Exploratory analysis of factors affecting the retention of medical students, doctors, and nurses in Lebanon during socioeconomic and political crises: a cross-sectional study

Abstract

Background

Due to the ongoing socioeconomic and political crisis in Lebanon, healthcare personnel face numerous challenges that are driving them toward emigration. This cross-sectional study examines the factors that affect Lebanese medical students’, doctors’, and nurses’ retention in the country.

Methods

Data were collected from August to December 2024 from Lebanese medical students, doctors, and nurses during a severe socioeconomic and political crisis. Exploratory factor analysis (EFA) was conducted on 819 participants via a structured 60-item online survey following a Likert-style scale to identify key retention determinants. Reliability analysis was performed via Cronbach’s alpha and McDonald’s omega, with a Kaiser–Meyer–Olkin (KMO) value of 0.964 and a statistically significant Bartlett’s test of sphericity (p < 0.05), confirming the data’s suitability for factor analysis.

Results

Twelve retention factors were identified and ranked by importance: compensation & benefits (mean = 4.58, α = 0.821), work‒life balance (mean = 4.55, α = 0.796), workplace sanitation & security (mean = 4.54, α = 0.813), disaster preparedness & crisis management (mean = 4.51, α = 0.919), job satisfaction & recognition (mean = 4.50, α = 0.911), political & economic stability (mean = 4.49, α = 0.844), autonomy (mean = 4.47, α = 0.809), support from colleagues (mean = 4.44, α = 0.889), career development (mean = 4.38, α = 0.773), support from management (mean = 4.38, α = 0.894), mental health & stress management (mean = 4.31, α = 0.862), and workplace comfort & amenities (mean = 4.20, α = 0.830).

Conclusion

This study highlights the multifaceted factors influencing healthcare professionals’ retention in Lebanon amid ongoing crises. These insights provide an opportunity for policymakers and healthcare leaders to focus on critical areas where targeted investments could strengthen the retention of doctors, nurses, and medical students in Lebanon and possibly in other countries experiencing similar crises.

Introduction

In previous years, Lebanon was met with a series of consecutive multifaceted crises. Prior to the onset of the COVID-19 pandemic, the country experienced a severe economic collapse through its liquidity crisis after years of unsustainable financial practices and dwindling foreign currency inflows, leading banks to impose severe withdrawal limits as dollar reserves evaporated [1]. Consequently, the United Nations reported (as of September 2021) that 82% of the Lebanese population is now experiencing multidimensional poverty [2]. COVID-19 exacerbated these conditions by exhausting the private healthcare system (Lebanese healthcare relies heavily on the private sector, which is characterized by inequalities in access according to socioeconomic classification) [3]. The situation further deteriorated as a result of the Beirut port blast that occurred on the fourth of August, 2020, which was deemed the “twenty-first century's most powerful nonnuclear explosion” [4]. It is believed that this ongoing socioeconomic and political crisis is considered to be the worst in Lebanon’s history [5]. Beirut blast alone killed at least 220 people and injured more than 6,000 people, displacing approximately 300,000 people and obliterating much of Beirut’s health infrastructure. It destroyed or incapacitated numerous hospitals (7 major hospitals were completely destroyed and 11 more were partially damaged), eliminating approximately 500 beds (approximately 36% of Lebanon’s health facility capacity) and leaving the remaining hospitals overwhelmed by the sudden surge in casualties. These mass-casualty events have repeatedly inundated emergency departments and stretched Lebanon’s medical response beyond its limits [6]. In addition to the multiple political and economic crises, there was a war that expanded geographically throughout the country in 2024 [7].

Given the prevailing economic conditions, limited resources, war, and constant pressure within the healthcare sector, professionals have increasingly sought emigration for better personal and professional stability abroad. The WHO estimated in September 2021 that approximately 40% of physicians and nearly 30% of nursing professionals had emigrated from Lebanon since October 2019 due to “brain drain” [8]. Living standards for doctors have severely declined as their fees have dramatically decreased: incomes have plummeted by more than 80% [9]. The health sector is also affected by the closure of various pharmaceutical companies, a shortage of medicines, and an increase in prices [10]. The mass exodus continues, as many young doctors and nurses choose to immigrate [11].

The emigration of nurses and physicians poses a serious challenge at different levels. First, this increases the already critical shortage of staff. The economic impact of this shortage is costly because health systems are unable to meet the needs of patients, leading to increased workloads, which are linked to diminished quality of care [12]. In addition, the exodus of newly graduated healthcare workers diminishes Lebanon's availability of skilled and motivated younger graduates who are eager to integrate into the workplace. For example, approximately one in five Lebanese nursing graduates emigrated within 1–2 years of graduation [13].

Therefore, this research does not seek to discourage healthcare professionals and medical students from pursuing opportunities abroad but rather to identify determinants that influence their retention in Lebanon, with the ultimate goal of strengthening the resilience of the sector amid ongoing socioeconomic and political crises. These findings might also influence policies and strategies geared toward improving the retention of healthcare professionals in countries experiencing similar crises.

“Retention” here is considered as “Employee Retention” which Frank et al. defined it as “… the effort by an employer to keep desirable workers in order to meet business objectives” [14]. Notably, there are many definitions of crises, we relied on the definition by Okumus and Karamustafa, who described a crisis as a significant and unexpected event that is likely to produce adverse outcomes, particularly for an organization (e.g., its employees and infrastructure) [15].

Importantly, we included medical students in our study because they exhibit strong internal drive and a genuine eagerness to prove themselves: one study reported that autonomous motivation was the predominant type across Lebanese medical students in the first five years [16]. We aimed to collect insights into how these students were experiencing during the study period and to identify the key retention factors that keep them working in Lebanon after graduation.

Methods

Study design

The study was conducted in 2024, from August to December, using Google Forms to administer a validated questionnaire for quantitative analysis with SPSS version 27. The questionnaire (see Appendix A) was available in both Arabic (the mother tongue language in Lebanon) and English, the two most common languages among Lebanese individuals, to ensure accessibility for all and reduce the language barrier.

Consent of the participants

The study included Lebanese medical students, doctors, and nurses aged 18 and above who were living in Lebanon at the time of the survey. The participants could complete the questionnaire in approximately 10 min. All participants were required to review an introduction outlining information about the study and provide electronic informed consent before beginning the survey. The respondents were assured of the confidentiality, anonymity, and security of their data on Google Forms. They were also given the option to withdraw from the survey at any time.

Questionnaire development and validation

The Google Form began with a detailed introduction outlining the objectives, purpose, and confidentiality measures of the study, followed by an informed consent section that participants were required to read and agree to before proceeding. The participants were also asked to confirm that they were at least 18 years old and were residing in Lebanon during the study. To ensure data integrity and prevent duplicate responses, the form was configured to accept only one submission per participant through the ‘Limit to one response’ feature, which required respondents to sign in with a unique email address. In addition, to prevent missing data, all Google Form questions were mandatory, and participants could not submit or move to the next section of the questionnaire without ensuring that all questions were answered.

