The Role of Emotional Intelligence in Nursing Leadership and Its Impact on Nursing Practice Outcomes: A Cross-Sectional Correlational Study in North-West Frontier Province (NWFP), Pakistan

Muhammad Shabir ¹, Muhammad Ishaq ¹, Muhammad Hilal ¹, Huzefa Amir ¹, Bilal Khan ¹, Waqar Alam ², Muhammad Awais ², Abdur Rahman ^2,*, Adnan Adnan ³

1. Department of Nursing, Elizabeth Rani College of Nursing and Allied Health Sciences, Mardan, Khyber Pakhtunkhwa, Pakistan.
2. Department of Management Sciences, Abasyn University Peshawar, Pakistan.
3. Department of Administeration, Mardan Medical Complex – Medical Teaching Institution (MMC-MTI), Khyber Pukhtunkhwa, Pakistan.

* Corresponding author: Abdur Rahman, Department of Management Sciences, Abasyn University Peshawar, Pakistan – E-mail: abdur28648@abasyn.edu.pk

Cite this article

ABSTRACT

Background: The core of the competency of nursing leadership is emerging as emotional intelligence (EI), which has implications for staff wellbeing and patient safety outcomes. But there is no quantitative study on the role of Nursing leader EI in Nursing Practice outcomes in District Mardan Khyber Pakhtunkhwa (KPK) Pakistan so far.

Objective: The objective of this study was to see the relationship between the EI of the nursing leaders and selected outcomes of nursing practice namely Job Satisfaction, Burnout, Medication Errors and Clinical Adherence in the Public and Private Hospitals of District Mardan.

Methods: Quantitative, Cross – Sectional, correlational design was used. A total of 235 participants (39 nursing leaders, 196 staff nurses) were recruited from six hospitals and completed the Wong and Law Emotional Intelligence Scale (WLEIS), the McCloskey/Mueller Satisfaction Scale (MMSS), the Maslach Burnout Inventory – Human Services Survey (MBI-HSS), and a researcher-developed clinical outcomes section. Data were analysed using descriptive statistics, Pearson correlation, independent samples t-test, one-way ANOVA with post hocTukey’s HSD and stepwise multiple linear regression with IBM SPSS Version 26.

Results: Overall results indicated that the nursing leaders had an average EI level of 4.35 (SD = 0.85). Significant positive correlations were found between leader EI and staff job satisfaction (r = 0.42, p < 0.01) and clinical adherence (r = 0.35, p < 0.01), and significant negative correlations with burnout (r = −0.38, p < 0.01) and medication errors (r = −0.31, p < 0.01). There were fewer medication errors reported in units with high-EI leaders (50.9%) and higher hand hygiene compliance (p < 0.001). The strongest predictors of clinical adherence were Others’ Emotion Appraisal (conceptually equivalent to empathy; β = 0.31) and Regulation of Emotion (conceptually equivalent to self-regulation; β = 0.27) within a three-predictor model that also included Use of Emotion (β = 0.20). The final model explained 38% of the variance in clinical adherence (R² = 0.38).

Conclusion: Nursing leader EI is significantly associated with the better staff outcomes and patient safety in district Mardan. In KPK, targeted EI development programmes are recommended, with the primary focus being on empathy and emotional regulation, for nursing leadership.

Keywords: emotional intelligence, nursing leadership, job satisfaction, burnout, patient safety, Pakistan, Khyber Pakhtunkhwa.

INTRODUCTION

Emotional intelligence (EI) is becoming an essential skill for health care management and is playing a key role in employee satisfaction and patient care quality [1]. EI is defined as the capacity to perceive, understand, manage and regulate one’s own feelings and the emotions of others, which will enable nurse leaders to respond to the emotional aspects of a clinical setting and perform effectively [2]. Various frameworks have been proposed to conceptualize EI, including the five-component model (self-awareness, self-regulation, motivation, empathy, and social skills) [3] and the four-dimensional model operationalized in the Wong and Law Emotional Intelligence Scale (WLEIS): Self-Emotion Appraisal, Others’ Emotion Appraisal (conceptually equivalent to empathy), Use of Emotion (relating to motivation and social skills), and Regulation of Emotion (corresponding to self-regulation) [2].

The present study adopts the WLEIS framework due to its established validity in the Pakistani nursing context [4] and its parsimonious four-dimensional structure, which has been shown to be culturally appropriate for healthcare professionals in South Asian settings. EI is especially critical in nursing leadership because leaders have to deal with a variety of daily pressures such as staffing shortages, heavy workload, stress and anxiety, and the need to make quick decisions [5]. Studies show that nurse managers who exhibit high EI are more likely to exhibit more transformational leadership behaviour, which subsequently leads staff to perform better and be engaged in the organization [6-8]. On the other hand, low EI among leaders has been linked to decreased team morale, conflict and intentions to leave [7].

The existing literature on EI and nursing outcomes can be organized into several thematic areas. First, regarding individual nurse outcomes, studies have consistently demonstrated that nursing leaders’ EI is positively associated with staff job satisfaction, with conflict management identified as a key mediating mechanism [9]. Higher EI among nurse managers has been linked to more effective conflict management strategies, which in turn create healthier work environments [10]. Furthermore, leaders’ EI positively influences staff nurses’ psychological empowerment and work engagement [11], suggesting that emotionally intelligent leadership fosters a more supportive and motivating work climate. Job satisfaction has been found to mediate the relationship between EI and job performance [12], while organizational commitment among nurses also shows positive correlations with EI [6].

Second, in the domain of leadership effectiveness, studies have shown that transformational leadership combined with EI enhances person-organization fit among nurses. During the COVID-19 crisis, EI emerged as an important factor influencing healthcare worker performance [13] reported that healthcare workers with higher EI demonstrated better adaptability, lower stress levels, and more effective coping strategies during the pandemic, suggesting that emotional competencies become particularly critical during periods of heightened stress and organizational disruption.

This finding has important implications for healthcare leadership, as it indicates that EI may serve as a protective factor during public health emergencies.

Gender disparities in EI among health managers have also been documented by Kitsios, F., (2022) [14], who reported significant differences in EI scores between male and female health managers, with female managers generally scoring higher on emotional appraisal and empathy dimensions. This finding underscores the need for gender-sensitive leadership development approaches that recognize and leverage the diverse emotional competencies that different groups bring to leadership roles. Such approaches should avoid reinforcing stereotypes while ensuring that all leaders, regardless of gender, have opportunities to develop the full range of EI competencies.

A concept analysis by Alsufyani, A. M., (2020) [15] established the multidimensional consequences of EI on nursing, identifying key outcomes across individual, team, and organizational levels. At the individual level, EI was associated with reduced burnout, increased job satisfaction, and enhanced coping abilities. At the team level, EI facilitated better communication, conflict resolution, and collaboration. At the organizational level, EI contributed to improved patient outcomes, reduced turnover, and enhanced quality of care. This multidimensional framework provides a useful lens for understanding the broad impact of EI in nursing settings [16]. Further reported a significant association between EI and quality of healthcare delivery, showing that nurses with higher EI were more likely to provide patient-centered care, engage in effective communication with patients and families, and demonstrate greater clinical competence.

Third, regional evidence from Pakistan, particularly Khyber Pakhtunkhwa, has provided important context-specific findings. A significant negative correlation between EI and burnout scores among nurses was found in research conducted in tertiary hospitals in KPK by Ullah et al. (2022) [4]. Similarly, EI acted as a buffer against psychological stress during the COVID-19 pandemic as reported by Alam et al. (2023) [13]. A positive association between emotional quotient and leadership effectiveness among healthcare leaders in Pakistan was identified by Zaman et al. (2021) [17]. Regional benchmarks for EI studies have been established through correlations between EI and job satisfaction among secondary school heads in KPK reported by Suleman et al. (2020) [18]. Additionally, EI is associated with academic success of undergraduate nursing students, indicating its significance throughout the nursing career pathway [19].

The nursing profession in Pakistan has its own set of challenges such as shortage of resources, large patient-to-nurse ratios, lack of professional development opportunities and emotional stress [20]. Research conducted in KPK has shown that EI is negatively correlated with burnout and positively correlated with life satisfaction among nurses [4,20]. The study by Bibi et al. (2025) [19] illustrated that EI is associated with academic success of undergraduate nursing students, indicating its significance throughout the nursing career pathway [21] assessed EI among medical students in Peshawar and correlated EI with job satisfaction among the secondary school head in KPK setting which set regional benchmarks for EI studies [18]. In Pakistan, recently Badshah et al. (2025) [22] identified a correlation between EI and burnout syndrome in clinicians at the workplace, further highlighting the relevance of this topic. Although these contributions have been made, no quantitative study has so far looked at the role of EI of nursing leaders on the outcome of nursing practice in District Mardan, KPK which is a large, semi-urban area of Pakistan with a vast network of healthcare infrastructure comprising of Mardan Medical Complex, District Headquarters Hospital and Tehsil Headquarters Hospitals.

The link between the level of emotional intelligence among nursing leaders and job satisfaction, burnout, medication errors, and compliance to clinical protocols—specifically among the staff nurses under their supervision—in District Mardan hospitals is not supported by empirical evidence. District Mardan is a large semi-urban district with a diverse healthcare infrastructure comprising Mardan Medical Complex (tertiary care), District Headquarters Hospital (secondary care), Tehsil Headquarters Hospitals (primary care), and private hospitals. This tiered healthcare system presents unique challenges including resource allocation disparities, varying patient-to-nurse ratios, and different staff development opportunities across hospital types. Without region-specific empirical evidence, healthcare administrators in District Mardan and policymakers in the KPK Health Department lack the necessary data to design contextually appropriate leadership development programmes that can enhance leadership effectiveness, improve staff wellbeing, and ultimately contribute to better patient safety outcomes.A critical review of the existing literature reveals several important gaps that this study addresses.

First, the majority of studies examining the relationship between nursing leader EI and staff outcomes have focused on single outcome measures, typically job satisfaction [9,12] or burnout [4,22], rather than adopting a comprehensive approach that includes multiple indicators of nursing practice quality. Only a limited number of studies have simultaneously examined both staff wellbeing outcomes and patient safety indicators such as medication errors and clinical adherence [16].

Second, while research has been conducted in various healthcare settings globally [9,16] and in some parts of Khyber Pakhtunkhwa [4,18], no empirical study has specifically investigated this relationship in District Mardan, a large semi-urban district with a diverse healthcare infrastructure comprising tertiary, secondary, primary, and private hospitals.

Third, the semi-urban context of District Mardan presents unique healthcare delivery challenges including resource constraints, varying patient-to-nurse ratios across different hospital tiers, and limited professional development opportunities, which may influence both leadership effectiveness and nursing practice outcomes in ways that differ from urban or fully rural settings. Furthermore, the District Mardan healthcare system operates within the broader context of Khyber Pakhtunkhwa’s nursing workforce challenges, which include shortage of trained personnel, heavy workloads, limited career advancement pathways, and emotional stress [17]. Understanding how nursing leader EI operates within this specific context is essential for developing targeted interventions that can improve leadership effectiveness and, consequently, staff wellbeing and patient safety. Without region-specific empirical evidence, healthcare administrators and policymakers in the KPK Health Department lack the necessary data to design contextually appropriate leadership development programmes, staff engagement policies, and retention strategies. This study therefore addresses these critical gaps by providing the first quantitative evidence on the relationship between nursing leader EI and a comprehensive set of nursing practice outcomes—including job satisfaction, burnout, medication errors, and clinical adherence—in the specific healthcare context of District Mardan.

