Associations Between Depressive Symptoms, Anxiety Disorders, Their Comorbidity and Health-related Quality of Life: a Large-scale Cross-sectional Study

Background: There were few studies exploring the associations between depressive symptoms, anxiety disorders and health-related quality of life (HRQoL) in the general population, especially in resource limited area. The aims of this study were to assess the associations between depressive symptoms, anxiety disorders, their comorbidity and HRQoL in rural area. Methods: A total of 23496 eligible participants from Henan rural cohort were included. The Patient Health Questionnaire-2 (PHQ-2) and Generalized Anxiety Disorder-2 (GAD-2) were employed to assess depressive symptoms and anxiety disorders, respectively. HRQoL was measured via European Quality of Life Five Dimension Five Level Scale (EQ-5D-5L). Tobit regression and generalized linear model were utilized to explore the associations between depressive symptoms, anxiety disorders, their comorbidity and HRQoL. Results: A total of 1320 individuals were identied as depressive symptoms with a prevalence of 5.62%, while 1198 participants were classied as anxiety disorders with a prevalence of 5.10%. After multiple adjustment, the regression coecients and 95% condence interval (CI) of utility index for depressive symptoms and anxiety disorders were -0.166 (-0.182, -0.149) and -0.132 (-0.149, -0.114), respectively. The regression coecients and 95% CI of VAS score for depressive symptoms and anxiety disorders were -7.65 (-8.60, -6.70) and -5.79 (-6.78, -4.80), respectively. Additionally, the comorbidity was strongest associated with low utility index and VAS score. These ndings were observed robustly in men and women. Conclusion: anxiety and their comorbidity were negatively associated with HRQoL in rural population, which needed further efforts on preventive and treatment

Health related quality of life (HRQoL) was a multidimensional concept contented by physical health status and subjective satisfaction with health, which was a reliable indicator to assess health status [8].
Most previous studies have focused on the effect of physical diseases on HRQoL [9][10][11], but ignored the effect of mental disorders. Previous studies have explored the association between depression, anxiety and HRQoL in speci c population [12][13][14], which indicated that depression and anxiety were negatively associated with HRQoL. However, there were few studies investigating the association in the general population, especially in rural population with limited resources. Additionally, multiple studies have documented that women had higher prevalence and were approximately twice as likely to suffer from depression and anxiety as men [15,16]. Thus, it is important to assess gender speci c associations between depression, anxiety and HRQoL.
The aims of this study were to explore the associations between depressive symptoms, anxiety disorders, their comorbidity and HRQoL in rural population. In addition, the difference between men and women was assessed.

Study population
The current study, designed as a cross-sectional study, was embedded in Henan rural cohort study, which was a population-based study with a large sample of rural people living in Yuzhou, Suiping, Tongxu, Xinxiang and Yima counties of Henan province in China. From July 2015 to September 2017, 39259 participants aged from 18 ~ 79 were included in the cohort study via multistage strati ed cluster sampling method, with a response rate of 93.7%. Detailed information of the cohort has been described previous publish [17].
Participants were excluded from the study if they had any of the following: (1) Did not participate in the EQ-5D-5L questionnaire survey (n = 15700); (2) Missing EQ-5D-5L data (n = 49); (3) Missing depressive symptoms and anxiety data (n = 14). Finally, a total of 23496 participants were included in this study.
The Henan Rural Cohort Study was approved by the Zhengzhou University Life Science Ethics Committee and conducted in accordance with the principles of the Declaration of Helsinki (Code: [2015] MEC (S128)). Before the study commenced, participants were informed of the study's purpose, health bene ts, and potential hazards. Participants were required to provide informed consent and both the researchers and respondents agreed to use the data for scienti c research purposes only.