The questionnaire was self-developed specifically for this study and was not adapted from any previously published tool. The questionnaire was developed on the basis of the researchers’ understanding of the major challenges facing Lebanese healthcare professionals during the ongoing crisis and was structured into four sections, each addressing a distinct dimension. In addition, we conducted an extensive review of the literature on healthcare workforce retention to identify key domains and determinants relevant to the study objectives. Based on these findings, an initial pool of questionnaire items was generated. This process was complemented by structured brainstorming sessions with a multidisciplinary team of experts, including hospital quality and accreditation officers, as well as healthcare professionals, to ensure the relevance, clarity, and contextual appropriateness of the items. Feedback from these sessions was used to refine and adapt the questionnaire prior to pilot testing and final deployment.

Section 1 of the questionnaire collected sociodemographic characteristics, stress-coping mechanisms, intentions to emigrate, and motivations to stay. Section 2 focused on professional background, including role, specialization, hospital type, work status, and work hours. Section 3 examines the factors and challenges affecting career development, such as pay, working conditions, burnout, and desired professional opportunities. Finally, Section 4 explores retention-related factors affecting healthcare workers’ decision to remain in Lebanon.

The questionnaire underwent a process of statistical validation and reliability testing, confirming its suitability and rigor for capturing our targeted variables. A pilot test with 10 participants confirmed the instrument’s clarity and internal consistency, after which minor linguistic refinements were made.

In total, 822 individuals accessed the questionnaire; however, only those who provided electronic informed consent were able to proceed, resulting in 819 valid responses retained for analysis. This high inclusion rate reflects the accessibility of the tool and the strong engagement of Lebanese healthcare professionals with issues of retention during crisis.

Inclusion and exclusion criteria

Participants were eligible if they were Lebanese medical students, doctors, or nurses aged 18 years or older and residing in Lebanon at the time of the study, ensuring that the sample reflected the target population experiencing the ongoing national crises. Individuals under 18 were excluded to comply with ethical standards for adult consent. Non-Lebanese participants and professionals outside Lebanon were excluded to maintain contextual relevance and focus on the Lebanese healthcare system. Respondents who were not medical students, doctors, or nurses were excluded to ensure alignment with the study’s objectives. Participation required informed consent; therefore, individuals who did not provide consent were excluded. Additionally, pilot study responses were excluded from the final analysis to avoid potential bias, as the pilot was conducted to test clarity and reliability of the questionnaire rather than to contribute to the main dataset.

Statistical analysis

Exploratory factor analysis (EFA) was conducted via principal axis factoring with Promax rotation to allow for correlated constructs. Sampling adequacy was confirmed by a Kaiser–Meyer–Olkin (KMO) value of 0.964, indicating excellent suitability for factor analysis, whereas Bartlett’s test of sphericity was statistically significant (χ² = 32,443.147, p < 0.001), confirming sufficient inter-item correlations for factor extraction. Items with factor loadings ≥ 0.4 were retained across almost all 12 extracted dimensions. However, some items with factor loading threshold of ≥ 0.30 were selected in accordance with established recommendations in exploratory factor analysis literature and because of their relevancy to our study. As noted by DeVellis and Thorpe [17], early stages of scale development prioritize identifying items that demonstrate at least a minimal but meaningful relationship with the underlying construct. Loadings around 0.30 are widely regarded as indicating a substantive, though modest, association between an item and a latent factor, particularly in studies with moderate-to-large sample sizes. DeVellis and Thorpe [17] emphasize that overly stringent cutoffs at preliminary stages may lead to the premature exclusion of items that could be theoretically important or perform better after refinement. Therefore, adopting a ≥ 0.4 threshold (with some with ≥ 0.3 threshold) supports an inclusive, theory-driven approach to item retention while still ensuring that retained items contribute meaningfully to the factor structure.

Reliability testing demonstrated high internal consistency, with Cronbach’s α values ranging from 0.773 to 0.919, and parallel McDonald’s ω coefficients supporting these findings.

Notably, regression analysis was not conducted in this study, as the primary objective was exploratory rather than predictive or inferential. Specifically, the study aimed to identify underlying latent constructs influencing the retention of healthcare professionals through the use of exploratory factor analysis (EFA). Since the factors were not predefined dependent variables but instead emerged from the data, applying regression modeling at this stage would not have been methodologically appropriate. The focus was therefore placed on uncovering the structure and dimensionality of retention-related factors rather than examining causal or predictive relationships between variables.

Sample size calculation

We adhered to the recommendations of Comrey and Lee, which suggest having 5–10 cases per variable in factor analysis [18]. With our questionnaire comprising 60 items related to retention factors, this guideline indicated a minimum sample size of 300–600 participants.

In addition, to further confirm the minimum sample size, we calculated it via the standard formula for estimating proportions in a survey-based study [19]:

$$\text{n}=\left({\text{Z}}^2\times \text{p}\times \left(1-\text{p}\right)\right)/{\text{e}}^2,$$

where:

• n is the required sample size,

• Z is the z score corresponding to the desired confidence level (1.96 for a 95% confidence interval),

• p is the estimated proportion of the population with the attribute of interest (set at 0.5 to maximize variability),

• e is the accepted margin of error (typically 0.05 for a 5% margin).

The values are substituted into the following formula:

$$\text{n}=\left(1.{96}^2\times 0.5\times 0.5\right)/0.05^2=384.16\text{.}$$

We decided to refine the estimate on the basis of a finite population by considering the number of doctors, nurses, and medical students in Lebanon according to the most recent statistics.

On the basis of the national physician density of 44.8 per 10,000 people [20] applied to the UN World Population Prospects 2024 population estimate for Lebanon [21], we estimated approximately 23,981 physicians in 2024 via this formula:

Number of Lebanese physicians in 2024 = (Lebanese physician density in 2024/10,000) × Lebanese population = (44.8/10,000) × 5,353,000 ≈ 23,981.

In addition, there were approximately 37,300 Lebanese nurses and midwives in 2024 based on a national nursing density of 63.2 per 10,000 people [22]. For the number of Lebanese medical students, we used data from Farhat et al., which reported that Lebanese medical schools had approximately 3,900 medical students in 2024 [23], resulting in a total finite population (N) of

$$\text{N}\approx 23,981+37,300+3,900=65,181\text{.}$$

Therefore, finite population correction (FPC) was applied:

$$\text{n}\_\text{adj}=\text{n}/\left[1+\left(\left(\text{n}-1\right)/\text{N}\right)\right]=384.16/\left[1+\left(\left(384.16-1\right)/65,181\right)\right]\approx 382\text{.}$$

Therefore, the minimum required sample size was approximately 382 participants.

Since we reached a sample size of 819, we far exceeded these requirements, confirming the validity, reliability, and generalizability of our study.