Aim

This study aims to explore the association of emotional intelligence of nursing leaders with selected indicators of nursing practice (job satisfaction, burnout, medication errors and compliance with clinical standards) in public and private hospitals of District Mardan, Khyber Pakhtunkhwa, Pakistan.

Research Questions

1. What is the mean score of Emotional Intelligence of the nursing leaders of District Mardan hospitals?
2. Do there appear to be meaningful differences among the mean scores for nursing leaders and staff nurses in their job satisfaction?
3. Do the scores of nursing leaders and staff nurses have a statistically significant correlation?
4. Are there significant differences in the number of medication errors and adverse events reported in hospital units led by a higher-EI leader?
5. Which of the four WLEIS dimensions (Self-Emotion Appraisal, Others’ Emotion Appraisal, Use of Emotion, and Regulation of Emotion) is most strongly associated with positive nursing practice outcomes?

Hypotheses

H1: Nursing leaders’ EI score is positively related with staff nurses’ job satisfaction.

H2: There will be a negative correlation between the EI scores of nursing leaders and the level of burnout among the staff nurses.

H3: Medication error rates are significantly lower in the units with nursing leaders in the top tertile of EI scores than in the units with nursing leaders in the bottom tertile of EI scores.

H4: Others’ Emotion Appraisal (conceptually equivalent to empathy) and Regulation of Emotion (conceptually equivalent to self-regulation) will be the most significant WLEIS dimensions related to increased clinical standard adherence.

The significance of this study is briefly mentioned. It will give the first baseline data of the emotional intelligence of nursing leaders in district Mardan. Second, the results could help guide local hospital leadership development initiatives, engagement policies, and retention strategies. Thirdly, findings can help inform the KPK Health Department’s leadership development policies for the specific context. Fourth, the study adds to the scarce research on EI in the healthcare context in Pakistan, and previous researchers [17,18,22] have called for more region-specific research.

MATERIALS AND METHODS

Research Design

A quantitative, cross-sectional, correlational design was used to explore the relationship between the emotional intelligence of nursing leaders and outcomes of nursing practice. This design was selected as it is appropriate for the specific aims of the study for several reasons.

First, it allows for efficient data collection across multiple hospitals within a defined timeframe, which is essential given the geographic distribution of the study sites in District Mardan.

Second, it enables the examination of relationships between variables without manipulation, which is suitable for studying naturally occurring phenomena such as emotional intelligence and its association with nursing practice outcomes.

Third, this design is widely used in nursing leadership research, facilitating comparison of findings with existing literature.

Fourth, it aligns with the post-positivist epistemological paradigm, which emphasizes objective measurement of constructs through validated instruments.

Setting

The study was carried out in six hospitals located in District Mardan of Khyber Pakhtunkhwa (KPK) Pakistan, which included two private hospitals, two Tehsil Headquarters hospitals (public), District Headquarters Hospital Mardan (secondary, public) and Mardan Medical Complex (tertiary, public). With this multi-site approach, the representativeness of the sample was improved and comparison across healthcare tiers was possible.

Population and Sample

The target population was all registered nurses (RN) working within the participating hospitals (nursing leaders such as head nurse, ward supervisor, deputy matron and matron, staff nurse who had been working under their leader for at least six months). There were around 600 nurses in the district Mardan.

The sample size was calculated using the standard formula for estimating a population proportion in a finite population, incorporating the finite population correction (FPC):

where: n = required sample size, Z = Z-statistic for a 95% confidence level (1.96), p = anticipated population proportion (0.50, assuming maximum variability in the population), e = margin of error (0.05), N = total population size (600 nurses). Substituting these values into previous formula:

This yielded a minimum required sample of 235 participants. The calculation was based on estimating a population proportion (the anticipated proportion of nurses with certain characteristics or outcomes), which is the standard approach for cross-sectional studies when the primary parameter of interest is a proportion. To account for an anticipated non-response and incomplete questionnaire rate of 10-15%, the target sample size was increased to 270 participants. This oversampling ensured that the final analyzable sample would meet or exceed the minimum required sample size of 235 even after excluding incomplete responses. Due to the limited number of nursing leaders in the participating hospitals (n = 39), they were recruited through purposive sampling, as all eligible nursing leaders across the six hospitals were invited to participate. For staff nurses, a stratified random sampling technique was employed to ensure proportional representation across clinical units. The strata were defined based on the clinical areas within each hospital: medical wards, surgical wards, intensive care units (ICUs), emergency departments, paediatrics units, and other specialized units (e.g., gynaecology, orthopaedics, and cardiac units). The sample size for each stratum was determined using proportional allocation, whereby the number of staff nurses selected from each clinical unit was proportionate to the total number of eligible staff nurses working in that unit across all six hospitals. For instance, staff nurses working in medical wards constituted approximately 30% of the total eligible staff nurse population across the six hospitals; accordingly, 30% (n = 59) of the staff nurse sample was drawn from medical wards. Similarly, surgical wards contributed 25% (n = 49), ICUs 15% (n = 29), emergency departments 13% (n = 25), paediatrics 10% (n = 20), and other specialized units 7% (n = 14). Within each stratum, staff nurses were selected using simple random sampling. A comprehensive list of all eligible staff nurses in each clinical unit was obtained from the nursing administration of each participating hospital. Each eligible staff nurse was assigned a unique identification number, and a computer-generated random number list was created using Microsoft Excel’s RAND function to select the required number of participants from each stratum. The selected staff nurses were then approached by trained research assistants for participation in the study. This approach ensured that every eligible staff nurse within each clinical unit had an equal probability of being selected, thereby minimizing selection bias and enhancing the representativeness of the sample.

The inclusion criteria were that all participants must be registered by the Pakistan Nursing Council and must give informed written consent. Nursing leaders had to have supervised others for at least 1 year; the staff nurses had to have been supervised by their current leader for at least 6 months. Students and student nurses/assisting nurses and those who were on approved leave during the time of data collection were excluded from the participants.

Instruments

Four instruments were used for data collection: one self-developed demographic sheet, two validated psychometric scales and one licensed burnout inventory as well as one self-developed clinical outcomes section.

Demographic Information Sheet: This information was gathered with a researcher-designed instrument that asked about age, gender, marital status, education level, professional designation, years in nursing practice, years in the current leadership position, hospital type and clinical unit.

Wong and Law Emotional Intelligence Scale (WLEIS): The emotional intelligence was assessed by the Wong and Law’s (2002) WLEIS [2], which is a 16-item self-report instrument. The scale is divided into four dimensions:  Self-Emotion Appraisal (SEA)—the ability to understand and express one’s own emotions (4 items);  Others’ Emotion Appraisal (OEA)—the ability to perceive and understand the emotions of others, which corresponds to what is commonly termed ’empathy’ (4 items);  Use of Emotion (UOE)—the ability to harness emotions to facilitate performance and social interactions, relating to motivation and social skills (4 items); and  Regulation of Emotion (ROE)—the ability to control and regulate one’s own emotions, corresponding to ‘self-regulation’ (4 items). All items are rated on a seven-point Likert scale ranging from 1 (totally disagree) to 7 (totally agree). The total scores range from 16 to 112, and higher scores indicate greater emotional intelligence. The WLEIS was utilized for the current study, as it has demonstrated strong internal consistency (Cronbach’s α = 0.89) in previous research conducted with nurses in tertiary care hospitals in Peshawar, KPK [4].

Table 1 presents the conceptual mapping between the WLEIS dimensions and commonly used terminology from Goleman’s model, clarifying the relationship between the four-dimensional framework adopted in this study and broader EI terminology.

WLEIS Dimension	Conceptual Equivalent	Commonly Used Terminology	Definition
Self-Emotion Appraisal (SEA)	Self-awareness	"Self-awareness" in Goleman’s model	The ability to understand and express one’s own emotions
Others’ Emotion Appraisal (OEA)	Empathy	"Empathy" in Goleman’s model	The ability to perceive and understand the emotions of others
Use of Emotion (UOE)	Motivation and social skills	"Motivation" and "Social Skills" in Goleman’s model	The ability to harness emotions to facilitate performance and social interactions
Regulation of Emotion (ROE)	Self-regulation	"Self-regulation" in Goleman’s model	The ability to control and regulate one’s own emotions

Table 1. Conceptual Mapping of WLEIS Dimensions.

Permission for Use

Formal written permission for use of the WLEIS was obtained directly from the WLEIS original developer, Dr. Chi-Sum Wong, The Chinese University of Hong Kong, via email correspondence. Dr. Wong confirmed that the scale can be used free of charge for non-profit academic research (pers. comm. C.-S. Wong, August 15, 2024).

McCloskey/Mueller Satisfaction Scale (MMSS): The instrument used to measure job satisfaction was the MMSS, a 31-item instrument created specifically for nursing populations [23]. It is rated on a 5-point scale (1 = very dissatisfied to 5 = very satisfied) across eight subscales: extrinsic rewards, scheduling, family-work balance, coworker relations, interaction opportunities, professional development opportunities, praise and recognition, and control and responsibility. Total scores range from 31 to 155. The MMSS has demonstrated satisfactory reliability (Cronbach’s α = 0.89-0.92) in previous research conducted with nurses in public sector hospitals in Punjab, Pakistan.

Table 2 summarizes which instruments were administered to each participant group, clarifying that the WLEIS was completed only by nursing leaders while the MMSS, MBI-HSS, and Clinical Practice Outcomes Section were completed by both groups.

Instrument	Nursing Leaders	Staff Nurses
Wong and Law Emotional Intelligence Scale (WLEIS)	✓	✗
McCloskey/Mueller Satisfaction Scale (MMSS)	✓	✓
Maslach Burnout Inventory – Human Services Survey (MBI-HSS)	✓	✓
Clinical Practice Outcomes Section	✓	✓

Table 2. Instrument Administration by Participant Group.

The WLEIS was administered only to nursing leaders to assess their emotional intelligence levels. The MMSS, MBI-HSS, and Clinical Practice Outcomes Section were completed by both nursing leaders and staff nurses, allowing for comparison of job satisfaction, burnout, and clinical outcome reporting across both groups.

The Maslach Burnout Inventory — Human Services Survey (MBI-HSS): The MBI-HSS is a 22-item questionnaire [24] with three subscale scores: Emotional Exhaustion (EE; 9 items; range 0-54), Depersonalization (DP; 5 items; range 0-30), and Personal Accomplishment (PA; 8 items; range 0-48). Each item is scored on a seven-point frequency scale (0 = never to 6 = every day). Higher scores on Emotional Exhaustion and Depersonalization indicate higher burnout, while higher scores on Personal Accomplishment indicate lower burnout.