Data collection
All participants were interviewed face-to-face by well-trained research staff via a standard questionnaire to collect data including information on demographic characteristics, lifestyle factors and individual history of chronic diseases. To ensure the accuracy and integrity of the collected data, trained investigators checked the integrity and logical errors of the questionnaire on the same day of questionnaire completion, and if there were any problems, they contacted the participants by phone and amended the responses.
Demographic covariates included gender, age, marital status, educational level, and average monthly income. Education level was divided into three levels: elementary school or below, junior high school, and senior high school or above. Average monthly income was also classi ed into three levels: <500 RMB, 500-999 RMB and ≥ 1000 RMB. Lifestyle factors covered smoking (never, former, and current), alcohol drinking (never, former, and current) and physical activity. The de nition of current smoking and current drinking have been described previous publish [18]. Physical activity was grouped into three levels (low, moderate, and high) according to the validated Chinese version of the International Physical Activity Questionnaire (IPAQ) [19]. Chronic diseases which included hypertension, dyslipidemia, T2DM, CHD and stroke were collected via physical examination, laboratory tests, or self-reports [20]. Body height and weight of the participants were measured twice with shoes and coats off and the readings were recorded to the nearest 0.1cm and 0.1kg, respectively. The average readings of the two measurements were taken for statistical analysis in this study. Body mass index (BMI, kg/m 2 ) was calculated as weight (kg) divided the square of height (m).

De nition of depressive symptoms and anxiety disorders
In this study, the Patient Health Questionnaire-2 (PHQ-2) and Generalized Anxiety Disorder-2 (GAD-2) were performed to assess depressive symptoms and anxiety disorders. The PHQ-2 and GAD-2 are an abbreviated version of the Patient Health Questionnaire-9 (PHQ-9) and the seven-item generalized anxiety disorder scale (GAD-7), met the demands of busy primary care practice and large population-based surveys. Both two scales were a reliable and valid screening tool for depressive symptoms and anxiety disorders in Chinese rural population [21,22].
Both two scales were consisted of two items and each item was consisted of four levels (not at all = 0, several days = 1, more than half the days = 2, and nearly every day = 3), with a total score ranged from 0 to 6. In this study, a cutoff of 3 was adopted to identify depressive symptoms and anxiety disorders [22,23]. In other words, participants who reported a score of 3 or above for PHQ-2 or GAD-2 scales were classi ed as having depressive symptoms or anxiety disorders, respectively.

Assessment of HRQoL
In this study, HRQoL was assessed by European Quality of Life Five Dimension Five Level Scale (EQ-5D-5L), a standardized measure of HRQoL developed by the EuroQol Group in order to provide a simple, generic measure of health for clinical and economic appraisal [24]. The ve dimensions of EQ-5D-5L were mobility (MO), self-care (SC), usual activities (UA), pain/discomfort (PD) and anxiety/depression (AD) and each dimension was consisted of ve levels (no problems, slight problems, moderate problems, severe problems, and extreme problems). The EQ-5D-5L utility index was calculated according to the recently available Chinese value set for the EQ-5D-5L instrument [25]. The calculated formula of EQ-5D-5L utility index can see in elsewhere [26]. The EQ-5D-5L also included a visual analogue scale (VAS), a vertical 0 to 100 point rating scale, which re ects the degree of satisfaction with their health status. The best and worst health states carry a score of 100 and 0, respectively.

Statistical analysis
Statistical description was presented as frequencies and percentages for categorical variables, while means and standard deviations (SD) were calculated for continuous variables. T-test or Kruskal-Wallis test was performed to compare differences between different groups for continuous variables, while Chi squared test was utilized for categorical variables.
Multivariate Tobit regression model [27] was performed to explore the associations between depressive symptoms, anxiety disorders and utility index, due to the distribution of the EQ-5D utility was skewed and the utility index was censored at 1. Due to the VAS score was abnormal distribution, a generalized linear model (GLM) was employed to assess the associations between depressive symptoms, anxiety disorders and VAS score. Model 1 was unadjusted. Model 2 adjusted age, gender, marital status, education level, average monthly income, physical activity, smoking status, drinking status, and BMI. Model 3 further adjusted hypertension, dyslipidemia, T2DM, CHD and stroke based on model 2.
Data were analyzed using SPSS 23.0 software package (SPSS Institute, Chicago) and STATA 15 for Windows. All P values were two-tailed with a statistical signi cance level of 0.05.