Data collection

This study used the “snowball sampling technique” for data collection, which is one of the most well-known forms of nonprobability practical sampling methods used by researchers when the population of interest is difficult to reach [24]. The method consists of giving the questionnaire to one person who then shares it with another until it has an increasing chain-link effect among participants [25]; it is also known to reach individuals who share similar characteristics [26]. Since questionnaires are distributed among medical students, doctors and nurses as initial respondents, they are distributed to other respondents who have characteristics similar to them (i.e., other doctors, medical students and nurses). The survey link was shared extensively through various social media platforms, including Facebook, Instagram, and WhatsApp. Notably, more than 10 individuals were involved in sharing the questionnaire and collecting the data.

In addition to online dissemination, the research team conducted onsite visits to several hospitals across Lebanon to increase participation and representativeness. The human resources departments of each institution were contacted to explain the study’s objectives and obtain formal approval. Upon approval, HR representatives facilitated the internal distribution of the survey by sharing it directly with their doctors and nurses, who were invited to participate voluntarily through the provided online link.

Results

Descriptive statistics

Table 1 reveals that the majority of participants were females, accounting for 59.0% of the total sample. The average age of the participants was 29.19 years. Most participants (27.0%) resided in North Lebanon. The majority of the participants (70.9%) dealt with stress by talking to friends or family. Additional details about the sample can be found in Table 1.

Table 1.

Description of the sociodemographic characteristics of the study sample (N = 819)

Variable	Frequency	Percentage
Gender
Female	483	59.0%
Male	336	41.0%
Educational level
Bachelor’s degree	329	40.2%
Doctor of Medicine (MD)	365	44.6%
Master’s degree	110	13.4%
PhD	15	1.8%
Residence
Baalbek-Herme	15	1.8%
Beirut	213	26.0%
Beqaa	29	3.5%
Mount Lebanon	213	26.0%
Nabatieh	41	5.0%
North Lebanon	225	27.5%
South Lebanon	83	10.1%
Marital Status
Divorced	13	1.6%
Married	232	28.3%
Single	573	70.0%
Widowed	1	0.1%
Do you have children?
No	619	75.6%
Yes	200	24.4%
Professional role
Doctor	199	24.3%
Medical student	276	33.7%
Nurse	344	42.0%
Which of the following coping mechanism(s) do you use to deal with stress?
Smoking	138	16.8%
Drinking alcohol	28	3.4%
Exercising	331	40.4%
Meditating	193	23.6%
Talking to friends or family	581	70.9%
Talking to a therapist	45	5.5%
I don’t have stress	3	0.4%
Do you have the intention to emigrate to another country?
No	262	32.0%
Not sure	269	32.8%
Yes	288	35.2%
What is/are the most important motivation(s) for you to stay in Lebanon?
Satisfied with current job	121	14.8%
Family commitment	579	70.7%
Financial stability	135	16.5%
Prefer the Lebanese lifestyle and culture	408	49.8%
Difficulty obtaining immigration documentation	132	16.1%
Lack of better opportunities abroad	64	7.8%
Emotional attachment to home country	307	37.5%
Commitment to contribute to Lebanon’s development	153	18.7%
Responsibility to care for aging parents or relatives	257	31.4%
Connection to Lebanese traditions and heritage	165	20.1%
Desire to raise children in Lebanon	266	32.5%
Concerns about political or social climate abroad	74	9.0%
Educational opportunities for children in Lebanon	55	6.7%
Commitment to the care of Lebanese patients	123	15.0%
	Mean ± SD
Age (in years)	29.19 ± 9.824

Table 2 illustrates the challenges faced by healthcare professionals in Lebanon as well as their career opportunities. The results showed that the majority of participants (70.6%) expressed that receiving better pay and benefits elsewhere would be a reason for leaving their current jobs. With respect to the greatest challenge they face as healthcare professionals in Lebanon, the majority of participants (74.1%) considered inadequate pay to be a major challenge, followed by 53.5% who considered high levels of stress to be a major challenge. Furthermore, research opportunities were the most desirable professional development opportunity for the majority of the participants (54.1%). Additional details can be found in Table 2.

Table 2.

Description of healthcare workers' challenges in Lebanon

Variable	Frequency	Percentage
What would most likely make you consider leaving your current job in Lebanon?
Better pay and benefits elsewhere	578	70.6%
Better work–life balance	366	44.7%
Opportunities for career growth	415	50.7%
Better working conditions	418	51.0%
Relocation	86	10.5%
Burnout	218	26.6%
Distance from home to workplace	120	14.7%
The culture	68	8.3%
What is the biggest challenge you face in your role as a healthcare professional in Lebanon?
Insufficient pay	607	74.1%
Long working hours	374	45.7%
Lack of resources and equipment	273	33.3%
Poor work–life balance	330	40.3%
Unsafe working conditions	251	30.6%
Limited opportunities for professional development	295	36.0%
High levels of stress	438	53.5%
Lack of support from management	295	36.0%
What type(s) of professional development opportunities would you like to have?
Workshops and seminars	430	52.5%
Advanced degree programs	412	50.3%
Certificate programs	332	40.5%
Online courses	214	26.1%
Mentoring programs	268	32.7%
Research Opportunities	443	54.1%

Exploratory factor analysis (EFA)

This section presents the results of the factor analysis using the principal axis factor method with Promax rotation to determine the optimal number of factors for assessing the retention of Lebanese doctors, nurses, and medical students.

Table 3 shows that the KMO value was 0.964, which is greater than the threshold of 0.5, and that the results of Bartlett’s test are statistically significant at the 0.05 level, indicating that the results are valid for factor analysis.

Table 3.

Assessment of sampling adequacy and sphericity for exploratory factor analysis

KMO and Bartlett's test		Value
Kaiser–Meyer–Olkin measure of sampling adequacy		0.964
Bartlett's test of sphericity	Approx. Chi-square	32,443.147
	df	1770
	Sig	0.000

Table 4 shows that the extracted communalities ranged from 0.298 to 0.855, indicating that most items shared a moderate to high proportion of variance with the factor solution. These findings suggest that the items were well represented in the factor model and suitable for retention in the analysis.

Table 4.

