Impact of Nurses' Knowledge, Self‐Efficacy and Clinical Reasoning Competency on Difficulties in Caring for Patients With Delirium in the Intensive Care Unit: A Cross‐Sectional Study

ABSTRACT

Aim

To examine the impact of critical care nurses' delirium knowledge, self‐efficacy and clinical reasoning competency on delirium care difficulties based on the information–motivation–behavioural (IMB) skills model from a behavioural perspective.

Design

Cross‐sectional study.

Methods

A total of 440 critical care nurses from five hospitals in China were selected using convenience sampling and invited to complete an online questionnaire for measurement. Data were collected in November 2024 and analysed using SPSS/AMOS with descriptive statistics, Pearson's correlation coefficient and multiple regression. Structural equation modelling was constructed to test the hypothesised relationships among the variables, with bootstrapping to assess mediation effects.

Results

The level of delirium care difficulties was moderated. Delirium care difficulties were negatively correlated with delirium knowledge, self‐efficacy and clinical reasoning competency. Clinical reasoning competency partly mediated delirium knowledge and self‐efficacy with regard to delirium care difficulties.

Conclusion

Delirium knowledge, self‐efficacy and clinical reasoning competency are essential for improving critical care nurses' delirium care competencies. The role of clinical reasoning competency in the relationship between the other two variables and delirium care difficulties was highlighted. Establishing multifaceted innovative delirium education programmes, emphasising individuals' sense of competence and enhancing clinical reasoning competency as behavioural skills were supported. Exploring these pathways using a nurse behaviour change‐based perspective is critical.

Implications for the Profession and/or Patient Care

Critical care managers should value nurses' delirium care competencies. Enhancing continuing professional development through system‐level support with high reliability and multiform professional education, including innovative theoretical and practical training; advancing policies that increase work motivation and self‐planning to stimulate self‐efficacy; and exercising critical and reflective thinking to improve clinical reasoning competency may enhance nurses' delirium recognition and care competencies, including prioritisation, potentially improving delirium care dilemmas and patient outcomes.

Reporting Method

The STROBE checklist was used as a guideline.

Patient or Public Contribution

Nurses completed questionnaires.

Trial Registration: Chinese Clinical Trial Registry (ChiCTR2400092177). https://www.chictr.org.cn/bin/project/edit?pid=249216

Summary.

• What does this paper contribute to the wider global clinical community?

  • ○
    This study revealed the factors influencing ICU nurses' delirium care difficulties, and the intrinsic relationships and interaction mechanisms between ICU nurses' delirium knowledge, self‐efficacy, clinical reasoning competency and delirium care difficulties.
  • ○
    The delirium knowledge, self‐efficacy and clinical reasoning of ICU nurses had a positive effect on alleviating delirium care difficulties and clinical reasoning competency partly mediated delirium knowledge and self‐efficacy with regard to delirium care difficulties.
  • ○
    From the perspectives of information, personal motivation and behavioural skills, these results implied factors that may influence behavioural changes related to ICU nurses' delirium care difficulties, thereby enhancing the understanding of the mechanisms linking these variables and pathways, and facilitating critical care managers to value and optimise delirium education, self‐efficacy and clinical reasoning training for ICU nurses.

1. Introduction

Delirium is a type of acute brain dysfunction characterised by inattention, confusion and cognitive dysfunction (American Psychiatric Association D ²⁰¹³). The incidence of delirium among hospitalised patients is 23%–35% (Gibb et al. 2020; NICE ²⁰²³), with the highest incidence (83.9%) occurring in the intensive care unit (ICU) (Miranda et al. 2023; van den Boogaard et al. ²⁰¹²; Inouye et al. ²⁰¹⁴). Long‐term consequences experienced by ICU patients with delirium include a decline in cognitive function, mental health and daily activities (Stollings et al. 2021), leading to caregiver burden for families (Boehm et al. 2021). Thus, the early diagnosis, prevention and management of ICU delirium are important (Devlin et al. 2018; Lewis et al. ²⁰²⁵).

An evidence‐based approach favours non‐pharmacological interventions for delirium management, focusing on delirium prevention and monitoring, pain management, reorientation, anxiety management, early enhanced mobilisation/rehabilitation over usual mobilisation/rehabilitation, environmental improvements such as improving circadian rhythms by reducing light and noise and increasing natural lighting to create rooms with ambient features and cognitive stimulation (Lewis et al. 2025; Chen et al. ²⁰²²; Deng et al. ²⁰²⁰; Sosnowski et al. ²⁰²³; Hamilton et al. ²⁰²⁵). Delirium monitoring in hospitals often relies on reports from patients at risk, nurses, or relatives (NICE ²⁰²³); because of the nature of the ICU, it mostly relies on nurses. As part of a multidisciplinary team, ICU bedside nurses, who are in constant contact with the patient throughout their delirium care, are optimally positioned to recognise symptom fluctuations, communicate and provide direct care (Karabulut and Yaman Aktaş ²⁰¹⁶); they play an essential role in all phases of delirium management (Donovan et al. 2018; Boot ²⁰¹²). However, ICU nurses face many challenges (Thomas et al. 2021; Partridge et al. ²⁰¹³; Dos Santos et al. ²⁰²²; Jiang et al. ²⁰²⁴), with a high degree of subjective burden due to the unpredictability and complexity of delirium, increased workloads, safety concerns and difficulty communicating with delirious patients (Zamoscik et al. 2017; Correya et al. ²⁰²⁵; Schmitt et al. ²⁰¹⁹; Mc Donnell and Timmins ²⁰¹²; Lee and Roh ²⁰²³; LeBlanc et al. ²⁰¹⁸). Meanwhile, the high incidence and severity of delirium in the ICU, perceived barriers and prolonged ICU stays exacerbate ICU nurses' suffering and sense of internal conflict (Elliott 2014; Bélanger and Ducharme ²⁰¹¹). Perceived burden may lead to burnout among nurses and hinder the effective practice of nursing strategies (Zamoscik et al. 2017; Wang et al. ²⁰²³), impacting patients' clinical outcomes and prognoses. Hence, identifying the factors that influence ICU nurses' challenges is critical for providing targeted, practical strategies that promote behavioural changes related to issues in delirium care, optimise the quality of care to improve patient outcomes and reduce nurses' burden.

The factors influencing behavioural changes during practice must be explored. Nurses' knowledge of delirium is fundamental for assessment and management (Lee and Roh 2023). Lack of knowledge about the causes of delirium and best‐practice approaches can cause ICU nurses to miss or misdiagnose delirium (Elliott 2014; Rowley‐Conwy ²⁰¹⁸), hindering the clinical implementation of established guidelines for delirium management and confidence in delirium care (Zamoscik et al. 2017; Emme ²⁰²⁰). Although studies have examined the current status of ICU nurses' delirium knowledge and perceived barriers (Lee and Roh 2023; Zhou et al. ²⁰²²; Mathew et al. ²⁰²⁴), few have used statistical methods to identify how the factors and action pathways of ICU nurses' enhanced delirium knowledge affect behavioural changes related to delirium care for intervention readiness. Consequently, ICU nurses report poor delirium recognition and management because of insufficient access to and poor quality of delirium education (Jiang et al. 2024; Elliott ²⁰¹⁴).