Characteristics of study participants
Characteristics of study participants according to depressive symptoms and anxiety disorders were presented in Table 1. A total of 1320 individuals were identi ed as depressive symptoms with a prevalence of 5.62%, while 1198 participants were classi ed as anxiety disorders with a prevalence of 5.10%. The mean age ± SD of participants (41.59% male and 59.41% female) was 55.28 ± 12.63. Compared with non-depressive group and non-anxiety group, depressive group and anxiety group were more prone to older, women, low education level, and low average monthly income, and less likely to be married/cohabiting, current smoking, current drinking, and obesity (all P < 0.05). The mean (SD) utility index and VAS score of the total sample were 0.954 (0.111) and 78.33 (14.80), respectively. Participants with depressive symptoms or anxiety disorders had lower utility index and VAS score than these without depressive symptoms or anxiety (all P < 0.001). Self-reported health problems of EQ-5D-5L Figure 1 demonstrated the self-reported health problems of EQ-5D-5L in total sample, depressive group, and anxiety group. Of all participants, the most frequently reported problem was pain/discomfort dimension (23.03%), followed by mobility (12.72%), anxiety/depression (7.76%), usual activities (6.44%), while the least report was the self-care dimension (3.66%). In depressive group and anxiety group, the most frequently reported problem was still reported in pain/discomfort dimension (52.43% and 51.75%, respectively), followed by anxiety/depression (42.41% and 41.13%, respectively). In both the depressive group and the anxiety group, the percentage of problems reported on all ve dimensions increased signi cantly, especially in anxiety/depression and pain/discomfort dimension. Compared with nondepressive group and non-anxiety group, depressive group and anxiety group had higher percentage of reported health problems in all of ve dimensions (all P < 0.001).

Associations between depressive symptoms, anxiety disorders and HRQoL
The results of Tobit regression and Generalized linear models analyses on utility index and VAS score were summarized in Table 2. After multiple adjustments, the Tobit regression model and GLM indicated that both utility index and VAS score were lower in participants with depressive symptoms or anxiety disorders. The regression coe cients and 95% con dence interval (CI) of utility index for depressive symptoms and anxiety disorders were − 0.166 (-0.182, -0.149) and − 0.132 (-0.149, -0.114), respectively. The regression coe cients and 95% CI of VAS score for depressive symptoms and anxiety disorders were − 7.65 (-8.60, -6.70) and − 5.79 (-6.78, -4.80), respectively. In addition, the regression coe cients and 95% CI of utility index associated with 1 score increase in PHQ-2 score and GAD-2 score were − 0.040 (-0.044, -0.036) and − 0.034 (-0.039, -0.030), respectively. The regression coe cients and 95% CI of VAS score associated with 1 score increase in PHQ-2 score and GAD-2 score were − 2.25 (-2.48, -2.03) and − 1.27 (-1.50, -1.04).  Figure 2 presented the gender speci c associations between depressive symptoms, anxiety disorders and utility index and VAS score. Both in men and women, the signi cantly negative associations between depressive, anxiety disorders and utility index and VAS score was observed. Additionally, there was no interaction between depressive symptoms, anxiety disorders and gender (all P for interaction >0.05).
Associations between comorbidity and utility index and VAS score.
The gender speci c associations between comorbidity and utility index and VAS score were showed in Fig. 3. The ndings indicated that comorbidity was strongest associated with low utility index and VAS score. Of total sample, the regression coe cients and 95% CI for comorbidity in utility index and VAS score were − 0.288 (-0.305, -0.271) and − 13.61 (-14.61, -12.60), respectively. Notably, the negative associations between depressive symptoms and HRQoL was stronger than anxiety disorders. These ndings were observed robustly in men and women.

