Loneliness During the NPIs for the COVID-19 Pandemic in Norway: Risk Factors and Associations With Psychopathology

Overview

The present study of loneliness during the COVID-19-related NPIs is part of a larger project aiming to investigate psychological reactions and symptoms associated with the current and ongoing governmental initiatives in place in Norway. The results will lead to a better understanding of the psychological effects on quarantine on the population and consequently will be relevant for the development of tailored prevention and intervention programs fit for pandemic crises. Objectives Investigate the levels of loneliness in a general population during the strict social distancing government-initiated non-pharmacological interventions (NPIs) for the COVID-19 pandemic. Investigate risk and resilience factors for loneliness and the associations between loneliness and psychopathology symptoms during the NPIs. Aims Inform the general public, policy makers, scientists, and health practitioners about the associations of the NPIs with the mental health problem of loneliness and its potential effect on psychopathology. Provide a foundation for policymakers and health-care professionals to employ interventions that protect the general public against increased psychological suffering and dysfunction during society's handling of pandemics.

Objectives Investigate the levels of loneliness in a general population during the strict social distancing government-initiated non-pharmacological interventions (NPIs) for the COVID-19 pandemic. Investigate risk and resilience factors for loneliness and the associations between loneliness and psychopathology symptoms during the NPIs. Hypotheses (Hs) H1: The social distancing measures used to cope with the pandemic are associated with increased loneliness. Exploratory: Investigate the differences in levels of loneliness across different demographic subgroups in the sample. H2. With regard to risk factors existing before the NPIs, lower age, lower educational level, not being in a permanent relationship, not having children, not being in work, and having a psychological diagnosis will be associated with more loneliness during the NPIs when the potential effects of gender and childlessness are controlled. H3: The NPIs-related state factors of worry about job and/or economy, worry about health (health anxiety), and worry and rumination in general are associated with more loneliness, above and beyond the influence of pre-existing risk factors. H4: More loneliness will be associated with more depressive symptoms and generalized anxiety symptoms. Because variables significantly related to loneliness may confound the relationship between loneliness and psychopathology symptoms, the variables confirmed for Hypothesis 2 and 3 are controlled. Presence of psychological diagnosis as an indicator of pre-NPI symptoms will be used as a covariate, whether or not it is supported for Hypothesis 2. Exploratory: Does emotional support - as partly a counterpart to loneliness - explain symptoms over and above loneliness? Method The present cross-sectional study is part of The Norwegian COVID-19, Mental Health and Adherence Project, and is conducted in accordance with the guidelines of the Strengthening the reporting of observational studies in epidemiology statement (STROBE).15 The pre-registered protocol of this study can be found at Clinicaltrials.gov (Identifier: NCT ). All elements of the submitted study adhere to the pre-registered protocol. Study design and participants In this cross-sectional and epidemiological study, we investigated the mental health status and compliance rates to NPIs of the general adult population across all 11 counties in Norway during the ongoing COVID-19 pandemic. These counties serve a geographically defined area of approximately 5·3 million individuals, the population of Norway, all serving under identical nationally initiated NPIs during the period of data collection. Eligible participants were, all adults including those of 18 years and above, who are currently living in Norway and thus experiencing identical NPIs, and who had provided digital consent to partake in the study. The period of data collection lasted seven days and was between March 31st 2020 and April 7th 2020, which encompasses a time-frame where all NPIs were held constant during the two weeks prior to data collection, as well as during the data collection week. Ethical approval of the study was granted by The Regional Committee for Medical and Health Research Ethics and the Norwegian Centre for Research Data (reference numbers: 125510 and 802810, respectively), where the study protocol and analysis plan was approved prior to data collection. Procedure The set of questionnaires in this study were hosted online on a platform belonging to the University of Oslo referred to as Services for Sensitive Data, which are nationally approved for dealing with person-sensitive material. The dissemination of the questionnaire was systematically conducted through various channels with the goal of obtaining a probability sample. Due to the time-sensitive nature of the study with the aim of measuring mental health during a period with identical NPIs, we could not disseminate our questionnaire through conventional methods such as access of registry data, as access to registry information involves a time-consuming and difficult approval process with regards to privacy regulations in Norway. Such an application timeline could therefore bias the NPI parameter, where it was highly