The health impacts of a 4-month long community-wide COVID-19 lockdown: Findings from a prospective longitudinal study in the state of Victoria, Australia.

OBJECTIVES: To determine health impacts during, and following, an extended community lockdown and COVID-19 outbreak in the Australian state of Victoria, compared with the rest of Australia.
METHODS: A national cohort of 898 working-age Australians enrolled in a longitudinal cohort study, completing surveys before, during, and after a 112-day community lockdown in Victoria (8 July- 27 October 2020). Outcomes included psychological distress, mental and physical health, work, social interactions and finances. Regression models examined health changes during and following lockdown.
RESULTS: The Victorian lockdown led to increased psychological distress. Health impacts coincided with greater social isolation and work loss. Following the extended lockdown, mental health, work and social interactions recovered to an extent whereby no significant long-lasting effects were identified in Victoria compared to the rest of Australia.
CONCLUSION: The Victorian community lockdown had adverse health consequences, which reversed upon release from lockdown. Governments should weigh all potential health impacts of lockdown. Services and programs to reduce the negative impacts of lockdown may include increases in mental health care, encouraging safe social interactions and supports to maintain employment relationships.

Introduction

Globally, many public health measures have been employed by governments to reduce viral transmission during the coronavirus pandemic. These include restrictions on gatherings, quarantines, movement and travel restrictions, curfews, business closures, the mandated use of personal protective equipment, mass community infection testing and contact tracing, and more recently vaccination efforts. Responses have varied globally in terms of the stringency of measures imposed, and they have also changed over time [1]. Some of the most stringent restrictions relate to mass quarantines or stay-at-home orders, herein described as lockdowns, which have often been coupled with the shutdown of parts of the economy. Lockdowns have been demonstratively effective at reducing the spread of coronavirus [2], although their use has been controversial due to the disruptions they cause to everyday life and adverse health consequences [3]. The World Health Organisation advises the use of lockdowns as short-term measures to regroup, rebalance resources, and protect health workers who are exhausted [4]. Despite this, the use of lockdowns has been widespread globally ranging from 3-day ‘snap’ or ‘circuit-breaker’ lockdowns (e.g. in Australia and New Zealand) to extended periods of more than 100 days (e.g. in Argentina, Azerbaijan, Bolivia, Nepal, UK, Peru, Saudi Arabia, Czech Republic, Greece, Germany, Ireland, and Australia). Approximately half of the world’s population (3.9 billion people) was under some form of lockdown at one stage in 2020 [5].

There are many known harmful side effects of lockdowns [6]. Stay-at-home orders can bring everyday activities to a standstill, leading to loss of work across non-essential industries, and can result in increased loneliness in the community, or social isolation [7]. Whilst some industries are able to adapt to an increased at-home workforce, many businesses have ceased operations during periods of lockdown. In some countries, financial supports such as wage subsidies were available for some eligible businesses or employees to ensure the maintenance of employment relationship during lockdown periods [8]. Personal finances and social interactions are important determinants of mental health, including during pandemic-induced periods of work cessation [9], and both are likely to be impacted during extended periods of lockdown. In summary, existing evidence suggests that lockdowns will have negative impacts on health, particularly mental health, and on several determinants of health including social interactions, engagement in work and financial resources. To date, very little is known about the persistence of these negative health consequences following the cessation of lockdown. Do communities and individuals recover rapidly as communities re-open, or do the adverse health impacts persist?

Our exploratory analyses focus on a group of working-aged individuals enrolled in a prospective longitudinal study of the health impacts of work loss during the COVID-19 pandemic. Participants completed surveys before, during, and following an extended community lockdown in the Australian state of Victoria [10]. Outcomes in this group were compared to those in a comparison group residing outside of the lockdown area over the same period. By utilising the longitudinal nature of our national dataset, and taking advantage of this unplanned natural experiment, we sought to answer the following research questions:

What were the impacts of an extended lockdown related to a COVID-19 outbreak on:

Mental and physical health?