Communalities estimates following principal axis factoring

Communalities
	Initial	Extraction
Q1	0.347	0.298
Q2	0.544	0.506
Q3	0.510	0.528
Q4	0.553	0.547
Q5	0.624	0.679
Q6	0.455	0.452
Q7	0.520	0.494
Q8	0.535	0.469
Q9	0.497	0.439
Q10	0.518	0.509
Q11	0.603	0.523
Q12	0.590	0.533
Q13	0.518	0.533
Q14	0.617	0.657
Q15	0.608	0.602
Q16	0.524	0.490
Q17	0.643	0.659
Q18	0.749	0.855
Q19	0.726	0.749
Q20	0.538	0.565
Q21	0.584	0.576
Q22	0.565	0.584
Q23	0.569	0.539
Q24	0.587	0.541
Q25	0.600	0.559
Q26	0.677	0.735
Q27	0.646	0.676
Q28	0.628	0.627
Q29	0.596	0.585
Q30	0.558	0.574
Q31	0.540	0.553
Q32	0.571	0.627
Q33	0.667	0.715
Q34	0.756	0.824
Q35	0.684	0.671
Q36	0.503	0.547
Q37	0.510	0.516
Q38	0.603	0.681
Q39	0.538	0.536
Q40	0.666	0.636
Q41	0.754	0.735
Q42	0.748	0.734
Q43	0.751	0.764
Q44	0.689	0.675
Q45	0.651	0.644
Q46	0.538	0.567
Q47	0.581	0.619
Q48	0.554	0.560
Q49	0.620	0.634
Q50	0.452	0.417
Q51	0.658	0.622
Q52	0.695	0.705
Q53	0.640	0.611
Q54	0.606	0.577
Q55	0.589	0.560
Q56	0.534	0.513
Q57	0.657	0.650
Q58	0.639	0.597
Q59	0.504	0.494
Q60	0.475	0.455

Extraction method: principal axis factoring

Table 5 shows that factor extraction process using Principal Axis Factoring identified 12 factors with eigenvalues greater than 1. Together, these factors accounted for approximately 59.2% of the total variance after extraction. The first factor explained the largest proportion of variance (37.5%), followed by subsequent factors contributing progressively smaller amounts. This indicates a multidimensional structure underlying the dataset.

Table 5.

Eigenvalues and total variance explained by extracted factors

Total variance explained
Factor	Initial eigenvalues			Extraction sums of squared loadings			Rotation sums of squared loadingsa
	Total	% of variance	Cumulative %	Total	% of variance	Cumulative %	Total
1	22.921	38.201	38.201	22.526	37.543	37.543	15.220
2	2.927	4.879	43.081	2.533	4.221	41.764	15.680
3	2.588	4.314	47.394	2.171	3.619	45.383	12.399
4	2.038	3.396	50.791	1.664	2.773	48.156	13.946
5	1.568	2.613	53.403	1.184	1.973	50.129	12.759
6	1.408	2.347	55.751	1.027	1.712	51.841	9.590
7	1.280	2.133	57.884	0.926	1.544	53.385	11.989
8	1.242	2.071	59.955	0.855	1.425	54.809	13.584
9	1.116	1.860	61.815	0.702	1.171	55.980	12.686
10	1.072	1.786	63.601	0.700	1.166	57.146	14.118
11	1.047	1.746	65.347	0.634	1.057	58.204	15.424
12	1.035	1.725	67.072	0.602	1.004	59.208	13.083
13	0.967	1.612	68.684
14	0.881	1.468	70.152
15	0.792	1.320	71.472
16	0.778	1.297	72.769
17	0.701	1.168	73.936
18	0.678	1.131	75.067
19	0.645	1.075	76.142
20	0.639	1.066	77.208
21	0.604	1.007	78.215
22	0.579	0.964	79.179
23	0.555	0.925	80.104
24	0.538	0.897	81.001
25	0.522	0.869	81.871
26	0.506	0.843	82.714
27	0.501	0.835	83.548
28	0.471	0.785	84.334
29	0.447	0.745	85.079
30	0.439	0.732	85.810
31	0.437	0.728	86.539
32	0.408	0.680	87.218
33	0.404	0.673	87.891
34	0.392	0.654	88.545
35	0.383	0.638	89.182
36	0.362	0.604	89.786
37	0.349	0.581	90.367
38	0.343	0.572	90.940
39	0.338	0.563	91.503
40	0.328	0.547	92.050
41	0.322	0.536	92.586
42	0.316	0.526	93.112
43	0.307	0.512	93.624
44	0.305	0.509	94.132
45	0.295	0.491	94.623
46	0.285	0.474	95.098
47	0.277	0.461	95.559
48	0.270	0.450	96.009
49	0.253	0.422	96.431
50	0.245	0.409	96.840
51	0.238	0.397	97.237
52	0.221	0.368	97.605
53	0.217	0.362	97.967
54	0.211	0.351	98.318
55	0.198	0.331	98.649
56	0.186	0.310	98.959
57	0.176	0.294	99.252
58	0.163	0.272	99.524
59	0.147	0.245	99.770
60	0.138	0.230	100.000

Extraction method: principal axis factoring

When factors are correlated, sums of squared loadings cannot be added to obtain a total variance

Table 6 shows that the pattern matrix revealed clear item loadings across 12 factors, with most items demonstrating strong loadings (> 0.4) on their respective factors and minimal cross-loadings. This supports the stability and interpretability of the factor structure. Items were then grouped conceptually based on their highest loadings to define meaningful constructs.

Table 6.

Rotated factor loadings from promax rotation (pattern matrix)

Pattern matrixa
	Factor
	1	2	3	4	5	6	7	8	9	10	11	12
Q1			0.545
Q2			0.765
Q3			0.779
Q4						0.650
Q5						0.795
Q6						0.606
Q7						0.358
Q8
Q9
Q10			0.526
Q11			0.633
Q12			0.665
Q13											0.758
Q14											0.784
Q15											0.659
Q16											0.426
Q17							0.736
Q18							0.930
Q19							0.791
Q20										0.732
Q21										0.683
Q22										0.676
Q23											0.416
Q24				0.485
Q25				0.466
Q26				0.895
Q27				0.850
Q28				0.676
Q29									0.435
Q30									0.637
Q31									0.490
Q32									0.775
Q33								0.924
Q34								0.938
Q35								0.673
Q36					0.732
Q37					0.594
Q38					0.835
Q39					0.607
Q40	0.596
Q41	0.855
Q42	0.828
Q43	0.939
Q44	0.893
Q45	0.728
Q46												0.655
Q47												0.636
Q48												0.489
Q49												0.553
Q50	0.377
Q51		0.740
Q52		0.854
Q53		0.647
Q54		0.577
Q55		0.484
Q56		0.502
Q57		0.540
Q58		0.417	0.302
Q59		0.481
Q60	0.311				0.344

Extraction method: principal axis factoring

Rotation method: promax with Kaiser normalization

a. Rotation converged in 9 iterations

Notably, items 8 and 9 exhibited factor loadings below 0.30 and were therefore excluded from the table to enhance clarity and readability. Items such as 50 and 60 demonstrated loadings close to the 0.40 threshold and were retained due to their conceptual relevance to the construct being measured. Additionally, items 58 and 60 showed cross-loadings across multiple factors; in these cases, each item was assigned to the factor on which it demonstrated the highest loading. Although Q11 (loading = 0.633) and Q16 (loading = 0.426) exceeded the minimum loading threshold, they were removed during the factor refinement stage due to conceptual inconsistency with their respective factors. Finally, Q11 (“my health and wellbeing”) did not align with factor 3 (Work–Life Balance), Q16 (“cost of living in Lebanon”) was not theoretically consistent with factor 11 (Workplace Sanitation), and Q7 (“finding my work meaningful and satisfying”) did not align with factor 6 (Career Development), therefore they were excluded from the study.

