The mediating effect of sleep quality on the relationship between social support and depressive symptoms among Chinese nurses during the omicron outbreak Authors: Yingying Gu, Pinglang Hu, Huanzeng Ren, Caijun Dai, Xuejiao He, Weizhen Cheng, Lihua Yu, Achang Fang, Xiaoling Meng, Meiyang Lou, Youying Chen, Danli Chi, Huasu Zhou, Qiaoge Chen, Shuhong Ni, Qiqi Huang Published: 2024 ABSTRACT BackgroundThe COVID-19 pandemic has led to increased depressive symptoms and diminished sleep quality among nurses. This study explores the relationships among social support, sleep quality, and depressive symptoms among Chinese nurses during the COVID-19 outbreak, with a focus on the mediating role of sleep quality.MethodsA cross-sectional study involving 2140 nurses was conducted from August to September 2022. The participants completed the Social Support Rating Scale, the Hospital Anxiety and Depression Scale, and the Pittsburgh Sleep Quality Index. The data were analyzed via serial multiple mediation analysis with the PROCESS macro in SPSS.ResultsDepressive symptoms affected an estimated 40.4% of the participants, with 69.2% reporting poor sleep quality. The present study revealed a significant negative correlation between social support and sleep quality and depressive symptoms (r = -0.229, p < 0.001; r = -0.322, p < 0.001, respectively). Furthermore, a substantial positive correlation was observed between sleep quality and depressive symptoms (r = 0.514, p < 0.001). Additionally, social support had a significant indirect effect on depressive symptoms through sleep quality (standardized effect = -0.0535, 95% CI = [-0.0648, -0.0424]). Sleep quality was found to mediate the relationship between social support and depressive symptoms in all subgroups, with variations based on years of experience, education levels, and hospital type.ConclusionDepressive symptoms were inversely associated with social support, which was moderated by sleep quality. Interventions targeting social support provision and sleep quality promotion are suggested for managing depressive symptoms among nurses during the COVID-19 pandemic. Introduction The emergence of the COVID-19 pandemic in December 2019 escalated into a global health crisis. By January 17th 2021, the pandemic had resulted in over 93 million confirmed cases and the loss of more than 2 million lives worldwide [1]. It has placed immense strain on healthcare systems and the mental health of healthcare workers, who have faced increased workloads and extensive exposure to infected individuals. In November 2021, the WHO reported a significant increase in COVID-19 cases due to the Omicron variant (B.1.1.529) [2]. This variant, while generally causing milder symptoms, has shown a high degree of transmissibility, presenting new challenges for pandemic mitigation strategies [3]. While the pandemic’s psychological repercussions have been widely acknowledged [4–7], this impact is particularly pronounced for healthcare professionals, especially nurses, who bear the brunt of the pandemic’s emotional and psychological toll [8, 9]. Research has reported a wide range of depressive prevalence rates among nurses during the COVID-19 outbreak, from 9 to 89%, with a pooled point prevalence of 35% [10–12]. This variability underscores the complexity of the issue and the need for a more nuanced understanding of the factors contributing to mental health outcomes in this high-risk group. The role of social support in mitigating depressive symptoms during crises, such as the COVID-19 pandemic, is well established [13, 14]. However, the mediating effect of sleep quality on this relationship remains underexplored. Studies have shown that increased social support is correlated with reduced psychological distress and serves as a buffer against stress [15, 16]. A significant negative association between social support scores and depressive symptoms has been confirmed [17]. Social support has also been highlighted as a promoter of better sleep quality [18]. Among healthcare professionals, higher levels of social support have been linked to improved self-efficacy and sleep quality during the COVID-19 epidemic [18, 19]. However, the mechanisms by which social support influences mental health, particularly through sleep quality, are not well understood. Sleep disturbance has emerged as a significant mental health concern during the COVID-19 pandemic, with sleep disturbance prevalence rates ranging from 12 to 87% and a pooled prevalence estimate of 43% [12, 