Several determinants of health such as employment, being out of work, social interactions and finances?

Are any effects persistent following the conclusion of lockdown?

Methods

Setting

Australia experienced two waves of coronavirus during the year 2020. The first smaller wave resulted in a similar set of restrictions across states and territories. The second larger wave was mostly localised to the state of Victoria, and more specifically the metropolitan Melbourne area. The state of Victoria is the second most populous Australian state with 6.68 million people (around 26% of Australia’s population), and correspondingly accounts for around a quarter (26.6% during December 2019) of the Australian labour force consisting of 2.90 million Victorian workers [11].

The Victorian state Government response to the second COVID-19 wave included a 112-day extended community lockdown [12]. The nature of the restrictions during the lockdown period changed over time (S1 Table), and between regional and metropolitan areas. During the most stringent period (96.2 / 100 on the government response index [1]), measures included a curfew between 8pm and 5am, a 5-kilometre (i.e. 3 mile) distance limit from home, maximum gathering limits of two people outside the home and no visitors to a home, and mandatory face coverings outdoors and indoors when leaving home [13]. Additionally, there were only four permitted reasons to leave home: shopping for essential items (limited to one person per household), caregiving, or work (where employers were required to support working from home if individuals could work from home), or a maximum of one hour per day of outdoor exercise. A staged easing of restrictions followed the lockdown period (S1 Table). In comparison, for people in the rest of Australia, there were significantly fewer restrictions in place throughout this period, with some differences between state and territory jurisdictions.

Participants, data and context

We report the findings from a prospective longitudinal Australia-wide cohort study on work and health [9] which was initiated soon after the first wave and, by serendipity, coincided with an extended community lockdown in the Australian state of Victoria. Participants were aged 18 year or older, living in Australia, and were employed in paid work prior to the COVID-19 pandemic. An online Qualtrics survey targeted people that had lost work after the first wave and was promoted via social and general media, and newsletters distributed by community sector and industry groups. The cohort also consisted of participants that completed the survey via a telephone interview conducted by a third-party market research company, and included participants that had lost work, and a group whose working hours had not reduced. Baseline findings from this cohort have been published previously including detail on the recruitment methods and measures [9].

Participants were enrolled in the study and completed a baseline assessment via an online or telephone survey between 27 March to 12 June 2020 [9]. Participants then also completed up to a further three surveys by December 2020, at intervals of one, three and six months following the baseline survey. These follow-up surveys included repeated measures of primary and secondary study outcomes. From these four surveys, responses were transformed into three periods based on the timing of the lockdown in Victoria: (1) Pre-lockdown: responses collected prior to 8 July 2020, (2) Lockdown: responses collected between 8 July and 27 October 2020 (inclusive), and (3) Post-lockdown: responses collected after 27 October 2020. The exposure group comprised people residing in the state Victoria during the lockdown period (Table 1). The comparison group comprised people residing elsewhere in Australia during the lockdown period. Note that the rest of Australia did not experience a lockdown during this period of time. Participants were included in the analysis if they completed at least one survey in each of the three periods. For participants completing multiple surveys during the lockdown period, the survey with the most recent data was included to reflect responses that maximised the duration of exposure to the lockdown, and correspondingly surveys completed earlier in the lockdown by the same participant were excluded from analysis. This resulted in three surveys at three-month intervals for each participant. It is important to note that the baseline (pre-lockdown) period coincided with the early stages of the COVID-19 pandemic in Australia, during which significant shifts in health were observed [9].

Table 1

Study design demonstrating survey timing and residential location.

	Pre-lockdown	Lockdown (in Victoria)	Post-lockdown
Exposed Group (State of Victoria, N = 305)	Data collected 27 March to 7 July 2020	Data collected 8 July to 27 October 2020	Data collected 28 October to 31 December 2020
Comparison Group (Rest of Australia, N = 593)

Health outcomes

Three health outcomes were assessed. Firstly, the 6-item Kessler Psychological Distress scale was used to evaluate scores of psychological distress ranging from 6 to 30, with moderate to high distress levels ranging from 11 to 30 [14]. The two remaining scores describe mental and physical health and were derived from the respective summary component scores from the 12-item Short-Form health survey (SF-12), where scores of 50 represent pre-pandemic population average, with a standard deviation of 10 [15].

