CLS score was calculated by summing and averaging the 15 items. Overall Cronbach’s alpha for the final instrument was 0.86. Two additional items (global clinical leadership (GCL) scale, α = 0.78) were added to assessconcurrent validity with the 15 itemCLSurveyinstrument during instrument development (4). For this research study, Cronbach’s alpha for the 15 item CL Survey and the GCL scale were .90and.89. Subscales were not analyzed separately.

Data Analysis

Data are presented as mean and standard deviation (SD) or frequency and percent, where appropriate. Preliminary analysis assessed assumptions of normality, linearity and multicollinearity. With the exception of age, all independent variables (Table 1) to be included in models met assumptions of normality with the exception of age.As a result, age was evaluated for transformation using 2 separate modeling approaches; either with age inversely transformed or untransformed. Given the similarity of results between both models and ease of interpretation, the untransformed value of age was used. Intraclass correlation coefficients (ICC) (Table 2) were calculated to determine the degree to which unique characteristics within each participating hospital contributed to the overall analysis. Primary nursing degree, when coded into bivariate variables for the regression equation (Table 3), failed assumptions for multicollinearity (tolerance and variance inflation factor). These variables were not predictors of CLS and therefore were not re-coded.

Following evaluation of correlation,a series of hierarchical regression modelswere developed with all study variables. In addition, effects of hospitals/healthcare systems on mean CLS scores were evaluated by the ICC. As institutional variation accounted for less than 10% of the total variation, no analysis comparing institutions was conducted. All analyses were conducted in PASW Statistics (ver. 18.0, Chicago, IL) and statistical significance was considered for p ≤ 0.05, two-tailed.

Results

There were 4617 RNs within the 23participating hospitals and/or healthcare systems meeting eligibility requirements. Following data cleaning and removal of incomplete surveys, a total of 306 surveys (6.6%) were analyzed (Figure 2).

Demographics

The mean age was 27.85 years (range 20 – 58), with 72% of respondents aged 20 – 29 years. The majority were female (90.5%), non-Hispanic (91.9%), and Caucasian/white (74.2%). Over 75% of respondents were prepared at the baccalaureate level. The remaining respondents were prepared by associate degree (20.9%) and diploma or master’s degree (1.3%) (Table 4).

Employment Characteristics

The mean length of experience as an RN was 10.91 months (range 1 – 24). Most were still employed in the same organization (86.3%). The majority (58.2%) did not self-report previous leadership experience in a formal or volunteer positiondefined as holding a role such as director, manager or similar position supervising others, but did self-report previous formal or volunteer healthcare experience (74.8%)in roles such as nursing aide, nurse tech, nurse intern/extern, clerical in medical setting, or emergency medicine technician/paramedic. Almost all were currently employed in a staff nurse position (99.7%) (Table 4).

Characteristics of the NGNTP

The average NGNTP length was 22.6 weeks (range 4 – 52) (Table 4). Approximately half of all NGNTPs included mentor support (49.7%) and almost all included classes to improve professional development skill (84.6%). Mean value for overall quality of the NGNTP was 3.98 (SD .92) as self-reported on a 5-point Likert scale (1 – 5) (Table 4).

Comparison of Groups by Length of NGNTP

To explore differences in outcomes related to NGNTP length, four categories were createdusing quartiles, then re-grouped to approximate program lengths typically used in the clinical setting (Table 4). RNs who participated in shorter NGNTPs were significantly more likely to be older when compared with RNs who participated in longer programs(p = .025). Longer NGNTPs were perceived to have higher overall quality when compared to NGNTPs that were ≤ 12 weeks in length (p < .05). When compared to NGNTPs that were ≤ 12 weeks in length, odds ratio calculation demonstrated that RNs participating in longer NGNTPs were significantly more likely to remain employed in the organization as originally hired (Table 5). RNs participating in NGNTPs > 24 weeks were 21 times more likely to remain employed in the organization when compared to NGNTPs ≤ 12 weeks in length.

