Profiles of innovative behavior and associated predictors among clinical nurses: a multicenter study using latent profile analysis

This study used convenience sampling with a multicenter cross-sectional study. The participants come from 6 general hospitals or specialist hospitals across 4 administrative districts in Shanghai, all of whom are nursing professionals working at the frontline of clinical care. The inclusion criteria for the study population were as follows: (1) aged ≥ 18 years, (2) holding a Nurse Professional Qualification Certificate, (3) with ≥ 6 months’ practice time, and (4) voluntarily agreeing to participate in this study. And the exclusion criteria were as follows: (1) internship or probationary nurses, and (2) nurses for sick or maternity leave.

This study utilized Chinese online survey platform (https://​www.​wjx.​cn/​) to generate questionnaire links and QR codes, subsequently distributing them to the sampled hospitals from April to May 2023. This study received support and cooperation from the nursing department directors of each survey site, while the nursing unit heads mobilized clinical nurses to participate in the survey. The online questionnaire can only be accessed, filled out, and submitted after obtaining confirmation of the participant’s informed consent.

Mistakes, such as missing or inaccurate responses, triggered prompts upon submission. Upholding survey integrity only fully and accurately filled questionnaires were accepted. Respondents had flexibility, utilizing desktops, tablets, mobiles, and other internet-enabled devices. Each account and device allowed completion of only one questionnaire to prevent duplication.

According to the Kendall criterion [47], multivariate analysis requires a sample size 5 ~ 10 times the number of variables. With 14 general data variables and 12 dimensions from three scales, a 15% buffer for potential invalid responses increased the estimated sample size to 144 ~ 288. Ultimately, 361 clinical nurses participated, and 354 valid questionnaires were received, achieving an effective response rate of 98.06%.

Mplus 7.0 software (Muthén & Muthén) was employed to conduct LPA, utilizing individual item scores of the IBSN as the observed variables. Multiple indicators were used to evaluate the quality of the different models. Firstly, fit indicators—Akaike information criterion (AIC), Bayesian information criterion (BIC), and adjusted BIC (aBIC)—were considered [48]. Additionally, the classification index, Entropy (ranging from 0 to 1), was examined. Finally, likelihood ratio test indicators, namely Lo-Mendell-Rubin likelihood ratio test (LMR), adjusted LMR (ALMR), and bootstrapped parametric likelihood ratio test (BLRT), were analyzed. The best model is characterized by the following criteria: when AIC, BIC, and aBIC in the model are minimized; Entropy value is greater than 0.8; the statistical significance of LMR, ALMR, and BLRT is less than 0.05; and the profiles are both concise and interpretable.

SPSS 26.0 software (IBM Corp.) was employed to conduct statistical analysis. Counting data were described with frequency and percentage, and measurement data with mean and standard deviation. The Pearson’s \(\:{\chi\:}^{2}\) test and one-way ANOVA identified differences in innovative behavior profiles based on socio-demographic and professional characteristics. In multivariate analysis, ordinal logistic regression analysis was used after passing the parallel line test. Statistically significant variables in univariate analysis served as independent variables, and latent class analysis results as the dependent variable. The relationship between clinical nurses’ innovative behavior profiles and research output indicators was assessed with Pearson’s \(\:{\chi\:}^{2}\) test. Differences within the same research output indicators subgroup were evaluated using adjusted standardized residuals (ASR) in post-hoc testing. An absolute ASR value greater than 3 indicated a statistically significant difference [49‐51].

Table 1 details the socio-demographic and professional characteristics of 354 clinical nurses in Shanghai. The age distribution indicates a significant proportion within the 25 to 35 age range (55.9%), with a predominant representation of females (94.6%). Marital status is balanced between married (56.2%) and single (43.8%). Educational attainment is largely at the undergraduate level (66.7%). The respondents are employed in either general hospitals (50.6%) or specialist hospitals (49.4%), and worked extensively in different units. The prevalent professional title is registered nurse (71.5%), with employment primarily under establishment system (54.0%). Work experience is concentrated in the less-than-5-years (41.8%). The majority of participants receive a monthly income of 10,000 CNY or more (62.1%). Night shift frequencies vary, with a substantial portion engaging in 1-to-4-night shifts per month (50.6%).

Table 2 outlines a model fit comparison of LPA with varying numbers of profiles. The 1-profile model serves as a baseline with 100% latent profile proportion. As shown in Fig. 1, as the number of profiles increases, fit indicators generally decrease, indicating improved model fit. The entropy value provides a measure of classification accuracy, with higher values indicating better classification. Likelihood ratio tests assess the significance of adding profiles, with p-values indicating statistical significance. Notably, the 3-profile model appears to strike a balance between model fit and simplicity, with significant improvements over the 2-profile model based on likelihood ratio tests.

