Coronavirus (COVID-19): mental health tracker study - wave 2 report

1. Background

1.1 Study overview and aims

In December 2019, a novel coronavirus was identified in Wuhan, China. Since then the associated disease COVID-19 has affected millions of people worldwide.

In addition to the physical health impact, there is growing evidence of the effects of the COVID-19 pandemic on mental health and wellbeing that will extend beyond those who have been directly affected by the virus (Holmes, O'Connor et al., 2020: O'Connor et al., 2020). As a result, it is important to continue to monitor population-based health and mental health outcomes to detect groups who may be most affected by the COVID-19 pandemic and associated restrictions. We know from the SARS outbreak in 2003 that anxiety increased, suicide rates also increased in some groups (e.g. Yip et al., 2010; Gunnell et al., 2020; Tsang et al., 2004) and that suicidal thoughts increased in the early phase of the pandemic in the UK (O'Connor et al., 2020). Although initial findings on the impact of the COVID-19 pandemic on suicide rates globally were reassuring, recent data highlight the need for vigilance (John et al., 2020). We need to act now, therefore, to understand and mitigate the mental health risk in Scotland as we continue to respond to the COVID-19 pandemic.

The Scottish COVID-19 (SCOVID) Mental Health Tracker Study is part of a UK-wide study ('Tracking the impact of the COVID-19 pandemic on mental health and wellbeing (COVID-MH) study') which started on 31^st March 2020 just after lockdown measures were imposed. Adults aged 18 years and older took part in this survey. In May 2020 the Scottish Government commissioned an additional Scottish sample to allow the tracking of the mental health and wellbeing of the Scottish population over a 12-month period. The Wave 1 survey ran from 28th May to 21st June 2020 which coincided with the Phase 1 easing of lockdown measures in Scotland^[10]. Findings from the Wave 1 survey were reported in the Scottish COVID-19 Mental Health Tracker Study: Wave 1 Report^[11]. The Wave 2 survey asked as many of the same respondents as possible about their mental health between 17^th July and 17^th August 2020. This allows us to track changes in mental health and wellbeing. The Wave 2 survey coincided with the Scottish Government's introduction of Phase 3 of the easing out of lockdown, which included an increase in the number of households that could meet indoors and outdoors, and the opening of indoor hospitality.

The Scottish survey measures are aligned with the COVID-MH study to allow direct comparisons with other regions of the UK. Findings^[12] were recently published from the UK COVID-MH study covering 3 waves of data from the start of the first lockdown (Wave = 31^st March to 9^th April 2020, Wave 2 = 10^th April to 27^th April 2020, and Wave 3 = 28^th April to 11^th May 2020). The results suggest that rates of suicidal thoughts increased over the waves, whereas rates of anxiety symptoms, and levels of defeat and entrapment decreased across waves, and rates of depressive symptoms did not change significantly. Additionally, positive mental well-being increased (O'Connor et al., 2020).

Wave 3 of the UK survey most closely corresponds to timing of Wave 1 of the SCOVID Mental Health Tracker study. By Wave 3 of the UK study, 9.8% of respondents reported suicidal thoughts in the past week, which closely resembles the 10.2% reported in Wave 1 of the Scottish survey. Similarly, rates of moderate to severe depressive symptoms were 23.7% in Wave 3 of the UK study, and 25.3% in the Scottish study. By Wave 3 rates of anxiety had decreased in the UK study to 16.8%, lower than the 19.1% reported in Wave 1 of the Scottish survey, although these rates had been higher in previous UK study waves (Wave 1 = 21.0%, Wave 2 = 18.6%). Overall, these findings suggest some consistency in rates of suicidal thoughts, an depressive and anxiety symptoms between Scotland and the UK.

The findings from the SCOVID Mental Health Tracker Study will help us to understand the impacts of the pandemic on the Scottish population's mental health and wellbeing, particularly the differential impacts on sub-groups of the population. The Wave 2 survey will aid with the tracking of these mental health outcomes as we navigate different levels of restrictions.

Key research aims for Wave 2 of the SCOVID Mental Health Tracker Study

1. To track changes in people's mental health and wellbeing in Scotland during the COVID-19 pandemic and easing of government restrictions. Specifically, changes in mental health and wellbeing from Phase 1 of restrictions (Wave 1 survey; 28^th May to 21^st June 2020) to the Phase 3 of restrictions (17^th July and 17^th August 2020).

2. To provide an overview of contextual factors during the COVID-19 pandemic and easing of government restrictions.

1.2 Methodology

Wave 2 recruitment for the SCOVID Mental Health Tracker Study occurred between 17^th July and 17^th August 2020, which coincided with the Phase 3 easing of Scottish Government restrictions in Scotland. These easing of restrictions included a significant relaxing of lockdown restrictions. In brief, on 10^th July 2020, five households (up to 15 people) were allowed to meet outdoors, and three households (up to 8 people) were allowed to meet indoors, and there were no longer any restrictions on travel. On 15^th July, childcare providers, indoor hospitality, hairdressers, museums, galleries, holiday accommodation, non-essential retail in shopping centres, and places of worship were allowed to open.