A lack of delirium knowledge is associated with low self‐efficacy in delirium care (Nie et al. 2024), which is important for helping ICU nurses measure and develop their proficiency in managing delirium (Yu et al. 2023). This potentially leads to difficulties in delirium care as self‐efficacy mediates the relationship between knowledge and behaviour (Plaza et al. 2002). Self‐efficacy is the ability to integrate and reshape one's comprehensive abilities to achieve a self‐growth goal (Hughes et al. 2011; Zulkosky ²⁰⁰⁹; Scholz et al. ²⁰⁰²; Schwarzer ²⁰¹⁴). Nurses' self‐efficacy is reflected in their confidence in their theoretical knowledge, practical skills and overall nursing competence (Nie et al. 2024). Specifically, delirium care for ICU nurses refers to their subjective judgements of their competence in delirium care (Caruso et al. 2016). Higher self‐efficacy in delirium care among ICU nurses could reduce obstacles in this area. Nurses with high self‐efficacy usually believe in their ability to address difficulties based on their knowledge and skills (Wang et al. 2022); thus, they can confidently recognise delirium and implement early management, which may reduce the extent of their challenges. Research exploring nurses' self‐efficacy for delirium care in critical care settings is limited.

Adequate knowledge and high levels of self‐efficacy in nursing can help nurses become more self‐assured in complex clinical decision‐making situations; this is known as ‘clinical reasoning’ (Banning 2008; Nafea ²⁰²²), which involves gathering information, understanding the problem, assessing the patient's condition, implementation and evaluation (Lee and Park 2019). Delirium in ICU patients is challenging to recognise (Stollings et al. 2021; Elliott ²⁰¹⁴; Fann ²⁰⁰⁰; Kotfis et al. ²⁰¹⁸) as they experience rapid and varied symptomatic changes (Lindroth et al. 2020), making it difficult for ICU nurses to manage the condition. Thus, nurses require strong clinical reasoning skills to assess and cope with the complex changes in patients' condition and needs, rather than relying on intuitive judgement and management (El Hussein and Hirst ²⁰¹⁶; Hoch et al. ²⁰²²). This can help nurses provide timely care to facilitate the potential for reducing issues with delirium care.

As for ICU nurses' challenges, current findings are mostly qualitative and based on self‐reports (Thomas et al. 2021; Partridge et al. ²⁰¹³; Dos Santos et al. ²⁰²²). In‐depth quantitative studies exploring the diverse factors affecting nurses' obstacles to providing delirium care are lacking. Several studies have attempted to analyse the individual relationships between the variables linking delirium knowledge with self‐efficacy, clinical reasoning and issues in providing delirium care among ICU nurses (Jiang et al. 2024; Nie et al. ²⁰²⁴). However, no study has examined their associations using one model grounded in a theoretical framework or how these variables collectively affect difficulties in offering delirium care in the ICU from the perspective of promoting behavioural change and connecting underlying processes rooted in behavioural theory.

1.1. Conceptual Framework

This empirical investigation is based on the information–motivation–behavioural (IMB) skills model proposed by Fisher and Fisher (1992) and Fisher et al. (2003, 2006). This model suggests that information, motivation and behavioural skills can directly influence behaviour, whereas information and motivation can indirectly influence behaviour through behavioural skills. In this context, information indicates the knowledge of disease management; motivation refers to what drives behavioural change; and behavioural skills are objective skills for managing behaviour. Information is the basic condition for behaviour implementation, motivation is the driving force for behaviour execution and behavioural skills are the facilitators that promote behavioural change. When a person possesses information, motivation and behavioural skills, behavioural changes are promoted to a certain extent.

2. The Study

2.1. Aims and Hypotheses

This study aimed to reveal the intrinsic relationships and interaction mechanisms between ICU nurses' delirium knowledge, self‐efficacy, clinical reasoning competency and delirium care difficulties, which is critical to improving the prognosis of ICU patients with delirium and reducing nurses' work stress and burnout. We explored the potential relationships according to the theoretical framework of the IMB skills model (Figure 1).

Hypothesis 1

Delirium knowledge is negatively associated with delirium care difficulties.

Hypothesis 2

Delirium knowledge is positively associated with clinical reasoning competency.

Hypothesis 3

Delirium knowledge is positively associated with self‐efficacy for delirium care.

Hypothesis 4

Self‐efficacy for delirium care is negatively associated with delirium care difficulties.

Hypothesis 5

Self‐efficacy for delirium care is positively associated with clinical reasoning competency.

Hypothesis 6

Clinical reasoning competency is negatively associated with delirium care difficulties.

Hypothesis 7

Clinical reasoning competency partially mediates delirium knowledge and delirium care difficulties.

Hypothesis 8

Clinical reasoning competency partially mediates self‐efficacy for delirium care and delirium care difficulties.

FIGURE 1.

Theoretical model and hypotheses.

3. Methods

3.1. Design

This study was a hospital‐based, multicenter, cross‐sectional survey using convenience sampling and followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting (von Elm et al. ²⁰⁰⁸) (File S1).

3.2. Study Setting and Sampling

This was a hospital‐based, multicentre, cross‐sectional survey of ICU nurses in five hospitals in Hangzhou Province, China, recruited using convenience sampling in November 2024. Because of the high prevalence of respiratory infectious diseases including influenza and COVID‐19 at the time of data collection for this study in winter, especially the high number of patients in ICUs of general hospitals who were admitted with critical illnesses directly or indirectly triggered by such respiratory infectious diseases, resulting in a wider and faster spread of such respiratory infectious diseases in ICUs, the researchers considered adopting widely applied convenience sampling method of non‐probability sampling method for the survey given the high efficiency of the data collection. However, it must be recognised that this sampling method may increase selection bias and limit the possibility of generalising the results to a wider geographical area or hospital grades (Bornstein et al. 2013).

According to Kendall's principle of sample estimation, the sample size was 5–10 times the number of variables (Ni et al. 2010), with 35 independent variables in this study. Considering an invalid response rate of 10%–20%, 389–438 participants were needed. Mueller (1997) estimated the sample size of a structural equation model (SEM) to be more than 200. Ultimately, 440 valid questionnaires met the statistical requirements.

3.3. Inclusion and Exclusion Criteria

We included nurses if they (a) were registered nurses working in the ICU, (b) had worked in the ICU for more than 1 year and (c) were willing to participate. We excluded informally employed nurses, such as interns and those not on duty owing to vacation, sick leave, or training.

3.4. Instrument With Validity and Reliability

3.4.1. Participants' General Characteristics

We designed a questionnaire, which included gender, age, marital status, education level, hospital grade, department, professional title, duty, ICU nursing experience, employment mode, average monthly income, delirium training and care experience, delirium assessment, management and learning approach.

3.4.2. Delirium Knowledge Test for the Intensive Care Nurses

The scale was used to assess ICU nurses' delirium knowledge acquisition. It was developed by Öztürk Birge et al. (2020) and adapted to the Chinese context by Chinese scholars Liu et al. (2022). The scale consists of five dimensions: delirium risk factors and causes (10 items), delirium characteristics and types (3 items), delirium outcomes (3 items), delirium screening (3 items) and pharmacological and non‐pharmacological management of delirium (7 items) – a total of 26 items, seven of which are reverse scored, each correct option is assigned a score of ‘1’, and I don't know, or the incorrect option is assigned a score of ‘0’, with a maximum score of 26. The higher the score, the better the nurse's knowledge of delirium. The Cronbach's α coefficient of this scale is 0.884, and the retest reliability is 0.753, which has good reliability in the Chinese population (Liu et al. 2022). Cronbach's α is 0.710 in the present study.

3.4.3. Delirium Care Self‐Efficacy Scale for ICU Nurses

The scale was developed by Chang et al. (2023) to assess nurses' confidence about caring for ICU patients, including 13 items across two dimensions: confidence in delirium assessment (7 items) and confidence in delirium management (6 items). Each response is converted to a numerical score from 1 to 5, where 1 indicates very uncertain and 5 indicates complete confidence, and the higher the score, the higher the perceived confidence in self‐efficacy for delirium care. The Cronbach's α for this scale was 0.94 (Chang et al. 2023). In the present study, Cronbach's α is 0.972.

