Influencing Factors And Network Characteristics Of Anxiety Symptoms In The Medical Staff During The COVID-19 Pandemic

BackgroundCOVID-19 is characterized by its high infectivity,strong pathogenicity,long incubation period and being infectious in incubation period.It has been classified by China Health Commission as Class B infectious diseases stipulated in Infectious Disease Prevention Act of People’s Republic of China,and been adopted preventive and control measures for Class A infectious disease.The outbreak of COVID-19 in early 2020 was raging across the country.It has become a major public health emergency,which has a serious adverse effect on people’s physical and psychological health.A large number of studies have confirmed that negative emotion is the most dominant and prominent psychological problem.And the anxiety is more serious with the highest detection rate.Medical staff are the high-risk group who work in the front line against the COVID-19 and have close contacts with the patients.Their excessive workload and psychological pressure are more likely to cause anxiety and directly affect their performance.Therefore,an accurate analysis of the front-line medical staff’s anxiety symptoms is of great significance for psychological assistance and negative emotions alleviation.As a new technology,network analysis provides a unique method to understand the pathology of mental and psychological disorders.The network perspective of symptoms focuses on the causal relationship between symptoms,and inspires researchers to consider the role of the relationship between symptoms in the development and maintenance of mental disorders.In addition,a unique advantage of using network analysis to analyze the symptoms of mental disorders is that it can reveal the most central(influential)symptoms in the symptom network.According to network theory,the intervening of the psychological symptoms with high centrality in the symptom network may transmit the effect to the whole network to the greatest extent,so as to effectively reduce and alleviate mental and psychological symptoms.This is of great significance for in-depth understanding of the pathological characteristics of mental and psychological disorders and providing effective intervention measures.Therefore,this study evaluated the anxiety of medical staff during the COVID-19 epidemic,and analyzed the relationship between anxiety symptoms from the perspective of network in exploration for the rules of symptoms.The study provides theoretical and feasible guidance for an optimized research into anxiety of front-line medical staff,which helps to timely identify problems,formulate appropriate mitigation and provide mental health services.MethodsFrom February 18 to April 3,2020,the online sampling method was applied to investigate and the entire research process was based on Zhongsheng Kaixin PEM medical staff mental health care platform.Medical staff were requested to fill in the questionnaire online by logging on the platform.The questionnaire consists of two parts: 1)Informed consent,which informs the subjects of the purpose,content,methods and precautions of the study,and defines the inclusion and exclusion criteria.2)Formal questionnaires,including COVID-19 general information questionnaire and generalized anxiety disorder scale(GAD-7)which were anonymously filled to eliminate the concerns and falsification of the subjects.Study 1: Anxiety scores of medical staff were compared with the scores of the public in the existing studies by a single sample t-test.Analysis of variance was conducted with“whether to participate in front-line rescue”,“gender",“marital status”,“education”,“position”,“professional title” and “working years” as class variables.The measurement data were expressed by mean add and subtract standard deviation(± s),with p <0.05 as the criterion of statistical significance.Study 2: First-line medical staff with GAD scale score ≥ 8 were screened,and the software R was used for network analysis and visual analysis.The Gaussian graphical models(GGM)was used to fit the data,and the graphical least absolute shrinkage and selection operator(glasso)in the qgraph of software R was used to select the model in combination with the Extended Bayesian information criterion(EBIC)to obtain the stable regularized partial correlation network(RPCN)and the centrality of nodes.The MGM package in software R was used to calculate the predictability of each node.The bootnet package was used to evaluate the accuracy and stability of the network.The Spinglass algorithm of igraph was applied to investigate the community structure in the network,and Network Comparison Test was used to compare the gender differences of anxiety symptom networks.The directed acyclic graph(DAG)of anxiety symptoms was evaluated by Bayesian hill-climbing algorithm of bnlearne.Results1.The condition of anxiety and its influencing factors for healthcare workers during COVID-19The total detection rate of anxiety symptoms of medical staff was 51.4%,with a total score of 5.15 ± 4.16,which was significantly higher than that of general population in same period studies.The anxiety score of first-line medical staff was significantly higher than that of non-first-line staff(P < 0.05).The anxiety score of female medical staff was significantly higher than that of male(P < 0.01).The anxiety score of widowed and divorced persons ranked the highest,followed by the married group and the unmarried group ranked the lowest,with statistically significant difference(p < 0.001).The anxiety score of nurses was significantly higher than that of doctors(p < 0.001).The anxiety score of medical staff with intermediate professional title was higher than that of senior,primary and non-professional title,and the difference was statistically significant(p < 0.001).The anxiety scores of medical staff with working time ≥ 10 years were higher than those with 4 ～ 10 years and ≤ 3 years,and the difference was statistically significant(p <0.05).2.Network analysis of anxiety symptoms of healthcare workers under COVID-19In the regularized partial correlation network,a wide correlation was detected between anxiety symptoms,A1 “feel nervous,anxious or anxious” and A2 “unable to stop or control worrying”,A6 “become easily irritable or irritable”and A7 “feel frightened that something terrible was about to happen”,A5 “unable to sit quietly because of anxiety” and A6,A2 and A3 “worry too much about various things”,A4 “hard to relax” and A5 which were strongly connected.A2 and A3 reported the highest expected impact,and A2 had the highest predictability.Community testing found one community composed of A1,A2,and A3 and another community composed of A4,A5,A6,and A7.The results of gender network comparison showed that the overall strength of female anxiety symptom network was greater than the total weight of the edge of male anxiety symptom network,yet with no significant difference found.Directed acyclic graph results showed that A2 symptoms had the highest probability priority,with A2 to A1,A2 to A7,A2 to A4 being the most important arrows for DAG structure.Conclusion1.The high anxiety scores and detection rates of medical staff in the early stage of COVID-19 indicate that the overall anxiety of medical staff is relatively serious.Six factors,including participation in front-line rescue,gender,marital status,post,professional title and working years,are related to their anxiety symptoms.2.The anxiety symptoms of front-line medical staff with GAD ≥ 8 are widely correlated.A2 “unable to stop or control worrying” is the central node of anxiety symptom network and has the highest predictability.A2 can be interfered to produce maximum effect on the overall GAD symptom network.A1,A2,A3 and A4,A5,A6,and A7 may aggregate into a “subsystem” respectively,which verifies the reliability of the symptom diagnostic criteria for GAD in DSM-5 from the perspective of network.The pattern of association between anxiety symptoms may be consistent across genders.