Depressive symptoms, loneliness and social support in healthcare employees: does the source of support matter?

Abstract

Purpose

Depressive symptoms are a risk factor for loneliness, while sources of perceived social support from family and friends are protective factors; however, the complex mechanisms behind these factors have not been examined among health-care employees in Portugal. The purpose of this study is to evaluate the indirect effect of perceived social support from family and friends on the association between depressive symptoms and loneliness.

Design/methodology/approach

Health-care employees (n = 279; 242 female and 37 male) from 12 health-care institutions in Portugal participated in this cross-sectional study. The SPSS program with PROCESS macro (Model 6) was used to test the hypothesis regarding the indirect effect analysis.

Findings

Bootstrap analysis found that family as a source of social support had an indirect effect in the relationship between depressive symptoms and loneliness. Similarly, depressive symptoms had a significant indirect effect on loneliness through the indirect role of social support from friends. Moreover, it was found that the relationship of family and friends as sources of social support mediated the association of the aforementioned variables.

Originality/value

The findings of this study underscore the important role of depressive symptoms on perceived social support from family and friends, leading to loneliness, which is a public health concern among health-care employees in Portugal.

Citation

Kulari, G. and Pereira de Castro, M. (2024), "Depressive symptoms, loneliness and social support in healthcare employees: does the source of support matter?", Journal of Public Mental Health, Vol. 23 No. 4, pp. 348-356. https://doi.org/10.1108/JPMH-06-2024-0066

Introduction

Increasingly, cross-sectional work shows that individuals with depressive symptoms are ten times more likely to feel lonely (Stubbs and Achat, 2022; Ho, 2024). One possible explanation is that individuals who experience depressive symptoms also report inability to maintain social networks, experience social disengagement and functional impairment which can lead to feelings of loneliness (Cacioppo and Cacioppo, 2018). This perceived loneliness loop is accompanied by feelings of hostility, pessimism and low self-esteem that can affect health-care employees’ professionalism and quality of care (Meese et al., 2024). A recent meta-analysis found a higher prevalence of depressive symptoms among physicians, nurses, technical assistants and other health-care employees than the general population (Al Maqbali et al., 2024). As in regard to Portugal, Duarte et al. (2022) reported that 28.7% among 1,535 health-care employees suffered substantial levels of depression, being thus classified the second country in Europe with the highest number of people using antidepressants. The effects of depressive symptoms among health-care employees can be so pervasive that it can lead to poor decisions and mistakes in the workplace, reduced empathy or even hostile attitudes towards patients and colleagues (Ferreira et al., 2021). Over time, these symptoms can manifest increased feelings of loneliness, leading to greater risk for physical health as well as suicidal ideation (Stubbs and Achat, 2022). A recent estimate suggests that one third of the population in industrialised countries experience loneliness to a problematic degree (Cacioppo and Cacioppo, 2018). Given the detrimental effects of depressive symptoms and loneliness among health-care employees, it is imperative to analyse the distribution and magnitude of these variables in this setting, thus laying groundwork for evidence-based prevention strategies for effective health-care institutions’ management.

Defined as the perception of quantity and quality of support that is potentially available from a person’s relationships and social contacts (Bone et al., 2023), perceived social support is classified in three subtypes: family, friends and significant other (Zimet et al., 1988). In the light of the precarious financial situation in Portugal related to an unstable labour market and low-paying contracts, the majority of health-care employees rely on their family, prolonging their co-residence with parents and late transition to conjugal union (Oliveira et al., 2014). Low perceived social support, in particular from co-workers, is associated with emotional exhaustion, poor psychological function and other physiological conditions (Kadar et al., 2020). The interpersonal theory of depression (Coyne, 1976) posits that individuals with depressive symptoms are more likely to see the social world as a void place, inducing negative effects in their relationships, which in turn elicit frustration and rejection from others. Hence, despite their access to considerable support from different sources, these individuals may present inappropriate or premature disclosure, expect negative feedback and social withdrawal, serving as vulnerable factors for loneliness (Stice et al., 2011). Hence, a growing number of studies have underlined the importance of establishing programs to provide social support as a core element among health-care employees (Beardon et al., 2020; Wigham et al., 2023). However, literature suggests that simply bringing family and friends together does not guarantee increased socialisation and better mental health (Al Maqbali et al., 2024; Ho, 2024). Rather, focusing on the individual’s condition can improve the perception of social support (Kulari, 2024). Our understanding of how depressive symptoms impact perceived social support among health-care employees and loneliness remains understudied and elusive in Portugal.