20–22]. Sleep quality is crucial for assessing health status, especially for healthcare professionals, as it is essential for their efficiency and critical in maintaining optimal immune function to prevent infections [23]. Depressive symptoms have shown the strongest correlation with poor sleep quality during the pandemic [24, 25]. Depression is associated with hyperactivation of the hypothalamic‒pituitary‒adrenal (HPA) axis, increased sympathetic nervous system reactivity, and elevated levels of norepinephrine and cortisol, leading to a heightened state of arousal and impaired sleep quality [26]. Sleep disturbances can both precede and trigger depressive episodes, exacerbating the onset and progression of depressive symptoms [27]. Despite the recognized correlation, the bidirectional relationship between sleep disturbances and depressive symptoms has not been fully elucidated, particularly in the context of the COVID-19 pandemic. Drawing on Thoits’s stress-buffering theory [28], we hypothesize that social support may mitigate the impact of COVID-19 related stress on nurses by improving their sleep quality, thereby reducing the risk of depressive symptoms. Empirical evidence supports this hypothesis, as studies have shown a significant negative association between the score of social support and depressive symptoms [29] and improved self-efficacy and sleep quality among healthcare professionals with greater social support during the COVID-19 epidemic [24, 25]. The interplay between sleep quality, social support, and depressive symptoms has been increasingly scrutinized during the COVID-19 pandemic [30, 31]. However, the mediating effect of sleep quality on the relationship between social support and depressive symptoms has not been investigated during the COVID-19 pandemic. The objective of this study was to fill this research gap and contribute to the evidence base for the effective prevention and management of depressive symptoms in nurses during crisis events. Consequently, we conducted an analysis focusing on the relationships among sleep quality, social support, and depressive symptoms among nurses. We hypothesized that sleep quality may have mediated the relationship between social support and the manifestation of depressive symptoms among nurses during the COVID-19 pandemic. Materials and methods Study design and participants A cross-sectional survey was conducted online via an online platform (www.wjx.com) from August to September 2022, recruiting a total of 2140 nurses from 11 different hospitals in Jinhua. Participants were included if they were 18 years or older and had the full capacity for independent questionnaire comprehension and completion. The exclusion criteria included refusal to participate in the survey. The study received ethical approval from the Research Ethics Committee of Jinhua Municipal Central Hospital (No. 20222220101), and all participants provided informed consent before participating in the online survey. Estimation of sample size The sample size for our study was determined via a Monte Carlo power analysis simulation, which is a recommended approach for complex mediation models, as outlined in the seminal work by. Schoemann et al. [32]. This method allows for a more accurate estimation of power and sample size requirements by simulating the sampling distribution of the indirect effect under the assumption that the alternative hypothesis is true. To determine the optimal sample size for achieving a desired power level of 0.95, we adopted the varying sample size approach within our Monte Carlo simulation. The simulation was meticulously designed with the following parameters: a total of 5,000 replications to bolster the reliability of our findings; 20,000 Monte Carlo draws per replication to ensure a precise approximation of the sampling distribution; and a confidence interval width set at the widely accepted 95% level. After running the simulation, we determined that a sample size of N = 780 was necessary to achieve adequate power for our study. Variables and measures The research variables were evaluated via validated questionnaires, which included demographic information collected through a custom demographic information sheet. The Pittsburgh Sleep Quality Index (PSQI) was used to measure sleep quality measured [33]. The Social Support Rating Scale (SSRS) was used to evaluate social support [34], whereas the Hospital Anxiety and Depression Scale (HADS) was used to evaluate depressive symptoms [35]. Demographic