Clinical Practice Outcomes Section

This researcher-developed section assessed unit-level adverse events over the past three months. Participants were asked to report: (1) the number of medication errors occurring in their unit, (2) the number of patient falls, (3) the number of hospital-acquired pressure ulcers, (4) the percentage of hand hygiene compliance, and (5) the percentage of documentation accuracy. All items were self-reported by nursing staff based on their unit’s records and observations. This section was developed specifically for this study and required no external permission for use.

The challenges and difficulties of translation and cultural adaptation

The procedure of forward-backward translation as recommended by Brislin [25] was followed. Two independent bilingual nursing educators (Master’s degrees, 10+ years of nursing education experience in KPK) performed forward translations from English to Urdu. A third translator (PhD candidate in Public Health, blinded to the original versions) performed back-translation. Discrepancies were resolved by consensus. An expert panel comprising: (1) a Professor of Nursing (15+ years, instrument development expertise), (2) a Senior Nursing Superintendent (tertiary care hospital), and (3) a Senior Lecturer in Nursing (cultural competence training) reviewed the reconciled Urdu versions for linguistic and cultural appropriateness.

Pre-testing with cognitive interviewing was conducted with 15 nurses (5 leaders, 10 staff) not included in the final sample. Participants verbalized their understanding of each item. Based on feedback, modifications were made to 3 items to clarify ambiguous phrasing. All changes were reviewed and approved by the expert panel.

Validity and Reliability of Previously Validated Instruments

Three instruments used in this study—the Wong and Law Emotional Intelligence Scale (WLEIS), the McCloskey/Mueller Satisfaction Scale (MMSS), and the Maslach Burnout Inventory – Human Services Survey (MBI-HSS)—are well-established instruments with documented psychometric properties from prior studies.

Wong and Law Emotional Intelligence Scale (WLEIS)

The WLEIS was originally developed and validated by Wong and Law (2002) [2], with content validity established through expert review during instrument development. The scale has demonstrated strong psychometric properties in multiple international studies. In the Pakistani nursing context, Ullah et al. [4] reported a Cronbach’s α of 0.89 for the WLEIS among nurses working in tertiary care hospitals in Khyber Pakhtunkhwa, indicating good internal consistency. The four-factor structure of the WLEIS (Self-Emotion Appraisal, Others’ Emotion Appraisal, Use of Emotion, and Regulation of Emotion) has been confirmed through confirmatory factor analysis in South Asian healthcare populations.

McCloskey/Mueller Satisfaction Scale (MMSS)

The MMSS was originally developed and validated by McCloskey and Mueller (1990) [23] specifically for nursing populations, with content validity established through extensive literature review and expert panel review during instrument development. In Pakistani nursing populations, the MMSS has demonstrated satisfactory reliability, with Cronbach’s α values ranging from 0.89 to 0.92 in public sector hospitals in Punjab, Pakistan. The eight-factor structure of the MMSS has been confirmed in multiple nursing populations internationally.

Maslach Burnout Inventory – Human Services Survey (MBI-HSS)

The MBI-HSS was originally developed and validated by Maslach and Jackson (1981) [24], with content validity established through extensive literature review and expert consultation. The instrument has been widely used internationally with documented reliability (Cronbach’s α = 0.88-0.92 across various studies). The three-factor structure (Emotional Exhaustion, Depersonalization, and Personal Accomplishment) has been confirmed through extensive factor analytic studies across diverse healthcare settings.

Validity and Reliability of the Researcher-Developed Clinical Practice Outcomes Section

The Clinical Practice Outcomes section was developed specifically for this study to assess unit-level adverse events over the past three months. As this section was not a previously validated instrument, its validity and reliability were evaluated through the following procedures:

Content Validity

The Clinical Practice Outcomes section was assessed for content validity by a three-member expert panel consisting of two Nursing faculty members from Elizabeth Rani College of Nursing and Allied Health Sciences and one Nursing Superintendent from Mardan Medical Complex. Each item was judged for its relevance and clarity by the panelists. The Content Validity Index (CVI) was calculated based on expert ratings of item relevance (rated on a 4-point scale: 1 = not relevant to 4 = highly relevant). Items were retained if they achieved a CVI of 0.78 or higher. The overall CVI for the section was 0.89, indicating good content validity.

Test-Retest Reliability

A test-retest reliability assessment was conducted with 20 nurses over a 2-week interval. The same Clinical Practice Outcomes section was administered twice, and the intraclass correlation coefficient (ICC) was calculated. The ICC value was r = 0.82, indicating good stability over time.

Internal Consistency Reliability of All Instruments in the Present Study

Cronbach’s alpha was used to assess the internal consistency of all instruments in the present study sample. The results are presented in Table 2 (for subscale-level reliability) and summarized below:

WLEIS: Overall α = 0.91; subscale α values ranged from 0.84 to 0.87

MMSS: Overall α = 0.90; subscale α values ranged from 0.82 to 0.88

MBI-HSS: Overall α = 0.88; subscale α values ranged from 0.84 to 0.89

All coefficient alpha values exceeded the recommended threshold of 0.70 by Nunnally & Bernstein, (1994) [26], indicating good to excellent internal consistency for all instruments in the present study sample.

Table 3 provides a comprehensive summary of the validity and reliability evidence for all instruments used in this study, including prior validation evidence and reliability coefficients obtained in the present sample.

Instrument	Type	Validity Evidence	Reliability Evidence (Prior Studies)	Reliability Evidence (Present Study)
WLEIS	Previously validated	Content validity established by original developers; four-factor structure confirmed in South Asian populations [2]	Cronbach’s α = 0.89 in KPK nurse samples [4]	Overall α = 0.91; subscales α = 0.84-0.87
MMSS	Previously validated	Content validity established by original developers; eight-factor structure confirmed in nursing populations [23]	Cronbach’s α = 0.89-0.92 in Pakistani nurse samples	Overall α = 0.90; subscales α = 0.82-0.88
MBI-HSS	Previously validated	Content validity established by original developers; three-factor structure confirmed in healthcare populations	Cronbach’s α = 0.88-0.92 in international studies	Overall α = 0.88; subscales α = 0.84-0.89
Clinical Practice Outcomes Section	Researcher-developed	Content Validity Index (CVI) = 0.89 (expert panel review)	Not applicable (newly developed)	Test-retest ICC = 0.82
WLEIS = Wong and Law Emotional Intelligence Scale; MMSS = McCloskey/Mueller Satisfaction Scale; MBI-HSS = Maslach Burnout Inventory – Human Services Survey; CVI = Content Validity Index; ICC = Intraclass Correlation Coefficient.

Table 3. Summary of Validity and Reliability Evidence for All Instruments

Data Collection Procedure

Formal approval was obtained from the Medical Superintendent and Nursing Superintendent of each of the hospitals involved prior to data collection. Two parallel methods were used to collect data from October 2025 to January 2026. In the case of paper-based surveys, trained research assistants sought out eligible respondents in their clinical units at times when they were not busy. An informed consent was obtained in writing prior to the distribution of questionnaires. Sealed envelopes were used for returned completed surveys to ensure confidentiality.

For electronic surveys, an anonymous Google Form was sent out via WhatsApp to those who gave electronic consent through the use of an embedded consent statement on the first page of the Google Form. None of the information collected was identifiable by personal data in either method. To ensure consistency between the two modes of administration and minimize potential mode-related bias, several steps were taken. First, both paper and electronic versions were identical in content, item order, and response options. Second, the same standardized instructions were provided in both formats. Third, a preliminary analysis was conducted comparing responses from paper-based (n = 168) and electronic (n = 67) formats; independent samples t-tests revealed no significant differences in mean scores on any of the three main instruments (all p > 0.05), suggesting minimal mode-related bias. Fourth, for electronic responses, completion times were reviewed to identify potential speeders; no respondents were excluded on this basis as all completion times were within acceptable ranges. Questionnaires were excluded from analysis if they had missing data of more than 10%. The responses to the paper-based format were entered manually in the IBM SPSS Statistics Version 26, while the responses filled in on electronic format were exported from Google Forms to Microsoft Excel and then imported into the SPSS.

Ethical Considerations

This study followed the ethical standards of the Declaration of Helsinki. The study was approved by the Institutional Review Board (IRB) of Elizabeth Rani College of Nursing and Allied Health Sciences, Mardan (Approval No. ERCON/IRB/2025/07; date of approval: 15 July 2025) prior to data collection. All participants gave informed consent. The participation was voluntary and no consequences would follow if participants wanted to drop out. There were no incentives provided. Names, employee identifiers or contact information have been kept anonymous throughout. Data was collected and kept on a password protected device which was only accessible to the principal investigator. Paper surveys were obtained in a locked cabinet in the institution. All data will be deleted 3 years after publication.

Data Analysis

Data were analyzed using IBM SPSS Statistics for Windows, Version 26.0 (IBM Corp., Armonk, NY, USA). All the demographic and study variables were analyzed using descriptive statistics, which included means, standard deviations, frequencies, and percentages. Cronbach’s alpha was used to assess the internal consistency of each instrument (Table 4).

Linkage of Research Questions and Hypotheses to Statistical Analyses

Table 4 presents the linkage between each research question/hypothesis and the corresponding statistical analysis, along with the rationale for each analytical choice.

Research Question/Hypothesis	Statistical Analysis	Rationale
RQ1: What is the mean EI score of nursing leaders?	Descriptive statistics (means, standard deviations)	To summarize the central tendency and variability of EI scores
RQ2: Are there differences in job satisfaction between nursing leaders and staff nurses?	Independent samples t-test (cluster-robust with robust standard errors)	To compare mean job satisfaction scores between two independent groups while accounting for nesting of staff nurses within leaders
RQ3: Are nursing leader EI scores correlated with staff outcomes?	Pearson’s product-moment correlation (cluster-robust using SPSS COMPLEX SAMPLES)	To examine the strength and direction of linear relationships between EI and continuous outcome variables while accounting for within-leader clustering
RQ4/H3: Are clinical outcomes different across EI tertile groups?	One-way ANOVA (cluster-robust using GEE with exchangeable correlation structure) with post hoc Tukey’s HSD; Chi-square tests	To compare means across three independent groups and examine differences in categorical adverse event rates while accounting for clustering
RQ5/H4: Which EI dimensions predict clinical adherence?	Stepwise multiple linear regression (cluster-robust using SPSS COMPLEX SAMPLES REGRESSION)	To identify the most important EI dimensions predicting clinical adherence while controlling for other dimensions and accounting for clustering

Table 4. Linkage of Research Questions and Hypotheses to Statistical Analyses

Tertile Creation and Unit of Analysis

The 39 nursing leaders were ranked by their overall WLEIS scores and divided into three tertiles based on the distribution of their scores: the lowest tertile (bottom 33.3%), the middle tertile (middle 33.3%), and the highest tertile (top 33.3%). The cut-off values for the tertiles were determined as follows: Low EI: WLEIS score < 3.85; Medium EI: WLEIS score 3.85–4.75; and High EI: WLEIS score > 4.75. Staff nurses (n = 196) were assigned to the EI tertile of their respective nursing leader, meaning that all staff nurses working under a nursing leader with a Low EI score were classified into the Low EI group, and similarly for Medium EI and High EI groups. This classification resulted in the following group sizes: Low EI (n = 78), Medium EI (n = 79), and High EI (n = 78), accounting for the full sample of 235 participants. The staff nurse served as the primary unit of analysis for all inferential analyses. Each staff nurse’s responses were linked to their respective nursing leader’s EI score. While staff nurses were nested within leaders, this clustering was accounted for by using robust standard errors to address potential non-independence of observations within leader groups. It is important to clarify the nature of the clinical outcome variables. Medication errors, patient falls, pressure ulcers, hand hygiene compliance, and documentation accuracy are unit-level indicators, reflecting the aggregated performance of the clinical unit (ward) over the past three months, rather than individual staff nurse behaviors. All staff nurses working on the same clinical unit reported identical values for these outcomes, as they were reporting their unit’s aggregated data. This creates a within-unit correlation (pseudoreplication) that violates the assumption of independence at the individual nurse level. This non-independence is explicitly addressed through the use of cluster-robust standard errors with nursing leader as the cluster variable, which corrects for the correlated nature of the data and provides valid inference for the relationships examined at the individual nurse level.

