The Effect of Lockdown on the Mental and Physical Health of Romanian Medical Undergraduates

As a response to the COVID-19 pandemic, medical universities were closed in Romania, between March 2020 and July 2020. The educational process was transferred to an on-line environment, in person contact becoming nonexistent outside the group of people from one’s household. The purpose of this study is to provide an overview of the mental health status in medical undergraduates in Romania during lockdown. A questionnaire was used in April 2020 to collect information regarding the life of students during social isolations and to assess the mental health of students using a self-reported 1-7 scale on 21 relevant endpoints. After data quality control, the remaining 1978 data entries were analysed in SPSS 20. We identifi ed multiple factors that may contribute to the mental health of undergraduate students, such as: partner status and offline contact, personality traits, social media and streaming platforms usage. There were differences regarding the mental health status indicated by endpoints between: genders, living environment, owning a pet, etc. This study describes how mental health, a neglected subject in the Romanian public health arena, is influenced by various factors during lockdowns. Further research should be conducted and appropriate measures should be taken to prevent the degradation of mental health in medical students under pandemic circumstances.

Th e fi rst part of the questionnaire consisted of questions regarding the university center, year of study, gender, the living environment after lockdown, the number of people/household, information on pet ownership and opinion on its eff ect on mental health, the use of social media and streaming platforms and partner's status during pandemic. All the above-mentioned questions had a pre-established list of possible answers, and the responders selected the option corresponding to their situation.
Personality traits, such as openness and optimism, were also evaluated using a scale from 1 to 7. For openness, a lower score is corresponding to introversion, while a higher score refl ects extroversion. Th e scale for optimism used lower scores for pessimist responders, and higher scores for the optimists. Th e inclusion in one of these categories was based on self-evaluation of openness and optimism.
Th e last 21 questions were representing the endpoints regarding the mental health of medical students. Each one of them consisted of an affi rmation associated with a scale from 1 to 7, 1 corresponding to the lowest agreement (disagreement) and 7 to the highest agreement. Th e responders selected the score that refl ects a subjective evaluation of their mental health status. Each affi rmation was corresponding to one of the endpoints, whose abbreviation and descriptive statistics can be found in Table 1.

INTRODUCTION
On the 11 th of March 2020, the World Health Organization recognised COVID-19 as a pandemic. Th is resulted in a series of measures designed to reduce the spread of the disease. Travel restrictions and social distancing were enforced.
As a response to the increasing cases of COVID-19, Romania declared lockdown, personal mobility being reduced to essential trips only. Another measure that was implemented with the purpose of reducing transmission was the closure of schools and universities, as this has been proven to have an eff ect on the amount of students infected 1 . Businesses and education were transferred to an online environment, in person contact becoming virtually nonexistent outside the group of people with whom one shared their home. While this resulted in a decrease in transmission, the lockdown might have had unintended consequences. As our activity turned to the internet, so did our social networks.
Th ough the current literature provides us with information about the psychological eff ects of an infectious viral outbreak 2 there is no information on the eff ect of prolonged lockdown during a pandemic. While the eff ects of perceived social isolation [3][4][5] and the association between social media usage and loneliness6 have been known for some time, the information is lacking on how these factors interweave in the context of an enforced lockdown. Th ere is also insuffi cient information available regarding the eff ect of streaming ser vices on mental wellbeing.
Th e purpose of this article is to provide an overview of the mental health situation in medical undergraduates in Romania and to uncover certain behaviors that have a positive infl uence on mental health during prolonged isolation periods.

EXPERIMENTAL (MATERIALS AND METHODS)
Our survey was carried out during the month of April 2020, one month after lockdown measures had been initiated in Romania. Th e study was conducted with the use of the Google Forms platform, and was distributed through the Facebook groups of medical students from across Romania with the help of the Romanian Federation of Medical Students' Association. Th e study aimed to assess the medical students' response to social isolation, to identify behaviors that had a positive eff ect on the mental health of the undergraduates and to provide recommendations should a situation that requires lockdown emerge again. 1022 responders confi rmed having a pet in their household. We have found no statistically signifi cant diff erence between people who have a pet in their household and those who don't on any of the endpoints. 1668 responders believed that a pet would make the isolation experience more manageable. Th ere was a highly signifi cant diff erence regarding the opinion on the infl uence of pets between the responders who owned a A detailed model of the questionnaire is provided in Appendix A. Given that the method of collecting questionnaires could not attest for the identity of the person completing, a couple of methods were put in place to prevent maliciousness. Using the option provided by the platform, the questionnaire limited responses to one per IP. Obvious malevolent responders had been eliminated based on their answers on the open ended questions. Th is removed 9 answers from the total 1987.
Th e remaining 1978 data entries were analysed in IBM SPSS Statistics 20. All of the endpoints (scale variables), except for Understanding, followed normal distribution. A T test has been employed to verify the existence of an association between two groups (defi ned by the answers to categorical questions) for the endpoints that presented normal distribution. Th e endpoints that did not follow normal distribution were analysed with the Mann Whitney test. For those variables showing signifi cance, the p value and the Mean Diff erence (MD) has been reported. For the scale questions (number 13-33) a cut-off of 0.3 (amounting to 5% of the possible range) has been chosen as the level of signifi cance for the Mean Diff erence (MD). MDs of more than 0.25 have been mentioned as tendencies.
Th e eff ects of Openness and Optimism on the endpoints, and the eff ect of diff erent endpoints on each other have been analysed with a linear regression. Only correlations with an R^2 of at least 0.05 have been included.  While the psychological or physical eff ects of extended periods of isolation have been studied on populations such as prison inmates housed in solitary housing units 7 , on ground spacefl ight simulations 8,9 and Antarctic research facilities 10 , the data is lacking on how a prolonged isolation period aff ects the population at large. Our study aims to provide an overview of the psychological eff ects of prolonged lockdown and to serve as a reference point for health policy decision makers.