characteristics The demographic characteristics of the participants included age, sex, marital status, education, years of nursing experience, different types of hospitals, night shifts and clinical achievement on the ladder. Pittsburgh Sleep Quality Index Sleep quality was measured by the PSQI [33, 36]. It comprises 19 items categorized into seven sections that evaluate the sleep quality of participants over the past month. The classifications included an individual’s subjective assessment of sleep quality, the time required to start sleep, the duration of sleep, the efficacy of sleep, disturbances experienced during sleep, the use of sleeping aids and the impact on daytime performance. The sum of scores from these subscales yields a global score ranging from 0 to 21; higher scores indicate poorer sleep quality. A global score of 5 or less is considered indicative of ‘good quality sleep’, whereas a score greater than 5 indicates ‘poor quality sleep’. For further interpretation of the scores: 0–5 points indicate good sleep, 6–10 points suggest regular sleep issues, 11–15 points reflect poor sleep, and 16–21 points denote very poor sleep quality. The PSQI has demonstrated strong reliability and validity in various populations and has been verified to have a reliability of 0.85 and a validity of 0.83, with a Cronbach’s α coefficient of 0.83 [33]. Social Support Rating Scale We employed a modified Chinese version of the SSRS to assess social support [34, 37]. The SSRS comprises ten items categorized into three dimensions: utilization of social support, objective support, and subjective support. The total score ranges from 12 to 66, with higher scores indicating increased levels of social support. Scores below 22 suggest low levels of social support, while scores between 23 and 44 indicate medium levels of social support, and scores between 45 and 66 denote high levels of social support. The reliability analysis yielded a Cronbach’s α coefficient of 0.81 for the full scale, 0.903 for subjective support, 0.878 for objective support, and 0.885 for the utilization of support. These coefficients indicate high internal consistency and reliability across the scale’s dimensions. The construct validity was confirmed through factor analysis, with all the items demonstrating clear loadings onto their respective factors [37]. The Hospital Anxiety and Depression Scale The HADS was used to assess emotional disorders in nonpsychiatric patients within a hospital setting [35, 38]. The HADS consists of 14 elements, with seven elements indicating depressive symptoms and the remaining seven elements indicating anxiety, each scored between 0 and 3. This yields scale values ranging from 0 to 21 for each domain. A higher cumulative score indicated a more pronounced manifestation of symptoms associated with depression. The precision of identifying major depression was determined to be 82% for sensitivity and 72% for specificity when a threshold score of > 7 out of a maximum total score of 21 was used [39]. Consequently, nurses’ HADS depression scores were categorized into two groups: > 7 (indicating depression) or < = 7 (indicating no depression). The HADS exhibits strong internal consistency, with Cronbach’s α coefficients of 0.879 for the total scale and 0.806 for both the anxiety and depression subscales [40]. These psychometric properties establish the HADS as a reliable and stable instrument for assessing anxiety and depression in medical settings. Data analysis Sociodemographic characteristics were initially subjected to descriptive analyses, with continuous variables presented as the means and standard deviations and categorical variables presented as frequencies and percentages. Pearson’s correlation was used to investigate the relationship between each pair of continuous variables. Furthermore, we use SPSS PROCESS macro version 3.5 (Model 4) to investigate whether there is an indirect effect of social support on depressive symptoms through sleep quality. Multiple mediation analyses were performed by bootstrapping (5,000 bootstrap samples) with 95% confidence intervals (CIs). An effect was considered statistically significant at the 0.05 level if the 95% bias-corrected bootstrap CI of the mediation effect did not include zero. Furthermore, we conducted a subgroup analysis on the basis of years of nursing experience, education level and type of hospital. All the statistical analyses were performed with SPSS 26.0, with two-tailed tests set at a significance