3.4.4. Clinical Reasoning Competency Scale

The scale, which assesses nurses' decision‐making and clinical reasoning skills, was developed by Bae et al. (2023) and revised for ICU nurses by Mohammadi et al. (2024). The scale consists of 22 items across three dimensions: plan setting (11 items), intervention strategy regulation (8 items) and self‐instruction (3 items). A 5‐point Likert scale was used, with total scores ranging from 22 to 110, with higher scores associated with higher clinical reasoning ability. The Cronbach's α for this scale was 0.91 (Mohammadi et al. 2024), which is 0.987 in the present study.

3.4.5. Difficulty Scale for Nurses Who Care for Patients With Delirium in the Intensive Care Unit

This scale was used to assess the difficulties faced by ICU nurses caring for patients with delirium, which was developed by Owaki et al. (2023) and adapted for the Chinese context by Jiang et al. (2024) in 2023. The scale consists of nine dimensions across 37 items. The eight dimensions of the main scale are delirium assessment (6 items), delirium management by multidisciplinary team (5 items), ensuring safety (5 items), dealing with and involvement in hypoactive delirium (4 items), dealing with stress and distress (5 items), adjustment of medication (4 items), involvement of family (2 items) and lack of resources (2 items). The additional scale was completed by nurses who had used the delirium screening tool, consisting of one dimension with four items: using delirium screening tools. A 4‐point Likert scale was used, ranging from ‘disagree’ to ‘agree’ on a scale of 1–4, with higher scores indicating greater difficulty in delirium care for ICU nurses. Cronbach's α of the total scale was 0.875, and that of the additional scale was 0.756 in the work of Jiang et al. (2024). In the present study, Cronbach's α of the total scale is 0.975, and that of the additional scale is 0.909.

3.5. Data Collection

We collected data since November 2024 using an online survey platform. Before the formal survey, we selected 20 ICU nurses who met the study requirements for the pre‐survey. We contacted the ICU head nurses to explain the aims and requirements of the study. After obtaining consent and support, we shared the link to the online questionnaire with ICU nurses. The online questionnaire was considered successful only when all questions had been answered and each participant received a red packet after submission; thus, no missing data were reported. Each participant submitted the questionnaire once to avoid duplicate submissions. Options in the same position were selected for all questions (three questionnaires). Questionnaires that took less than 1 min to answer (one questionnaire) were considered invalid and deleted. Meanwhile, to enhance the consistency of the sample and control for the influence of cultural and systemic healthcare factors, we excluded four questionnaires from secondary hospitals, retaining 440 questionnaires exclusively from tertiary hospitals in Hangzhou, Zhejiang Province. Given Hangzhou's relatively homogeneous medical resources and economic development, the sample demonstrates good consistency in cultural and systemic healthcare factors, which helps to minimise external variations affecting variables in this study. Consequently, there were 440 valid questionnaires.

3.6. Statistical Analysis

SPSS version 26.0 (SPSS Inc.) and AMOS version 24.0 (IBM Corp.) were used for statistical analysis. The count data were described by frequency and percentage, and the measurement data were described by the mean ± standard deviation. All variables were assessed for common method bias using Harman's single‐factor test. Spearman correlation analysis was performed to analyse the relationship among ICU nurses' delirium knowledge, self‐efficacy for delirium care, clinical reasoning competency and delirium care difficulties. Multiple regression analysis was used to analyse the factors affecting ICU nurses' delirium care difficulties. A structural equation model (SEM) was employed to identify both direct and indirect relationships in the model. Standardised and unstandardised path coefficients, variances, R‐squared values, standard error and p values were reported. Variables with nonsignificant coefficients were removed from the model. The structural model fit was evaluated according to the following standards: χ ²/DF ≤ 5.00, comparative fit index (CFI) ≥ 0.90, relative fit index (TLI) ≥ 0.90, normative fit index (NFI), incremental fit index (IFI) ≥ 0.90 and root mean square error of approximation (RMSEA) ≤ 0.08 (Abd‐El‐Fattah ²⁰¹⁰). In this study, α = 0.05, whereas tests were two‐tailed. Bootstrapping was performed to test the mediation effect; estimates without zero in the 95% confidence interval (CI) indicated that the mediation effects were significant.

3.7. Ethical Considerations

This study was approved by the Clinical Research Ethics Committee on 6 November 2024 (No. 20240602) and adhered to the principles of the Declaration of Helsinki. All eligible potential participants were informed of the details of the study before completing the electronic questionnaire and completed an informed consent form through a unique link in the online questionnaire, selecting ‘Agree’ to be enrolled in the study. Participation was voluntary with an option to withdraw, and participants were assured of the anonymity and confidentiality of responses to protect privacy strictly. The tools used in this study were authorised by the original authors. This study has been registered in the Chinese Clinical Trial Registry (ChiCTR2400092177).

4. Results

4.1. Characteristics of the Sample

The general characteristics of 440 ICU nurses are shown in Table 1. 81.1% of the participants were female (n = 357), with a higher percentage of 20–29 (n = 219, 49.8%) and 30–39 (n = 196, 44.5%) age groups. Most participants had a Bachelor's degree (n = 423, 96.1%). Most participants had received ICU delirium training (n = 411, 93.4%), cared for patients with ICU delirium (n = 431, 98.0%) and used both pharmacological and non‐pharmacological delirium management strategies (n = 380, 86.4%). The main delirium learning approach was specialist training (n = 334, 75.9%).

TABLE 1.

Demographic characteristics of ICU nurses (N = 440).

Variable	N (%)
Gender
Male	83 (18.9)
Female	357 (81.1)
Age (years)
20–29	219 (49.8)
30–39	196 (44.5)
40–49	19 (4.3)
≥ 50	6 (1.4)
Marital status
Married	211 (48.0)
Unmarried	226 (51.4)
Divorced/Widowed	3 (0.7)
Education level
Junior college	3 (0.7)
Bachelor's degree	423 (96.1)
Master's degree	11 (2.5)
Doctoral degree and above	3 (0.7)
Department
General ICU	366 (83.2)
Medical ICU	18 (4.1)
Surgical ICU	56 (12.7)
Professional title
Nurse	52 (11.8)
Junior Nurse	238 (54.1)
Supervisory Nurse	135 (30.7)
Associate Nurse	11 (2.5)
Chief Nurse	4 (0.9)
Duty
Nursing Team Leader	79 (18.0)
Nursing instructor	49 (11.1)
Head Nurse	13 (3.0)
None	299 (68.0)
Nursing experience in the ICU (years)
1–5 years	203 (46.1)
6–10 years	113 (25.7)
11–20 years	112 (25.5)
More than 21 years	12 (2.7)
Employment mode
Formal establishment	211 (48.0)
Labour dispatch	229 (52.0)
Average monthly income (RMB)
≤ 5000	25 (5.7)
5001–10,000	181 (41.1)
10,001–15,000	205 (46.6)
> 15,000	29 (6.6)
Delirium training
Yes	411 (93.4)
No	29 (6.6)
Delirium care experience
Yes	431 (98.0)
No	9 (2.0)
Delirium assessment approach in the department
Based directly on nurses' clinical experience	43 (9.8)
Validated delirium screening tools	376 (85.5)
ICU physician assessment	19 (4.3)
Psychiatric consultation assessment	2 (0.5)
Delirium management approach in the department
Only non‐pharmacological management Strategies such as bundle care	35 (8.0)
Only pharmacological management Strategies such as analgesia and sedation	15 (3.4)
Both strategies above	380 (86.4)
None of the above management strategies	4 (0.9)
Don't know	6 (1.4)
The main delirium learning approach
Specialist training	334 (75.9)
Continuing education	37 (8.4)
Knowledge manual/Literature/Lectures and exchange sessions	24 (5.5)
Accumulation of work experience/Colleague exchange	45 (10.2)

4.2. Descriptive Statistics of Variables

The total mean score of ICU nurses' delirium care difficulties was 77.432 ± 21.492, which was at the low‐middle level. Among the subdimensions, dealing with stress and distress had the highest scores. The mean total scores of ICU nurses' delirium knowledge, self‐efficacy for delirium care and clinical reasoning were 19.631 ± 3.889, 45.636 ± 11.697 and 88.267 ± 15.269, respectively, at the upper‐middle level (Table 2).