Discussion
In this large population-based study, we investigated the association between depressive symptoms, anxiety disorders, their comorbidity and HRQoL, and assessed the difference between men and women.
The mean (SD) utility index and VAS score of the total sample were 0.954 (0.111) and 78.33 (14.80), respectively. Of all participants, the most frequently reported problem was pain/discomfort dimension, while the least report was the self-care dimension. In both the depressive group and the anxiety group, the percentage of problems reported on all ve dimensions increased signi cantly, especially in anxiety/depression and pain/discomfort dimension. In addition, our results suggested that comorbidity was strongest associated with low HRQoL. Notably, the negative associations between depressive symptoms and HRQoL was stronger than anxiety disorders. These ndings were observed robustly in men and women.
Participants in previous studies conducted in medical institutions had low utility index and VAS score [28,29]. However, most participants in this study were satisfactory with their HRQoL, which was consistence with the previous studies conducted in China [10,30]. It may be due to the participants of this study were recruited from rural areas and lived a normal life at home, which hinted HRQoL is better than those who live in medical institutions for professional care. Of all participants, the most frequently reported problem was reported in pain/discomfort dimension, while the least report was the self-care dimension. This was in line with previous studies in China [9,10]. In both the depressive group and the anxiety group, the percentage of problems reported on all ve dimensions increased signi cantly, especially in anxiety/depression and pain/discomfort dimension. Depressive symptoms and anxiety disorders increased the self-reported problems in pain/discomfort dimension, it may be due to mental disorders and chronic pain tend to further aggravate the severity of both disorders [31]. Self-reported problems in anxiety/depression dimension increased in depressive group and the anxiety group, which indicated that subjectively perceived depression and anxiety were similar to those detected by the PHQ-2 and GAD-2 scale.
In this study, depressive symptoms and anxiety disorders were associated with low HRQoL. Previous study conducted in speci c population with chronic diseases have found stable associations between mental disorders and low HRQoL [12][13][14]32]. In addition, serval studies conducted in old adults also found the negative associations between mental disorders and HRQoL [33][34][35][36]. The negative association between depression and HRQoL was also observed in postmenopausal women in Korea [37]. The ndings of these studies were consistent with the current study. However, these studies were all conducted in speci c subgroup of population, and may not fully re ect the association between mental health and HRQoL. The current study conducted in general rural population contributed new evidence in rural area and may provide a better understanding of the mental determinants of improving the HRQoL. Nonetheless, it should be noted that our ndings based on cross-sectional study cannot con rm causal relationship between depressive symptoms, anxiety disorders and HRQoL. Certainly, a study suggested that low HRQoL was associated with depression [38].
To the best of our knowledge, this is the rst study to investigate the association between comorbidity and HRQoL, whcih suggested that comorbidity was strongest associated with low HRQoL. A prior study conducted in Swedish general population found that comorbidity was associated with higher symptom severity and lower health-related quality of life [39]. However, this study only explored the relationship between comorbidity and each dimension of EQ-5D. The current study has explored the association between comorbidity and utility index calculated according to the recently available Chinese value set, which can better re ect the HRQoL. Our study also found that the negative associations between depressive symptoms and HRQoL was stronger than anxiety disorders. However, we have not found any other research evidence to support our results and it need further exploration.
There were several limitations in this study. Firstly, the results only can indicate association and cannot establish causal relationship, because this study was cross-sectional design. Prospective studies on mental disorders and HRQoL are needed. Secondly, the PHQ-2 and GAD-2 are useful screening measures rather than diagnostic tools, thus the prevalence of depressive symptoms and anxiety disorders may be overestimate, which can induce bias. However, in busy primary care or large population studies, these two scales were quite suitable for saving time while still providing accepted diagnostic performances. Thirdly, some information of participants in this study was collected based on self-reported, but higher test-retest reliability, effective training of study workers and good eld implementation will ensure the accuracy and reliability of the information.

Conclusions
In summary, this study demonstrated that depressive symptoms and anxiety disorders were negatively associated with HRQoL. In addition, comorbidity was strongest associated with low HRQoL and the negative associations between depressive symptoms and HRQoL was stronger than anxiety disorders.
These ndings were observed robustly in men and women. However, large-scale prospective studies are needed to prove our ndings and provide more information about the causal relationship and internal mechanisms of this association.

Declarations
Ethics approval and consent to participate Ethics approval was obtained from the "Zhengzhou University Life Science Ethics Committee", and written informed consent was obtained for all participants. Ethics approval code: [2015] MEC (S128). The study is conducted in accordance with the principles of the Declaration of Helsinki Consent for Publication Not applicable.

Availability of data and material
The data analyzed during current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests. Authors' contributions CW and FW conceived and designed the study. WL, ZL, XD, XW and YM analyzed the data. WL, YL, NC, NK, XL and WH drafted the manuscript. ZL, XD, XW, YM, YL and CW collected data. XL, WH, FW, and CW provided technical direction and writing assistance in the preparation of this manuscript. All authors critically revised the manuscript and approved the nal version for publication. Figure 1 Self-reported health problems of EQ-5D-5L in total sample, depressive group, and anxiety group. Gender speci c associations between depressive symptoms, anxiety disorders and utility index and VAS score (Adjusted age, gender, marital status, education level, average monthly income, physical activity, smoking status, drinking status, BMI, hypertension, dyslipidemia, T2DM, CHD and stroke).