level of p < 0.05. Results Descriptive analyses A questionnaire survey was conducted, with a total of 2140 nurses participating, as presented in Table 1. Most of the study population was female (96.8%), with an average age of 30.15 years (standard deviation (SD) of 7.70), and a significant proportion (85.7%) fell within the age range of 18–39 years. Among the participants, a substantial percentage (62.7%) possessed at least a university degree or higher qualification, whereas nearly half (49.4%) reported being married. Furthermore, 60.9% had less than 10 years of professional experience in nursing practice. More than half of the nurses expressed concerns about contracting infections during their duties and approximately one-fourth reported high exposure to cases of COVID-19 in their workplace settings. The distribution of nurses across hospital types was as follows: general hospitals account for 75.7% of the sample, while specialized hospitals comprise 26.3%. In terms of social support, 80.5% indicated moderate levels, whereas only 3.5% reported low levels of social support. Table 1Demographic characteristics of the study subjects (n = 2140)Variablesn%GenderFemalemale20726896.83.2Age(year) (M, SD)< 2930–3940–49> 5030.151172667245567.7054.831.211.42.6Marital status, n (%)MarriedSingleOther105810443849.448.81.8Education level, n (%)College degree or lowerUndergraduate degreeMaster or higher7981333937.362.30.4Professional title, n (%)PrimaryIntermediateSenior102986724448.140.511.4Exposure level, n (%)HighLow555158525.974.1Years of work experience (year)< 10≥ 10127486659.540.5Fear of infection, n (%)YesNo123290857.442.6Nursing workloadReducedNormalIncreased4618119132.18.589.4Hospital typesGeneral hospitals161975.7Specialty hospital52124.3Night shiftYes166477.8No47622.2Sleep quality (M, SD)GoodRegularPoor / very poor7.59659 (30.8)1049 (49.0)432 (20.2)3.5230.849.020.2Depressive symptom (M, SD)> 7≤ 76.8486612744.0140.459.6Social support (M, SD)≤ 2223–4445–6636.137417233438.143.580.516.0Primary: “Nurse” and “Junior Nurse Practitioner”; Intermediate: “Charge Nurse”; Senior: “Deputy Chief Nurse Practitioner” and “Chief Nurse Practitioner” Table 1Demographic characteristics of the study subjects (n = 2140)Variablesn%GenderFemalemale20726896.83.2Age(year) (M, SD)< 2930–3940–49> 5030.151172667245567.7054.831.211.42.6Marital status, n (%)MarriedSingleOther105810443849.448.81.8Education level, n (%)College degree or lowerUndergraduate degreeMaster or higher7981333937.362.30.4Professional title, n (%)PrimaryIntermediateSenior102986724448.140.511.4Exposure level, n (%)HighLow555158525.974.1Years of work experience (year)< 10≥ 10127486659.540.5Fear of infection, n (%)YesNo123290857.442.6Nursing workloadReducedNormalIncreased4618119132.18.589.4Hospital typesGeneral hospitals161975.7Specialty hospital52124.3Night shiftYes166477.8No47622.2Sleep quality (M, SD)GoodRegularPoor / very poor7.59659 (30.8)1049 (49.0)432 (20.2)3.5230.849.020.2Depressive symptom (M, SD)> 7≤ 76.8486612744.0140.459.6Social support (M, SD)≤ 2223–4445–6636.137417233438.143.580.516.0Primary: “Nurse” and “Junior Nurse Practitioner”; Intermediate: “Charge Nurse”; Senior: “Deputy Chief Nurse Practitioner” and “Chief Nurse Practitioner” Demographic characteristics of the study subjects (n = 2140) Gender Female male 2072 68 96.8 3.2 Age(year) (M, SD) < 29 30–39 40–49 > 50 30.15 1172 667 245 56 7.70 54.8 31.2 11.4 2.6 Marital status, n (%) Married Single Other 1058 1044 38 49.4 48.8 1.8 Education level, n (%) College degree or lower Undergraduate degree Master or higher 798 1333 9 37.3 62.3 0.4 Professional title, n (%) Primary Intermediate Senior 1029 867 244 48.1 40.5 11.4 Exposure level, n (%) High Low 555 1585 25.9 74.1 Years of work experience (year) < 10 ≥ 10 1274 866 59.5 40.5 Fear of infection, n (%) Yes No 1232 908 57.4 42.6 Nursing workload Reduced Normal Increased 46 181 1913 2.1 8.5 89.4 Hospital types General hospitals Sleep quality (M, SD) Good Regular Poor / very poor 7.59 659 (30.8) 1049 (49.0) 432 (20.2) 3.52 30.8 49.0 20.2 Depressive symptom (M, SD) > 7 ≤ 7 6.84 866 1274 4.01 40.4 59.6 Social support (M, SD) ≤ 22 23–44 45–66 36.13 74 1723 343 8.14 3.5 80.5 16.0 Primary: “Nurse” and “Junior Nurse Practitioner”; Intermediate: “Charge Nurse”; Senior: “Deputy Chief Nurse Practitioner” and “Chief Nurse Practitioner” Prevalence of depressive symptoms and sleep quality The mean depressive symptom score, as derived from the HADS, was 6.84 ± 4.01, as shown in Table 1. Approximately 