likely for NPIs to change during or closely following the application process. Consequently, we disseminated our questionnaire systematically through various national, regional and local platforms, with the goal providing the entire population with an equal opportunity to participate in the survey, and thus approximating a probability sample. Therefore, the survey was disseminated in the following six ways; 1) through broadcasting on the national news channel of Norway with approximately 1·1 million viewers at the time of broadcast; 2) through random targeting of any individual residing in Norway that is 18 years or above through Facebook-advertisement, with imputed parameters obtaining a population of 3·6 million, and the number of individuals reached once advertisement was over encompassing a random selection of 174 885 of these 3·6 million individuals. We also informed about the survey on 3) national radio stations; 4) regional and local radio stations across the country; 5) in national newspapers; and 6) in regional and local newspapers across the country. The adult population of Norway is approximately 4·2 million. Consequently, only the Facebook-advertisement includes nearly 85% of this population, reaching a random sample of 174 885 individuals with the survey. We estimate to have reached the latter 15% (600 000 individuals) of the adult population that are not on Facebook through the five other dissemination procedures, such as broadcasting on national television with over a million viewers, broadcasting on national, regional and local radio stations, as well as newspapers. The stopping rule for data collection was designed to ensure that the identical NPIs were held constant across all counties for two weeks prior to and the week during the data collection period, as well as controlling for expectation effects by stopping data collection instantly once information to the general population concerning modification of NPIs were given. Measurement Participants were asked to fill out a survey including demographic variables, validated measures of psychopathology symptoms, trait variables, situational factors related to the consequences of the COVID-19 virus, cognitions, beliefs and fears related to COVID-19, as well as adherence to government-initiated non-pharmacological epidemiological measures (NPI's). The demographic variables include sex, self-reported identification with sex, age, education, ethnicity, partner relationship status, number of children and employment status. Also presence of psychological diagnosis was assessed. Symptoms of health anxiety was measured with two items from the validated Health Anxiety Inventory (HAI)18 as well as an item measuring specific fear of being infected with coronavirus and an item measuring fear of dying from the coronavirus. Worry about job and economy the last two weeks was measured by the items: "I am worried that I will lose my job" and "I am worried about my economy", using a 0 (never) to 3 (almost every day) Likert-type scale. A general worry and rumination item was taken from the Cognitive Attentional Syndrome-1 (CAS-1; Wells, 2009): "How much time in the last week have you found yourself dwelling on or worrying about your problems? (0-8 Likert-type scale). Power analysis The mentioned Norwegian COVID-19, Mental Health and Adherence Project involve multiple studies, where some involve a complex systems (network analysis) approach. These multivariate analyses require large samples and power analysis was conducted accordingly. Following power analysis guidelines by Fried \& Cramer (2017), it is recommended that the number of participants be at the very least three times larger than the number of estimated parameters. However, more conservative recommendations by Roscoe (1975) for multivariate research, recommends sample size that is ten times larger than the number of estimated parameters. In a network analysis, the number of estimated parameters follows the following formula: (N x N-1) / 2, where N = number of nodes (questions asked). Our most extensive planned network analysis includes about 40 nodes, needing a sample size about 8000 according to the more conservative estimates of Roscoe. To ensure a required amount in the subgroup of health care professionals, the required size of the total sample was set to 10 000. Statistical analyses Hypothesis 1 The levels of loneliness (ULS-8) will be compared to studies measuring loneliness by the ULS-8 scale in general and clinical populations using independent t-tests. Hypothesis 2 Hierarchical regression will be conducted using loneliness as the dependent variable. In the first step, the relatively stable characteristics existing before the NPIs are included: age, gender, partnership status, work status, having children, and having a psychological diagnosis. Hypothesis 3 In the second step, the NPI-related situational variables worry about job and/or economy, health anxiety, and the variable worry and rumination in general will be included. Hypotheses 4 Two separate hierarchical linear regression analyses will be conducted using depressive symptoms (PHQ-9) and generalized anxiety symptoms (GAD-7) as dependent variables. In both analyses, the variables significantly predicting loneliness in the analysis of Hypothesis 2 as well as presence of psychological diagnosis will be included. In all regression analyses, multicollinearity and other assumptions will be checked; in particular if the multicollinearity assumption is violated (if VIF \> 5 and Tolerance \< 0.2; Hocking, 2003; O'Brian, 2007).