Determinants of health

Data on several determinants of health were also captured as secondary outcomes. Engagement in work was defined as a binary variable coded from responses describing whether individuals had worked any hours during the prior week, or not. Similarly, a separate binary variable was used to describe whether individuals were either employed or unemployed during each time-point. Two binary measures of social interaction were derived from components of the Social Interaction sub-scale of the Duke Social Support Index describing interactions in the prior week [16]. These included whether people had spent any time with anyone (excluding household members) or not, and whether people engaged in calls (online or telephone) to fewer than seven people during the past week or spoken with seven or more people. The social interaction sub-scale itself was not derived following participant feedback on the question regarding taking part in group meetings (such as sports or other clubs). Many of these activities were not legally permitted under lockdown restrictions and therefore if included the overall interaction sub-scale would instead partially reflect compliance with lockdown restrictions. Financial resources were evaluated with the question ‘If all of a sudden you had to get $2000 for something important, could the money be obtained within a week?’ [17]. Responses of ‘yes’ were categorised as having more financial resources and responses of ‘no’ and ‘don’t know’ as having less financial resource. Financial stress was measures on a 10-point scale from 1 (not at all stressed) to 10 (as stressed as can be). Financial stress was categorised as responses greater than 5.

Analytical approach

Descriptive statistics for exposure and comparison groups were calculated. Mean average health outcomes were calculated and graphically represented for the two groups across three time-points: before, during, and after lockdown. Similarly, group percentages were calculated and graphically represented by location and time-point for the determinants of health.

Mixed linear regression models were used to account for repeated measures of continuous health outcomes with fixed demographic effects. A variable was included describing the timing of lockdown as a random effect using a first-order autoregressive structure with homogeneous variances. The exposure for linear regression models was an interaction term between residential state describing whether individuals resided in Victoria or in the Rest of Australia with a variable describing whether responses were recorded before, during or after lockdown. Fixed effects were included for gender, age group, survey mode, residential location and survey timepoint. For models examining the mental health and psychological distress outcomes pre-existing (prior to the first survey) anxiety and depression were also included as fixed effects, and for the physical health outcome model the total number of pre-existing health conditions were included with three categories describing no conditions, one condition and two or more conditions. The reference group was defined as the outcomes after the Victorian lockdown period for individuals living in Australian states and territories other than Victoria.

Generalised estimating equations were used with binary logistic models for binary outcomes describing work status, social isolation, and financial resources, which captured the correlated outcomes within each individual. Each participant was included as a subject variable, and the ternary lockdown context variable was set as a within-subject variable. Models included an interaction term between time-point and location. Main effects were included for gender, age group, survey mode, residential location and survey timepoint.

Statistical tests described as the ‘The Lockdown Impact’ were performed for each outcome to investigate pairwise differences for changes in people living in Victoria and people living in the rest of Australia. This analysis aims to describe health impacts of the lockdown whilst accounting for overall changes in health measures over time nationally. More specifically, changes in outcomes were compared between exposed and comparison groups upon two temporal transition categories: (1) During Lockdown: differences during the lockdown period controlling for differences pre-lockdown, and (2) After Lockdown: post-lockdown differences controlling for pre-lockdown differences.

Results

A total of 898 participants completed surveys, prior to, during and following the Victorian lockdown period and were included in analyses. There were 161 (18.0%) participants completing two surveys during the lockdown period, and in these cases only the latter survey was included. Cohort demographics and attrition is described and characterised (S2 Table), and all regression models adjust for listed differences between the groups of included or excluded participants. Within the national cohort, 305 (34.0%) participants resided in the state of Victoria with the remaining 593 (66.0%) of participants residing in the rest of Australia and they had similar baseline characteristics (S3 Table).