40% of the nurses met the criteria for depressive symptoms on the basis of their scores on the HADS scale scores. A substantial proportion of the nursing population, 69.2% (n = 1481), reported poor sleep quality as indicated by their Pittsburgh Sleep Quality Index (PSQI) scores (> 5). The average sleep quality score was recorded as 7.59 ± 3.52. In terms of the nurses’ views on sleep quality (Table 2), a total of 808 nurses (37.8%) expressed dissatisfaction with their sleep quality. For approximately 60% of the nurses, it took more than half an hour to fall asleep and a significant majority (92.1%) reported sleeping less than or equal to seven hours per night. A total of 839 (39.2%) nurses reported habitual sleep efficiency less than 85%, while the vast majority (91%) experienced subjective disturbance during the initial or maintenance phases of sleep. Furthermore, approximately 12.2% of nurses used medication aids for better sleep, whereas most people in this profession (82.5%) acknowledged experiencing daytime dysfunction. Table 2Participants in perspectives on sleep quality (n = 2140)Itemsn%Subjective sleep quality Satisfied133262.2 Unsatisfied80837.8Sleep latency 0–30 min88141.2 31–60 min75535.3 >60 min50423.6Sleep duration > 7 h1707.9 ≤ 7 h197092.1Habitual sleep efficiency 85–100%130160.8 75–84%70532.9 0–74%1346.3Subjective sleep disturbances No1929.0 Yes194891.0Use of sleep medication Use26112.2 Not use187987.8Daytime functioning Yes176582.5 No37517.5Total scores of sleep disturbances > 5148169.2 ≤ 565930.8 Table 2Participants in perspectives on sleep quality (n = 2140)Itemsn%Subjective sleep quality Satisfied133262.2 Unsatisfied80837.8Sleep latency 0–30 min88141.2 31–60 min75535.3 >60 min50423.6Sleep duration > 7 h1707.9 ≤ 7 h197092.1Habitual sleep efficiency 85–100%130160.8 75–84%70532.9 0–74%1346.3Subjective sleep disturbances No1929.0 Yes194891.0Use of sleep medication Use26112.2 Not use187987.8Daytime functioning Yes176582.5 No37517.5Total scores of sleep disturbances > 5148169.2 ≤ 565930.8 Participants in perspectives on sleep quality (n = 2140) Correlation analyses The results of the correlation analysis revealed that social support was significantly negatively correlated with depressive symptoms and sleep quality (r = −0.322, p < 0.001; r = −0.229, p < 0.001), respectively). Furthermore, a substantial positive correlation was observed between depressive symptoms and sleep quality (r = 0.514, p < 0.001) (Table 3). Table 3Correlations between key study variablesVariablesSSRSHADSPSQISSRS1HADS−0.322***1PSQI−0.229***0.514***1***P < 0.001SSRS Social Support Rating Scale, PSQI Pittsburgh Sleep Quality Index, HADS Hospital Anxiety and Depression Scale Table 3Correlations between key study variablesVariablesSSRSHADSPSQISSRS1HADS−0.322***1PSQI−0.229***0.514***1***P < 0.001SSRS Social Support Rating Scale, PSQI Pittsburgh Sleep Quality Index, HADS Hospital Anxiety and Depression Scale Correlations between key study variables ***P < 0.001 SSRS Social Support Rating Scale, PSQI Pittsburgh Sleep Quality Index, HADS Hospital Anxiety and Depression Scale Mediation model Table 4; Fig. 1 present the results of the mediation analysis. Mediation analysis indicated that social support was negatively related to sleep quality (standardized effect = −0.0991, 95% CI = [−0.1169, −0.0812]) and that sleep quality positively mediated depressive symptoms (standardized effect = 0.5397, 95% CI = [0.4975, 0.5818]). Social support had a significant indirect effect on depressive symptoms through sleep quality (standardized effect = −0.0535, 95% CI = [−0.0648, −0.0424]). Furthermore, the analysis revealed statistically significant total and direct effects of social support on depressive symptoms (standardized effect = − 0.1620, 95% CI [−0.1821, −0.1418]; standardized effect = − 0.1085, 95% CI [−0.1267, −0.0903], respectively).Fig. 1Mediation model. ***p < 0.001Table 4Results of the mediation modelingPathStandardized EffectSELLCIULCITotal effect−0.16200.0103−0.1821−0.1418Direct effect−0.10850.0093−0.1267−0.0903Total indirect effects−0.05350.0058−0.0653−0.0421SE standard error, LLCI and ULCI lower level and upper level of the 95% bias-corrected bootstrap confidence interval Mediation model. ***p < 0.001 Table 4Results of the mediation modelingPathStandardized EffectSELLCIULCITotal effect−0.16200.0103−0.1821−0.1418Direct effect−0.10850.0093−0.1267−0.0903Total indirect effects−0.05350.0058−0.0653−0.0421SE standard error, LLCI and ULCI lower level and upper level of the 