TABLE 2.

Descriptive analyses of ICU nurses' delirium knowledge, self‐efficacy for delirium care, clinical reasoning competency and delirium care difficulties (N = 440).

Variables	Range	Mean ± SD
ICU nurses' delirium knowledge	0–26	19.631 ± 3.889
Delirium risk factors and causes	0–10	6.755 ± 1.917
Delirium characteristics and types	0–3	2.164 ± 0.799
Delirium outcomes	0–3	2.441 ± 0.734
Delirium screening	0–3	2.323 ± 0.773
Pharmacological and non‐pharmacological management of delirium	0–7	5.950 ± 1.203
ICU nurses' self‐efficacy for delirium care	13–65	45.636 ± 11.697
Confidence in delirium assessment	7–35	24.046 ± 6.406
Confidence in delirium management	6–30	21.591 ± 5.594
ICU nurses' clinical reasoning competency	22–110	88.267 ± 15.269
Plan setting	11–55	43.475 ± 7.754
Intervention strategy regulation	8–40	32.580 ± 5.766
Self‐instruction	3–15	12.173 ± 2.240
ICU nurses' delirium care difficulties	37–148	77.432 ± 21.492
Delirium assessment	6–24	11.864 ± 3.751
Delirium management by multidisciplinary team	5–20	10.423 ± 3.307
Ensuring safety	5–20	10.809 ± 3.582
Dealing with and involvement in hypoactive delirium	4–16	8.811 ± 2.779
Dealing with stress and distress	5–20	11.407 ± 3.581
Adjustment of medication	4–16	7.589 ± 2.730
Involvement of family	2–8	3.825 ± 1.409
Lack of resources	2–8	3.759 ± 1.392
Using delirium screening tools	4–16	8.675 ± 2.818

4.3. Common Method Bias Assessment

All data in this study were self‐reported and collected at a single point in time, which may lead to common method bias, a type of systematic error that can confound the results. Although the methods of questionnaire anonymisation, questionnaire design to conceal the meaning of the questions and design of partially reverse‐scored questions were adopted in this study, further to reduce the adverse effects of common method bias, data were assessed for common method bias using Harman's single‐factor test. The results showed that the first unrotated factor explained 26.790% of variance, less than the 40% threshold (Hair Jr. et al. ²⁰¹⁰), so this study had no serious common method bias.

4.4. Correlations Between Variables

As shown in Table 3, significant correlations were observed between ICU nurses' delirium knowledge, self‐efficacy, clinical reasoning and delirium care difficulties (p < 0.01). Delirium knowledge and self‐efficacy were significantly and positively correlated with clinical reasoning (r = 0.306, p < 0.01 and r = 0.442, p < 0.01, respectively) and significantly and negatively correlated with delirium care difficulties (r = −0.346, p < 0.01 and r = −0.361, p < 0.01, respectively). Clinical reasoning was significantly and negatively correlated with delirium care difficulties (r = −0.456, p < 0.01).

TABLE 3.

Correlations among study variables (N = 440).

	1	2	3	4
ICU nurses' delirium knowledge	1
ICU nurses' self‐efficacy for delirium care	0.332**	1
ICU nurses' clinical reasoning competency	0.306**	0.442**	1
ICU nurses' delirium care difficulties	−0.346**	−0.361**	−0.456**	1

^** p < 0.01; all tests were two‐tailed.

4.5. Multiple Regression Analysis of the Factors Influencing Nurses' Delirium Care Difficulties

We performed multiple stepwise regression analyses with ICU nurses' delirium care difficulties as the dependent variable; demographic factors as the control variables; and ICU nurses' delirium knowledge, self‐efficacy and clinical reasoning as the independent variables (Table 4). The variance inflation factors were less than 5, indicating no multicollinearity in the multiple regression model (James et al. 2013; Thompson et al. ²⁰¹⁷). The regression model was significant (F = 8.628, p < 0.001). The adjusted R ² was 0.335, denoting that all variables collectively explained 33.5% of the variance in nurses' delirium care difficulties.

TABLE 4.

Regression analysis of the influence factor of ICU nurses' delirium care difficulties (N = 440).

		B	SE	β	t	p	95% CI [lower, upper]	Tol	VIF
(Constant)		149.676	12.411		12.060	< 0.001	[125.279, 174.072]
Age (years)		−2.523	2.810	−0.076	−0.898	0.370	[−8.047, 3.001]	0.214	4.671
Education level		4.827	4.250	0.054	1.136	0.257	[−3.528, 13.182]	0.665	1.503
Professional title		−0.707	1.966	−0.024	−0.360	0.719	[−4.572, 3.157]	0.337	2.968
Nursing experience in the ICU (years)		3.445	2.059	0.144	1.673	0.095	[−0.602, 7.493]	0.205	4.870
Average monthly income (RMB)		−0.472	1.490	−0.015	−0.317	0.752	[−3.400, 2.457]	0.637	1.570
Gender (ref: Male)	Female	−3.605	2.299	−0.066	−1.568	0.118	[−8.124, 0.915]	0.863	1.159
Duty (ref: None)	Nursing team leader	−1.201	2.986	−0.021	−0.402	0.688	[−7.071, 4.669]	0.531	1.882
	Nursing instructor	−2.088	3.049	−0.031	−0.685	0.494	[−8.081, 3.906]	0.759	1.318
	Head nurse	−1.093	6.216	−0.009	−0.176	0.861	[−13.312, 11.126]	0.630	1.587
Employment mode (ref: Formal establishment)	Labour dispatch	−1.503	2.181	−0.035	−0.689	0.491	[−5.790, 2.784]	0.588	1.700
Delirium training (ref: Yes)	No	0.098	4.036	0.001	0.024	0.981	[−7.835, 8.031]	0.696	1.436
Delirium care experience (ref: Yes)	No	−7.429	7.211	−0.049	−1.030	0.304	[−21.604, 6.746]	0.670	1.493
Delirium assessment approach (ref: Validated delirium screening tools)	Directly based on nurses' clinical experience	7.331	3.271	0.101	2.242	0.026	[0.902, 13.761]	0.740	1.351
	ICU physician assessment	8.319	4.353	0.079	1.911	0.057	[−0.237, 16.875]	0.892	1.121
	Psychiatric consultation assessment	−80.924	27.125	−0.254	−2.983	0.003	[−134.246, −27.601]	0.210	4.770
Delirium management approach (ref: both strategies above)	Only non‐pharmacological management strategies such as bundle care	−3.815	3.358	−0.048	−1.136	0.257	[−10.417, 2.786]	0.845	1.183
	Only pharmacological management strategies such as analgesia and sedation	−0.643	4.829	−0.005	−0.133	0.894	[−10.135, 8.850]	0.909	1.100
	None of the above management strategies	32.839	9.086	0.145	3.614	< 0.001	[14.977, 50.701]	0.939	1.065
	Don't know	15.255	9.843	0.082	1.550	0.122	[−4.094, 34.603]	0.536	1.867
The main delirium learning approach (ref: Specialist Training)	Continuing education	−2.205	3.136	−0.029	−0.703	0.482	[−8.369, 3.959]	0.922	1.085
	Knowledge manual/Literature/Lectures and exchange sessions	−4.995	3.841	−0.053	−1.300	0.194	[−12.545, 2.555]	0.918	1.090
	Accumulation of work experience/Colleague exchange	0.558	3.159	0.008	0.177	0.860	[−5.651, 6.767]	0.762	1.312
ICU nurses' delirium knowledge		−1.012	0.253	−0.183	−4.006	< 0.001	[−1.508, −0.515]	0.726	1.378
ICU nurses' self‐efficacy for delirium care		−0.256	0.086	−0.139	−2.985	0.003	[−0.425, −0.087]	0.693	1.442
ICU nurses' clinical reasoning competency		−0.537	0.066	−0.381	−8.143	< 0.001	[−0.666, −0.407]	0.691	1.448
R 2					0.379
Adjusted R 2					0.335
F					8.628***

^*** p < 0.001; all tests were two‐tailed.

