This study investigated the change in depressive, anxiety and attention deficit symptoms of college students who had previously applied to a university mental health center before the onset of COVID-19 in Turkey. Compared to the pre-pandemic period, students were found to have decreased symptoms of depression, anxiety, and attention deficit. Since high levels of stress, anxiety and depression in college population is a frequent finding in the literature, a modest decrease in symptomatology after the onset of COVID-19 pandemic is a valuable finding. Decrease in stress levels and symptoms of social media use disorder were the predictors of this change, which is also another novel finding of our study. Although the screen time increased during pandemic, the level of social media use disorder decreased. The levels of depression and attention deficit are affected by social media use disorder both directly and indirectly through stress level. The level of anxiety is affected by fear of COVID-19 both directly and indirectly through stress level. So, stress thermometer was mediating the relationship between i) social media use disorder and depression scores, ii) social media use disorder and attention deficit scores, iii) fear of COVID-19 and anxiety scores.
At the time of the recruitment for the study, the campus had been closed for about three months and the Turkish government had further implemented a number of social distancing measures. Despite many regulations restricting social life, the decrease in psychiatric symptomatology was an unexpected finding. Majority of studies reported increased mental health problems and a large scale nation-wide survey from China reported that massive media exposure and prior mental health problems are associated with increased risk of college mental health problems in the early phase of the pandemic 15. Only few studies such as a study from USA conducted with Hispanic adolescents have reported decreased mental health problems, similar to our results 19. Despite all its drastic effects, some individual stressors related to academic life such as rigid study time tables or social stressors such Fear of Missing Out (FoMO) and competitive environment may be reduced at the time of crisis 34 and these factors might be related to decreased stress levels in the early phase of pandemic.
It is expected that the closure of the campus will result in disconnection from friends and romantic partners, thus creating risks of loneliness and isolation. However, in our sample, the majority of the students were living with their families or friends during lockdown which possibly minimized the negative effects of isolation. Even though the frequency of alcohol use and living with parents predicted the psychiatric symptoms in the first regression model, their effects became insignificant when SMDS and stress thermometer were added in the model. The reunion of the students, who stayed in the dormitories during the pre-pandemic period, with their families may have been beneficial in terms of social support and meeting basic needs. On the other hand, the fact that young adults may be asymptomatic carriers, might have increased the students’ anxiety about transmission of COVID-19 to their elderly family members35. Eventually, going back to the family house may increase social support and decrease household responsibilities but this protective effect might have been mixed with the stress of infecting the family members and possible adverse effects of the negative environment in some families. Of course, not every family is supportive, some families may have restrictive attitudes. There may also be concerns about covering online classes due to the lack of proper internet access and distraction due to a crowded family environment. A cross-sectional study from USA reported that over 40% of the students were worried about their families who were more vulnerable and nearly half of them mentioned that their home is a distractive environment 36. Therefore, increased stress due to the pandemic can only to some extent be compensated by returning to the family home. The insignificant effect of living with family on mental health in this study might be a result of its mixed negative and positive effects.
According to mixed effect regression analysis, decrease in stress and social media use disorder compared to pre-pandemic period, were found significant predictors of decrease in anxiety, depression and attention deficit symptoms. A mediator analysis revealed that stress thermometer had a mediator effect on the relationship between SMDS and PHQ-9 as well as between SMDS and ASRS but not on SMDS and GAD-7. Thus, decrease in problematic social media use yielded a decrease in depressive and attention deficit symptoms by both directly and indirectly by decreasing the stress levels. This finding is in line with previous studies which reported that, internet and smart phone addiction increased the depression levels during pandemic 37,38. Although there are several studies showing a mediator effect of addictive social media use on anxiety 14,39, social media disorder did not mediate the relationship between the decrease in stress levels and anxiety scores in this study. It can be predicted that the use of social media will increase in a period when physical interaction is restricted. In our sample, the decrease in the SMDS score despite the increase in screen time can be explained by a possible change in perception of social media use. Social media use behavior may have been somewhat normalized during the pandemic, and the possibility of perceiving intense use as pathological may have decreased. Moreover, increased screen time may be due to the time spent for online education, instead of the time spent on social media.