Based on the identification of latent profile models, Figs. 2 and 3 illustrate the distribution probability of innovative behavior among clinical nurses across various items and dimensions. In the overall sample, Class 1, Class 2, and Class 3 exhibited probabilities of 30.51% (n = 108), 42.09% (n = 149), and 27.40% (n = 97), respectively. The mean IBSN scores for Class 1, Class 2, and Class 3 were (23.01 ± 3.75), (31.70 ± 2.72), and (42.07 ± 4.22), respectively. Significant differences were further identified through pairwise comparisons within each dimension (all p < 0.001). Therefore, categories Class 1, Class 2, and Class 3 were aptly named as low-level innovative behavior (LIB), moderate-level innovative behavior (MIB), and high-level innovative behavior (HIB).

Table 3 lists the socio-demographic and professional characteristics associated with latent profiles of innovative behavior. Significant associations were found in marital status (p < 0.001), with a higher proportion of single individuals in the LIB group. Education level exhibited differences (p = 0.002), with more than 70% master’s or above located in the HIB group. Also, variations in professional titles (p < 0.001), employment methods (p = 0.023), working experience (p = 0.026), income levels (p < 0.001), and night shift frequencies (p < 0.001) were evident across the profiles.

Table 4 presents the scores for future time perspective and work engagement across different latent profiles of innovative behavior. Significant differences were observed in all variables. Future time perspective scores significantly varied among the profiles (p < 0.001), with the HIB group demonstrating the highest scores. Similar patterns were observed for work engagement (p < 0.001).

Table 5 displays the results of ordinal logistic regression analysis predicting latent profile membership. Marital status emerged as a significant predictor, with married individuals being more likely to belong to higher innovative behavior profiles (OR = 5.186, p < 0.001). Master’s degree or above significantly predicts higher profile membership (OR = 7.561, p = 0.012). Longer working experience (≥ 10 years), higher average monthly income (≥ 10,000 CNY) and no night shifts per month were associated with increased odds of higher profile membership (OR = 2.575, p = 0.043; OR = 5.590, p < 0.001; OR = 2.438, p = 0.037, respectively). Future time perspective (OR = 1.057, p < 0.001), and work engagement (OR = 1.059, p < 0.001) also emerged as significant positive predictors of profile membership.

Statistically significant associations were detected between clinical nurses’ innovative behavior profiles and their research output indicators (all p < 0.001) by Pearson’s chi-squared test revealed in Table 6. Moreover, the post-hoc test demonstrate that clinical nurses in the HIB group have a significantly higher likelihood of publishing academic papers (ASR = 3.5), registering patents (ASR = 3.8), and winning Sci. & Tech awards (ASR = 5.2) compared to expectations, while the LIB group exhibits lower-than-expected numbers in those areas. These results suggest that higher levels of innovative behavior are positively correlated with increased research productivity and recognition.

Based on their respondent, we categorized clinical nurses into three groups via LPA: a high-level innovative behavior group (27.40%), a moderate-level innovative behavior group (42.09%), and a low-level innovative behavior group (30.51%). The results showed an “olive-shaped” distribution structure with small percentages at both extremes and a larger percentage in the middle. Similar to the study conducted by Lin [38] et al., our research indicates that the current level of innovative behavior among clinical nurses is predominantly at a moderate level, with less than 30% demonstrating a high level of innovation. This suggests that there is still much to improve regarding clinical nurses’ innovative behavior. We also noted a descending trend in the mean scores of the items in the three dimensions of innovative behavior: generating ideas, obtaining support, and realizing ideas. This suggests that nurses often generate innovative ideas during their work, but there may be insufficient support during the innovation process, potentially hindering the implementation of these ideas. Therefore, managers’ attention to nurses’ innovative efforts and supportive policies serves as the cornerstone for fostering innovation among nurses in the workplace [52].

Meanwhile, we should also recognize the intricate and multifaceted nature of innovative behavior. Clinical nurses work within a dynamic interplay of personal, organizational, and environmental factors. These factors collectively shape and influence the extent to which nurses engage in innovative practices [3, 40, 53]. Acknowledging this complexity is crucial for tailoring interventions and support mechanisms that address the specific needs and challenges faced by clinical nurses in their pursuit of innovation. A nuanced approach that considers individual aspirations, organizational culture, and the broader environmental context will undoubtedly contribute to a more effective and sustainable promotion of innovative behavior among clinical nurses.

After our study divided innovative behavior into three profiles, and further regression analysis revealed found that marital status, education level, work experience, monthly income, night shifts, future time perspective, and work engagement were associated with innovative behavior profiles.

Although using different instruments to measure innovative behavior, several studies [38, 54] have proved a similar result that better socio-demographic status and professional characteristics positively influenced innovative behavior among clinical nurses. Some potential mechanisms can explain the relationship. In terms of marital status, married status may provide a more stable and supportive family environment, and this stability and support may give individuals more confidence and motivation to pursue innovation [55]. Furthermore, nurses with longer work experience and higher education levels demonstrate significantly higher innovation capabilities compared to those with shorter work experience and lower education levels. This pattern may be attributed to the accumulation of rich clinical experience as work tenure increases, laying the foundation for innovation among nurses [56]. And nurses with higher educational qualifications are more adept at utilizing existing research tools and conditions to drive innovation [12]. The association with monthly income points to the potential influence of financial stability on a nurse’s ability to innovate [55]. In addition, too much night work may interfere with nurses’ biological clocks, leading to fatigue and low energy, reducing innovative thinking and creativity [55]. These finding prompts considerations about how financial incentives, job satisfaction, and innovative behavior might be interconnected.