4.6. Structural Model

We constructed an SEM with ICU nurses' delirium care difficulties as the dependent variable based on the IMB model (Table 5 and Figure 2). The model demonstrated an acceptable fit: χ ²/df = 3.178 < 5, RMSEA = 0.070, CFI = 0.951, TLI = 0.943, IFI = 0.951 and NFI = 0.931. The standardised path coefficients from nurses' delirium knowledge (β = −0.220, t = −3.849, p < 0.001), self‐efficacy (β = −0.149, t = −2.843, p = 0.004) and clinical reasoning (β = −0.328, t = −6.360, p < 0.001) to delirium care difficulties were negatively significant. The standardised path coefficients from delirium knowledge (β = 0.214, t = 3.801, p < 0.001) and care self‐efficacy (β = 0.384, t = 7.604, p < 0.001) to clinical reasoning were positively significant; the correlation between delirium knowledge and self‐efficacy was significant (β = 0.412, t = 6.337, p < 0.001).

TABLE 5.

Pathway analysis between ICU nurses' delirium knowledge, self‐efficacy for delirium care, clinical reasoning competency and delirium care difficulties (N = 440).

Path	Estimate (95% CI [lower, upper])	Std. estimate (95% CI [lower, upper])	SE	C.R.	p
ICU nurses' delirium knowledge → Delirium care difficulties	−0.549 [−1.015, −0.153]	−0.220 [−0.391, −0.057]	0.143	−3.849	***
ICU nurses' delirium knowledge → Clinical reasoning competency	1.254 [0.285, 2.316]	0.214 [0.054, 0.354]	0.330	3.801	***
ICU nurses' delirium knowledge ↔ Self‐efficacy for delirium care	3.099 [2.161, 4.187]	0.412 [0.296, 0.517]	0.489	6.337	***
ICU nurses' self‐efficacy for delirium care → Delirium care difficulties	−0.077 [−0.172, −0.002]	−0.149 [−0.320, −0.001]	0.027	−2.843	0.004
ICU nurses' self‐efficacy for delirium care → Clinical reasoning competency	0.465 [0.281, 0.649]	0.384 [0.249, 0.513]	0.061	7.604	***
Clinical reasoning competency → Delirium care difficulties	−0.139 [−0.209, −0.062]	−0.328 [−0.477, −0.154]	0.022	−6.360	***

^*** p < 0.001.

FIGURE 2.

Final model and standardised model paths. Solid lines indicate significant paths; path coefficients are standardised coefficients. **p < 0.01; ***p < 0.001.

Nurses' clinical reasoning was a partial mediator (−0.070) between delirium knowledge and delirium care difficulties. Clinical reasoning partially mediated (−0.126) the relationship between self‐efficacy and delirium care difficulties (Table 6).

TABLE 6.

Results of mediation effect analysis (N = 440).

Variable	Total effect	Direct effect	Indirect effect	SE	Bias‐corrected		p	Per
					95% CI
ICU nurses' delirium knowledge → Clinical reasoning competency → Delirium care difficulties	−0.290	−0.220	−0.070	0.035	−0.148	−0.015	0.009	24.14%
ICU nurses' self‐efficacy for delirium care → Clinical reasoning competency → Delirium care difficulties	−0.275	−0.149	−0.126	0.035	−0.203	−0.068	0.002	45.82%

5. Discussion

The delirium knowledge, self‐efficacy and clinical reasoning of ICU nurses had a positive effect on alleviating delirium care difficulties and clinical reasoning mediated the relationship between delirium knowledge or self‐efficacy and delirium care difficulties. From the perspectives of information, personal motivation and behavioural skills, these results implied that such factors may influence behavioural changes related to ICU nurses' delirium care difficulties, thereby enhancing our understanding of the mechanisms linking these variables and pathways.

5.1. Levels of Delirium Knowledge, Self‐Efficacy, Clinical Reasoning and Delirium Care Difficulties

In this study, ICU nurses' scores indicated moderate levels of difficulty in providing delirium care, which was not ideal but better than the results of Jiang et al. (Jiang et al. 2024). The heterogeneity of samples may explain this discrepancy. Nurses in this study were from first‐tier cities with adequate resources for learning and well‐stocked hospitals; 99.3% had a bachelor's degree or higher, as indicated by lower scores on the ‘lack of resources’ sub‐dimension. We identified ‘dealing with stress’ as the most challenging sub‐dimension of delirium care, which is consistent with previous studies (Jiang et al. 2024) that found that nurses often experienced continuous physical and emotional exhaustion when facing patients with delirium, which may lead to stronger perceived delirium care difficulties. The complexity of delirium assessment and care often leads to heavy workloads and high stress among ICU nurses, thus generating negative emotions such as nervousness, frustration and helplessness (LeBlanc et al. ²⁰¹⁸; Wang et al. ²⁰²³), as demonstrated by this study's higher scores in the ‘delirium assessment’ dimension.

Meanwhile, ICU nurses displayed vulnerability when irritable delirium patients verbally blamed or violently attacked them (Thomas et al. 2021; Yue et al. ²⁰¹⁵; Luo et al. ²⁰²⁴); they had difficulty dealing with this emotional damage and obtaining external support (Zamoscik et al. 2017). Because delirium patients behave unpredictably, nurses are often concerned about their and other patients' safety (Zamoscik et al. 2017; LeBlanc et al. ²⁰¹⁸; Yue et al. ²⁰¹⁵); this is reflected in the high scores for the ‘ensuring safety’ dimension. Nursing managers can strengthen nurses' emotional regulation skills and resilience through group interventions and cognitive–behavioural stress management strategies. Nursing managers can also develop better safety strategies to prevent nurses from working in an emotionally charged state and compromising patient care.

The nurses' delirium knowledge, self‐efficacy and clinical reasoning levels were moderate. A study investigating Chinese nurses' levels of delirium knowledge found that ICU nurses scored higher than nurses in general wards (Zhou et al. 2022), implying that ICU nurses may be more driven to learn how to manage delirium because ICU patients are more susceptible to delirium. This may explain the upper‐to‐middle levels of delirium knowledge observed in this study. The ability of ICU nurses to translate knowledge into clinical skills may have been limited, which is consistent with other findings (Chen et al. 2016; Chappell and Richards ²⁰¹⁵; Wangensteen et al. ²⁰¹²). This may explain why ICU nurses' subjective judgements of their ability to competently provide delirium care and their reasoning capacity to identify patients' problems and provide care were not advanced. Moreover, 75.9% of the ICU nurses in this study received specialised delirium training, which occurs in groups and is interspersed with busy clinical work. This approach is frequently adopted in China and is less practical and interactive (Guan et al. 2019). However, nurses prefer contextual educational approaches that stimulate clinical critical thinking and reasoning skills (Xm et al. 2022; Sinvani et al. ²⁰²¹; Makhija et al. ²⁰²³; Speed ²⁰¹⁵). The transition from a quantitative to a qualitative understanding of changes in competence requires continuous learning and personal perceptions. Thus, focusing on more flexible training (e.g., role‐playing and digital teaching for junior nurses) may help nurses maintain a continuum of knowledge, self‐efficacy and clinical skills in their daily operations.