Current study

Often in research, family and friends’ relationships are not included in the list of variables that can relate to depressive symptoms and loneliness (Al Maqbali et al., 2024). Paradoxically, people spend much of their time at work, and consequently, friends’ relationships built in the workplace become central for the undergoing of their daily work. The association between depressive symptoms and loneliness has been thoroughly documented, and various studies have evidenced the causal influence of depression on loneliness over time (Erzen and Çikrikci, 2018; Ho, 2024). Although it is not a new topic, the scope of the effects of depressive symptoms is not yet fully known, with few effective strategies addressing loneliness in place, partially because of the lack of empirical studies testing its outcomes among a variety of health-care employees in Portugal. Thus, understanding whether and how this association between the aforementioned variables influences hospital staff accomplish their work goals while adhering to service fundamentals may help inform health-care administrators to support the sustainability of their workforce as demand increases. For this purpose, this study formulated the following hypothesis:

H1.

Depressive symptoms are positively associated with loneliness among health-care professionals in Portugal.

H2.

Perceived social support from family plays an indirect effect between depressive symptoms and loneliness among health-care professionals.

H3.

Perceived social support from friends plays an indirect effect between depressive symptoms and loneliness among health-care professionals.

H4.

Perceived social support from family and friends plays an indirect effect in the relationship between depressive symptoms and loneliness among health-care professionals through their indirect effect chain. The proposed model is shown in Figure 1.

Methodology

Participants and procedure

The present study followed a cross-sectional design including 279 health-care employees recruited by convenience sampling among 12 health-care institutions in Portugal. The study participants included a broad occupational group such as nurses, doctors, psychologists, health-care assistants, administration assistants and others who performed care and support activities in said institutions. The following inclusion criteria was considered: being in professional practice in the health-care setting at the time of data collection. The exclusion criteria included: professional practice absent for any reason at the time of data collection period.

After approval by the University’s Ethical Committee, this study approached 29 large health-care institutions in Portugal. To gain approval, an email was sent to the HR department of each institution. The heads of administration of 12 health-care institutions subsequently gave permission to distribute the link directly to their staff via email. All questionnaires were anonymous and proceeded online via MS Forms including the informed consent which presented participants with the purpose of this study. All the privacy rights of the participants were respected. Participants completed a 20-min self-administered questionnaire from March 2023 to January 2024. Majority of participants were nurses (n = 56; 20.1%) and operational assistant (n = 56; 20.1%), female (n = 247; 88.5%) pertaining to the group age of 40–50 years old (n = 95; 34.1%) with a permanent contract (n = 176; 63.15%). The detailed sociodemographic data are presented in Table 1.

Measures

Dependent variable.

UCLA Scale (University of California, Los Angeles, Version 3; Russell et al., 1980) was used to measure loneliness. The scale is composed of 20 items rated on a four-point Likert scale from 1 (never) to 4 (always). All items were summed to provide a loneliness score ranging from 20 to 80, whereby a higher score indicates greater loneliness. The instrument has a very good reliability coefficient of 0.94 and is the most worldwide used scale. The scale was translated and validated in Portuguese language by Pocinho et al. (2010). In our study, Cronbach’s alpha of the instrument was 0.95.

Independent variable.

