Network structure of depression and anxiety symptoms in Chinese female nursing students

Depression and anxiety are common mental disorders. A retrospective study of prevalence rates in 44 countries found that the worldwide prevalence of anxiety disorders was estimated at 7.3%, which is equivalent to one in 14 people worldwide suffering from anxiety disorders at any given time [1]. The one-year prevalence of major depressive disorder varies among countries worldwide, but the overall level is about 6% [2]. Comorbidity between depressive and anxiety is common. For example, in the Netherlands Study of Depression and Anxiety, 67% of individuals diagnosed with primary depression had a current comorbid anxiety diagnosis. Similarly, 63% of individuals diagnosed with primary anxiety had a current comorbid depression diagnosis [3]. The level of comorbidities is usually related to the severity of the illness, chronicity, and inability to function well in daily life [4–6]. As we all know, the presence of one mental disorder of depression or anxiety often acts as a trigger for the other [7, 8]. Recently, a meta-analysis study shows that symptoms of depression and anxiety can be predicted from one another during a period of time such as weeks and months [9]. As a result, when symptoms of one disorder are present, the risk of onset of symptoms of the second disorder may rise correspondingly.

In recent years, the network model to psychopathology has been advanced as an alternative way of conceptualizing mental disorders. This model can analyze the relationship between complex variables from a mathematical point of view and display it intuitively [10, 11]. It is data-driven rather than relying on previous assumptions about causality between variables [10, 12]. The network is composed of two parts, one is the node, which represents the variable, and the other is the edge, which represents the relationship between the variables [13]. In this model, mental disorders arise from the direct interactions of different symptoms [14–16]. In other words, mental disorders are emergent phenomena caused by the direct interactions between their corresponding symptoms rather than unobserved latent entities that cause the emergence of symptoms. An accurate description of these interactions is key to interpreting psychopathological mechanisms and developing targeted intervention strategies. Compared with the mere correlational approaches, network model can provide the corresponding centrality and predictability index for each node to examine its importance and controllability in the whole network [17, 18]. Central symptoms in mental disorders may be potential targets for clinical interventions. In addition, network model also provides a new perspective on the understanding of comorbidities [19, 20]. When a person suffers from a certain disorder, the corresponding symptoms of the disorder may increase the risk of other disorders, thus leading to the diagnosis of comorbidities, and the symptoms that increase the risk of other disorders are considered as bridge symptoms [19]. The network model to psychopathology suggests that bridge symptoms may play a role in the development and maintenance of comorbidities. Therefore, it is possible for clinicians to prevent and treat comorbidities from the perspective of bridge symptoms [19, 20].

In the studies of depression and anxiety symptoms, researchers found different and common network characteristics in different samples by applying the network model [10, 21–24] (Wei Z, Ren L, Wang X, Liu C, Cao M, Hu M, et al. Network of depression and anxiety symptoms in patients with epilepsy 2021. (under review).). For example, the relations between “fatigue” and “sleep difficulties”, between “anhedonia” and “depressed or sad mood”, and between “nervousness or anxiety” and “uncontrollable worry” were strong in both migrant Filipino domestic workers and a psychiatric sample [10, 22]. In addition, “depressed or sad mood” and “worry too much” were considered as the most central symptoms in both migrant Filipino domestic workers and a psychiatric sample [10, 22]. “Fatigue” was disclosed as bridge symptom which increases risk of comorbidity between depression and anxiety in migrant Filipino domestic workers [22]. These network studies have provided new insights into the relations and comorbidities of depression and anxiety symptoms. Moreover, network analysis can help us better understand how symptoms related to suicidal ideation. “Thoughts of death” was closely related to “psychomotor agitation/retardation” in migrant Filipino domestic workers [22], while it was related to “depressed or sad mood” and “feeling of worthlessness” in a psychiatric sample [10].