Statistical analysis

Prior to conducting inferential analyses, statistical assumptions were assessed. Normality was evaluated through visual inspection of Q-Q plots and Kolmogorov-Smirnov tests, which indicated no substantial departures from normality. Homogeneity of variance was examined using Levene’s test, with no significant violations observed across groups for t-tests and ANOVA (all p > 0.05). For regression analyses, multicollinearity was assessed using Variance Inflation Factors (VIFs), which ranged from 1.85 to 2.10, indicating no evidence of problematic multicollinearity. Scatterplots indicated approximately linear relationships between predictors and outcome variables. Standardized residual plots showed random dispersion around zero, supporting the assumption of homoscedasticity. The normality of regression residuals was further assessed using the Shapiro-Wilk test and visual inspection of residual distributions, with no substantial deviations from normality observed (p > 0.05). Statistical significance was set at α = 0.05, and all tests were two-tailed.

Handling of Hierarchical Data Structure

The study design involved a hierarchical data structure in which staff nurses (n = 196) were nested within nursing leaders (n = 39), with an average cluster size of 5.03 staff nurses per leader (range: 2-8 per leader). This clustering creates non-independence of observations, as staff nurses within the same unit may be more similar to each other than to staff nurses in other units due to shared leadership, work environment, and unit culture. Failure to account for this non-independence can lead to underestimated standard errors and inflated Type I error rates.

To address this clustering, robust standard errors (Huber-White sandwich estimators) were used for all inferential analyses. All reported inferential statistics (t, F, r) presented in the Results section are derived from these cluster-robust procedures, not from conventional analyses that ignore the nested data structure. The degrees of freedom reported throughout the Results section reflect the total sample size (N = 235) adjusted for the parameters estimated in each analysis, consistent with standard output from SPSS COMPLEX SAMPLES and GEE procedures. In SPSS COMPLEX SAMPLES, degrees of freedom are calculated based on the residual degrees of freedom from the design-adjusted analysis, which accounts for the clustering structure while using the total sample size as the basis for the denominator degrees of freedom. Similarly, for GEE analyses, degrees of freedom are based on the total number of observations minus the number of estimated parameters, with robust standard errors providing valid inference despite the nested data structure. The use of N = 235 in the reported degrees of freedom (e.g., t(233), F(2,232), r(233)) reflects the total sample size minus the appropriate number of parameters estimated, and is consistent with standard practice in cluster-robust analyses where the effective sample size for hypothesis testing remains the total number of observations, while standard errors are adjusted for clustering. This approach is widely used in nursing and healthcare research when the number of clusters is limited and multilevel modelling is not feasible (McNeish et al., 2017) [28].

Robust standard errors adjust the standard errors of parameter estimates to account for within-cluster correlation, providing valid inference even when observations are not independent. These were implemented as follows:

For Pearson correlations, robust standard errors were estimated using IBM SPSS COMPLEX SAMPLES procedure with the nursing leader as the cluster variable. This procedure applies the Huber-White sandwich estimator to correct standard errors for the nested data structure.

For independent samples t-tests and one-way ANOVA, generalized estimating equations (GEE) with an exchangeable correlation structure were used to compare group means while accounting for within-cluster correlation. GEE provides population-averaged estimates that are robust to misspecification of the correlation structure.

For stepwise multiple linear regression, robust standard errors were implemented using the SPSS COMPLEX SAMPLES REGRESSION procedure, which applies the Huber-White sandwich estimator to produce cluster-robust standard errors and confidence intervals for regression coefficients.

Justification for Not Using Multilevel Modelling

While multilevel modelling (MLM) is the theoretically preferred approach for nested data, it was not feasible in this study due to the limited number of level-2 units. Simulation studies by Maas and Hox (2005) [27] demonstrated that multilevel models require a minimum of 50 clusters at Level 2 for reliable estimation of random effects and standard errors. Similarly, McNeish et al., (2017) [28] showed that with fewer than 50 clusters, multilevel models can produce biased estimates of random effects and standard errors, particularly for cross-level interactions. With only 39 nursing leaders, we determined that multilevel modelling would have been underpowered for reliable random effects estimation. The use of robust standard errors provides a conservative and appropriate alternative that corrects for non-independence without imposing potentially unstable random effects specifications. The staff nurse served as the primary unit of analysis for all inferential analyses. Each staff nurse’s responses were linked to their respective nursing leader’s EI score, and all analyses were conducted at the individual staff nurse level with clustering accounted for through robust standard errors.

Hierarchical Structure Considerations

The hierarchical structure of the data (staff nurses nested within nursing leaders) was explicitly addressed in all statistical analyses through the use of robust standard errors. This approach was selected because multilevel modelling, while theoretically appropriate, would have been underpowered given the limited number of nursing leaders (n = 39). Simulation studies by Maas and Hox (2005) and McNeish (2017) [27,28] indicate that multilevel models require a minimum of 50 clusters at Level 2 for reliable random effects estimation. With only 39 nursing leaders, we opted for robust standard errors to correct for non-independence without imposing potentially unstable random effects specifications. The intraclass correlation coefficient (ICC) was estimated to assess the proportion of variance in key outcome variables attributable to between-leader differences. The ICC values ranged from 0.12 to 0.18 across outcome variables, indicating moderate clustering effects that warranted correction for non-independence. All analyses were conducted at the individual staff nurse level (n = 196) with clustering accounted for through robust standard errors using the nursing leader as the cluster variable.

RESULTS

Response rate, and participant flow.

A total of 305 nursing personnel were invited to participate with 77.0% (257) giving informed consent, and 235 of them returned completely filled questionnaires. Nursing leaders had a higher response rate (86.7%, n = 39) than did staff nurses (75.4%, n = 196). As per methodology, questionnaires with incomplete responses more than 10% were excluded from the analysis.

Demographic Characteristics

The demographic characteristics of the 235 participants are presented in Table 5.

Characteristic	Category	Nursing Leaders (n = 39) n (%)	Staff Nurses (n = 196) n (%)	Total (N = 235) n (%)
Age (years)	20–29	2 (5.1)	68 (34.7)	70 (29.8)
	30–39	18 (46.2)	89 (45.4)	107 (45.5)
	40–49	14 (35.9)	32 (16.3)	46 (19.6)
	50+	5 (12.8)	7 (3.6)	12 (5.1)
Gender	Male	9 (23.1)	32 (16.3)	41 (17.4)
	Female	30 (76.9)	164 (83.7)	194 (82.6)
Marital Status	Married	33 (84.6)	138 (70.4)	171 (72.8)
	Unmarried	6 (15.4)	58 (29.6)	64 (27.2)
Qualification	Diploma	10 (25.6)	101 (51.5)	111 (47.2)
	BSN	14 (35.9)	62 (31.6)	76 (32.3)
	Post RN	11 (28.2)	28 (14.3)	39 (16.6)
	MSN	4 (10.3)	5 (2.6)	9 (3.8)
Hospital Type	Public	29 (74.4)	149 (76.0)	178 (75.7)
	Private	10 (25.6)	47 (24.0)	57 (24.3)
Clinical Unit	Medical	10 (25.6)	62 (31.6)	72 (30.6)
	Surgical	9 (23.1)	48 (24.5)	57 (24.3)
	ICU	7 (17.9)	33 (16.8)	40 (17.0)
	Emergency	6 (15.4)	27 (13.8)	33 (14.0)
	Paediatrics	4 (10.3)	16 (8.2)	20 (8.5)
	Other	3 (7.7)	10 (5.1)	13 (5.5)

Table 5. Demographic characteristics of the participants (N = 235).

The sample was predominantly female (82.6%), married (72.8%), and aged between 30 and 39 years (45.5%). Most were employed in public hospitals (75.7%) while the most frequently mentioned clinical setting was medical wards (30.6%). The credentialling system in KPK public hospitals was largely based on a diploma system, with nursing leaders having more likely qualifications of either a BSN or a Post registration qualification (28.2% and 28.2%).

Reliability Analysis

Table 6 presents Cronbach’s alpha coefficients for all study instruments.

Instrument	Subscale	Items (n)	Cronbach’s α
WLEIS	Self-Emotion Appraisal	4	0.86
	Others’ Emotion Appraisal	4	0.84
	Use of Emotion	4	0.85
	Regulation of Emotion	4	0.87
	Overall WLEIS	16	0.91
MMSS	Extrinsic Rewards	4	0.88
	Scheduling	6	0.86
	Family–Work Balance	3	0.82
	Coworkers	3	0.85
	Interaction Opportunities	4	0.84
	Professional Opportunities	3	0.87
	Praise and Recognition	3	0.86
	Control and Responsibility	5	0.88
	Overall MMSS	31	0.9
MBI-HSS	Emotional Exhaustion	9	0.89
	Depersonalization	5	0.85
	Personal Accomplishment	8	0.84
	Overall MBI-HSS	22	0.88

Table 6. The Cronbach’s Alpha Reliability Coefficients of all instruments are presented.

All instruments were pre tested for internal consistency with Cronbach’s alpha before inferential analysis. All alpha coefficients were above the suggested cut-off of 0.70 [26]: Overall, the WLEIS score was 0.91, the MMSS was 0.90, and the MBI-HSS was 0.88, which meant that these scores indicated excellent internal consistency on all measures.

Research Question 1 (RQ1): How do the EI scores of nursing leaders compare?

The mean WLEIS scores obtained for all four dimensions of the nursing leader are shown in Table 7, presents the mean WLEIS scores for all four dimensions among nursing leaders. It was revealed that the EI score of the nursing leaders of District Mardan was at moderate level with mean score of 4.35 (SD=0.85) on a seven point scale. Nursing leaders’ appreciation of their own emotional states was relatively more evident in the dimension of self-emotion appraisal (mean = 4.82, SD = 0.92). Ability to regulate emotional responses (mean = 3.98, SD = 1.15) was the lowest-scoring dimension, suggesting this to be a relative weakness in this sample.