5.2. Factors Influencing Delirium Care Difficulties Analysed by Regression

The delirium assessment approach using validated delirium assessment tools or psychiatric consultation, and the delirium management approach using both pharmacological and non‐pharmacological strategies, were beneficial in reducing delirium care difficulties according to the regression model, which will be a vital reference for healthcare managers to optimise and centrally deploy specific strategies to assess and manage delirium in the critical care context.

Specifically, psychiatric consultation assessment and delirium care difficulties were negatively correlated, and this assessment approach appeared to have a more positive effect on reducing ICU nurses' delirium care difficulties than using delirium screening tools. However, delirium assessment based directly on nurses' clinical experience is detrimental to mitigating delirium care difficulties. Identifying barriers is important for effective delirium management (Almoliky et al. 2025). Studies have shown that ICU nurses' familiarity with delirium is generally not high (Zhou et al. 2022; Almoliky et al. ²⁰²⁵; Asmar et al. ²⁰²¹). Therefore, an approach to delirium assessment that relies on experience‐driven rather than validated tools may increase the probability of misdiagnosis and be detrimental to subsequent delirium management, especially for novice nurses with limited clinical experience who have not been able to gain full exposure to the complexities of delirium (Hebeshy et al. 2024). This may explain the effect of assessments based directly on nurses' clinical experiences of delirium care difficulties. Although delirium screening tools are widely recommended (Lewis et al. 2025; Miranda et al. ²⁰²³; Palakshappa and Hough ²⁰²¹), nurses often reported that the screening tools were complex and lacked confidence in their use (Zhou et al. 2022; Yue et al. ²⁰¹⁵; Dechant et al. ²⁰²³; Özsaban and Acaroglu ²⁰¹⁶). Therefore, they preferred seeking psychiatric consultations to increase the accuracy of delirium assessment results, aligning with the findings of Özsaban et al. (Özsaban and Acaroglu 2016). Support from professional teams may help to reduce the burden of delirium assessment on ICU nurses, reducing the possibility of missed opportunities to provide optimal delirium care to patients, which may affect their delirium care difficulties.

Additionally, the explanatory power of the variables (R ² = 33.5%) was relatively low, suggesting that future studies may need to include other potential variables for further analysis, such as workload, policy and protocol development and teamwork. Workload is negatively associated with nursing quality, and the heavy workload of ICU nurses may affect delirium care delivery and patient outcomes (Lee and Roh 2023; LeBlanc et al. ²⁰¹⁸; Almoliky et al. ²⁰²⁵; Boehm et al. ²⁰¹⁷). The lack of specific delirium management protocols in the unit may leave nurses confused or uninformed (Chow and Chan 2024; Palacios‐Ceña et al. ²⁰¹⁶). Effective collaboration with other healthcare professionals is important for delirium management (LeBlanc et al. ²⁰¹⁸; Balasanova and Park ²⁰²¹), as ICU nurses reported their expectations of receiving feedback and suggestions from physicians or specialists (Jiang et al. 2024).

In the future, delirium assessment training for ICU nurses should refine the explanation of screening tool items to increase nurses' understanding and increase hands‐on training to seek assessment strategies better suited to clinical work rhythms; for example, by embedding the screening tool into the electronic nursing record system to facilitate nurses' effective application and practice improvement. In addition to reinforcing ICU nurses' confidence and mastery in delirium screening tool use at the individual level, the organisational level should also be considered. Standardised delirium management protocols should be developed and included in nurses' professional assessment index (Xie et al. 2022). Delirium specialist teams should be contacted to enhance clinical support for ICU nurses to promote judgement development, mitigate delirium care difficulties and improve delirium management quality.

5.3. Correlations Between Delirium Knowledge, Clinical Reasoning and Delirium Care Difficulties

The nurses' delirium knowledge was negatively associated with perceived delirium care difficulties. Adequate delirium knowledge positively affected mitigating delirium care difficulties; however, clinical reasoning played a partial mediating role. Thus, a higher level of delirium knowledge among nurses influenced the demonstration of greater clinical reasoning competency. Through clinical reasoning, nurses' adequate delirium knowledge indirectly reverses and reduces perceived delirium care difficulties. These results validate those of other qualitative studies on nurses' self‐reported inability to manage delirium efficiently due to their lack of knowledge about delirium assessment and management (Thomas et al. 2021; Zamoscik et al. ²⁰¹⁷). They also explain the possible action pathways regarding the correlation between knowledge and delirium care difficulties in a cross‐sectional study by Jiang et al. (2024). Studies using ICU nurses' delirium knowledge as an entry point for future interventions have primarily focused on describing the current situation (Lee and Roh 2023; Zhou et al. ²⁰²²; Mathew et al. ²⁰²⁴); despite calls for extensive education and training (Fowler 2019; Collinsworth et al. ²⁰¹⁶; Negro et al. ²⁰²²), the effective pathway to enhance delirium knowledge to influence nurses' delirium care remains unclear. Hence, discovering this action pathway may help identify the means of increasing delirium knowledge to tangibly improve difficulties and patient outcomes, according to nurses' perspectives on behavioural change.

Additionally, the ICU nurses' clinical reasoning competency partly mediated the relationship between delirium knowledge and delirium care difficulties. Applying knowledge to clinical situations facilitates the development of a process for solving patient problem scenarios known as clinical reasoning, which informs decisions about identifying and managing the disease and is essential for nursing competence (Banning 2008; Simmons et al. ²⁰⁰³). Accordingly, ICU nurses with less knowledge about delirium have difficulty making professional clinical judgements about delirium. They have poor competence in assessing the quality of evidence to address and manage delirium, mainly due to a lack of clinical reasoning (Hamilton et al. 2025). Clinical reasoning involves connecting cues, indicators and recognition (El Hussein and Hirst ²⁰¹⁶) using thinking skills to adopt flexible coping strategies for complex situations such as delirium trends, trajectory capture and coping strategies for symptomatic fluctuations (Benner et al. 2009). This will help enhance nurses' high sensitivity in their ability to recognise delirium, especially low‐activity delirium (Hoch et al. 2022), and high‐quality delirium care, which may negatively impact delirium care difficulties.

Thus, according to the relationship between information, behavioural skills and behaviour in the IMB theoretical model, specialised and specific knowledge is more important than general knowledge in education (Chen et al. 2016). ICU nurses' specialised knowledge of delirium management is linked to factors such as unit‐based policies for delirium, financial support and self‐planning (Lange et al. 2023). Hence, our results underscore the need for departmental attention and optimisation of training policies to alleviate the challenges in delirium care and reduce adverse patient outcomes. Healthcare providers should focus on highly reliable delirium education for ICU nurses through diverse pathways to promote a practical understanding of the latest authoritative knowledge on delirium, such as guidelines (Trogrlić et al. 2017).