Nursing has historically been a female-dominated profession and a large number of previous studies have shown that the incidence of depression and anxiety is high among nursing students [25, 26]. A previous study, which collected 2111 undergraduate nursing students from various institutions in the United States, showed that 16% of them expressed symptoms consistent with moderate major depressive disorder (MDD), and about 10% of them reported symptoms consistent with moderate generalized anxiety disorder [27]. Unlike other undergraduate majors, nursing students have a heavy study load, including clinical practice, skill exams and lots of other assignments. In addition to the intense study load, they also have to deal with financial burdens, interpersonal problems and other factors that may put them at increased risk of depression, anxiety and suicidal ideation [25]. In addition, previous studies have shown that women are almost twice as likely to suffer from depression and anxiety disorders as men [28, 29]. Suicide is the second most common cause of death among young people aged 15–29, and women have higher rates of suicidal ideation than men [30–32]. For students majoring in nursing, Goetz held opinion that they may have a high risk of suicide compared with college students of other majors, and the reason may be related to the high pressure of academic curriculum for nursing students [33]. A study of nursing students from Greece showed that the incidence of lifetime suicide ideation was about 10.6%. The results also showed that 1.7% female participants reported the thoughts to actually die though suicide if they had an opportunity [34]. Therefore, it is necessary to investigate symptoms of depression and anxiety in female nursing students from a network perspective to provide related clinical implications.

The current study is the first to apply network model to investigate how symptoms of depression and anxiety relate to each other in Chinese female nursing students. We hypothesize that the symptoms network structure, central symptoms, and bridge symptoms in the present study may have similar and specific results when compared with previous network studies. Particularly, we focus on the symptoms that are directly related to “thoughts of death”. We hypothesize that the symptoms “psychomotor agitation/retardation”, “depressed or sad mood” and “feeling of worthlessness” are closely related to “thoughts of death”.

This study provides a complex network of depression and anxiety symptoms among Chinese female nursing students. Through this network, we find that the strongest edges exist within each disorder, which is consistent with the results of previous network researches investigating comorbidity of depression and anxiety symptoms [10, 21–24]. Our findings are similar to those of Garabiles et al. and Beard et al. who observed that strongest edges were between “sleep difficulties” and “fatigue”, between “psychomotor agitation/retardation” and “thoughts of death”, between “nervousness or anxiety” and “uncontrollable worry”, and between “uncontrollable worry” and “worry too much” [10, 22].

The current study finds a second strongest edge between “fatigue” and “anhedonia”. To our knowledge, this result has not been discovered in previous studies. This finding may be unique to the specific sample in current study (i.e., Chinese female nursing students), which needs to be further investigated. In fact, according to a recent report on reviewing the similarities and differences between fatigue and anhedonia, approximately 40% of the articles considered fatigue and anhedonia as related or overlapping constructs [48]. Several studies have proposed potential common mechanisms or pathways to elucidate the connection between fatigue and anhedonia. For example, Capuron et al. provided the underlying mechanism by which inflammatory cytokines influence fatigue and anhedonia [49]. In addition, we also find a strong relation between “nervousness or anxiety” and “worry too much”. Indeed, in the fifth edition of Diagnostic and Statistical Manual of Mental Disorders (DSM-5) [50], these two symptoms act as one core symptom of generalized anxiety disorder (i.e., excessive anxiety and worry about various events). The average predictability of present network is 62%, implying that the present network consisted of depression and anxiety symptoms are more likely to be self-determined.

Node expected influence centrality may play an important role in finding symptoms that activate or maintain psychopathological networks as well as providing potential targets for intervention. Depression symptom “fatigue” has the highest centrality which indicates this symptom play the most important role in activating and maintaining psychopathology network of depression and anxiety. Therefore, interventions targeting “fatigue” might generally alleviate both anxiety and depression symptoms in Chinese female nursing students. In addition, this centrality result is consistent with previous studies which investigated symptoms network of depression and anxiety among Filipino immigrant family workers and patients with major depressive disorder [22, 24]. A previous study has shown that moderate to severe fatigue was a very common phenomenon (83.5%) among nursing undergraduate students [51]. And 44 % of the nursing students mentioned that the primary cause of fatigue are related to the characteristics of the nursing course. The excess of activities and sleep disorders are the second and third causes of fatigue [51]. In addition, as the editorial board member suggested, culture may favor physical rather than verbal expression as coping strategy and then leads to fatigue. Researchers have found that Filipinos tend to somaticize to express distress rather than confronting their problems directly [52]. In fact, “fatigue” is an overlapping symptom of major depressive disorder and generalized anxiety disorder, which exist in the DSM-5 diagnostic criteria [50]. The overlapping symptom could explain comorbidity rates. However, the bridge expected influence of “fatigue” is lower than average level of bridge centrality. This may because the edges between “fatigue” and “sleep difficulties”, and “fatigue” and “anhedonia” are the strongest edges. Depression symptom “feeling of worthlessness” and anxiety symptom “irritable” also have high centralities which suggests targeting these symptoms may be efficient to decrease severity of depression and anxiety symptoms in Chinese female nursing students. According to previous studies, irritability is defined as a low threshold for experiencing anger in response to frustration [53, 54]. A recent study showed that symptom “anger” was highlighted the importance to symptomatology in idiographic dynamic network of mood and anxiety symptom [55].