WLEIS Dimension	Mean	SD	Min	Max
Self-Emotion Appraisal	4.82	0.92	2.5	6.75
Others’ Emotion Appraisal	4.15	1.08	2	6.5
Use of Emotion	4.43	0.96	2.25	6.75
Regulation of Emotion	3.98	1.15	1.75	6.5
Overall Emotional Intelligence	4.35	0.85	2.38	6.56

Table 7. The Mean Emotional Intelligence Score of all Nursing Leaders (n = 39) were calculated.

Research Question 2 (RQ2): Job Satisfaction Differences between Nursing Leaders and Staff Nurses

Independent samples t-tests were used to compare the nursing staff and the nursing leaders on MMSS. As shown in Table 8, nursing leaders reported significantly higher overall job satisfaction (mean = 2.92, SD = 0.72) compared to staff nurses (mean = 2.64, SD = 0.78), t(233) = 2.08, p = 0.039. Significant between-group differences were also found for extrinsic rewards, t(233) = 2.04, p = 0.043; scheduling, t(233) = 2.16, p = 0.032; family-work balance, t(233) = 2.21, p = 0.028; praise and recognition, t(233) = 2.13, p = 0.034; and control and responsibility, t(233) = 2.75, p = 0.006. No significant differences were found for interaction opportunities, t(233) = 0.41, p = 0.682, or coworker relations, t(233) = 0.46, p = 0.646. Interestingly, the overall mean of both groups were in the low to moderate range on a five-point scale, which indicates a general dissatisfaction among the nursing profession in District Mardan.

Table 8 compares job satisfaction scores between nursing leaders and staff nurses across all MMSS subscales.

MMSS Subscale	Nursing Leaders (n = 39) Mean (SD)	Staff Nurses (n = 196) Mean (SD)	Mean Difference	t	p-value
Extrinsic Rewards	2.68 (0.85)	2.35 (0.92)	0.33	2.04	0.043*
Scheduling	2.52 (0.92)	2.18 (0.88)	0.34	2.16	0.032*
Family–Work Balance	2.71 (0.96)	2.34 (0.94)	0.37	2.21	0.028*
Coworkers	3.45 (0.82)	3.38 (0.86)	0.07	0.46	0.646
Interaction Opportunities	3.38 (0.78)	3.32 (0.84)	0.06	0.41	0.682
Professional Opportunities	2.98 (0.88)	2.72 (0.91)	0.26	1.62	0.107
Praise and Recognition	2.84 (0.95)	2.48 (0.97)	0.36	2.13	0.034*
Control and Responsibility	2.78 (0.92)	2.32 (0.96)	0.46	2.75	0.006*
Overall Job Satisfaction	2.92 (0.72)	2.64 (0.78)	0.28	2.08	0.039*
*=significant test.

Table 8. Job Satisfaction Score (MMSS) was compared between nursing leaders and staff nurses.

Note to Table 8

Items were rated on a 5-point Likert scale (1 = very dissatisfied, 5 = very satisfied). Data are presented as mean (SD). Differences between nursing leaders and staff nurses were examined using independent-samples t-tests with cluster-robust standard errors, with nursing leader specified as the clustering variable to account for the nested data structure.

Research Question 3 (RQ3): Correlations between Emotional Intelligence and Outcomes of Practice

Pearson’s product-moment correlation coefficients were calculated to check for relationship between the overall EI score of nursing leaders with staff nurse practice outcomes. Table 9 shows the results. Staff job satisfaction was moderately positively correlated with leader EI,  r(233) = 0.42, p < 0.001; that is, employees’ job satisfaction was higher when their leader had a higher EI. There was a moderate negative correlation between EI and Burnout,  r(233) = -0.38, p < 0.001, indicating that staff members in units with higher EI leader competence reported lower burnout levels. There was also a significant negative correlation between EI and medication errors,  r(233) = -0.31, p < 0.001, and a significant positive correlation between EI and clinical adherence,  r(233) = 0.35, p < 0.001. All correlations were statistically significant at the p < 0.01 level (two-tailed).

Variable	Overall EI	Job Satisfaction (MMSS)	Burnout (MBI-HSS)	Medication Errors	Clinical Adherence
Job Satisfaction (MMSS)	r = 0.42, p < 0.001*	—
Burnout (MBI-HSS)	r = −0.38, p < 0.001*	r = −0.55, p < 0.001*	—
Medication Errors	r = −0.31, p < 0.001*	r = −0.32, p < 0.001*	r = 0.28, p < 0.001*	—
Clinical Adherence	r = 0.35, p < 0.001*	r = 0.38, p < 0.001*	r = −0.26, p < 0.001*	r = −0.29, p < 0.001*	—
*=significant test.

Table 9. The Pearson correlation between each nursing leader EI and each practice outcome is shown. (N = 235).

Note to Table 9

All correlations are Pearson’s product-moment correlation coefficients. N = 235 for all analyses, and all p-values are two-tailed. Correlation coefficients were estimated using cluster-robust procedures (SPSS Complex Samples) with nursing leader as the clustering variable.

Research Question 4 (RQ4): Clinical Outcomes by Emotional Intelligence Tertiles

According to the overall WLEIS scores of the leaders, the participants were divided into three tertiles: low EI (n = 78), medium EI (n = 79), and high EI (n = 78). The clinical outcome indicators were compared between tertiles using one-way ANOVA. Statistically significant differences were found for all five clinical outcome variables as presented in Table 10: medication errors, F(2, 232) = 32.45, p < 0.001; patient falls, F(2, 232) = 18.92, p < 0.001; pressure ulcers, F(2, 232) = 22.14, p < 0.001; hand hygiene compliance, F(2, 232) = 28.56, p < 0.001; and documentation accuracy, F(2, 232) = 30.12, p < 0.001.

Post-hoc Tukey’s HSD tests confirmed that units led by high-EI leaders reported significantly fewer medication errors (mean = 4.15, SD = 2.38) compared to units led by low-EI leaders (mean = 8.45, SD = 3.82), mean difference = -4.30, p < 0.001; patient falls (mean = 3.15, SD = 1.92 vs. mean = 5.68, SD = 2.75), mean difference = -2.53, p < 0.001; and pressure ulcers (mean = 2.25, SD = 1.58 vs. mean = 4.32, SD = 2.14), mean difference = -2.07, p < 0.001. Units led by high-EI leaders also demonstrated significantly higher hand hygiene compliance (76.8% vs. 58.2%), mean difference = 18.6%, p < 0.001, and documentation accuracy (79.6% vs. 62.5%), mean difference = 17.1%, p < 0.001, compared to units led by low-EI leaders.

Clinical Outcome	Low EI (n = 78) Mean (SD)	Medium EI (n = 79) Mean (SD)	High EI (n = 78) Mean (SD)	F	p-value
Medication Errors (3 months)	8.45 (3.82)	5.72 (2.94)	4.15 (2.38)	32.45	<0.001*
Patient Falls (3 months)	5.68 (2.75)	4.22 (2.18)	3.15 (1.92)	18.92	<0.001*
Pressure Ulcers (3 months)	4.32 (2.14)	3.15 (1.82)	2.25 (1.58)	22.14	<0.001*
Hand Hygiene Compliance (%)	58.2 (14.5)	68.4 (12.8)	76.8 (11.2)	28.56	<0.001*
Documentation Accuracy (%)	62.5 (15.2)	71.3 (13.4)	79.6 (10.8)	30.12	<0.001*
*=significant test.

Table 10. Medical outcomes are compared based on the nursing leader’s EI tertile.

Note to Table 10

Tertile classification was based on nursing leaders’ overall WLEIS scores: Low EI (< 3.85), Medium EI (3.85-4.75), and High EI (> 4.75). Staff nurses were assigned to the EI tertile of their respective nursing leader. Group comparisons were performed using cluster-robust procedures (GEE with exchangeable correlation structure) with nursing leader as the clustering variable.

Table 11 presents the post-hoc Tukey HSD pairwise comparisons for all clinical outcomes across EI tertiles.

One-way ANOVA with Tukey’s HSD post hoc tests confirmed significant differences between all three tertile groups for both job satisfaction scores (F(2, 232) = 28.56, p < 0.001) and burnout scores (F(2, 232) = 31.42, p < 0.001). Error bars represent standard deviations.

Clinical Outcome	Comparison	Mean Difference	95% CI	p-value
Medication Errors	High vs. Low	-4.3	[-5.78, -2.82]	<0.001*
	Medium vs. Low	-2.73	[-4.21, -1.25]	<0.001*
	High vs. Medium	-1.57	[-3.05, -0.09]	0.038*
Patient Falls	High vs. Low	-2.53	[-3.85, -1.21]	<0.001*
	Medium vs. Low	-1.46	[-2.78, -0.14]	0.03*
	High vs. Medium	-1.07	[-2.39, 0.25]	0.112
Pressure Ulcers	High vs. Low	-2.07	[-2.98, -1.16]	<0.001*
	Medium vs. Low	-1.17	[-2.08, -0.26]	0.012*
	High vs. Medium	-0.9	[-1.81, 0.01]	0.053
Hand Hygiene Compliance	High vs. Low	18.6	[13.45, 23.75]	<0.001*
	Medium vs. Low	10.2	[5.05, 15.35]	<0.001*
	High vs. Medium	8.4	[3.25, 13.55]	0.002*
Documentation Accuracy	High vs. Low	17.1	[11.85, 22.35]	<0.001*
	Medium vs. Low	8.8	[3.55, 14.05]	0.001*
	High vs. Medium	8.3	[3.05, 13.55]	0.002*
*=significant test; All pairwise comparisons are based on Tukey’s HSD post hoc tests following significant one-way ANOVA results.

Table 11. Post-hoc Tukey HSD Pairwise Comparisons for Clinical Outcomes

Each tertile of nursing leader EI is correlated with staff outcomes in Figure 1.

Figure 1. Job Satisfaction and Burnout Scores by Nursing Leader EI Tertile (N = 235).

The trend was clearly stepwise: Staff nurses working with high-EI leadership had higher job satisfaction scores than those working with medium-EI leaders, who had higher scores than those working with low-EI leaders. As a result, the scores of the Burnout scale fell as the leader EI increased. This trend suggests that there are important improvements in staff wellbeing outcomes associated with relatively small improvements in the EI of nursing leaders.

One-way ANOVA with Tukey’s HSD post hoc tests confirmed significant differences between all three tertile groups for medication errors (F(2,232) = 32.45, p < 0.001) and hand hygiene compliance (F(2,232) = 28.56, p < 0.001). Error bars represent standard deviations.

The pattern of clinical safety indicators is similar as seen in Figure 2.

Figure 2. Medication Errors and Hand Hygiene Compliance by Nursing Leader EI Tertile (N =235).