The Victorian lockdown resulted in significant changes in health (Figs 1 and 2, Table 2), work and determinants of health over time (S1 Fig, S4 Table), and also between individuals in Victoria and the rest of Australia. Amidst relatively high levels of psychological distress in general during the pandemic, distress levels remained elevated for the exposed group during the lockdown period whilst distress levels were lower in the comparison (non-exposed) group. Following the lockdown period, there were no significant differences in psychological distress or mental health of participants in the exposed and comparison groups. During the lockdown period, the mental health of the comparison group significantly improved from pre-lockdown, whereas the no significant differences were observed in the exposure group between pre-lockdown and lockdown period. Physical health deteriorated across each time interval nationally, and there were no significant differences identified arising from the lockdown (Table 2, S2 Fig).

Fig 1

Changes in psychological distress before, during, and after the community lockdown in Victoria (during 8 July–27 October 2020) compared to the Rest of Australia.

Data describe group mean values and 95% confidence intervals. The shaded region indicates moderate-high distress.

Fig 2

Changes in mental health before, during, and after the community lockdown in Victoria (during 8 July–27 October 2020) compared to the Rest of Australia.

Data describe group mean values and 95% confidence intervals. The shaded regions indicate below (pre-pandemic) average mental health.

Table 2

Health impacts of the lockdown in Victoria, Australia.

	Adjusted estimates for differences in health score[95% Confidence Interval]
	Psychological distress(b* < 0: worse distress)	Mental health(b < 0: poorer health)	Physical health(b < 0: poorer health)
The Lockdown Impact
During lockdown	-0.77 [-1.25–0.29]	-1.90 [-3.06, -0.74]	0.04 [-0.82, 0.90]
After lockdown	0.03 [-0.50, 0.56]	0.27 [-1.03, 1.58]	-0.32 [-1.28, 0.64]
Changes in health over location and time
VIC * pre-lockdown	-1.30 [-1.85, -0.74]	-3.23 [-4.48, -1.98]	1.48 [0.49, 2.47]
VIC * lockdown	-1.36 [-1.92, -0.80]	-3.23 [-4.48, -1.98]	0.99 [0.00, 1.98]
VIC * post-lockdown	-0.22 [-0.77, 0.34]	-0.48 [-1.73, 0.77]	0.51 [-0.48, 1.50]
RoA * (pre-lockdown)	-1.05 [-1.36, -0.74]	-2.48 [-3.24, -1.71]	0.65 [0.09, 1.21]
RoA * (lockdown)	-0.35 [-0.63, -0.06]	-0.58 [-1.26, 0.10]	0.12 [-0.38, 0.63]
RoA* (post-lockdown)	0.00 (ref.)	0.00 (ref.)	0.00 (ref.)

Estimates with p < .05 shown in bold.

*Coefficients and confidence intervals listed for psychological distress within the table have been negated (i.e. -b) to ease comparisons with corresponding coefficients for mental health. VIC–the Australian state of Victoria (i.e. lockdown location during 8 July– 27 October 2020). RoA–Rest of Australia (i.e. not Victoria). Models were adjusted for gender, age group, survey mode and by pre-existing medical conditions. The Lockdown Impact describes health differences of working-age Victorians to the Rest of Australia, controlling for health differences pre-lockdown.

Engagement in work (i.e. having worked in the prior week) declined in the exposed group during the lockdown and recovered following the lockdown period. Whilst unemployment levels were significantly higher prior to and during the lockdown nationally compared to after the Victorian lockdown, we did not observe any significant impacts resulting from the lockdown.

During lockdown there was an increase in social isolation, which decreased after the lockdown resulting in no significant post-lockdown effects. Virtual interactions were more common pre-lockdown, and there were no significant changes resulting from the impact of lockdown. Financial stress and people with fewer financial resources decreased over time in the cohort nationally, and there were no significant differences resulting from Victoria’s lockdown.