Depressive symptoms were measured using Depression, Anxiety, Stress scale (DASS-21; Lovibond and Lovibond, 1995). DASS-21 has three subscales: depressive symptoms, anxiety and stress, composed of 21 items. Each subscale is composed of seven items rated on four-point Likert scale from 0 (did not apply to me at all) to 3 (applied to me very much or most of the time). In the present study, we are going to use only the depressive symptoms dimension which assesses lack of interests/involvements, devaluation of life, self-depreciation and hopelessness. Items of depressive symptoms subscale were summed to provide a total score ranging from 0 to 21. Higher scores indicated higher levels of depressive symptoms. The scale has a good reliability coefficient for depressive symptoms subscale 0.94 and, in this study, indicated a good Cronbach’s alpha of 0.92. The instrument was translated and validated in Portuguese by Pais-Ribeiro et al. (2004).

Mediator variable.

Perceived social support was measured through Multidimensional Scale of Perceived Social Support (MSPSS; Zimet et al., 1988). The MSPSS has three dimensions: perceived social support from family, friends and significant other. Each dimension is composed by four items rated on a Likert scale from 1 (very strongly disagree) to 7 (very strongly agree), ranging from 4 to 28. Higher scores for each dimension indicate higher level of perceived social support from friends, family and significant others. The scale was translated and validated in Portuguese by Carvalho et al. (2011). For the purpose of this study, perceived social support from friends and family are going to be considered as two mediators of the proposed model. The original scale has a good reliability coefficient for family 0.92 and friends 0.91. In this study, the reliability coefficient for both dimensions was 0.96.

Data analysis

First, correlation analysis was conducted to examine the association among loneliness, perceived social support from friends and family and depression symptoms. Then, independent indirect effect analysis was conducted to test whether social support from family and friends significantly impacted the relationship between loneliness and depressive symptoms. To determine indirect effect analysis, bootstrapping (5,000 samples) was performed using the Process model (V.3.4, SPSS; Hayes, 2022). Two-sided bias-corrected 95% CIs were constructed to evaluate the indirect effect. As recommended by the author Hayes (2022), these values are neither set nor standardised.

Results

Descriptive analysis

Correlations analysis for the sample is reported in Table 2. Loneliness was negatively correlated with social support from family (r = −0.497 and p < 0.01) and social support from friends (r = −0.610 and p < 0.01) but had a positive correlation with depressive symptoms (r = 0.638 and p < 0.01). Similarly, depressive symptoms showed a negative correlation with social support from family (r = −0.390 and p < 0.01) and social support from friends (r = −0.537 and p < 0.01). Finally, social support from family had a positive correlation with social support from friends (r = 0.677 and p < 0.01). Therefore, the significant correlation between research variables provides a good foundation for the subsequent research hypothesis and indirect effect analysis.

Indirect effect analysis

Model indices are depicted in Table 3. First, in the path of a1→b1, depressive symptoms had a significant negative effect on social support from family (β = −0.886 and p < 0.001), and social support from family had a significant positive negative on loneliness (β = −0.045 and p < 0.01). In the path of a2 → b2, depressive symptoms had a significant negative effect on social support from friends (β= −0.701 and p < 0.001), and social support from friends had a significant negative effect on loneliness (β = −0.100 and p < 0.001). While in the path of a1 → d → b2, social support from family had a significant positive effect on the social support from friends (β = 0.530 and p < 0.001). The present results supported our study hypothesis.

For the prediction of loneliness as shown in Table 4, depressive symptoms were statistically significant and a positive predictor (effect of c = 0.4848 and p < 0.001) in the total effect model without consideration of the mediators. The direct effect remained significant [effect of c′ = 0.3275 and p < 0.001; 95% CI 0.2531 and 0.4019)]. The total indirect effect was positive and significant [total indirect effect = 0.1574 and p < 0.001; 95% CI (0.1031, 0.2155)]. Correspondingly, all two possible indirect effect were significant [effect of a1 → b1 = 0.0398 and p < 0.001; 95% CI (0.0023, 0.0812); effect of a2→ b2 = 0.0704 and p < 0.001; 95% CI (0.0343, 0.1121); effect of a1 → d → b2 = 0.0471 and p < 0.001; 95% CI (0.0223, 0.0783)].