The number of medication errors/unit decreased from 8.45 in the low-EI tertile to 5.72 in the medium-EI tertile to 4.15 in the high-EI tertile, a 50.9% reduction between the lowest and highest EI tertiles, for the 3-month reporting period. Hand hygiene compliance increased from 58.2% in low-EI units, to 68.4% in medium-EI units and 76.8% in high-EI units. Overall, the results of Figures 1 and 2 indicate an incremental improvement pattern, whereby higher nursing leader EI levels were associated with progressively better staff wellbeing and patient safety indicators. This pattern of incremental improvements, while not establishing causality due to the cross-sectional design, suggests a consistent association between nursing leader EI and multiple domains of nursing practice outcomes.

Research Question 5 (RQ5): What are the predictors of positive nursing practice outcomes?

Patients’ clinical adherence was the dependent variable of the stepwise multiple linear regression model, which included the four dimensions of the WLEIS as the independent variables: Self-Emotion Appraisal, Others’ Emotion Appraisal, Use of Emotion, and Regulation of Emotion. Table 12 presents the final three-step multiple linear regression model predicting clinical adherence from the WLEIS dimensions. In Step 1, the highest predictor (β =.42, p < 0.001) was  Others’ Emotion Appraisal (which corresponds to what is commonly termed ’empathy’), explaining a large amount of variance in clinical adherence.  Regulation of Emotion (corresponding to ‘self-regulation’) was added in Step 2 (β = 0.31, p < 0.001), and  Use of Emotion (relating to motivation and social skills) was added in Step 3 (β = 0.20, p = 0.011).  Self-Emotion Appraisal did not meet the criteria for inclusion in the final model.

The final model explained 38% of the variance in clinical adherence (R² = 0.38, Adjusted R² = 0.36), and was statistically significant, F(3, 231) = 47.12, p < 0.001. There were no multicollinearity issues, as the variance inflation factor for all of the predictors ranged from 1.85 to 2.10.

Step	Predictor (WLEIS Dimension)	B	SE	β	t	p	VIF
1	Constant	2.45	0.32		7.66	<0.001*
	Others’ Emotion Appraisal	0.38	0.08	0.42	4.75	<0.001*	1.85
2	Constant	1.98	0.35		5.66	<0.001*
	Others’ Emotion Appraisal	0.32	0.07	0.36	4.57	<0.001*	1.92
	Regulation of Emotion	0.28	0.06	0.31	4.67	<0.001*	1.88
3	Constant	1.65	0.38		4.34	<0.001*
	Others’ Emotion Appraisal	0.28	0.07	0.31	4	<0.001*	2.1
	Regulation of Emotion	0.24	0.06	0.27	4	<0.001*	2.05
	Use of Emotion	0.18	0.07	0.2	2.57	0.011*	1.95
*=significant test; †=Others’ Emotion Appraisal is conceptually equivalent to empathy; ‡=Regulation of Emotion is conceptually equivalent to self-regulation; §=Use of Emotion relates to motivation and social skills.

Table 12. Descriptive statistics for stepwise multiple linear regression of clinical adherence predictors are presented (N = 235).

Note to Table 12

Self-Emotion Appraisal did not meet the inclusion criteria for the final model. The final model explained 38% of the variance in clinical adherence (R² =.38, Adjusted R² =.36; F(3, 231) = 47.12, p < 0.001). Outcome variable: clinical adherence composite score. All statistics were estimated using cluster-robust regression procedures with nursing leader as the clustering variable.

DISCUSSION

Overview

This section discusses the results of this present study with reference to the literature reviews on emotional intelligence and Nursing Practice Outcomes. This study addressed a critical gap in the nursing leadership literature by providing the first quantitative evidence on the relationship between nursing leader emotional intelligence and comprehensive nursing practice outcomes in the specific healthcare context of District Mardan, Khyber Pakhtunkhwa, Pakistan. Prior research in the region had established relationships between EI and individual outcomes such as burnout [4,22] and job satisfaction [9,12,18] but had not examined the broader array of nursing practice measures including medication errors and clinical adherence in a single study. Furthermore, no previous investigation had focused specifically on District Mardan’s semi-urban healthcare setting, which presents unique challenges including resource constraints, varying hospital tiers (tertiary, secondary, primary, and private), and specific workforce dynamics documented in the KPK nursing literature [4,18,20].

The mechanisms through which nursing leader EI may influence staff outcomes are multifaceted. First, through emotional contagion, leaders with high EI can transmit positive emotional states to their staff, creating a more supportive and less stressful work environment. Second, emotionally intelligent leaders are better equipped to provide social support, recognize staff distress, and intervene before burnout develops. Third, high EI leaders model effective emotional regulation, which staff may adopt through observational learning, thereby enhancing their own coping capacities. Fourth, emotionally intelligent leadership fosters psychological safety, where staff feel comfortable reporting errors and discussing concerns without fear of blame—a critical factor for patient safety. These mechanisms are particularly relevant in resource-constrained settings like District Mardan, where staff face high workloads and limited institutional support.

The discussion is presented in terms of the five research questions and four hypotheses that were explored in this study. The results are presented in context with the existing international evidence and regionally in the context of Khyber Pakhtunkhwa (KP), Pakistan.

The level of emotional intelligence amongst nursing leaders in District Mardan

The overall mean score of emotional intelligence of all the nursing leaders in District Mardan was 4.35 (SD = 0.85) which represents moderate level of EI on seven point scale. Self-Emotion Appraisal was the highest dimension (mean = 4.82) and Regulation of Emotion was the lowest (mean = 3.98). The results of this study coincide with previous research that reported moderate EI levels among nurses in tertiary health care hospitals in KPK [4], as well as studies that found moderate EI levels among nursing staff in Peshawar [4]. This moderate EI as observed in the present study could be attributed to the fact that formal leadership development programmes are not easily available in District Mardan hospitals, a challenge that has been highlighted in the context of nursing workforce challenges in Pakistan [17]. It is noteworthy that the scores on Regulation of Emotion are relatively low. This dimension shows how well individuals control and regulate their own emotions, especially important in clinical settings in which there is a high degree of stress. This is consistent with observations that the roles of nursing leaders are often emotionally challenging tasks, such as staffing, workload, and time-sensitive decision-making, which all require the nursing leader to have a strong ability to regulate their emotions [1,3]. The implication is that, in this context, targeted leadership development interventions that focus specifically on emotional regulation may result in positive outcomes that lead to increased nursing leadership effectiveness.

Differences in Job Satisfaction between Nursing Leaders and Staff Nurses Overall job satisfaction was significantly higher among the nursing leaders than staff nurses (mean = 2.92 vs. mean = 2.64; p = 0.039); but on the 5-point scale overall satisfaction is low to moderate for all the nurses of District Mardan. The result is similar to previous research which found that nurses in Pakistan were generally unhappy with the low levels of resources, pay, patient-nurse ratio, and professional development [18]. A high level of satisfaction in this context is related to the factors of structural/organizational factors, which are found to be the main causes of dissatisfaction in this context, because there are significant differences in the subscales of extrinsic rewards, scheduling, family–work balance, praise and recognition, and control and responsibility. By comparison, there are no significant differences for coworkers’ relations or coworker interaction opportunities, which shows that interpersonal aspects of the work environment are relatively more satisfactory. The pattern indicates that policies aimed at remuneration, workload distribution, and recognition systems within the teams will be more effective than team-building efforts. The results were consistent with previous research which concluded that organizational factors in the nursing workforce mediated between EI and job satisfaction [12].

Correlations between Emotional Intelligence and Practice Outcomes

As hypothesized, there was a significant positive moderate correlation between nursing leader EI and job satisfaction among staff, r(233) = 0.42, p < 0.001. This result is in line with the general literature, which has found positive correlations between EI and organizational commitment among nurses [6,9,11], and demonstrated that leaders’ EI positively correlates with nurses’ psychological empowerment and work engagement [11,29]. The results of the correlations found in the present study support the mediation of job satisfaction between EI and job performance identified by Chauhan, R. et al., (2022) [12].

To test Hypothesis 2, a significant moderate negative correlation was found between nursing leader EI and staff burnout, r(233) = -0.38, p < 0.001. This finding aligns with the significant negative correlation between EI and burnout syndrome among KPK nurses [4,22], and is consistent with findings that there is a significant association between EI and burnout syndrome among Pakistani clinicians [22]. Potentially, leader EI can help to diminish staff burn out by fostering more emotionally support work environments, which would be in line with research noting that high EI nurse managers employ more effective conflict management strategies, which in turn lead to healthier ward climates [9,10]. Medication errors were also negatively correlated with EI, r(233) = -0.31, p < 0.001, and positively correlated with clinical adherence, r(233) = 0.35, p < 0.001. The emotional competencies of nursing leaders are not only related to the wellbeing of nurses, but also to patient safety outcomes, as demonstrated in previous research on EI and delivery of healthcare quality [16].

Clinical Outcomes by Emotional Intelligence Tertile

The units with high-EI nursing leaders had scores that were significantly higher than the units with low-EI nursing leaders on each of the five indicators of clinical safety. The gradient between the lowest and the highest tertiles resulted in a 50.9% reduction in medication errors and a 76.8% improvement in hand hygiene compliance compared to 58.2% in low-EI units. The incremental improvement pattern supports Hypothesis 3 and suggests a significant association between nursing leader EI and patient safety indicators at the unit level. The clinical significance of these findings cannot be overstated. In a healthcare system where medication errors are a leading cause of preventable harm, a 50.9% reduction represents a substantial improvement in patient safety. Similarly, the 18.6 percentage-point increase in hand hygiene compliance has direct implications for infection prevention, particularly in a context where healthcare-associated infections are underreported and contribute significantly to morbidity and mortality. These improvements are not merely statistical artifacts but represent tangible benefits for patients and staff alike.

These results are consistent with the international literature. In the study conducted by Turjuman, F. et al., (2023) and Liu, N., et al., (2023) [29,30], high EI nurse managers were more likely to demonstrate transformational leadership behaviours, which are linked with better staff performance and patient outcomes relating to safety. A similar study [13] found EI was also significant in predicting the quality of work in healthcare workers during the COVID-19 pandemic. However, the present study extends these findings in novel ways by demonstrating the practical significance of leader EI in a semi-urban hospital setting in Pakistan, where healthcare resources are limited and the burden of preventable errors is high.

At the health system level, these findings have important implications for the KPK Health Department’s ongoing efforts to strengthen nursing leadership and improve healthcare quality. Our findings suggest that investing in EI development for nursing leaders could be a cost-effective strategy to improve both staff wellbeing and patient safety outcomes, complementing existing efforts to address structural challenges such as staffing ratios and infrastructure deficits. Unlike system-level reforms that require substantial capital investment, EI development programs can be implemented at relatively low cost through workshops, mentorship, and reflective practice groups. However, it is important to recognize that leader EI is not a panacea. The low-to-moderate job satisfaction scores observed across both leaders and staff nurses indicate pervasive dissatisfaction with structural factors such as compensation, workload, and recognition. These systemic issues require policy interventions at the organizational and governmental levels. EI development should be viewed as a complementary strategy that can enhance the effectiveness of leaders within existing constraints, but cannot replace the need for structural reforms. The most effective approach is likely to combine EI development for nursing leaders with systemic improvements in working conditions, compensation, and professional development opportunities.