Based on the rotated factor structure, 12 distinct factors were identified and labeled according to the conceptual meaning of their items. Factor loadings ranged from 0.344 to 0.939, indicating generally strong relationships between items and their respective constructs. The distribution of items across factors, along with their corresponding loadings and mean scores, is presented in detail in Table 7.

Table 7.

Distribution of items and factor loadings across retention domains

Factor	Original item code	Item	Factor loading	Mean score (mean of the items)	Possible score range	Number of items	N
Disaster preparedness and crisis management	Q40	Feeling that my workplace is well prepared for potential disasters	0.596	4.509	1.00–5.00	7	819
	Q41	Having clear protocols for dealing with emergencies	0.855				819
	Q42	Availability of resources and support for disaster response	0.828				819
	Q43	Dealing effectively with major events such as the COVID-19 pandemic	0.939				819
	Q44	Adequate support during crises such as the Beirut explosion	0.893				819
	Q45	Preparing for future major events	0.728				819
	Q50	Protection from Israeli bombing and attacks	0.377				819
Job satisfaction and recognition	Q51	Feeling valued and appreciated at work	0.74	4.503	1.00–5.00	9	819
	Q52	Satisfaction with my work environment	0.854				819
	Q53	The extent to which my job meets my career goals	0.647				819
	Q54	The recognition I receive for my accomplishments at work	0.577				819
	Q55	The alignment of my job role with my skills and interests	0.484				819
	Q56	Having a manageable workload	0.502				819
	Q57	Equitable distribution of working hours	0.54				819
	Q58	Adequate rest and breaks	0.417				819
	Q59	Feeling respected and appreciated by patients	0.481				819
Work–life balance	Q1	Flexible working hours and schedules	0.545	4.547	1.00–5.00	5	819
	Q2	My work schedule allows me to maintain a healthy work–life balance	0.765				819
	Q3	My working hours are manageable while considering my personal responsibilities	0.779				819
	Q10	Balancing family commitments and responsibilities with work	0.526				819
	Q12	Having sufficient time for personal activities and rest	0.665				819
Mental health and stress management	Q24	Receiving support when dealing with patient deaths	0.485	4.314	1.00–5.00	5	819
	Q25	Being able to deal effectively with rude or difficult patients	0.466				819
	Q26	Access to mental health support services	0.895				819
	Q27	Prioritizing mental health in my workplace	0.85				819
	Q28	Having the resources to manage work-related stress	0.676				819
Workplace comfort and amenities	Q36	Access to healthy food options in the cafeteria	0.732	4.197	1.00–5.00	5	819
	Q37	Adequate parking and transportation facilities	0.594				819
	Q38	Availability of well-equipped break rooms (e.g., sofas, TV, microwave, refrigerator)	0.835				819
	Q39	Availability of internet and good phone signals	0.607				819
	Q60	Availability of lockers and personal storage space	0.344				819
Career development	Q4	Opportunities for promotion and career growth	0.65	4.376	1.00–5.00	3	819
	Q5	Access to professional development and training	0.795				819
	Q6	Support to participate in research activities	0.606				819
Support from management	Q17	Establishing a positive relationship with management	0.736	4.377	1.00–5.00	3	819
	Q18	Feeling supported by management	0.93				819
	Q19	Communicating effectively with management	0.791				819
Support from colleagues	Q33	Establishing good working relationships with my colleagues	0.924	4.445	1.00–5.00	3	819
	Q34	Feeling supported and cooperative with colleagues	0.938				819
	Q35	A strong sense of teamwork among employees	0.673				819
Compensation and benefits	Q29	Being satisfied with my compensation and benefits	0.435	4.576	1.00–5.00	4	819
	Q30	Receiving fair pay for my work	0.637				819
	Q31	Having an adequate retirement plan	0.49				819
	Q32	Receiving regular salary increases	0.775				819
Autonomy	Q20	Enjoying sufficient autonomy to make decisions in my role	0.732	4.466	1.00–5.00	3	819
	Q21	The ability to practice my profession without undue interference	0.683				819
	Q22	Having my opinions and decisions respected by colleagues and supervisors	0.676				819
Workplace sanitation and security	Q13	Feeling protected from exposure to infection at work	0.758	4.543	1.00–5.00	4	819
	Q14	Availability of necessary medical equipment and resources	0.784				819
	Q15	Safety and cleanliness of my workplace	0.659				819
	Q23	Feeling safe from violence and harassment from patients or their families	0.416				819
Political and economic stability	Q46	The political stability in Lebanon	0.655	4.491	1.00–5.00	4	819
	Q47	Feeling safe at my workplace despite the political situation	0.636				819
	Q48	Government policies that support healthcare professionals	0.489				819
	Q49	Economic stability in the country	0.553				819

Figure 1 displays the scree plot showed a pronounced initial decline in eigenvalues followed by a gradual leveling off, indicating a clear elbow in the curve. To facilitate interpretation, a red horizontal line was added to represent the eigenvalue threshold of 1 (Kaiser criterion). Visual inspection shows that 12 factors lie above this threshold and were therefore retained for further analysis, as they account for the meaningful shared variance before the curve flattens.

Fig. 1.

Results of factor extraction

According to Table 8, the results showed that the factor analysis extracted 12 factors on the basis of 60 items, with factor loadings ranging from 0.344 to 0.939. These included: Disaster Preparedness and Crisis Management, Job Satisfaction and Recognition, Work–Life Balance, Mental Health and Stress Management, Workplace Comfort and Amenities, Career Development, Support from Management, Support from Colleagues, Compensation and Benefits, Autonomy, Workplace Sanitation and Security, and Political and Economic Stability.

Table 8.

