Psychological distress is common in AYA cancer patients.1,7,8 A meta-analysis showed that approximately 32% of cancer patients experience some type of psychological distress during active treatment,9 and a review reported a significant association between younger age and greater psychological distress during the cancer trajectory.10 In China, a study found that approximately 75% of AYA cancer patients reported anxiety and 90% of them reported depression. of life and may lead to a higher risk of mortality.13–15 As a complex symptom, pain is prevalent in the cancer population.16,17 A recent cross-sectional study found that 66% of cancer patients reported moderate to severe pain,18 while another study showed that 42.1% of adult cancer patients reported pain.12 Previous studies have found that age, sex, genetic predisposition, and cognitive and/or emotional processing around pain could significantly influence the perceived level of pain.19,20 Inappropriate pain management could has negative consequences for the patients’ emotions and cognitive function, and their quality of life. 21 Anxiety, depression, and pain are prevalent and highly interactive in AYA cancer patients.22 Individuals with comorbid anxiety and depression tend to respond more slowly to treatment and are more likely to commit suicide and relapse.23 In addition, the pain often co-occurs with anxiety and depression and together leads to a significant social and economic burden.24 Pain related to the oncological procedure can be accompanied by anxiety about medical procedures and hospitalization, separation anxiety and psychological stress.22 The presence of pain is associated with also with more depressive symptoms and worse outcome, such as lower quality of life, lower work performance and higher use of health services.19 In a randomized controlled trial, patients with increased pain were more likely to report higher levels of anxiety, fatigue, and depression.25 On the other hand, fatigue, anxiety, and depression also have a significant effect on pain in cancer patients.25 The study showed that compared to non-depressed pain patients, patients with comorbid depression and pain tended to have more pain-related complaints, more intense pain symptoms, and larger lesions. and highly interactive with each other, the interplay between these conditions has not been fully explored. To our knowledge, no study has examined the comorbidity of anxiety, depression, and pain in the cancer population from a network analysis perspective. Network analysis is an alternative new approach to examining the comorbidity of two or more disorders.27 In network analysis, comorbidity is believed to occur when symptoms from different disorders are directly related to each other.28 It estimates the unique associations between each pair of measured symptoms while controlling for all other symptoms and shrinking potentially spurious associations to zero.29 In this study, 1) network analysis was applied to assess the connectivity between anxiety, depression, and pain symptoms in a Chinese AYA cancer sample; 2) the prevalence of anxiety, depression and pain in AHAs was also investigated.
Methods
Participants
The participants of this study consecutively visited the outpatient units of Nanfang Southern Medical University Hospital and Guangdong Provincial People’s Hospital. Participants were recruited from January 1, 2018 to November 30, 2018. To be eligible, patients had to: 1) have a cancer diagnosis within the past 6 months; 2) are aged between 15 and 39 years. 3) understand Chinese. Those with impaired consciousness were excluded.
Study Process
This study was conducted in accordance with the Helsinki standard, and the study protocol was approved by the Ethics Committee of Nanfang Hospital (Reference No.: NFEC-2018-038) and the Ethics Committee of Guangdong Provincial People’s Hospital (Reference No. 2018295H( R1)). All participants were approached in the waiting room by trained research assistants. For those patients (adults), who show interest in participating; they were asked to provide a written informed consent form. At the same time, parental/legal guardian approval is also required for participants under the age of 18. All recruited participants were asked to complete a personal information collection form and a series of scales after which they were immediately returned to the research assistants. The recruitment and assessment process was supervised by a licensed psychiatrist.
Meters
Basic Demographic and Clinical Data
Basic demographic and clinical data of the participants (such as gender, age, cancer site, comorbidities, and family history of cancer) were collected using a custom-designed case recording form.
Patient Health Questionnaire (PHQ-9)
The PHQ is a self-rating scale that measures patients’ depression-related symptoms. The PHQ contains 9 items, each item using a four-level score from 0 to 3, with a total score of 0 to 27. A total score of 5 or more indicates depressive symptoms.30 The Chinese PHQ-9 has shown good psychometric properties,31 and internal consistency of PHQ-9 was 0.802 in Chinese AYA cancer patients.32
Generalized Anxiety Disorder (GAD-7)
The GAD is a 7-item self-report scale used to assess an individual’s anxiety symptoms. Available response options are scored from 0 to 3, and a total score of 5 or more indicates anxiety symptoms.33 The Chinese GAD-7 showed satisfactory psychometric properties,34 and the Cronbach’s alpha of the GAD-7 was 0.883 among Chinese AYA cancer patients. 32
McGill Pain Questionnaire-Visual Analogue Scale (MPQ-VAS)
The McGill Pain Questionnaire-Visual Analogue Scale (MPQ-VAS) is one of the most widely used tests to measure pain,35 and has been validated in the Chinese language in 2013.36 The score for the VAS ranges from 0 (no pain) to 100 (worst possible pain) and a total score of 40 or more indicates that the patient is currently suffering from pain.36
Sample size estimation
The sample size (N) was calculated using the formula: 37 where Z is the significance test statistic, alpha is the significance level, P is the prevalence, and d is the allowable error. In this study, alpha was set at 0.05, Za was set at 1.96, and d was 0.1 P. Based on a previous finding that approximately 75% of AYA cancer patients reported anxiety and 90% of them reported depression in China.11 Therefore, we assume the prevalence to be 75% and to allow further analyses, we increased the expected sample size by 50%. Finally, at least 192 participants were recruited into this study.
Statistical analysis
Network Analysis
First, we computed the network using the R packages “bootnet” and “qgraph”,38 with “EBICglasso” (i.e., the Extended Bayes Information Criterion combined with the Graphical Least Absolute Shrinkage and Selection Operator method) as the default method.29 A The network is a graphical representation of variables (nodes) and their correlations are depicted as edges.39 In the network, thicker and more saturated edges represent stronger correlations, green lines represent positive correlations, and red lines represent negative correlations. Second, to quantify the importance of each node in the network, we calculated the centrality indices of Strength, Betweenness, and Closeness. Higher values of the centrality index are representative of greater importance within the network, and symptoms with high centrality scores may be important as potential targets for further therapeutic interventions.40 Furthermore, to identify the bridge symptoms connecting different communities, we calculated the centrality bridge position (Bridge Strength, bridge between and bridge proximity). Recent studies have suggested that both Betweenness, Bridge Betweenness, Closeness, and Bridge Closeness indicators may not be robust in psychological networks.41 Therefore, in the following network analysis, we mainly focused on Strength42 and Bridge Strength.43,44 Strength means the sum of the absolute value of the correlations of a node with other nodes in the structure, while the bridge strength refers to the sum of the absolute edge weight values of all intercommunity edges.43,44 Centrality diagrams were created to represent these indices. Third, to examine the stability and accuracy of the networks29, a bootstrap procedure with dropout was performed to calculate the correlation stability coefficient (CS-C). CS-C is required to be above 0.25,29 Also, a non-parametric bootstrapping method was used to estimate the accuracy of edge weights by calculating confidence intervals (CIs). This means that larger CIs indicated less accuracy in limb estimation, while narrower CIs indicated a more reliable network.42 In addition, to examine the possible effect of age and gender on the emotional disturbances of the cancer patient, the network model was made and the indicators of local structure. reassessed, after controlling for age and sex.