Predictors of CLS

The strongest predictors of CLS were overall quality of a NGNTP, NGNTP length, and months of clinical experience as an RN (Table 3). Previous literature suggested that CLS improves over time regardless of any intervention; therefore, regression modeling was conducted using months of clinical experience as a control variable (9). Four hierarchical regression models were developed (Table 3). Months of clinical experience as an RN (Table 3; Model 1) accounted for only 1.1% of the variability in CLS (R2 = .014, R2adj = .011, F = 4.296, p = .039). When NGNTP characteristicswere added to the model (Table 3;Model 2), these variables accounted for 6.9% of the variability in CLS (R2 = .084, R2adj = .069, F = 5.761, p = .000).

Additional Findings

Additional analyses were conducted using sublevel variables: quality of mentor support(n = 132) and perceived improvement in professional development skill (n = 258) (Table 3). Both quality of mentor support (r = .278) and perceived improvement in professional development skill (r = .201) correlated positively with CLS (p = .01). Hierarchical regression modeling with months of clinical experience as an RN and NGNTP characteristics including sublevel variables (quality of mentor support and perceived improvement in professional development skill)(n = 122) improved overall model prediction to 12.6% (R2 = .162, R2adj = .126, F = 5.203, p = .001) (Table 3).

Discussion

Based on study results, RNswith ≤ 24 months of clinical experience have higher levels of CLS when they participate in NGNTPs that they perceive to be of high quality. It is unclear what specific NGNTP characteristics result in the perception of high quality. NGNTP length correlated negatively with CLS (r = -.138, p = .05) and comparison of scores demonstrated lower CLS associated with NGNTPs > 16 weeks when compared to NGNTPs ≤ 16 weeks, however the difference was not statistically significant. This is consistent with previous research findings describing a pattern variation in CLS over the 1st post-graduate year with higher scores at baseline, dipping at 6 months, then increasing but not returning to baseline at 12 months (9).It may be that younger RNs in this study also had longer NGNTPs, which is consistent with the higher number of traditional and 2nd degree baccalaureate prepared nurses in the cohort of NGNTPs > 24 weeks and eligibility criteria for the UHC/AACN Nurse Residency™ program. Lower CLS has also been correlated in the literature with baccalaureate prepared nurses when compared to associate degree nursing (ADN) prepared nurses which may be an additional explanation. Months of clinical experience as an RN correlated positively with CLS (r = .118, p = .05) and was a statistically significant predictor of CLS (p = .039) which corresponds with previous research findings describing improvement in CLS over the 1st post-graduate year.

In sublevel analysis, overall quality of a NGNTP remained the strongest predictor of CLS. Quality of mentor support correlated positively with perceived overall NGNTP quality(r = .434, p = .000) and with CLS (r = .278, p = .01) suggesting that mentors may play an important role in perception of quality and CLS.Evidence in the literature supports the role that mentors play in supporting NGNs (17, 32 – 35). The need for mentor support may also differ for nurses prepared differently by academic degree, or the need for mentor support may vary over time. These areas were not explored in this study. The number of months of clinical experience and characteristics of NGNTPs predicted only 6.9% of the variance in CLS for this study cohort. Among RNs with an assigned mentor, months of clinical experience and characteristics of NGNTPs predicted only 12.6% of the variance in CLS for this study cohort. Much of the variance in CLS, therefore, remains unexplained.

Results from this study cohort also support previous research findings that have demonstrated a relationship between longer NGNTPs and increased NGN retention. When NGNs participated in NGNTPs that were > 24 weeks in length, they were 21 times more likely to remain employed within the organization as compared to NGNs participating in NGNTPs that were ≤ 12 weeks. The cost-benefit of retention is critically important for NGNs, preceptors/mentors, nurse managers, nurse leaders, organizations, and the nursing profession. High turnover is costly from human resource and economic perspectives. Retaining 1 NGN in an organization has an estimated cost savings of 1.2 – 1.3 times the average salary of an RN (36). The human resource costs on preceptors/mentors and nurse managers to continually orient and support new employees are burdensome and can lead to burnout further exacerbating the “revolving door” effect (36). Doubling the length of a NGNTP from 12 to 24 weeks can result in a significant return in investment.