Extracted factors and item loadings

Item	Item loading
Factor 1—disaster preparedness and crisis management
Dealing effectively with major events such as the COVID-19 pandemic	0.939
Adequate support during crises such as the Beirut explosion	0.893
Having clear protocols for dealing with emergencies	0.855
Availability of resources and support for disaster response	0.828
Preparing for future major events	0.728
Feeling that my workplace is well prepared for potential disasters	0.596
Protection from bombing and attacks	0.377
Factor 2—Job Satisfaction and Recognition
Satisfaction with my work environment	0.854
Feeling valued and appreciated at work	0.740
The extent to which my job meets my career goals	0.647
The recognition I receive for my accomplishments at work	0.577
Equitable distribution of working hours	0.540
Having a manageable workload	0.502
The alignment of my job role with my skills and interests	0.484
Feeling respected and appreciated by patients	0.481
Adequate rest and breaks	0.417
Factor 3—Work–Life Balance
My working hours are manageable while considering my personal responsibilities	0.779
My work schedule allows me to maintain a healthy work–life balance	0.765
Having sufficient time for personal activities and rest	0.665
Flexible working hours and schedules	0.545
Balancing family commitments and responsibilities with work	0.526
Factor 4—Mental Health and Stress Management
Access to mental health support services	0.895
Prioritizing mental health in my workplace	0.850
Having the resources to manage work-related stress	0.676
Receiving support when dealing with patient deaths	0.485
Being able to deal effectively with rude or difficult patients	0.466
Factor 5—Workplace Comfort and Amenities
Availability of well-equipped break rooms (e.g., sofas, TV, microwave, refrigerator)	0.835
Access to healthy food options in the cafeteria	0.732
Availability of internet and good phone signals	0.607
Adequate parking and transportation facilities	0.594
Availability of lockers and personal storage space	0.344
Factor 6—Career Development
Access to professional development and training	0.795
Opportunities for promotion and career growth	0.650
Support to participate in research activities	0.606
Factor 7—Support from Management
Feeling supported by management	0.930
Communicating effectively with management	0.791
Establishing a positive relationship with management	0.736
Factor 8—Support from Colleagues
Feeling supported and cooperative with colleagues	0.938
Establishing good working relationships with my colleagues	0.924
A strong sense of teamwork among employees	0.673
Factor 9—Compensation and Benefits
Receiving regular salary increases	0.775
Receiving fair pay for my work	0.637
Having an adequate retirement plan	0.490
Being satisfied with my compensation and benefits	0.435
Factor 10—Autonomy
Enjoying sufficient autonomy to make decisions in my role	0.732
The ability to practice my profession without undue interference	0.683
Having my opinions and decisions respected by colleagues and supervisors	0.676
Factor 11—Workplace Sanitation and Security
Availability of necessary medical equipment and resources	0.784
Feeling protected from exposure to infection at work	0.758
Safety and cleanliness of my workplace	0.659
Feeling safe from violence and harassment from patients or their families	0.416
Factor 12—Political and Economic Stability
The political stability in Lebanon	0.655
Feeling safe at my workplace despite the political situation	0.636
Economic stability in the country	0.553
Government policies that support healthcare professionals	0.489

Reliability analysis

Cronbach’s alpha and McDonald’s omega were computed to assess the reliability and internal consistency of the factors, as shown in Table 9.

Table 9.

Reliability analysis using Cronbach's alpha and McDonald's omega

Variable	Cronbach's alpha (α)	McDonald's omega (ω)
Disaster preparedness and crisis management	0.919	0.920
Job satisfaction and recognition	0.911	0.909
Work–life balance	0.796	0.795
Mental health and stress management	0.862	0.855
Workplace comfort and amenities	0.830	0.833
Career development	0.773	0.777
Support from management	0.894	0.896
Support from colleagues	0.889	0.893
Compensation and benefits	0.821	0.822
Autonomy	0.809	0.809
Workplace sanitation and security	0.813	0.816
Political and economic stability	0.844	0.846

The internal consistency of a scale refers to the extent to which items complement each other to measure the same construct. Cronbach’s alpha measures the internal consistency of items, assessing their stability. Hair et al. [27] suggested that values greater than 0.7 are accepted as reliable.

The Cronbach’s alpha values for all factors in this study ranged from 0.773 to 0.919, exceeding the threshold of 0.7. Similarly, McDonald’s omega values ranged from 0.777 to 0.920, also exceeding the threshold of 0.7. Therefore, the results indicated that the items used to assess each factor were reliable and valid.

Descriptive statistics for the extracted variables

In this section, the means and standard deviations are presented for each of the extracted factors.

According to Fig. 2, compensation and benefits had the highest mean score (mean = 4.58), with a standard deviation of 0.55, followed by work‒life balance, workplace sanitation and safety, and disaster preparedness and crisis management (mean = 4.55, 4.54, and 4.51, respectively). This finding indicates that the factor considered most important by the participants was compensation and benefits. In contrast, workplace comfort and amenities had the lowest mean scores (mean = 4.20). Additional details about the sample can be found in Fig. 2.

Fig. 2.

Descriptive statistics for the extracted variables

Association between professional role and retention factors

Table 10 reveals statistically significant differences across professional roles on nine retention factors (p-value < 0.05). In particular, nurses showed higher scores than doctors and medical students on disaster preparedness and crisis management, job satisfaction and recognition, mental health and stress management, workplace comfort and amenities, support from management, support from colleagues, compensation and benefits, workplace sanitation and safety, and political and economic stability. Additional information can be found in Table 10.

Table 10.

Association between professional role and factors affecting retention of Lebanese healthcare workers

Factor	Doctor	Medical student	Nurse	p-value
	Mean (SD)	Mean (SD)	Mean (SD)
Disaster preparedness and crisis management	4.37 (0.67)	4.49 (0.63)	4.61 (0.57)	0.000
Job satisfaction and recognition	4.48 (0.56)	4.45 (0.52)	4.56 (0.50)	0.024
Work–life balance	4.53 (0.55)	4.53 (0.58)	4.57 (0.49)	0.503
Mental health and stress management	4.24 (0.75)	4.21 (0.73)	4.44 (0.65)	0.000
Workplace comfort and amenities	4.15 (0.76)	4.13 (0.74)	4.27 (0.74)	0.048
Career development	4.29 (0.77)	4.39 (0.71)	4.41 (0.68)	0.132
Support from management	4.26 (0.82)	4.30 (0.72)	4.51 (0.67)	0.000
Support from colleagues	4.31 (0.74)	4.42 (0.66)	4.54 (0.66)	0.000
Compensation and benefits	4.55 (0.56)	4.48 (0.56)	4.67 (0.52)	0.000
Autonomy	4.45 (0.63)	4.42 (0.62)	4.51 (0.59)	0.134
Workplace sanitation and security	4.46 (0.66)	4.49 (0.63)	4.64 (0.52)	0.000
Political and economic stability	4.48 (0.62)	4.42 (0.67)	4.55 (0.59)	0.047

Discussion

To the best of our knowledge, this study is the first to use exploratory factor analysis (EFA) to identify the main elements affecting the retention of Lebanese medical students, physicians, and nurses. Although current studies on healthcare workforce retention in Lebanon have focused mostly on nurses, this study expands the scope by examining the viewpoints of doctors and medical students.

This study revealed that, in line with the systematic studies of Willis-Shattuck et al. [28] and De Vries et al. [29] and in Lebanon by Kawwas [30], with respect to the retention factors of healthcare professionals, most of the widely recognized healthcare workforce retention variables are also observed in Lebanon.

The nomenclature of related concepts varies somewhat. For example, while this study emphasizes "support from management" overall, the literature describes "support from a supervisor" as a retention factor. Nevertheless, the differences in nomenclature do not alter the fundamental idea we intend to preserve. Although it highlights “the workplace”, the literature does not specifically address "sanitation and security", which was found to be a retention factor for Lebanese doctors, nurses, and medical students and can be considered part of “the workplace”.

