Synergistic effect of social media use and psychological distress on depression in China during the COVID‐19 epidemic

The COVID‐19 pandemic is expected to have long‐term effects on mental health with implications at a population health level. While limiting the transmission of the virus, lockdown measures subject individuals to significant psychological distress and interpersonal isolation, which may increase risk for depression, a chronic and disabling disease associated with tremendous societal, individual, and economic costs (e.g., workplace productivity loss, unemployment, work absence, and long‐term disability). ¹ In addition to the elevated risk of depression and loneliness attributable to frequent and prolonged social media (SM) use outside the context of epidemics, frequent exposure to fearful and exaggerated information through SM can exacerbate psychological and emotional distress. ² , ³

We investigated the impact of SM use and psychological and emotional distress on depression in 3064 adults in Mainland China. A national convenience sample of 2574 health‐care workers and 490 non‐medical workers in China was surveyed cross‐sectionally by telephone or WeChat between 29 January and 11 February 2020. Our study participants consisted of physicians (n = 783), nurses (n = 1587), and other medical staff (n = 204) employed in health‐care settings providing direct care for patients in hospitals, as well as 490 adults not employed in a health‐care setting (Table S1). The study was approved by the Institutional Review Board at Renmin Hospital of Wuhan University (No. WDRY2020‐K004). Detailed methods and results are available in the Supplementary Information.

We assessed the effect of SM use and psychological and emotional distress (according to the Hyperarousal, Intrusion, and Avoidance subscales of the 22‐item Impact of Event Scale – Revised [IES‐R]) on depressive symptom severity (according to the 9‐item Patient Health Questionnaire [PHQ‐9]). Greater IES‐R and PHQ‐9 scores indicate greater severity. Participants were asked about their use of SM to obtain information about COVID‐19.

We analyzed PHQ‐9 score as a continuous outcome variable using generalized linear models with a negative binomial distribution and as a dichotomous outcome variable using binomial logistic regression models (reported in Supplementary Information). We evaluated the synergistic effect of prolonged SM use to obtain information about COVID‐19 and psychological and emotional distress as a result of the epidemic on the risk for depression in Mainland China. We evaluated whether the odds of depressive symptoms with more prolonged SM use and greater psychological and emotional distress were significantly greater than the sum of the odds of depressive symptoms with more prolonged SM use alone and with greater distress alone. We calculated a synergy index and relative excess risk due to interaction to model interaction effects, with adjustments for age, sex, educational attainment, marital status, living arrangements, and health‐care/non‐health‐care‐worker status separately for each IES‐R subscale. ⁴ , ^{5 .}

The mean (standard error) PHQ‐9 score among study participants was 5.2 (0.1), denoting the presence of clinically significant depressive symptoms. Approximately 18.1% (n = 554) of all participants reported spending less than 1 h per day on an SM platform in the past week, 41.6% (n = 1306) reported spending 1–2 h per day, 22.5% (n = 689) reported spending 3–4 h per day, and 16.8% (n = 515) reported spending more than 5 h per day on an SM platform. Greater time spent on SM predicted greater depressive symptom severity (Fig. S1). IES‐R Intrusion and Hyperarousal subscale scores significantly predicted PHQ‐9 scores, while the Avoidance subscale scores did not (Table S1).

Individuals reporting both prolonged SM use (i.e. ≥3 h per day) and significant symptoms of distress, particularly hyperarousal, had significantly higher odds of having depressive symptoms or probable depression relative to individuals with either factor alone (Fig. 1). That is, the odds of depression with prolonged SM use and significant hyperarousal symptoms were significantly greater than the sum of the odds of depression with prolonged SM use (in the absence of significant hyperarousal) and hyperarousal (with reduced SM use), as instantiated by a positive synergistic effect (Table S2).

Figure 1.

The co‐occurrence of prolonged social media (SM) use and significant distress increases odds for depression. Odds ratio, relative excess risk due to interaction (RERI), and synergy index were calculated with adjustment for age, sex, educational attainment, marital status, living arrangements, and health‐care/non‐health‐care‐worker status. CI, confidence interval; IES‐R‐A, Avoidance subscale of the 22‐item Impact of Event Scale – Revised; IES‐R‐H, Hyperarousal subscale of the 22‐item Impact of Event Scale – Revised; IES‐R‐I, Intrusion subscale of the 22‐item Impact of Event Scale – Revised; PHQ‐9, 9‐item Patient Health Questionnaire. () Relative Excess Risk due to Interaction (RERI). () Synergy Index.

SM networks can be used to provide reassurance, increase public awareness about effective ways to reduce risk of infection, and communicate practical information to curb public panic and reduce the mental health burden of public health crises. ⁶ However, SM use is also associated with elevated risk for depression: greater symptoms of depression and loneliness are observed in young adults who use SM extensively. ⁷ , ⁸ Moreover, during public health crises, SM can aggravate public fear and panic: for example, SM networks have been implicated in the spread of false information and amplification of risk and harm during the 2014 Ebola outbreak. ⁹ There is an urgent and unmet need to address the impact of COVID‐19 on the mental health of affected individuals.

Data are available on request from the authors.

Disclosure statement

R.S.M. has received research grant support from the Stanley Medical Research Institute, CIHR/GACD/Chinese National Natural Research Foundation; speaker/consultation fees from Lundbeck, Janssen, Shire, Purdue, Pfizer, Otsuka, Allergan, Takeda, Neurocrine, Sunovion, and Minerva. All other authors declare no competing interests.

Supporting information

Appendix S1 Supplementary information.

Figure S1 Mean 9‐item Patient Health Questionnaire (PHQ‐9) scores are significantly higher among individuals with more prolonged social media use. Marginal means reported after adjustment for age, sex, educational attainment, marital status, living arrangement, and health‐care/non‐health‐care‐worker status.

Table S1 Demographics and summary of model effects on depressive symptom severity (according to the 9‐item Patient Health Questionnaire [PHQ‐9] total score as a continuous variable).

Table S2 Predictors of depressive symptoms.