Associations between Depression and Self-Reported COVID-19 Symptoms among Adults: Results from Two Population-Based Seroprevalence Studies in Switzerland

Abstract

(1) Mental health may modulate the perceived risk of SARS-CoV-2 infection. However, it is unclear how psychological symptoms may distort symptom perception of COVID-19 and SARS-CoV-2 infection. We assessed whether depressive symptoms predicted self-reported COVID-19 symptoms, independently of serologically confirmed SARS-CoV-2 infection. (2) Participants (aged 20-64) in the Geneva (N = 576) and Ticino (N = 581) Swiss regions completed the Patient Health Questionnaire before being tested for anti-SARS-CoV-2 IgG antibodies and recalled COVID-19-compatible symptoms on two occasions: April-July 2020 (baseline), and January-February 2021 (follow-up). We estimated prevalence ratios for COVID-19 symptoms by depression scores in interaction with serological status. (3) At baseline, in Geneva, higher depression predicted higher probability of reporting systemic, upper airways, and gastro-intestinal symptoms, and fever and/or cough; in Ticino, higher depression predicted systemic, upper airways, and gastro-intestinal symptoms, fever and/or cough, dyspnea, and headache. At follow-up, in Geneva, higher depression predicted higher probability of reporting systemic symptoms and dyspnea; in Ticino, higher depression predicted higher probability of reporting systemic and upper airways symptoms, dyspnea and headache (all p values < 0.05). (4) We found positive associations between depressive symptoms and COVID-19-compatible symptoms, independently of seropositivity. Mental wellbeing has relevant public health implications because it modulates self-reported infection symptoms that inform testing, self-medication, and containment measures, including quarantine and isolation.

Keywords: COVID-19; depressive symptoms; population-based observational study; seropositivity.

Figure A1

Prevalence ratios (95% confidence intervals) from Poisson regressions estimating likelihoods of reporting COVID-19 symptoms at follow-up across studies. Note: values are prevalence ratios with 95% confidence intervals from Poisson regressions with robust standard errors, adjusted for all listed variables. For Geneva, analyses were further adjusted for the year of PHQ-9 completion. a—excluded from the model when there were not enough positive responses to compute the statistics.

Figure 1

Flow chart for the selection of participants in Geneva and Ticino, Switzerland.

Figure 2

Results of Poisson regressions estimating likelihoods of reporting COVID-19 symptoms at baseline across studies. Note: values are prevalence ratios with 95% confidence intervals from Poisson regressions with robust standard errors, adjusted for all listed variables. For Geneva, analyses were further adjusted for the year of PHQ-9 completion.