Notably, this study revealed new retention variables for Lebanese healthcare professionals that will be discussed later:

• Workplace Sanitation and Security

• Disaster preparedness and crisis management

• Political and Economic Stability

• Mental Health and Stress Management

• Workplace Comfort and Amenities

Payment as a primary retention factor and economic stability

The EFA results revealed "compensation and benefits" as the most important retention factor (with a mean score of 4.58). In addition, most of the participants in this study (74.1%) pointed to low remuneration as a serious difficulty. In line with recent research in Lebanon by Kawwas [30], remuneration remains one of the most important determinants of retention in Lebanon and comparable settings, which is expected given the country’s ongoing economic collapse.

Disaster preparedness as an emerging retention factor in Lebanon

The 2020 Beirut Port explosion revealed profound deficiencies in hospital emergency readiness. At LAUMC-RH, staff faced patient surges that overwhelmed capacity despite an existing emergency plan [31]. Further studies highlighted the absence of triage systems and emergency operating centers [32]. Strengthening preparedness through protocols, regular drills, and specialized response teams not only improves safety but also reinforces staff confidence and willingness to continue working during crises. Sultan et al. emphasized that confidence gained through proper preparedness training increases healthcare workers’ readiness to remain on duty during emergencies [33].

Importance of workplace sanitation

In a qualitative study of rural primary care centers in Niger, chronically unsanitary conditions (unsafe water points, overflowing bins, and sporadic cleaning) were cited by nurses as daily stressors that pushed staff to seek transfers or resign [34]. Multiple survey data from 14 low- and middle-income nations likewise reveal that healthcare providers working in facilities with reliable handwashing stations, clean toilets, and regular surface disinfection report significantly higher job satisfaction, which the authors note is a key precursor to retention [35]. The World Health Organization echoes this link, concluding that protecting health worker safety through adequate WASH contributes to improving the motivation and retention of health workers [36].

Amenities as an overlooked retention factor

Studies have linked comfortable and well-designed environments to increased motivation and productivity [37]. Elements such as break areas, access to natural light, and restorative spaces foster relaxation and mental well-being [38]. Ergonomic design, esthetic appeal, and safety features (such as secure parking and adequate lighting) improve satisfaction and efficiency while mitigating stress [39]. Additional amenities, including onsite childcare and meal services, further alleviate daily pressures and contribute to better retention [40].

Healthcare workplace security and protection against violence

A meta-analysis of over 330,000 healthcare workers revealed that 62% encountered some form of workplace violence in a single year, including 24% who experienced physical assault [41]. In Lebanon, the problem mirrors global trends but is intensified by the country’s economic collapse and political instability. A national survey revealed that 62% of nurses experienced verbal abuse and that 10% experienced physical violence within a year, predominantly from patients or their relatives [42]. Qualitative studies report that violent incidents, sometimes involving weapons, have increased during the economic crisis and the COVID-19 pandemic, fueled by patient frustration, financial distress and cultural acceptance of aggressive behavior [43]. Such violence significantly affects mental health and job satisfaction: over half of exposed nurses reported emotional exhaustion, and nearly one-third expressed an intention to quit [44].

Addressing this issue requires comprehensive prevention strategies, including zero-tolerance policies, incident reporting systems, and visible managerial support. Regular training in conflict resolution, de-escalation, and communication has proven effective in lowering violence rates [45].

Importance of work–life balance

The sample included 330 (40.3%) participants who mentioned poor work–life balance and 374 (45.7%) who mentioned long working hours as their greatest challenge. In addition, the EFA analysis identified work–life balance as a retention factor. Our results are consistent with the literature. A study in Ireland also showed that doctors were struggling to achieve work–life balance [46].

Flexible working hours help female nurses increase their job and life satisfaction, thus improving retention rates [47]. A study in a South African training hospital has shown that reducing working hours and creating a flexible work schedule is a way to retain doctors [48].

Lebanese healthcare workers’ need support from management

The exodus of doctors and nurses in Lebanon’s struggling healthcare system shows how managerial behavior shapes stay-or-go decisions [49]. Nurses whose managers coached resilience and maintained clear communication reported higher job satisfaction and lower intent to leave, according to evidence from the main COVID-19 referral hospital in the country [50]. Strong leadership soft skills are also linked to lower burnout across many health professions [51]. Structured leadership training and mentoring are among the most cost-effective approaches for reducing burnout and stabilizing the workforce in low- and middle-income countries, according to a 2025 rapid review synthesizing global evidence [52].

Lebanese healthcare workers’ need support from colleagues

Stronger social support from colleagues is linked to a modest but meaningful decrease in nurses’ turnover intention [53]. Hospitals that foster open communication and mutual support show significantly greater job satisfaction in the Eastern Mediterranean Region, highlighting the regional relevance of collegial bonds [54]. A systematic review identified encouragement from teammates among the key drivers that keep nurses in their roles during the COVID-19 pandemic [55]. An integrative review revealed that co-worker support buffers stress, increases motivation and increases the intention to stay (especially for newly graduated nurses) [56].

Importance of autonomy

In the literature, autonomy is researched mostly among nurses, where clinical autonomy is described as the capacity to go beyond standard procedures and make patient-specific care decisions [57]. A Norwegian study revealed that autonomy was an essential factor that positively impacted nurses’ intention to stay [58]. A study among Canadian nurses revealed that when they work in autonomy-supportive environments, they show more job satisfaction and are less inclined to leave [59].

The Lebanese healthcare workforce requires more recognition for their work

Low turnover rates among healthcare professionals are strongly linked to high degrees of job satisfaction, which often results from supportive work environments, fair compensation, and professional development opportunities [60]. Nurses who believe that their superiors value and support them are more likely to remain dedicated to their institution, lowering burnout and enhancing retention [61]. Retention also depends on recognition (formal and informal alike). Simple gestures such as employee awards, public praise or financial incentives can increase morale and strengthen the connection between healthcare professionals and their institutions [62].

Importance of mental health

Working on mental health is crucial for Lebanese doctors, nurses and medical students because research shows they face markedly elevated levels of anxiety, depression and burnout in the country’s turbulent social and economic climate. For instance, a meta-analytic study found that anxiety, depression, post-traumatic stress and insomnia were among the most common mental health issues among Lebanese healthcare workers [63]. Without timely psychological support and systemic interventions, these professionals are at increased risk of reduced quality of life [64] impaired academic [65] and clinical performance [66].

Recent evidence from conflict-affected settings in the Eastern Mediterranean showcased the importance of social relationships and peer support as resilience-related resources for healthcare professionals working under chronic threat. In a cross-sectional study of nurses affected by the October 7, 2023 war, social relationships emerged as a comparatively preserved quality-of-life domain despite substantial post-traumatic stress symptoms, and the authors noted social support as a protective factor while recommending structured peer-support programs within hospitals [67]. Similarly, qualitative accounts from Gaza during the ongoing war highlight community solidarity and peer support among healthcare workers as key coping strategies that help sustain endurance and continued caregiving under extreme adversity [68]. In addition, evidence from conflict-affected Northwest Syria indicates that mental health and psychosocial support responses have incorporated peer-to-peer support initiatives to assist distressed providers facing heavy workloads and insecurity [69]. Consistently, Lebanese nurses describing wartime care emphasize shared identity and mutual support as elements of collective resilience that may buffer psychological strain during repeated crises [70]. In line with the protective role of social and peer support identified in previous research, our results suggest that strong social networks may buffer against the detrimental psychological effects of conflict and crisis among healthcare workers.