95% bias-corrected bootstrap confidence interval Results of the mediation modeling SE standard error, LLCI and ULCI lower level and upper level of the 95% bias-corrected bootstrap confidence interval Subgroup analysis results of mediation modeling We conducted a subgroup analysis on the basis of years of nursing experience, education level, night shift status and hospital types (Table 5). Owing to the sample size considerations, we have grouped individuals with an undergraduate degree and those with a master’s degree or higher into a single category, referred to as an “undergraduate degree or higher”. The subgroup analysis revealed that sleep quality significantly mediates the relationship between social support and depressive symptoms across various subgroups. The indirect effects of social support on depressive symptoms through sleep quality were consistent across subgroups defined by years of nursing experience, education level, night shift status, and hospital types. Table 5Subgroup analysis results of the mediation modelingSubgroupDirect effect (SE)95%confidence intervalIndirect effect (SE)95%confidence intervalTotal effect(SE)95%confidenceintervalYears of work experience < 10−0.1108(0.0129)−0.1362,−0.0855−0.0712(0.0083)−0.0878,−0.0557−0.1820(0.0142)−0.2098,−0.1542 ≥ 10−0.1602(0.0151)−0.1898,−0.1306−0.0540(0.0082)−0.0710,−0.0390−0.2142(0.0160)−0.2457,−0.1828Hospital types General hospitals−0.1168(0.0106)−0.1376,−0.0961−0.0461(0.0061)−0.0580,−0.0342−0.1629(0.0116)−0.1856,−0.1403 Specialty hospital−0.0750(0.0193)−0.1129,−0.0372−0.0798(0.0139)−0.1068,−0.0531−0.1548(0.0224)−0.1989,−0.1108Education level College degree or. lower−0.1142(0.0166)−0.1468,−0.0816−0.0689(0.0099)−0.0891,−0.0501−0.1830(0.0181)−0.2185,−0.1476 Undergraduate. degree or higher−0.1175(0.0116)−0.1403,−0.0947−0.0551(0.0070)−0.0689,−0.0412−0.1726(0.0127)−0.1976,−0.1477Night shift Yes−0.1107(0.0108)−0.1319,−0.0895−0.0605(0.0068)−0.0743,−0.0477−0.1712(0.0119)−0.1946,−0.1478 No−0.0908,(0.0190)−0.1281,−0.535−0.2080(0.0102)−0.0490,−0.0086−0.1189(0.0212)−0.1605,−0.0772 Table 5Subgroup analysis results of the mediation modelingSubgroupDirect effect (SE)95%confidence intervalIndirect effect (SE)95%confidence intervalTotal effect(SE)95%confidenceintervalYears of work experience < 10−0.1108(0.0129)−0.1362,−0.0855−0.0712(0.0083)−0.0878,−0.0557−0.1820(0.0142)−0.2098,−0.1542 ≥ 10−0.1602(0.0151)−0.1898,−0.1306−0.0540(0.0082)−0.0710,−0.0390−0.2142(0.0160)−0.2457,−0.1828Hospital types General hospitals−0.1168(0.0106)−0.1376,−0.0961−0.0461(0.0061)−0.0580,−0.0342−0.1629(0.0116)−0.1856,−0.1403 Specialty hospital−0.0750(0.0193)−0.1129,−0.0372−0.0798(0.0139)−0.1068,−0.0531−0.1548(0.0224)−0.1989,−0.1108Education level College degree or. lower−0.1142(0.0166)−0.1468,−0.0816−0.0689(0.0099)−0.0891,−0.0501−0.1830(0.0181)−0.2185,−0.1476 Undergraduate. degree or higher−0.1175(0.0116)−0.1403,−0.0947−0.0551(0.0070)−0.0689,−0.0412−0.1726(0.0127)−0.1976,−0.1477Night shift Yes−0.1107(0.0108)−0.1319,−0.0895−0.0605(0.0068)−0.0743,−0.0477−0.1712(0.0119)−0.1946,−0.1478 No−0.0908,(0.0190)−0.1281,−0.535−0.2080(0.0102)−0.0490,−0.0086−0.1189(0.0212)−0.1605,−0.0772 Subgroup analysis results of the mediation modeling −0.1108 (0.0129) −0.1362, −0.0855 −0.0712 (0.0083) −0.0878, −0.0557 −0.1820 (0.0142) −0.2098, −0.1542 −0.1602 (0.0151) −0.1898, −0.1306 −0.0540 (0.0082) −0.0710, −0.0390 −0.2142 (0.0160) −0.2457, −0.1828 −0.1168 (0.0106) −0.1376, −0.0961 −0.0461 (0.0061) −0.0580, −0.0342 −0.1629 (0.0116) −0.1856, −0.1403 −0.0750 (0.0193) −0.1129, −0.0372 −0.0798 (0.0139) −0.1068, −0.0531 −0.1548 (0.0224) −0.1989, −0.1108 −0.1142 (0.0166) −0.1468, −0.0816 −0.0689 (0.0099) −0.0891, −0.0501 −0.1830 (0.0181) −0.2185, −0.1476 −0.1175 (0.0116) −0.1403, −0.0947 −0.0551 (0.0070) −0.0689, −0.0412 −0.1726 (0.0127) −0.1976, −0.1477 −0.1107 (0.0108) −0.1319, −0.0895 −0.0605 (0.0068) −0.0743, −0.0477 −0.1712 (0.0119) −0.1946, −0.1478 −0.0908, (0.0190) −0.1281, −0.535 −0.2080 (0.0102) −0.0490, −0.0086 −0.1189 (0.0212) −0.1605, −0.0772 Discussion The present study provides empirical evidence supporting the mediating role of sleep quality in the association between social support and depressive symptoms and provides a nuanced understanding of the mental health dynamics among Chinese nurses during the Omicron outbreak. This relationship is not confined to the pandemic context; rather, it extends to various populations and periods, highlighting the universal significance of sleep quality for emotional well-being. Previous studies have