The positive association between future time perspective and innovative behavior among nurses unveils a compelling avenue for fostering innovation within the nursing profession. Future time perspective, characterized by an individual’s consideration and planning for future events, emerges as a significant predictor of nurses’ innovative behavior. Individuals with a forward-looking outlook are more likely to engage in activities that contribute to long-term goals, such as adopting innovative practices in their professional endeavors. The significance of future time perspective in fostering nurse innovation is diverse, as nurses with this mindset are inclined to invest in skills acquisition, explore novel patient care approaches, and embrace healthcare advancements. This forward-thinking approach catalyzes proactive problem-solving and creative solutions generation in nursing challenges [57]. The positive correlation underscores the importance of nurturing a visionary mindset among nurses. Educational interventions emphasizing the long-term impact of innovation on patient outcomes and healthcare advancements can enhance future time perspective. By instilling purpose and highlighting the potential benefits of innovation, nurses may be more motivated to actively seek and implement innovative solutions in their practice.

We also explored the relationship between work engagement and nurses’ innovative behavior, recognizing the pivotal role that an engaged workforce plays in fostering innovation within healthcare settings. The findings revealed a significant positive association between work engagement and various dimensions of innovative behavior. Work engagement, characterized by vigor, dedication, and absorption in one’s work, emerged as a crucial factor influencing nurses’ propensity for innovation [58]. The positive correlation suggests that nurses who are actively involved and enthusiastic about their work are more likely to exhibit innovative behavior. The aligns with Zhou [53] et al. emphasizing the link between employee engagement and organizational innovation. The observed association can be attributed to several factors. Engaged nurses may feel a heightened sense of purpose and commitment, motivating them to actively seek and implement innovative solutions to enhance patient care and workflow efficiency. Furthermore, a work environment that fosters engagement likely encourages open communication, collaboration, and idea-sharing among nurses, creating a conducive atmosphere for innovation. Considering these findings, healthcare organizations should consider strategies to enhance work engagement among nursing staff. This may involve creating a supportive work culture that values and recognizes nurses’ contributions, providing opportunities for skill development and professional growth, and ensuring that nurses have the autonomy and resources needed to implement innovative ideas.

In this study, the factors incorporated into the regression equation collectively accounted for 43.7% of the variance in the innovative behavior of clinical nurses. This suggests that innovative behavior may be subject to the influence of additional important factors. Further research is warranted to explore and identify these factors, contributing to a more comprehensive understanding of the determinants of nurses’ innovative behavior.

Innovative behavior is an outward manifestation of innovative capability [59]. Our results indicated the reciprocal relationship between innovative behavior profiles and research output. Post-hoc analyses revealed more insight into noteworthy patterns within groups of innovative behaviors. Specifically, clinical nurses classified under the HIB group demonstrated a markedly higher likelihood of engaging in scholarly activities. In contrast, the LIB group displayed a contrasting trend, with lower-than-expected numbers in these key areas of research productivity. This suggests that a deficiency in innovative behavior might be a limiting factor for clinical nurses in achieving notable success in terms of academic publications and patent registrations. These findings underscore the importance of cultivating innovative behavior in clinical nurses, not only for their personal and professional growth but also for advancing scientific knowledge and technological innovation in nursing. The observed positive correlation between higher levels of innovative behavior and increased research productivity aligns with existing literature [60], emphasizing innovation’s crucial role in healthcare advancement.

Furthermore, our study highlights the potential benefits of acknowledging and rewarding innovative behavior. The notably higher likelihood of winning Sci. & Tech awards among the HIB group suggests that celebrating innovation could motivate nurses to actively participate in research. Considering this, nursing administrators may consider implementing incentive programs and awards not only to recognize achievements but also to foster a culture of continuous improvement and innovation.

Nursing administrators play a vital role in implementing effective incentive systems and fostering an innovative culture among nurses. Based on our findings, several targeted interventions can be introduced [61‐65]:

By implementing these targeted interventions, healthcare organizations can unlock the innovative potential within clinical nurses and cultivate a work culture that actively encourages innovation. These measures not only enhance patient care but also improve job satisfaction and retention among nursing staff, making innovation a strategic investment in the future of healthcare.

This study may possess several limitations. Firstly, being a web-based survey, the sampling process and self-reporting strategy probably introduce certain biases. Secondly, our survey exclusively targeted clinical nurses in Shanghai, and caution should be exercised in generalizing the findings to broader populations. Lastly, the study’s design constrains making causal inferences from the results.

To address these limitations, future research should adopt more robust study designs. Longitudinal studies could track changes and trends in innovative behavior over time, providing insights into its development and influencing factors. Mixed-methods approaches could offer a deeper understanding by combining quantitative data with qualitative perspectives. Expanding the scope of research to include diverse geographic regions and healthcare settings would further enhance the generalizability of the findings.