Node bridge expected influence centrality may provide guidance for searching bridge symptoms that play important roles in the development and maintenance of comorbidity of mental disorders. There are two bridge symptoms in the current network, including depression symptom “depressed or sad mood” and anxiety symptom “irritable”. These findings indicate that when depression presents, treating “depressed or sad mood” may decrease risk of contagion to anxiety and when anxiety presents, treating “irritable” may decrease risk of contagion to depression. In previous studies on the bridge symptoms between anxiety and depression, “depressed or sad mood” was also found as a bridge symptom [20, 22]. Further, by conceptualizing mental state of depression and anxiety as a dynamic network, researchers found that “feeling down” was the strongest bridge mental state in both depression-anxiety comorbid and anxiety-only groups [23]. As for “irritable”, it performs as a diagnostic criterion for not only generalized anxiety disorder but also depression in children and adolescents in the DSM-5 [50]. However, the recent research has shown that on mental state level, “irritable” is not relevant as bridge mental states in personal which are prone to suffer from comorbidity of depression and anxiety [23]. This needs to be further investigated.

Even though “thoughts of suicide” ranked the lowest centrality which was quite similar to previous studies [10, 22], it has always been a very important clinical manifestation of depression. In our study, “thoughts of suicide” had the lowest mean and standard deviation, which may have attenuated its expected influence [10]. In addition, what is worth mentioning is that increasing studies present suicide ideation from a network perspective and might provide new insights into this problem [56–59]. In our flow network, “thoughts of death” is directly connected with most of symptoms of depression and anxiety, which to some extent reflects the suicide ideation is actually a complex phenomenon. Our finding that the connection between “thoughts of death” and depression symptoms “psychomotor agitation/retardation” has the strongest correlation coefficient among all direct connections is well aligned with previous study [22]. Among patients with a major depressive episode (MDE), studies have shown that “psychomotor agitation and impulsivity” was one of the most frequently variables related to previous suicide attempts (SA) [60]. The connection between “thoughts of death” and “feeling of worthlessness” also have a great correlation coefficient. Using directed acyclic graphs to explore the relations among symptoms of alcohol use disorder and MDD and suicidal behaviors, researchers found that “worthlessness/guilt” was the symptom directly associated with suicide ideation for both men and women groups [61]. Moreover, “feelings of worthlessness” during MDE was the only symptom that predicted the increase of SA after the remission of MDE [62]. A recent study found that only 3 symptoms from the first 8 PHQ-9 questions (i.e., the second symptom “depressed or sad mood”, the sixth symptom “feeling of worthlessness”, and the eighth symptom “psychomotor agitation/retardation”) were significant explanatory variables for suicidal ideation among adult primary care patients [63]. Previous studies also found that negative beliefs about the self (e.g., low self-worth) is highly prevalent in those who entertain suicidal thoughts [64]. In addition, the relations between “thoughts of death” and “anhedonia” and “depressed or sad mood” are weak which indicate inadequacy of the PHQ-2 (i.e., the first 2 questions of PHQ-9) for identifying Chinese female nursing students with suicide ideations. Overall, these findings suggest that we may have to pay particular attention to any indication that Chinese female nursing students present psychomotor agitation/retardation (i.e., moving or speaking so slowly, or being so fidgety or restless) and feeling of worthlessness (i.e., bad self-feeling, or feel like a failure, or like they have let themselves or their family down). Meanwhile, these findings may also provide several potential pathways for interventions of suicidal ideation. The predictability of “thoughts of death” is 0.53, indicating that “thoughts of death” is moderately affected by its neighboring symptoms in the current network. This result indicates that we could intervene on “thoughts of death” not only through other relevant variables that are not belong to the current network or itself but also by its strong neighboring symptoms (i.e., “psychomotor agitation/retardation” and “feeling of worthlessness”). It is important to note that predictability is the upper bound estimation.