Furthermore, the mediating role of clinical reasoning competency highlights the importance of behavioural techniques for training critical and reflective thinking during the educational process of behavioural change. Based on self‐regulated learning theory, developing ICU nurses' clinical reasoning competency based on preparatory nursing education is essential, which may help ensure and control the diverse and evolving realities of delirium (El Hussein and Hirst ²⁰¹⁶; Bae et al. ²⁰²³). Delirium education for ICU nurses should emphasise and utilise behavioural skills, integrating thinking and practical training in clinical reasoning into delirium knowledge training. Specifically, delirium training should not be limited to traditional approaches, such as speciality training and fixed theoretical knowledge, but should develop implementable innovative teaching strategies to exercise clinical reasoning competency. For example, clinical case‐based practices and brainstorming should be prioritised. ICU nurses should be involved in the purposive contexts of interacting with cases and exercising their ability to collect and integrate information into contexts to judge different patients' symptoms and make decisions. Additionally, structured, planned and mutually responsive teaching of theory and clinical reasoning practice is needed to bridge the gap between theory and practice and enhance nurses' in‐depth knowledge and flexibility in coping with delirium. Adapting an effective training programme that focuses on developing knowledge‐facilitated clinical reasoning competency is essential for enhancing nurses' competencies in ICU delirium practice (Flinkman et al. 2017).

5.4. Correlations Linking Delirium Knowledge With Self‐Efficacy

ICU nurses' delirium knowledge was positively associated with self‐efficacy. The results indicated that higher levels of delirium knowledge positively affected the strengthening of self‐efficacy, which is consistent with the findings of Nie et al. (2024) and supported by evidence from studies in other fields on the role of nurses' knowledge of self‐efficacy (Kim et al. 2020; Papaioannou et al. ²⁰²³). Meanwhile, the results also indicated that high self‐efficacy positively affected the mastery of acquired delirium knowledge, as suggested by studies finding that nurses' educational intervention outcomes were related to their self‐efficacy in patient management (Ylimäki et al. 2015). Hence, ICU managers should prioritise improving nurses' delirium knowledge and self‐efficacy simultaneously because adequate knowledge and perception of delirium can contribute to nurses' ability to provide care, which is associated with improved outcomes in clinical practice.

5.5. Correlations Between Self‐Efficacy, Clinical Reasoning and Delirium Care Difficulties

Self‐efficacy was negatively correlated with perceived delirium care difficulties, and clinical reasoning partially mediated this relationship. The results indicated that higher self‐efficacy positively influenced ICU nurses facing fewer delirium care difficulties. As proposed by Bandura's social cognitive theory, individuals' beliefs about their abilities underlie their thoughts and actions (Bandura 1997). Nurses with high self‐efficacy possess a deep level of confidence in their ability to achieve tasks and take positive actions in responding to complex or challenging situations (Nie et al. 2024; Caruso et al. ²⁰¹⁶). Those with greater self‐efficacy are willing to put in effort and demonstrate high levels of motivation and perseverance when faced with setbacks (Lunenburg 2011). Thus, when treating a patient with delirium, a complex issue that needs to be addressed urgently in the ICU (Nie et al. 2024; Chang et al. ²⁰²³), they always show positive practical dispositions towards engaging with and solving this complex clinical challenge (Ma et al. 2023; Kurniawan et al. ²⁰¹⁹). This influences their behaviour such that their perceived delirium care difficulties may gradually decline.

Additionally, ICU nurses' clinical reasoning competency partly mediated the relationship between self‐efficacy and delirium care difficulties. Self‐efficacy in delirium care positively influenced clinical reasoning competencies. Nurses' level of self‐efficacy is considered critical in shaping their thinking patterns and performance in manipulating specific skills to complete tasks, which are essential in the reasoning process (Caruso et al. 2016; Bandura ¹⁹⁷⁷; Durmus and Akpinar ²⁰²⁵). According to the social cognitive theory developed by psychologist Albert Bandura (Bandura 1986; Zimmerman ¹⁹⁸⁹), self‐regulation through self‐efficacy in the learning process impacts clinical reasoning competency (Moran et al. 2021; Hong et al. ²⁰²¹). Self‐regulation is a cyclical adaptive process of thinking about clinical skills by synthesising different contexts and successfully finding solutions (Moran et al. 2021; Hong et al. ²⁰²¹), involving self‐observation, self‐judgement and self‐response (Zimmerman 1989), which are relevant to the connotation of clinical reasoning competency (Alfaro‐LeFevre ²⁰¹⁶). Clinical reasoning is a dynamic cyclical nursing process for thinking about and dealing with complex clinical problems (Kuiper and Pesut 2004; Tanner ²⁰⁰⁶), involving many skills such as information synthesis, judgement, decision‐making and problem‐solving (Ebrall 2008). Therefore, activating nurses' metacognition through self‐regulation to facilitate their immediate judgement of problems promotes the development of clinical reasoning competency (Bae et al. 2023). Several studies on education and training have confirmed that an individual's self‐efficacy positively impacts the development of clinical reasoning competency (Nafea 2022; Hong et al. ²⁰²¹; Croy et al. ²⁰²⁰). These findings further confirm the positive impact of self‐efficacy on clinical reasoning found in this study and help elucidate its logical mechanisms from a conceptual and theoretical perspective. Thus, the high self‐efficacy of ICU nurses in this study may have improved their ability to take the required actions in the clinical reasoning process, as nurses with high self‐efficacy usually believe that they can manipulate and practise the skills required for effective symptom management (Kurniawan et al. 2019) and are adept at finding coping strategies to address clinical challenges through self‐regulation, including assessing and caring for delirium cases. Therefore, ICU nurses' high self‐efficacy in delirium care may positively affect their development of clinical reasoning competency. Additionally, nurses with high self‐efficacy were confident and adept at using critical thinking in their work to identify patient needs, facilitating stronger nurse–patient relationships (Bu et al. 2022), which is essential for identifying symptoms and caring for patients with severe delirium, whose communication is impaired and whose behaviours are unpredictable. According to this pathway, ICU nurses with low self‐efficacy lacked subjective thinking and the ability to analyse, accurately perceive and flexibly apply skills to care for patients' delirium‐related problems in several situations (Eller et al. 2018). This may negatively impact the development of clinical reasoning competency, with a focus on problem‐solving, which affects the quality of delirium care and increases delirium care difficulties.

Thus, according to the relationship between motivation, behavioural skills and behaviour in the IMB theoretical model, critical care managers should prioritise nurses' self‐efficacy in delirium care. Nursing managers should use reasonable tools to assess ICU nurses' levels of self‐efficacy (Chang et al. 2023), thereby developing tailored interventions to enhance their professional confidence and self‐efficacy by providing them with additional education, external support and incentives that are conducive to enhancing their well‐being. Moreover, encouraging participation in self‐reflection and assessment will help nurses become intentional and directed towards clarifying pathways for their professional growth. This will help nurses develop greater autonomy and self‐efficacy, thereby improving patient outcomes.

Furthermore, the mediating role of clinical reasoning competency highlights the importance of behavioural techniques for clinically reflective and critical self‐regulated learning in the growth of self‐efficacy in behavioural change. Developing clinical reasoning habits should be emphasised when developing ICU nurses' self‐efficacy. Nurses' uncertainty about their assessment outcomes and coping strategies has been identified as a significant barrier to delirium care (Chow and Chan 2024). Implementing clinical reasoning competency training based on simulated real‐life situations may be beneficial, particularly during the initial stages of training because, for nurses who need to improve their self‐efficacy, this relatively controlled, low‐risk environment may be more conducive to focused judgement and decision‐making in different learning cases, thus helping them develop their reasoning competency and nursing confidence. Second, long‐term strategies should also be emphasised. Nurses should be organised to participate in dynamic reasoning on a cyclical and continuous basis. Long‐term, intuitive, heuristic simulated case training may help develop nurses' cognitive and metacognitive clinical reasoning habits. When transitioning to real clinical cases, nurses may be confident in adapting and coping. Additionally, critical care managers can use case vignettes to test the effectiveness of nurses' staged clinical reasoning training, promote efficient learning and improvement of their reflective clinical reasoning skills and motivate them to take on challenges. These strategies may help activate nurses' confidence and reasoning skills, facilitating their prioritisation and effective judgements and decisions in delirium care and the ability to reason about delirium‐patient needs to respond quickly and logically to their changing circumstances (Correya et al. 2025). Clinical reasoning plays a vital role in delirium care quality. Therefore, focusing on the effective growth of ICU nurses' self‐efficacy and clinical reasoning competency is likely to improve delirium care and patient prognosis.