Discussion

Our findings demonstrate that the extended community lockdown in the state of Victoria during the southern hemisphere winter and spring of 2020 had negative mental health consequences for people of working age including increases in psychological distress as assessed using the Kessler-6 scale, and decreases in mental health assessed using the SF-12 mental component score. However, these effects were short-lived. Following an abrupt national decline in mental health observed during the early stages of the pandemic, people exposed to an extended lockdown experienced a delay in the recovery of mental health observed in the rest of Australia. Following the easing of lockdown measures, we then observe a resolution of the negative mental health impacts of lockdown within a two-month period, bringing the mental health of Victorian residents back in line with that of the rest of Australia. We also observed a pattern of deterioration during lockdown, followed by recovery post lockdown, in two determinants of health: social interactions and engagement in paid work. Whilst nationally we observed significant reductions in financial stress, physical health and unemployment throughout the pandemic period, along with increases in virtual interactions with others, significant impacts of the extended Victorian lockdown were not elucidated.

The negative psychological consequences of enforced quarantine are well described [18]. A recent meta-analysis of psychological health in COVID-19 induced lockdowns reported an overall small effect on mental health symptoms among those exposed to lockdown conditions, with substantial heterogeneity among studies [19]. The studies in this review reported lockdown periods ranging from 1 to 60 days, and included a range of study designs including comparisons between groups exposed and not exposed to lockdowns, or within person designs comparing mental health before and after exposure to lockdown orders. Our study utilises both a within subject (pre, during and post-lockdown) measurement and incorporates a contemporaneous comparison group also assessed longitudinally. The duration of exposure to lockdown conditions in our study is also much longer than those reported in this meta-analysis, at 112 days. We also identified changes in several determinants of health occurring during the lockdown period, including a reduction in social interactions and engagement in work, both of which may have contributed to the increased psychological distress and poorer mental health observed.

A number of strategies for mitigating the negative psychological consequences of lockdown have been proposed [18]. These include limiting the duration of quarantine restrictions, effective public health messaging, ensuring ongoing access to basic supplies, and actions that enable social interaction [18]. In the state of Victoria during the 2020 lockdown studied in this paper, the official public health messaging occurred on an ongoing basis via daily press conferences with the Premier and senior government health officials, and release of information and statistics via social media and health department websites. One feature of the Victorian lockdown was regular changes in the restrictions being enforced. For example, initially the lockdown rules allowed exemptions to movement restrictions for visiting intimate partners [20], and this was later extended to include other forms of social bubbles for people living alone [21]. These regular changes may have contributed to confusion among those required to enforce compulsory restrictions [22], and among members of the public.

The burden of mental ill health was apparent from the early stages of the COVID-19 pandemic, and prior to the Victorian lockdown period, demonstrating the need for additional mental health supports and services during the pandemic [9,22,23]. Our findings suggest that during periods of extended lockdown this need becomes more acute, coinciding with increased loss of work and social isolation. As a result, dedicated mental health supports and services are encouraged during periods of lockdown, in addition to actions that are anticipated to ameliorate some of the negative impacts of lockdowns on determinants of health such as minimising loss of work and reducing loneliness.

Financial loss has been highlighted as a potential post-lockdown stressor [18]. Our findings demonstrated that financial stress was highest during the early stages of the pandemic, and decreased over time nationally. Furthermore, we did not identify any significant differences in financial stress due to the lockdown within our sample. This may be due, in part, to several financial supports for businesses suffering significant losses or closures during the Victorian lockdown such as dedicated lump sum payments [24], and extended eligibility for a national wage subsidy program due to the lockdown measures [25]. Nationally, economic support for households and businesses was estimated to amount to around 15% of GDP [26], of which wage subsidies and temporary doubling of payments to income support recipients (e.g. unemployment allowance) constituted 6% of GDP [27]. These supports are anticipated to have contributed to reduced financial stress for people experiencing loss of work due to the lockdown, and consequently ameliorated further deteriorations in mental health.