Strengths and limitations

The large sample size (n = 819) and strong psychometric indicators—with internal consistency reliability ranging from α = 0.773—0.919 and ω = 0.777—0.920, as well as a Kaiser‒Meyer–Olkin (KMO) measure of sampling adequacy of 0.964 and a significant Bartlett’s test (p < 0.001)—showcase the robustness of our measurement approach. Conducted amid Lebanon’s severe socioeconomic and political collapse, this study provides timely real-world insights into the healthcare workforce retention challenge under socioeconomic and political crises, offering evidence that may inform policy and management decisions in similarly stressed health systems.

Despite its strengths, the study has important limitations. First, the use of snowball sampling and social media recruitment might not have yielded a probability sample and underrepresented older clinicians, those with limited internet access, and staff working in resource-constrained facilities. Consequently, selection bias may have been introduced, as those who are more digitally connected or engaged online were more likely to participate. This may affect the generalizability of the findings to the broader population of healthcare workers in Lebanon and should be considered when interpreting the results.

Second, a key limitation of this study is the reliance on self-reported data, as participants independently identified their professional category (e.g., doctor, nurse, or medical student), socioeconomic status, and other background characteristics without external verification. While this approach is common in online survey research, it introduces the possibility of misclassification, reporting bias, or inaccuracies that cannot be cross-checked. Consequently, the validity and robustness of the findings may be affected, as discrepancies between reported and actual characteristics could influence the interpretation of results.

Third, our sample composition was imbalanced: nurses are overrepresented, physicians are underrepresented, and most participants were from private hospitals, which may limit the generalizability of findings to the broader Lebanese healthcare workforce, particularly in public or rural settings.

Fourth, a formal content validity index (CVI) was not calculated prior to data collection, which may limit the assessment of item-level content relevance and expert agreement. Nevertheless, this was partially mitigated through the development of items based on an extensive literature review, consultation with experts in the field and academic staff, and pilot testing to refine clarity and contextual appropriateness. Future research is recommended to incorporate formal CVI procedures to further strengthen the content validity of the instrument. In addition, the psychometric robustness of the instrument was supported through exploratory factor analysis (EFA), including assessment of sampling adequacy using the Kaiser–Meyer–Olkin (KMO) measure and Bartlett’s test of sphericity, extraction of meaningful factors with appropriate item loadings, and evaluation of internal consistency using both Cronbach’s alpha and McDonald’s omega. These analyses demonstrated a coherent and reliable factor structure.

Fifth, although sampling adequacy (KMO) and Bartlett’s test supported factorability, the distributional properties of item responses were not formally assessed. Given the ordinal nature of the data, deviations from normality may have influenced the factor structure. Although Likert-scale items may deviate from normality, exploratory factor analysis using Pearson correlations has been shown to be robust under moderate violations, particularly with adequate sample sizes. As noted by Tabachnick and Fidell [71], such violations do not substantially affect the stability of factor solutions in adequately sized samples. Accordingly, the use of Pearson correlations was considered appropriate for the present study. Nonetheless, future studies may consider the use of polychoric correlations or robust estimation methods.

Finally, the study includes only healthcare workers and students who were residing in Lebanon at the time of data collection, thereby excluding individuals who had already emigrated. This was an intentional design choice, as the study aimed to capture the experiences and perspectives of those who remained in the country during ongoing socioeconomic and political crises, rather than examining factors that might attract emigrated professionals back. This may introduce bias, as the experiences, motivations, and perspectives of those who left could differ substantially from those who remained. As a result, the findings related to retention and migration intentions may not fully capture the broader dynamics of healthcare workforce mobility, and this limitation should be considered when interpreting the results.

Recommendations and implications

Clear, context-specific actions for practice should be emphasized, particularly those that are immediately feasible within Lebanon’s resource-constrained setting. Short-term, actionable interventions include implementing non-financial retention strategies such as flexible scheduling to improve work–life balance, strengthening managerial support through regular feedback and communication, and introducing low-cost mental health support initiatives. Additionally, hospitals can prioritize basic workplace improvements (such as ensuring sanitation, safety, and essential amenities) which require relatively limited financial investment. While broader structural reforms (e.g., salary adjustments and national policy changes) remain critical, distinguishing between immediate, feasible actions and long-term strategies can enhance the practical applicability of the study’s recommendations.

Conclusion

This study provides the first comprehensive factor analysis of healthcare workforce retention in Lebanon during socioeconomic and political crises, identifying 12 interrelated determinants: compensation and benefits, work‒life balance, sanitation and safety, disaster preparedness, and recognition emerged as central drivers, alongside political and economic stability.

These findings show that retention is shaped not only by financial incentives but also by organizational support, mental health resources, and broader macropolitical conditions. Strengthening these domains is essential to reduce emigration, stabilize the healthcare workforce, and safeguard Lebanon’s already fragile health system. Policymakers, hospital leaders, and educators should prioritize evidence-based reforms—such as competitive pay structures, supportive management, disaster readiness planning, and career development opportunities—to foster resilience and sustainability in the health sector.

It is recommended that future research build on these findings through longitudinal designs to track retention trends over time, intervention-based studies to evaluate the effectiveness of targeted policies, and qualitative follow-up research to capture healthcare workers’ lived experiences and evolving motivations in crisis contexts.

Abbreviations

EFA

Exploratory factor analysis

KMO

Kaiser–Meyer–Olkin

IRB

Institutional review board

ULS

Université La Sagesse

CPD

Continuing professional development

WHO

World Health Organization

UNDP

United Nations Development Programme

WASH

Water, Sanitation, and Hygiene

LAUMC-RH

Lebanese American University Medical Center Rizk Hospital

BMC

BioMed Central

COVID-19

Coronavirus disease 2019

SPSS

Statistical Package for the Social Sciences

CVI

Content validity index

FPC

Finite population correction

SD

Standard deviation

Author contributions

HB led the study conception, design, data analysis, data collection, literature review research, and manuscript writing. MB, JK, HF, AS, MAS, and FAM contributed to data collection, editing, and preliminary analysis. Dr. DEK supervised the project, provided methodological guidance, and critically reviewed the manuscript. All the authors read and approved the final manuscript.

Funding

The authors report that this research did not receive any specific grant from funding agencies in the public, commercial, or nonprofit sectors.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Ethical approval was granted by the La Sagesse University Ethics Committee for Research (ULS/IRB – 2024 – 6). The participants were provided with an explicit explanation of the objectives of the study before providing their consent to participate. Informed consent was obtained from all participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.