This is the only study to our knowledge to investigate network structure of depression and anxiety symptoms in Chinese female nursing students. The current study provides several possible implications for clinical prevention and intervention to meet the needs of mental health in Chinese female nursing students. First, “fatigue”, “feeling of worthlessness”, and “irritable” have the highest expected influence. From a network perspective, targeting these symptoms might generally alleviate both anxiety and depression symptoms in Chinese female nursing students. Second, we find that “depressed or sad mood” and “irritable” are bridge symptoms. Therefore, depression symptom “depressed or sad mood” may put one at risk of anxiety and anxiety symptom “irritable” may put one at risk of depression. To prevent or to treat comorbidity of depression and anxiety in Chinese female nursing students, bridge symptoms may be the efficient targets. Third, there are strong relations between “thoughts of death” and “psychomotor agitation/retardation” and “feeling of worthlessness”. Through observing and alleviating the “psychomotor agitation/retardation” and “feeling of worthlessness” in Chinese female nursing students, it might be efficient to detect and intervene suicidal ideation in them.

There are some limitations in our study. First, we recruited Chinese female nursing students who majoring in school of nursing and reporting symptoms of depression and anxiety that span the full range of normal to abnormal, which likely limits the generalizability of our findings. For example, depression and anxiety symptoms network in men or clinical sample may be different from the network structure in the current study. In addition, as suggested by a reviewer, the potential influences of depressive disorders, anxiety disorders, and/or other psychiatric disorders of the female nursing students on the estimated network structure have not been evaluated in the present study. These potential influences should be further explored in future studies. Second, the cross-sectional data applied to construct the network structure of depression and anxiety symptoms preclude claims about causality. Therefore, we cannot clarify the causality between the most central symptom and the other symptoms, because there are many possibilities, such as the central symptom activates the other symptoms, or the other symptoms activates the central symptom, or both. Future studies could use intensive longitudinal data to investigate the causality of these symptoms. Third, the network structure constructed here investigated between-subject effects on a group level. This means that within a single individual, the network structure may not be replicated in the same way. Fourth, in this study, the symptoms were single-item, self-reported assessments, which may be limited to capture clinical phenomena. More items and methods could be used in future research. Finally, the network structure in the current study is specific to the questionnaires we used. There are often some differences among self-report tools for assessing symptoms of depression and anxiety. Thus, different self-report tools could result in different network structures.

In conclusion, this study is the first article investigating network structure of depression and anxiety symptoms in Chinese female nursing students. For centrality indices, “fatigue”, “feeling of worthlessness”, and “irritable” have highest expected influence. Results reveal two bridge symptoms: “depressed or sad mood” and “irritable”. Results also identify “psychomotor agitation/retardation” and “feeling of worthlessness” as critical priority due to they are related to “thoughts of death”. Clinical implications such as treating particular symptoms as targets for preventions and interventions are discussed.

We would like to thank all the individuals who participated in the study. We also thank all the administrative staff and teachers in the university who help us with the recruitment.

LR, YFW, LW, ZHW, LBC, XYW, XYH, JXP, YCJ, FZL, QY and XFL conceived and designed the study and interpreted the study results. LR analysed the data. LR, YFW and LW wrote the paper. ZHW, LBC, XYW, XYH, JXP, YCJ, FZL, QY and XFL critically reviewed drafts of the paper. All authors approved the final version of the manuscript.

This study was supported by the Key Research and Development Program of Shaanxi Province (2020SF-245).

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

Qun Yang, Email: yangqun1125@hotmail.com.

Xufeng Liu, Email: lxf_fmmu@163.com.