RESULTS AND DISCUSSIONS
Social isolation can be defi ned either as social isolation per se (the lack of social relationships) and percei-pet and those who did not (p<0.0001). As expected, the majority of `No` responders to this question were those who did not own a pet.
People who have reported that they believe a pet to be useful in making their lockdown experience more manageable have reported higher scores on Overeating (p=0.03, MD=0.30  scores. Th is leads us to believe that the real determinant of perceived social isolation is not the existence of an intimate relationship, but much rather the possibility of frequent offl ine contact with one's signifi cant other. Th e markers of perceived social isolation (Loneliness, Friends and belief in the positive eff ect of a pet in the household) have been linked to a poorer response on our endpoints. Th ese fi ndings support the results found in the literature [14][15][16] . However, given the context in which the survey was realised, the causal relationship between stress and loneliness 5 is worth considering. Given the results we conclude that perceived social isolation is a more important determinant of mental health during a prolonged lockdown than actual physical isolation.
In the current literature there is a known correlation between social media usage and perceived social isolation, but the causal relationship between the two has not been clearly established 6,13 . Our fi ndings suggest that social media usage isn't the cause of the increase in loneliness but much rather an eff ect of it. Th is hypothesis is supported by the existing literature 13 . However, given the disparity between people who have maintained their SMU habits and those who have either reduced or increased it, we propose that social media does not necessarily cause perceived social isolation, and in some cases might actually decrease the feeling of isolation one might face during a prolonged isolation period. Th is disparity between social media users is supported by the literature 17 .
ved social isolation (the feeling of lacking meaningful relationships) 11 .
Even though the two aspects are related, one can perceive a feeling of isolation while having strong social ties, and vice versa 11 . Out of the objective indicators of social isolation during the pandemic, the number of people in the household provided no signifi cant correlations. Th is might be explained by the normal living situation of medical undergraduates. Th e majority of medical undergraduates study away from home, seeing their families and home at a frequency ranging from a few times every month to a few times a year. Th e only signifi cant diff erences between students spending their lockdown with their family as opposed to away from them was an increase in anger which can be explained by the infringement of autonomy and intimacy that students experience with their families 12 . From this we can conclude that the major eff ects of isolation were not caused by the undergraduates' ability or inability to see their family, but much rather the physical separation from their other relationships.
Th is fact is corroborated by the diff erences in Loneliness and Interaction experienced by CoLocked and SoLocked undergraduates. Th is decrease in perceived loneliness by students who were living with their signifi cant other is corroborated by literature 13 . One interesting fi nding is that SoLocked undergraduates reported no signifi cant diff erence in Depression, Interactions and Loneliness when compared to their Single counterparts, but reported higher Frustration and Anger  According to our study, using social media more during lockdown had both positive and negative infl uence on the mental health of responders, but the negative aspects were dominant. Th e undergraduates who used social media more than before COVID-19 pandemic considered their mental stimulation less adequate, perceived a higher level of loneliness, frustration, anger, stress, sedentarism, sleep problems, time wasting, overeating, concentration problems, time management diffi culties, and lower energy, self-determination and perception of having enough distracting activities. However, this category of students felt less depressed, more active and reported a lower level of purposelessness and they claimed to have enough people to discuss with during lockdown.
Th e relationship between streaming platforms usage and the endpoints follows the pattern of social media usage, the diff erence is that higher streaming platforms use is associated with lower activity and there was no statistically signifi cant diff erence with Friends endpoint. Th ough the literature concerning SPU is rather lacking, we are proposing that the causal eff ect between time spent on streaming platforms and feelings of loneliness, concentration problems, overeating and sedentarism follow in part the same pattern as SMU, though the degree to which it infl uences the various endpoints diff ers. We believe that loneliness causes an initial increase in SMU and SPU, but consequently, as the use becomes more passive rather than active, the time spent in the on-line medium contributes to a worsening in mental health outcomes, creating a vicious cycle of loneliness and passive online escapism. Th e negative eff ect of passively using social media is supported by the existing literature 6 .
Our study also highlighted the importance of personality traits on the mental health of medical students during COVID-19 lockdown. Introverts were more satisfi ed with the level of interaction, while extroverts felt less depressed. Introverts also reported to have fewer activities to distract them and to perceive less self-determination. Extroverts claimed to feel more active and