suggested that sleep quality is a crucial factor in the link between social support and depressive symptoms during the Omicron outbreak [41–43]. Our results corroborate this finding, showing that social support is negatively associated with sleep quality, which in turn is positively associated with depressive symptoms. This chain of influence is consistent with the stress-buffering model proposed by Thoits [28], where social support acts as a buffer against the adverse effects of stress, including sleep disturbances and depressive symptoms. Beyond the pandemic, sleep quality has been established as a critical mediator in the relationship between social support and mental health. For example, a study among early pregnant women [44] revealed that better sleep quality mediated the positive impact of social support on reducing depressive symptoms. This is particularly relevant as pregnancy is a period of significant hormonal and emotional changes, which can be exacerbated by sleep disturbances. Similarly, a study among people living with HIV [45] showed that sleep quality mediated the relationship between social support and depressive symptoms, indicating that the link between sleep quality and depressive symptoms is not limited to the general population or those facing acute stressors such as a pandemic. The subgroup analysis results highlight the significant mediating role of sleep quality in the relationship between social support and depressive symptoms among nurses with varying years of work experience. Both nurses with less than 10 years of experience and those with 10 or more years of experience demonstrated significant mediating effects of sleep quality, with the total effect being more pronounced among those with more experience. This finding may be attributed to several factors. Nurses with more experience are likely to have been exposed to more high-stress situations, potentially leading to higher levels of depressive symptoms. However, they also have greater opportunities to build and rely on social support networks, which can act as a buffer against these stressors. The mediating effect of sleep quality suggests that the quality of rest and recovery during off-duty hours is crucial for mental health, particularly for those who have been in the profession longer and may be more adept at managing stress through social support and other coping mechanisms. This aligns with previous research showing a positive correlation between perceived social support and psychological resilience, as well as a negative association between perceived social support and sleep quality [46]. This finding indicates that individuals perceiving greater social support generally exhibit higher levels of resilience and better sleep quality, which can mitigate the impact of stress and reduce depressive symptoms [47]. A study among Chinese nurses revealed that resilience was positively associated with perceived social support, supporting the notion that social support can enhance resilience and improve sleep quality, thereby reducing depressive symptoms [48]. The social support theory was developed by House [49] can elucidate several mechanisms through which sleep quality mediates the relationship between social support and depressive symptoms. According to this theory, social support is multifaceted and includes emotional, instrumental, informational, and appraisal support. Each of these dimensions can contribute to an individual’s sense of belonging, self-esteem, and perceived control over their environment, all of which can influence mental health outcomes. First, individuals with higher levels of social support often report greater satisfaction with their social relationships [50], leading to a deeper sense of meaning in life, which is inversely associated with depressive symptoms. Second, the emotional and instrumental support from social networks can directly impact sleep quality by providing a sense of security and reducing stressors that might otherwise disrupt sleep [51]. Third, informational and appraisal support can help individuals develop effective coping strategies for stress, which can improve sleep and reduce depressive symptoms [50, 52]. The theory also aligns with the allostatic load theory [53, 54] which posits that repeated adaptation or maintaining stability through change can lead to the ability of stress response systems to properly downregulate and accumulate wear and tear on the body [54], potentially triggering depressive symptoms. High levels