Discussion

Main findings

The current study extends research on the role of perceived social support in the association between depressive symptoms and loneliness. Our focus is on two specific sources of perceived social support, family and friends, to examine their indirect effect between the aforementioned variables. These issues were studied in an occupational diverse sample of health-care employees in Portugal. In line with previous studies (Stubbs and Achat, 2022; Wigham et al., 2023), this study found a strong positive relationship between depressive symptoms and loneliness. As reported by Fond et al. (2023), elevated levels of depressive symptoms in a health-care setting are highly correlated with workplace factors such as low wages, long work hours, heavy workload and lack of access to paid leave. These factors have worsened since the recent pandemic and were reflected in health-care employees’ increased feelings of loneliness (Duarte et al., 2022). In return, loneliness is manifested through low motivation, reduced energy and lack of self-worth (Meese et al., 2024). Considering the public health concern regarding depressive symptoms and loneliness, several health-care institutions in Portugal have implemented psychological intervention programs. However, because of the scarcity of empirical studies with this population, it is unclear whether these interventions are suited in this setting. Hence, this study raises awareness among health-care administrators, understanding the implications of these variables among health-care employees.

Moreover, results showed that depressive symptoms have a significant indirect effect on loneliness through the effect of perceived social support from family. In accordance with existing literature, lack of perceived family support has been linked to depressive symptoms and, therefore, loneliness (Bone et al., 2023). Furthermore, social support from family in Portugal often takes the form of financial assistance, particularly in regard to accommodation, with individuals remaining in the family home until later in life, leading to prolonged family ties (Oliveira et al., 2014). Hence, these financial conditions may exacerbate depressive symptoms and, in turn, leave health-care employees vulnerable to perceive social support from family as an obstacle for the attainment of a sense of autonomy and self-reliance.

Similarly, friends as a source of social support also played a significant indirect effect on loneliness. Health-care employees’ identity is socially constructed through interactions in the workplace (Mikkola et al., 2018). Several studies have reported relationships between social support from friends in the workplace and the quality of care, professional performance and job satisfaction in the health-care setting (Beardon et al., 2020; Orgambídez-Ramos and de Almeida, 2017). Moreover, Torres et al. (2024) evidenced that health-care employees can demonstrate stigmatising attitudes toward colleagues who suffer depression and preferably avoid them. As a result, health-care employees struggling with depressive symptoms may fear judgement, therefore distancing themselves from others, which exacerbates feelings of loneliness. To this aim, health-care institutions can help modify health-care employees’ attitude toward mental health through psychoeducation programs raising awareness on its symptomatology and promoting initiatives to seek help.

Finally, the relationship between perceived social support from family and friends had an indirect effect in the association of depressive symptoms and loneliness. The present study found that the two variables had a chain indirect effect in the process of depressive symptoms, affecting loneliness. It has been suggested that, while friends may provide voluntary support, family members may feel a greater sense of obligation to provide support (Kulari, 2024). Therefore, as evidenced in this study, depressive symptoms have a more pronounced effect on perceived social support from friends than from family. It is possible that the social withdrawal and passivity associated with depressive symptoms drives away friends’ support, while having less of an impact on family support, possibly because of the unconditional nature of family support, which is not evoked for friends.

Practical and theoretical implications

This study contributes to the general body of literature regarding depressive symptoms, loneliness and sources of social support among health-care employees. Hence, faced with the global increase of depressive symptoms and loneliness becoming a public health concern, it is critical to understand factors that can buffer its effects among vulnerable populations. From a practical perspective, the results of this study can be helpful to health-care administrators to implement empirically based interventions to assist their staff. First, interventions that effectively address depressive symptoms include mindfulness and meditation training, self-care and self-compassion, positive psychology group-based programs that incorporate elements of hope and gratitude. Organising seminars on related topics to raise awareness of mental health issues (signs, consequences, coping strategies and stigmatisation) to incentivise seeking help. Second, interventions aiming to improve colleagues’ social support may include seminars promoting communicative competences and social skills, training sessions to improve coping skills and formal and informal social networks fomenting activities among colleagues. Finally, policymakers should provide programs that facilitate health-care employees’ reception of support from family. These could take the form of flexible work schedules that permit meaningful quality time with family, as well as financial aid that removes some burden from economic reliance on family.