Methodological Considerations Regarding Hierarchical Data Structure

An important methodological consideration in this study is the hierarchical structure of the data, with staff nurses nested within nursing leaders. The use of robust standard errors addresses the non-independence of observations arising from this clustering. The intraclass correlation coefficients (ICC) ranging from 0.12 to 0.18 across outcome variables indicate that 12-18% of the variance in staff outcomes was attributable to differences between nursing leaders. This moderate clustering effect underscores the importance of accounting for the hierarchical structure in the analysis. While robust standard errors provide valid inference for the relationships examined, the inability to use multilevel modelling due to the limited number of nursing leaders (n = 39) means that we could not formally partition variance into between-leader and within-leader components. The consistent finding that leader EI significantly predicts staff outcomes, even after accounting for clustering through robust standard errors, strengthens confidence in the robustness of these relationships. Nevertheless, replication with larger samples of leaders that would permit multilevel modelling is warranted to confirm these findings and to explore potential cross-level interactions.

Predictors of Positive Nursing Practice Outcomes

Three WLEIS dimensions emerged as significant predictors of clinical adherence in the final regression model: Others’ Emotion Appraisal (conceptually equivalent to empathy; β = 0.31), Regulation of Emotion (conceptually equivalent to self-regulation; β = 0.27), and Use of Emotion (β = 0.20). Self-Emotion Appraisal (self-awareness) did not meet the inclusion criteria for the final model and was excluded. The final model, comprising these three predictors, explained 38% of the variance in clinical adherence (R² = 0.38, Adjusted R² = 0.36). Consistent with Hypothesis 4, Others’ Emotion Appraisal and Regulation of Emotion emerged as the strongest predictors, while Use of Emotion made a smaller but statistically significant contribution to the model.

This emphasis on empathy (Others’ Emotion Appraisal) as a predictor aligns with Alsufyani, A.M. et al., (2020) and Aseery, M., et al., (2023) [10,15] who observed that empathy promotes communication, conflict resolution, and teamwork among healthcare professionals, which are all interrelated with the ability to make adherence to clinical protocols. In the context of District Mardan, where healthcare resources are limited and patient-to-nurse ratios are often high, leaders who can accurately perceive and understand the emotions of their staff (high Others’ Emotion Appraisal) are better equipped to address staff concerns, reduce tension during periods of high workload, and foster a supportive environment that facilitates adherence to clinical protocols. The importance of self-regulation (Regulation of Emotion) as a predictor is consistent with the theoretical work of Saha, S., et al., (2023) and Wong, C. S., et al., (2002) [2,3] who highlighted the ability to comprehend, control and regulate emotions as being crucial for successful nursing leadership. In the high-stress clinical environment of District Mardan hospitals, leaders who can regulate their own emotional responses—particularly during crises such as staffing shortages, equipment failures, or medical emergencies—are better able to maintain composure, make sound decisions, and model appropriate emotional responses for their staff, thereby promoting a calm and focused clinical environment that supports protocol adherence. The predictive contribution of social skills further suggests that leaders who can effectively use their emotions to facilitate communication and build relationships (Use of Emotion) are able to foster team cohesion and collaboration, which are essential for maintaining consistent clinical standards. The findings from these show that it is important to emphasize empathy and self-regulation training, as the two most important elements of leadership development programmes using EI in this context.

Theoretical Implications

The findings of this study extend existing theoretical frameworks on emotional intelligence and nursing leadership in several important ways. First, while previous research has established EI as a predictor of individual outcomes such as job satisfaction [9] and burnout [22], the present study demonstrates that leader EI has cascading effects across multiple domains of nursing practice, simultaneously influencing staff wellbeing and patient safety indicators. This suggests that EI should be conceptualized not merely as an individual competency but as an organizational resource with system-level consequences. Second, the identification of empathy (Others’ Emotion Appraisal) and self-regulation (Regulation of Emotion) as the strongest predictors of clinical adherence extends the theoretical understanding of which EI components matter most in healthcare leadership. While some frameworks emphasize all EI components equally, our findings suggest that in the high-stress, resource-constrained environment of semi-urban Pakistani hospitals, the ability to understand others’ emotions and regulate one’s own responses may be particularly salient. This aligns with conservation of resources theory, which posits that individuals in resource-scarce environments prioritize emotional competencies that help preserve psychosocial resources [32]. Third, the incremental improvement pattern observed between leader EI tertiles and both staff outcomes and patient safety indicators supports a threshold model of EI effectiveness. Rather than a simple presence/absence effect, incremental improvements in leader EI were associated with incremental improvements in outcomes, suggesting that leadership development efforts should aim for continuous improvement rather than binary competence.

CONCLUSION

Summary of Key findings

This study was quantitative cross sectional study which explored the relationship between EI of Nursing Leaders and Nursing Practice outcome in public and private hospital of district Mardan, Khyber Pakhtunkhwa, Pakistan. There are five main findings from the analysis.

The first results which were obtained by District Mardan nursing leaders showed moderate level of overall EQ (mean = 4.35, SD = 0.85), the strongest dimension was Self-Emotion Appraisal and the weakest dimension was Regulation of Emotion.

Secondly, nursing leaders reported significantly higher overall job satisfaction than staff nurses [t(233) = 2.08, p = 0.039]; however, both groups demonstrated overall job satisfaction levels within the low-to-moderate range on the five-point scale (leaders: mean = 2.92, SD = 0.72; staff: mean = 2.64, SD = 0.78), indicating a generally unsatisfactory work environment for the nursing workforce in District Mardan.

Third, nursing leader EI was strongly and positively associated with staff job satisfaction and clinical adherence, and strongly and negatively associated with staff burnout and medication errors, highlighting the wide effects of nursing leader EI on nursing practice outcomes.

Fourth, the clinical safety performance of units with high-EI nursing leadership was significantly better, with 50.9% fewer medication errors, and high compliance rates and documentation accuracy for hand hygiene compared to units with low-EI nursing leadership.

Fifth, The final regression model, which included three predictors—Others’ Emotion Appraisal (empathy; β = 0.31), Regulation of Emotion (self-regulation; β = 0.27), and Use of Emotion (β = 0.20)—explained 38% of the variance in clinical adherence (R² = 0.38, Adjusted R² = 0.36). While Use of Emotion was a statistically significant predictor in the model, Others’ Emotion Appraisal and Regulation of Emotion emerged as the strongest independent predictors. The findings of all four a priori research hypotheses were supported. Nursing leader EI was significantly associated with staff wellbeing and patient safety outcomes in District Mardan.

Theoretical Contributions

This study offers several contributions to the existing literature. First, it provides baseline quantitative data regarding EI levels of nursing leaders in District Mardan, addressing a missing dimension in the regional literature identified in previous studies [17,18,22]. Second, it extends the evidence on the EI-nursing outcomes relationship to a new semi-urban healthcare context in Pakistan, complementing and building upon prior research conducted in other parts of KPK. Third, unlike most previous studies that focused on single outcomes (job satisfaction or burnout), this study presents a more comprehensive picture of leader EI effects by simultaneously examining multiple nursing practice outcomes—including medication errors, patient falls, pressure ulcers, hand hygiene compliance, and documentation accuracy—alongside staff wellbeing indicators. Fourth, by identifying empathy and self-regulation as the strongest predictors of clinical adherence, the study provides specific, actionable targets for leadership development interventions in resource-constrained settings. This adds to the existing evidence on the EI-nursing outcomes relationship in a new semi-urban healthcare context in Pakistan and builds on previous evidence from the KPK region [4,18]. Additionally, this study presents a more complete picture of the effects of leader EI than most studies to date, as it considers a variety of nursing practice outcomes that are not typically assessed separately in other studies – such as medication errors, patient falls, pressure ulcers, hand hygiene compliance, and documentation accuracy – along with job satisfaction and burnout.

Practical Implications

Based on the findings of this study, there are some important practical implications for the nursing leadership in District Mardan and KPK in general.

The study offers empirical evidence to support the value of investing in EI for leadership development programmes focused on nursing leaders for hospital administrators and nursing superintendents. Development programmes should target the competencies of empathy and self-regulation in particular, as these were the strongest correlates of positive outcomes, and the weakest area of EI identified was regulation of emotion. Similar interventions, including reflective practice groups, emotional regulation workshops, and mentorship programs, have been shown to have a positive impact on enhancing EI-related skills among nursing leaders in similar contexts as reported by Saikia et al. (2024) [31].

The KPK Health Department discovered that nursing staff and leaders had lower to moderate levels of job satisfaction, suggesting a need for systemic policy interventions related to compensation, workload, scheduling, and recognition. These structural factors seem to affect job satisfaction even when people have low or high levels of EI and are a concern at the population level that can’t be solved by leadership development.

The findings support the integration of EI development to pre-registration and post-registration nursing education programs, as suggested by Bibi et al. (2025) [19] who have shown that there was a significant relationship between EI and academic achievements of the undergraduate nursing students in Pakistan.

Limitations

A few restrictions of the present study should be noted. First, the cross sectional design does not allow causal inferences to be made – the correlational results indicate association, but do not imply causation or that other unmeasured confounding variables have not had an impact. To determine causal relationships between nursing leader EI and practice outcomes, a longitudinal study design would be necessary. Secondly, the clinical outcomes data were self-reported by nursing staff and not derived from hospital record, so there was a risk of social desirability bias and recall error, including for the frequency of adverse events over 3 months. Future research should aim to complement the self-reported results with objective administrative data in order to establish a triangulation. Third, the study was conducted only in District Mardan and results of the study may not be generalizable to other districts of KPK or other provinces of Pakistan, as they could vary in healthcare infrastructure, staffing norms and cultural context. Fourth, the relatively small number of nursing leaders (n = 39) that were available may be representative of the available sample, but it is not sufficient to provide enough statistical power for subgroup analyses. Fifth, EI of staff nurses was not evaluated, and thus, the potential moderating or mediating effect of staff nurse EI on the relationship of leader EI and outcomes was not examined. Sixth, although robust standard errors were used to address clustering, the hierarchical structure of the data (staff nurses nested within nursing leaders) represents an important methodological consideration. The limited number of nursing leaders (n = 39) precluded the use of multilevel modelling, which would have been the preferred approach with a larger sample of leaders. With only 39 level-2 units, multilevel models would have been underpowered for reliable random effects estimation [27,28].

Future studies with larger samples of nursing leaders across multiple districts or provinces should employ multilevel modelling to fully account for the hierarchical structure and to examine cross-level interactions between leader-level characteristics (e.g., EI dimensions) and staff-level outcomes (e.g., job satisfaction, burnout). This would also allow for the examination of mediating and moderating effects at both levels of the hierarchy.