of social support can buffer against these effects by enhancing the ability of stress response systems to downregulate stress, thus protecting mental health [54]. Furthermore, poor sleep quality increases the vulnerability to allostatic loads, making people more susceptible to stress impacts during the COVID-19 pandemic [55]. In addition, stress and allostasis affect the hypothalamic‒pituitary‒adrenal (HPA) system, which is critical for understanding depressive and treating depressive [56, 57]. In such circumstances, where nurses have increased access to supportive resources, leading to improvements in sleep quality, it could mitigate their susceptibility to developing severe depressive symptoms over time. Given the severity of the COVID-19 outbreak as one of the main life stressors, people with higher levels of social support tend to perceive better sleep quality and lower levels of stress, thus protecting against depressive symptoms [58]. This finding underscores the importance of social support as a buffer against the adverse mental health impacts of stress. These findings also suggest that interventions aimed at improving social support could potentially enhance sleep quality and reduce stress and depressive symptoms, offering a novel perspective on mental health promotion during times of crisis. Limitations A significant strength of our study lies in its large sample size and multicenter design, which allows for a more comprehensive examination of the relationship among social support, sleep quality, and depressive symptoms among Chinese nurses. However, it is important to acknowledge some limitations of this study. First, the subjective nature of self-report, questionnaires that assess social support, sleep quality, and depressive symptoms has raised concerns. A more convincing criterion for psychiatric diagnoses would involve structured clinical interviews. Second, we used the HADS to assess depressive symptoms among nurses. While it is not specifically tailored for healthcare professionals such as nurses, it may not fully capture the specific stressors and mental health issues that nurses might experience in their work environment. We suggest that future research employ pertinent scales, particularly those designed to assess the specific challenges faced during pandemic conditions [59, 60]. Third, although allostatic load is a well-researched theory that significantly contributes to the development of depression, this study was unable to collect relevant physiological data related to the allostatic load model due to the pandemic situation. Additionally, although we controlled for several demographic variables, unmeasured confounders such as income and night shift frequency that could influence the relationship among social support, sleep quality, and depressive symptoms. Finally, the cross-sectional design of our study could be considered a limitation since it did not establish causality in the observed associations. Thus, we cannot establish a cause‒and‒effect relationship between the variables. Longitudinal studies are needed to explore the temporal dynamics of these relationships. Conclusions This study offers significant contributions to the understanding of mental health dynamics among nurses during the COVID-19 pandemic. This study revealed that depressive symptoms were inversely associated with social support, with sleep quality playing a mediating role. These findings underscore the importance of social support and sleep hygiene in mitigating depressive symptoms among nurses, particularly during times of crisis. The mediating effect of sleep quality was significant across different subgroups, highlighting the universal need for interventions that promote optimal sleep quality and social support in healthcare settings. The large sample size and multicenter design of this study increase its generalizability. To ensure early detection and prevention of depressive symptoms among nurses during stressful periods such as the COVID-19 pandemic, providing adequate social support and ensuring good sleep quality are crucial. Future research should build upon these findings by exploring the longitudinal effects of social support and sleep quality on mental health outcomes and by developing targeted interventions on the basis of these relationships. Future research should explore the longitudinal effects of social support and sleep quality on mental health outcomes and develop targeted interventions on the basis of these relationships.