Compared to traditional media, social media has played a multitude of positive roles in information exchange during the COVID-19 crisis, including disseminating health-related recommendations, enabling connectivity and psychological first aid 40 showing public attitudes, experience, and perception of the disease as well as sentiment to the government 41. So, when more people use social media to seek and share health information, social media usage can provide an informational, emotional and social support 42. A Chinese study which reported similar rewarding effect of social media use concluded that, social media usage helped manage stressors and health risks and it can be predicted by the informational, emotional, and peer support they received from the shared health information 43. A recent study showed the dual impact of social media on mental health: using social media for reduced loneliness and entertainment was associated with poorer mental health, and using social media for personal contact and relationship maintenance was associated with better mental health 44.
According to the mediator analysis, not the SMDS but the fear of COVID-19 mediated the relationship between stress thermometer and GAD-7 scores. Thus, low levels of COVID-19 fear yielded a decrease in anxiety both directly and indirectly by reducing the stress levels. This finding can be easily explained by a high correlation between fear of COVID-19 and general anxiety levels, which is supported by our correlation analysis. Studies that measured the levels of COVID-19 fear often report that it is a determinant of participants’ anxiety levels 45,46.
Except for problematic social media use, we could not show any effect of other measured vulnerability factors on stress level. There are variables that were not assessed in this study and are still thought to be possible determinants of the reduction in stress levels. These may include elimination of peer pressure and attendence obligation as well as moving away from academic competition. Despite high levels of uncertainty at the beginning of the pandemic period, worries of a fatal infection and restricted social life; factors such as the elimination of physical participation in classes, update of grading system at Koc University in the form of pass fail under pandemic conditions, elimination of the negative consequences of peer interaction may have made students perceive less stress compared to the active school period. Actions taken by professors, such as reducing course loads, open book exams and other allowances on grading requirements, could also have contributed to alleviating or reducing stress. For students receiving counselling services on campus no access to those services could lead to exacerbation of psychiatric symptoms. So, prompt telecommunication with students and warranting the continuation of mental health services as tele-mental health services can prevent increase of anxiety.
This study has several limitations. First, observational research design does not allow us to make causal claims. Students tend to apply when their symptoms are most intense so the decrease in psychopathology may not be only related to the changing conditions during pandemic but also certain interventions before pandemic. In fact, some symptoms might decrease over time. Although symptoms are likely to subside in some way prior to the pandemic, it has been hypothesized that a significant mental health-threatening stressor, such as a global pandemic, may be a triggering factor in a population with a prior history of psychiatric problems. Since the response rate was 50% and varied by gender and psychopathology severity, the attrition bias can not be excluded. We ran additional analysis to compare the baseline characteristics of respondents and non-respondents and did not find significant differences in terms of pre-pandemic psychiatric symptom scales between them. Second, the scales, used to measure dependent variables, are not diagnostic and are used for screening purposes. Third, relative small sample sizes, sex imbalance and the timing of study (study was conducted 3 months after the lockdown) limits the generalizability of our results. Because the pandemic is still unfolding, social isolation and economic impacts may have worsened over the past 1year period of time. Moreover, the time period between the first and the last evaluation was not constant among participants. All the analyses were done based on mean score that subgroup analysis could not be performed due to the small sample size. Last, not all stress factors such as social support and online education conditions could be measured.
This study has several strengths such as longitudinal design, development of regression models to measure the effect of risk factors. Besides, there were no difference among responders and non-responders in terms of psychiatric measurements and the study was carried out in a well-planned timeframe according to academic calender, out of exam period but before the final grades are submitted that academic concerns were still on going. There could be a relieving effect of termination of final exams, however, since the grades have not been announced yet, we can not say the academic stress is over during recruitment.
Consequently, despite fear, uncertainty and restrictions in the early stages of the pandemic, it may be possible to reduce stress and protect mental health with some interventions. These in-clude providing mental health services as online promptly and making adaptive arrangements for online education and grading system such as switching to the pass/fail system and encour-aging the use of social media for communication and cooperation.