5.6. Limitations

This study has several limitations. First, because the data were cross‐sectional, this limited the longitudinal exploration of the results and inferences about the causal relationships between variables. Although we adopted the theoretical framework that would enhance the directionality of the relationships among the variables in the model, future research should conduct further longitudinal or intervention studies to better understand the dynamic mechanisms of action between these variables and provide more reliable evidence for confirming causality. Second, despite an adequate sample size and multi‐centre design, our findings' generalisability may be limited because we conducted this study in several hospitals in one region and used convenience sampling to recruit participants. Convenience sampling is a non‐probability sampling method, which, although widely used for its efficiency, may introduce selection bias (Bornstein et al. 2013). This meant that our sample may not represent ICU nurses in other regions or countries; meanwhile, because nurses in the sample analysed for this study were from tertiary hospitals, this may hinder the possibility of generalising the results of this study to different hospital grades. Therefore, future studies should consider adopting a more representative sampling method to balance the selection of the study population by conducting a more comprehensive survey of ICU nurses from a wider range of regions, different hospital levels and cultural backgrounds, to expand the impact of the study and improve the generalisability of results. Third, the data were self‐reported by nurses, which may introduce potential social desirability and self‐report bias (Song et al. 2015; Podsakoff et al. ²⁰¹²). For example, nurses may overestimate or underestimate their actual ability and tend to present themselves with more favourable responses, conforming to the expectations of hospital leaders and society. Future research should combine self‐report with objective assessments (e.g., clinical observations, case records) and integrate data from multiple sources (e.g., unit leaders' assessments) through triangulation to improve the reliability of the results. Finally, we explored the factors affecting delirium care difficulties and the relationships between the variables among ICU nurses, but future studies should investigate other relevant populations, such as patients, physicians and hospitals.

5.7. Implications for Practice and Future Research

This study found that ICU nurses' delirium‐care knowledge, self‐efficacy and clinical reasoning competency positively influenced their delirium‐care behaviour performance based on the IMB model. Based on the roles of information (delirium knowledge), motivation (delirium care self‐efficacy) and behavioural skills (clinical reasoning competency) that may influence nurses' behavioural changes in delirium care, we propose several potentially practical clinical strategies. First, the education system should be optimised to narrow the educational gaps. Hospital managers should tailor training modalities and content to nurses' specific academic needs, integrate various teaching techniques, such as demonstrations and interactive discussions and innovate diverse pedagogical strategies and tools (e.g., experiential learning, especially for younger nurses) to promote delirium knowledge and self‐efficacy in critical care organisations. Organised continuing education that integrates nursing theoretical knowledge and practice should be emphasised, including relevant physiological and pharmacological knowledge. Online and offline multiform education should also be offered for nurses to consolidate repeatedly. Second, self‐efficacy is cultivated through cumulative experience and perceptions. Thus, ICU nursing managers should implement interventions that focus on self‐efficacy levels. For example, nurses should be offered a supportive organisational environment and introduced to new resources to provide them with confidence, support and feedback to enhance self‐efficacy. Additionally, nurses should be encouraged to engage in self‐reflection and assessment, and clinical support groups organised for individual feedback and reflective listening are required. Such support may enhance nurses' sense of purpose and self‐efficacy for professional growth, especially for nurses with low delirium care self‐efficacy.

Third, because of the mediating role of behavioural skills in the IMB model, competency training focusing on nurses' clinical reasoning is crucial, along with information and motivation education. For example, intuitive and heuristic simulated case studies should be conducted on a long‐term basis, and the results of stage‐by‐stage reasoning skills training can be tested directly or assisted through case illustrations and advanced computer vision techniques and then gradually transitioned to bedside training. ICU nurses should also be encouraged to adopt a reflective, self‐regulated learning approach to enhance cognitive (critical thinking) and metacognitive (reflective thinking) development through reflective self‐regulated learning and develop behavioural skills that enable them to use multidimensional strategies to solve clinical problems. In addition, forming groups within the unit for problem‐based learning and coaching may be useful in motivating nurses to share problems with others following a staged assessment and reflection on delirium reasoning practice. Establishing supportive relationships may advance nurses' confidence and abilities to develop their clinical reasoning competencies. Increased awareness and competence in clinical reasoning, such as delirium recognition based on the nuances and decisions surrounding task prioritisation, may help ICU nurses address delirium in critical and life‐threatening situations. Finally, critical care managers should develop policies that incorporate the assessment and promotion of nurses' delirium care skills into critical care plans to facilitate strategic uptake.

6. Conclusions

In complex and demanding critical care environments, ICU nurses frequently experience global challenges in assessing and managing delirium. As immediate healthcare needs and long‐term prognostic requirements for patients with delirium continue to increase, promoting behavioural changes among ICU nurses has become urgent. A theoretically informed understanding of the variables that influence ICU nurses' delirium care difficulties is vital to implementing effective strategies. This study confirmed that ICU nurses' delirium knowledge and self‐efficacy influenced delirium care difficulties directly and indirectly through the partial mediation of clinical reasoning, and the results support the IMB model. According to the IMB model context of enhancing the role of information, motivation and behavioural skills in behavioural change, in future practice, ICU nurses should be provided with effective delirium knowledge as information stock (e.g., tailoring needs‐based training modalities and content; integrating teaching techniques; innovating teaching strategies and tools; conducting structured continuing education integrating nursing theory and practice); and their self‐efficacy should be stimulated as action motivation (e.g., providing supportive organisational environment and introduced resources; organising clinical mutual support groups to support nurses in strengthening their sense of professional purpose through personal feedback and listening); simultaneously, when clinical reasoning competency, a behavioural skill required to facilitate behavioural change, is also improved (e.g., conducting long‐term training in intuitive and heuristic simulated cases; organising staged exercises and tests with human or technological assistance; encouraging nurses to practise and master reflective self‐regulated learning methods; setting up in‐unit support groups to provide problem‐based learning and coaching), they can jointly initiate behaviours that may reduce ICU nurses' delirium care difficulties and facilitate positive patient outcomes.

Future research should continue to focus on dilemmas around delirium care, analysing nurse‐level factors and seeking to better understand the impact of team and organisational factors, patients and family members on the difficulties involved in providing delirium care and ICU patients' long‐term recovery and well‐being. This will provide an empirical basis for interventions.

Author Contributions

Jing Dong: writing – original draft, writing – review and editing, methodology, investigation, formal analysis, software, data curation, conceptualisation. Weijing Sui: methodology, formal analysis, data curation, conceptualisation, project administration. Xiaoyan Gong: resources, project administration, investigation, conceptualisation. Li Wang: project administration, supervision. Qiqi Ni: supervision, conceptualisation. Ran Yan: validation, supervision. Jia Yi: visualisation, validation. Ying Ding: was responsible for visualisation and validation. Yiyu Zhuang: resources, writing – review and editing, validation, supervision.

Ethics Statement

This study was approved by the Clinical Research Ethics Committee of Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University on 6 November 2024 (No. 20240602) and adhered to the principles of the Declaration of Helsinki.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

File S1.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.