Recommendations for Future Research

Given the above limitations, the following suggestions for future studies are offered. Further research is required to find causal relationships and determine effectiveness, through the implementation of longitudinal and experimental designs, especially randomized controlled trials of EI-based leadership development interventions. Multi-district and multi-provincial studies would enable regional variations to be explored in nursing leader EI and determinants across Pakistan. Objective administrative adverse event and clinical outcome data should be added to the self-reported data in future studies. Investigating the moderator or mediator of the leader EI–outcomes relationship for staff nurse EI is warranted. Finally, the qualitative research component to explore nursing leaders experiences of emotional intelligence in a clinical context would lend context to an evidence base set by this study and related studies.

Conflicts of interest

The authors have no conflict of interest to declare with regard to the research, authorship or publication of this article.

Funding Statement

The research did not receive specific funding from any public, commercial or not-for-profit funding agencies. This study was carried out as independent academic research project in Elizabeth Rani College of Nursing and Allied Health Sciences, Mardan, Khyber Pakhtunkhwa, Pakistan.

Author Contributions

Shabir Muhammad, Muhammad Awais: Conceptualization, Methodology, Data Collection, Formal Analysis, Writing – Original Draft, Writing – Review and Editing, Project Administration.

Muhammad Ishaq, Bilal Khan: Supervision, Conceptualization, writing – Review and Editing.

Muhammad Hilal: Data Collection, Resources, Writing – Review and Editing.

Huzefa Amir, Adnan: Data Collection, Resources, Writing – Review and Editing.

Abdur Rahman, Waqar Alam: Supervision, Writing – Review and Editing, Corresponding Author responsibilities.

Ethical Approval

The procedures of this study were carried out in compliance with the declaration of Helsinki. The study was approved by the Institutional Review Board (IRB) of Elizabeth Rani College of Nursing and Allied Health Sciences, Mardan, Khyber Pakhtunkhwa, Pakistan (Approval No. ERCON/IRB/2025/07; date of approval: 15 July 2025).

Data Availability Statement

The data underlying this article cannot be shared publicly because it contains potentially identifying information about participants and should only be disclosed with their written informed consent.Data that support the findings of this study are available from the corresponding author upon reasonable request, subject to ethical restrictions. The Raw data of the participants cannot be shared freely to ensure the confidentiality and anonymity of the participants as approved by the Institutional Review Board of Elizabeth Rani College of Nursing and Allied Health Sciences, Mardan under approval No ERCON/IRB/2025/07.

Acknowledgements

The authors are thankful to the Nursing Heads and staff Nurses of all the participating hospitals in District Mardan for their cooperation and time. Thanks are due to the Medical Superintendents, Nursing Superintendents of Mardan Medical Complex, District Headquarters Hospital, Mardan, Tehsil Headquarters Hospitals and other private hospitals for allowing us to carry out this study. The authors also appreciate the contribution of the expert panel members for the content validity review and the research assistants for data collection.

REFERENCES

1. Lambert, S. (2021). Role of emotional intelligence in effective nurse leadership. Nursing Standard, 36(12), 45-49. https://doi.org/10.7748/ns.2021.e11782
2. Wong, C. S., & Law, K. S. (2002). The effects of leader and follower emotional intelligence on performance and attitude: An exploratory study. The Leadership Quarterly, 13(3), 243-274. https://doi.org/10.1016/S1048-9843(02)00099-1
3. Saha, S., Das, R., Lim, W. M., Kumar, S., Malik, A., & Chillakuri, B. (2023). Emotional intelligence and leadership: Insights for leading by feeling in the future of work. International Journal of Manpower, 44(4), 671-701. https://doi.org/10.1108/IJM-12-2021-0690
4. Ullah, S., Rafique, S., Sultan, A., Ahmed, F., & Khan, A. (2022). Emotional intelligence and burn out among nurses working in tertiary hospitals of KPK: Emotional intelligence and burn out. Pakistan Journal of Health Sciences, 3(2), 11-16. https://doi.org/10.54393/pjhs.v3i02.68
5. Alonazi, W. B. (2020). The impact of emotional intelligence on job performance during COVID-19 crisis: A cross-sectional analysis. Psychology Research and Behavior Management, 13, 749-757. https://doi.org/10.2147/PRBM.S263656
6. Al-Oweidat, I., Shosha, G. A., Baker, T. A., & Nashwan, A. J. (2023). The relationship between emotional intelligence and organizational commitment among nurses working in governmental hospitals in Jordan. BMC Nursing, 22(1), 195. https://doi.org/10.1186/s12912-023-01361-2
7. Majeed, N., & Jamshed, S. (2020). Nursing turnover intentions: The role of leader emotional intelligence and team culture. Journal of Nursing Management, 28, 1-11. https://doi.org/10.1111/jonm.13144
8. Alwali, J., & Alwali, W. (2022). The relationship between emotional intelligence, transformational leadership, and performance: A test of the mediating role of job satisfaction. Leadership & Organization Development Journal, 43(6), 928-952. https://doi.org/10.1108/LODJ-10-2021-0486
9. Soriano-Vázquez, I., Cajachagua Castro, M., & Morales-García, W. C. (2023). Emotional intelligence as a predictor of job satisfaction: The mediating role of conflict management in nurses. Frontiers in Public Health, 11, 1249020. https://doi.org/10.3389/fpubh.2023.1249020
10. Aseery, M., Mahran, S., Felemban, O., Aseery, M. Y., & Mahran, S. M. (2023). The relationship between emotional intelligence and conflict management strategies from the nurse managers’ perspective. Cureus, 15(3). https://doi.org/10.7759/cureus.35669
11. Alotaibi, S. M., Amin, M., & Winterton, J. (2020). Does emotional intelligence and empowering leadership affect psychological empowerment and work engagement? Leadership & Organization Development Journal, 41(8), 971-991. https://doi.org/10.1108/LODJ-07-2020-0313
12. Chauhan, R., Kaul, V., & Maheshwari, N. (2022). Impact of emotional intelligence on job performance of nurses with mediating effect of job satisfaction. Asia Pacific Journal of Health Management, 17(2), 1-8. https://doi.org/10.24083/apjhm.v17i2.1257
13. Alam, F., Yang, Q., Rūtelionė, A., & Bhutto, M. Y. (2023). Virtual leadership and nurses’ psychological stress during COVID-19 in the tertiary hospitals of Pakistan: The role of emotional intelligence. Healthcare, 11(11), 1537. https://doi.org/10.3390/healthcare11111537
14. Kitsios, F., Papageorgiou, E., Kamariotou, M., Perifanis, N. A., & Talias, M. A. (2022). Emotional intelligence with the gender perspective in health organizations managers. Heliyon, 8(11). https://doi.org/10.1016/j.heliyon.2022.e11488
15. Alsufyani, A. M., Baker, O. G., & Alsufyani, Y. M. (2020). Consequences of emotional intelligence in nursing: A concept analysis. Erbil Journal of Nursing and Midwifery, 3(1), 82-90. https://doi.org/10.15218/ejnm.2020.10
16. Oweidat, I., Alzoubi, M., Shosha, G. A., Ta’an, W., Khalifeh, A., Alzoubi, M. M., Al-Mugheed, K., Alabdullah, A. A. S., Abdelaliem, S. M. F. (2024). Relationship between emotional intelligence and quality of healthcare delivery. Frontiers in Psychology, 15, 1423235. https://doi.org/10.3389/fpsyg.2024.1423235
17. Zaman, N., Memon, K. N., Zaman, F., Khan, K. Z., & Shaikh, S. R. (2021). Role of emotional intelligence in job performance of healthcare providers working in public sector hospitals of Pakistan. Journal of Mind and Medical Sciences, 8(2), 245-251. https://doi.org/10.22543/7674.82.P245251
18. Suleman, Q., Syed, M. A., Mahmood, Z., & Hussain, I. (2020). Correlating emotional intelligence with job satisfaction: Evidence from a cross-sectional study among secondary school heads in Khyber Pakhtunkhwa, Pakistan. Frontiers in Psychology, 11, 240. https://doi.org/10.3389/fpsyg.2020.00240
19. Bibi, S., Khurshid, S., Jamil, A., Gul, A., & Parveen, K. (2025). Emotional intelligence and academic success among undergraduate nursing students in Pakistan: A cross-sectional study. NURSEARCHER, 5(3), 9-14. https://doi.org/10.54393/nrs.v5i3.172
20. Aisha, Ahmad, R., & Ali, U. (2024). Role of emotional intelligence in life satisfaction among hospital nurses in Pakistan. NURSEARCHER, 4(1), 47-50. https://doi.org/10.54393/nrs.v4i01.74
21. Zahir, S., Khan, K., Khan, M., Zahid, M., Arshad, K., Shahbaz, M., Khattak, A. A., Mazhar, S., Khan, A. S., & Malalai, Z. (2024). Evaluating emotional intelligence among undergraduate medical students: A cross-sectional study from Peshawar, Pakistan. Pakistan BioMedical Journal, 7(6). https://doi.org/10.54393/pbmj.v7i6.1114
22. Badshah, A., Yousaf, A., Mehboob, U., & Khan, M. B. (2025). Correlation of burnout syndrome with emotional intelligence among clinicians at workplace. Pakistan Journal of Medical Sciences, 41(1), 83. https://doi.org/10.12669/pjms.41.1.9725
23. McCloskey, J. C., & Mueller, C. W. (1990). The McCloskey/Mueller Satisfaction Scale. St. Louis, MO: Mosby.
24. Maslach, C., & Jackson, S. E. (1981). The measurement of experienced burnout. Journal of Occupational Behaviour, 2(2), 99-113. https://doi.org/10.1002/job.4030020205
25. Brislin, R. W. (1970). Back-translation for cross-cultural research. Journal of Cross-Cultural Psychology, 1(3), 185-216. https://doi.org/10.1177/135910457000100301
26. Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric Theory (3rd ed.). New York, NY: McGraw-Hill.
27. Maas, C. J. M., & Hox, J. J. (2005). Sufficient sample sizes for multilevel modeling. Methodology, 1(3), 85-91. https://doi.org/10.1027/1614-2241.1.3.86
28. McNeish, D. M., & Stapleton, L. M. (2016). The effect of small sample size on two-level model estimates: A review and illustration. Educational Psychology Review, 28(2), 295-314. https://doi.org/10.1007/s10648-014-9287-x
29. Turjuman F, Alilyyani B. Emotional intelligence among nurses and its relationship with their performance and work engagement: A cross-sectional study. Journal of Nursing Management. 2023;2023:5543299. https://doi.org/10.1155/2023/5543299
30. Liu N, Wang L, Yin J. Impact of transformational leadership on the person-organization fit of Chinese nurses: The moderating effect of emotional intelligence. Nursing Open. 2023;10(7):4586-4596. https://doi.org/10.1002/nop2.1706
31. Saikia M, George LS, Unnikrishnan B, Nayak BS, Ravishankar N. Thirty years of emotional intelligence: A scoping review of emotional intelligence training programme among nurses. International Journal of Mental Health Nursing. 2024;33(1):37-51. https://doi.org/10.1111/inm.13235
32. Hobfoll SE. Conservation of resources: A new attempt at conceptualizing stress. American Psychologist. 1989;44(3):513-524. https://doi.org/10.1037/0003-066X.44.3.513