Recruitment was conducted by Taylor McKenzie, a social research company. At Wave 1, members of an existing online UK panel (Panelbase.net) were invited by email to take part in an online survey on health and wellbeing. These respondents also agreed to be followed up over subsequent waves initially timed at around 6, 12, 24 weeks, and 12 months following Wave 1 but with the flexibility to be responsive to policy changes related to the COVID-19 pandemic response. Consistent with this, the Wave 2 survey was launched approximately 6 weeks after the Wave 1 survey to coincide with the Phase 3 easing of lockdown on 10^th July 2020. Figure 1.1 provides an overview of key events/policy decisions in the UK in relation to the COVID-19 mental health tracker studies.

At the Wave 1 survey recruitment, a quota sampling methodology was employed to recruit a close to national sample of adults (n= 2,604) from across Scotland. Quotas were based on age, gender, housing tenure, and highest educational qualification. To gain insight into the mental health and wellbeing of those living in urban/rural areas and within different NHS Health Boards, further quotas based on location within Scotland were also recruited. The majority of the quotas were met (see Wave 1 report for further detail) however, individuals without educational qualifications were underrepresented in the sample.

The Wave 2 recruitment was launched on 17^th July 2020 and all Wave 1 respondents were invited to take part by email. A total of 65.4% respondents from Wave 1 took part in this survey (n=1703). This attrition rate (i.e., loss to follow-up) was higher than anticipated, and meant that a number of demographic groups are under-represented in the Wave 2 findings. For example, many of the young men from Wave 1 did not take part in the Wave 2 survey. The differences in profiles of the sample in Wave 1 to Wave 2 are outlined in Table 1.1. As with the Wave 1 data, the Wave 2 data was weighted to reflect the Scottish population, and this accounted for the loss of respondents at follow-up between Wave 1 to Wave 2. Unweighted data is provided in the study annex (Tables A-D2). Although overall trends were the same with or without weighting applied, we do note that some subgroup findings should be interpreted with caution, in particular young men, due to their small sample sizes.

Within the Wave 2 survey, respondents were asked to complete questions on mental health and wellbeing including measures of anxiety, depression, distress, mental wellbeing, loneliness, defeat, entrapment, and self-harm as well as social support. A range of questions exploring contextual factors such as sources of emotional and social support and lifestyle factors were included along with perceptions, experiences of, and the impact of COVID-19 related restrictions.

Within the report, inferential statistical tests^[13] were used to investigate changes in mental health and wellbeing from Wave 1 to Wave 2, as well as differences between key subgroups. Due to loss to follow-up from Wave 1 to Wave 2 it was not possible to conduct some of subgroups analyses for the Wave 2 report. This is because their samples would be too small for robust and reliable analyses, and/or the Wave 2 subgroup no longer reflected the composition of the original subgroup at Wave 1.

Specifically, the Black, Asian and Minority Ethnic (BAME) group lost 60% of respondents to follow-up, leaving just 49 people in the Wave 2 sample, which was deemed too small for statistical analysis. Additionally, the '5+ hours unpaid carers group' and 'those with dependents under 5 years' were removed (those with '5+ hours caring responsibilities' reduced by 40%; 'those with under 5's' reduced by 50%) (see Table 1.1). Finally, sub-group analyses for the shielding (n=88) group was not conducted. The high-risk group^[14] was not included in analysis, instead we have conducted analyses for those with a pre-existing health condition (n=516), as this group will include many of those who are at high risk of COVID-19. Finally, although nearly 70% of young adults (18-29 years) were lost to follow-up, they were retained in the analysis as they represented over 10% of the sample (n=177). Despite this, some of the analysis with young adults (particularly young men) should be interpreted with caution.

^a These groups were dropped from the Wave 2 analysis; ^b This group was added to the Wave 2 analysis

Layout of Report Findings

The subgroups included within the Wave 2 analyses were: age, sex, socio-economic grouping, a pre-existing mental health condition, a pre-existing physical health condition, those with dependents under 16 years, carers, living alone, living rural or urban, key worker, and change in working status groups. The report focusses on the statistically significant changes for these subgroups from Wave 1 to Wave 2 on the various mental health outcomes, and the differences between key subgroups at Wave 2, rather than discussing findings for each of these subgroups according to each study measure.

In addition, this report uses particular terms to describes the rates of particular mental health outcomes reported by subgroups within the overall sample, and the degree to which an outcome is being experienced. The term 'rates' refers to the proportion of respondents within a named subgroup who have reported a particular outcome; it does not describe the degree of a particular outcome. For example, an increased rate of men reporting moderate to severe depressive symptoms means that a higher proportion of men have reported these symptoms; it does not mean that men as a subgroup are experiencing more severe depressive symptoms overall. The term 'level' refers to the degree to which a particular mental health or wellbeing measure is being experienced. For example, stating that older adults reported higher levels of mental wellbeing than younger age groups means that the average mental wellbeing score for older adults was higher than the average score for younger groups.

The main body of the report focuses on the changes from Wave 1 to Wave 2 on the core mental health outcomes of depressive symptoms, anxiety symptoms, suicidal thoughts, mental distress and mental ill-health, and mental wellbeing for the subgroups outlined above. Contextual measures, such as lifestyle factors and employment status, are reported briefly. However, as they are not main outcomes only a selection of subgroup analyses are reported. The annex contains more detailed information on contextual factors.

Ethical approval was obtained on 21^st May 2020 from the University of Glasgow's Medical, Veterinary and Life Sciences ethics committee to add a Scottish only sample to the existing UK study being led by the University of Glasgow (UK COVID-MH Ethics approval: 200190146).