Limitation and future studies

The results of the present study should be interpreted with some considerations. Firstly, the cross-sectional design prevents drawing conclusions regarding causality between the variables as indicated in the model. However, the empirical evidence on the effect of the resources of social support on the relationship between depressive symptoms and loneliness allowed establishing this relationship between the variables of the study. Secondly, all the variables in the study were assessed using self-reports questionnaires, which increases the risk of biases in the responses. Hence, further longitudinal studies should be conducted to establish directionality in the relationships proposed in this model. Furthermore, family and friends were the only mediators of this study. Thus, other variables such as sex, professional category, social isolation and emotional resilience should be examined in future studies as mediators. Fourthly, this study consisted of a variety of health-care employees of different institutions, which cannot be generalised to all health-care employees. Subsequent research could include a wider number of participants for this understudied population and analyse differences among public and private health-care institutions.

Figures

Figure 1

Hypothesised model

Table 1

Sociodemographic factors (n = 279)

Respondent characteristics	N	%
Sex
Male	32	11.5
Female	247	88.5
Age
18–28 years	46	16.5
29–39 years	74	26.5
40–50 years	95	34.1
51–61 years	58	20.8
62–66 years	6	2.2
Employment contract
Self-employed	49	17.6
Fixed term contract	30	10.8
Permanent contract	176	63.1
Other	24	8.6
Profession
Doctor	13	4.7
Nurse	56	20.1
Physiotherapist	17	6.1
Psychologist	47	16.84
Oral hygienist	47	16.84
Operational assistant	56	20.1
Technical assistant	9	3.2
Other	34	12.18
Institution
Public hospital	157	56.27
Private medical clinic	62	22.22
National health system local Centre	42	15.05
Private hospital	18	6.4

Source: Authors’ own work

Table 2

Means, standard deviation and correlations between study variable (n = 279)

Variables	M	SD	1	2	3
Loneliness	32.80	10.33	–	–	–
Depressive symptoms	4.55	4.757	0.638**	–	–
Social support from family	20.94	6.175	−0.497**	−0.390**	–
Social support from friends	20.09	5.932	−0.610**	−0.537**	0.677**

Note:

**p < 0.01

Source: Authors’ own work

Table 3

Results of regression analysis (n = 279)

Model	Outcome	Predictors	β	SE	t	LLCI	ULCI
Model 1	SS family	DS	−0.886	0.1257	−49.3017***	−1.133	−0.6385
			R2 = 0.153, F(1,227) = 49.6837***
Model 2	SS friends	DS	−0.701	0.0961	−7.3062***	−0.8910	−0.5127
		SS family	0.530	0.0423	12.5360***	0.4469	0.6133
			R2 = 0.546, F(2,276) = 166.2730***
Model 3	Loneliness	DS	0.327	0.0378	8.6671***	0.2531	0.4019
		SS family	−0.045	0.0191	−2.3534**	−0.0824	−0.0073
		SS friends	−0.100	0.0217	−4.6313***	−0.1430	−0.0577
			R2 = 0.517, F(3,275) = 98.2324***

Notes:

SS = social support; DS = depressive symptoms; ***p < 0.001; **p < 0.01

Source: Authors’ own work

Table 4

Total, direct and indirect effects of depressive symptoms on loneliness

Model	Effect	SE	LLCI	ULCI
Total effect (c)	0.4848	0.0352	0.4156	0.5541
Direct effect (c’)	0.3275	0.0378	0.2531	0.4019
Total indirect effect	0.1574	0.0292	0.1031	0.2155
DS → SS Family → Loneliness (a1 → b1)	0.0398	0.0196	0.0023	0.0812
DS → SS Friends → Loneliness (a2 → b2)	0.0704	0.0199	0.0343	0.1121
DS → SS Family → SS Friends → Loneliness (a1 → d → b2)	0.0471	0.0143	0.0223	0.0783

Notes:

SS = social support; DS = depressive symptoms

Source: Authors’ own work