The Victorian lockdown shaped the nature of work over this period. Many businesses had to temporarily pause operations, whilst others adapted to working from home arrangements in line with Government directives [12]. Our findings suggest that engagement in work increased nationally from the early stages of the pandemic, and that this increase was hampered during the lockdown period, yet recovered thereafter. Given the decrease in work over the lockdown period, encouraging businesses to develop pandemic response plans [28] or operating models that incorporate adaptive work options for employees during periods of lockdown are likely to support the mental health of employees. Some changes to work, such as increased working from home arrangements for people in Victoria, are expected to have long-lasting effects over the coming years. [29] The health impacts of increased working from home arrangements in the workforce resulting from lockdowns, and the pandemic more generally, are yet to be fully uncovered, and warrant further longitudinal research.

Due to our focus on individuals experiencing loss of work during the pandemic, several demographic groups are either underrepresented or absent in our sample, including older adults, children and people that were not engaged in paid work prior to the pandemic. It has been recognised that the experiences of lockdown by these groups will differ due to impacts of school closures resulting in debilitating effects on children’s development, education and behavioural problems [30], and disproportionate consequences for older adults whose primary social contact is outside the home [31]. Whilst our sample of individuals is not nationally representative, we accounted for age-related and gender-specific differences within the working-age cohort. Additionally, the use of a comparison group assisted to account for other forms of bias, such as the differences in cohort retention based on survey mode. However, this comparative approach has limitations for forms of attrition bias resulting changes in health due to the experience of lockdown, such as increased attrition for individuals in distress. The risk of the Victorian COVID-19 outbreak spreading elsewhere nationally may have contributed to negative shifts mental health for people in the rest of Australia. In such circumstances, our findings may be underreporting the magnitude of mental health consequences, and of determinants of health, during lockdowns.

While our findings describe group-level changes in day-to-day psychological distress and several determinants of health during and after lockdown, the impacts of lockdown are unlikely to be shared equally. We have previously reported on the negative mental health impacts of pandemic-related loss of work, and how individuals with pre-existing mental health conditions, or those with fewer social or financial supports, are most at-risk of poor mental health [9]. Additional supports are warranted for these groups upon exposure to lockdown measures. We were limited in our ability to investigate the health consequences of lockdowns in other potentially vulnerable groups such as disadvantaged communities, Aboriginal and Torres Strait Islanders or culturally and linguistically diverse communities due to their small sample size. It is noted that people who experienced elevated levels of anxiety either before or after lockdown report symptoms of poor physical health [32], although this was not explored within the context of our analysis. Whilst our findings describe a recovery in aspects of mental health such as symptoms of anxiety and psychological distress, mental disorders such as post-traumatic stress disorder may re-emerge over longer periods of time [33], or upon repeated exposure to lockdowns or crises.

The nature of lockdowns has substantially differed globally, in terms of the stringency of restrictions, durations, the repeated nature of lockdowns and government responses to them [1]. Our findings describe the health impacts during and after a stringent 112-day community lockdown in Victoria, Australia. However, the experiences we have observed may vary in other contexts. Thus, a need for similar studies across a range of scenarios and contexts is required to truly understand the impact of lockdowns on health and its determinants.

Conclusion

The extended, 112-day community lockdown in the state of Victoria during the winter and early spring of 2020 had a negative impact on mental health, engagement in work and social interactions. After the removal of lockdown restrictions, the mental health of those exposed to the extended lockdown returned to be equivalent to those in the non-exposed comparison group. Restrictions that promote safe forms of social interactions and engagements in work may mitigate some negative psychological consequences during periods of lockdown. Governments and employers can help to minimise the negative mental health effects of lockdowns by taking actions and implementing policy that encourages maintenance of employment relationships, supports engagement in work and promotes social interaction and addresses the mental health needs of the community. Longer-term follow-up is required to identify any longer-term persistent health effects of community lockdowns.

Changes in determinants of health prior to, during, and after the community lockdown in Victoria, compared to the Rest of Australia.

(DOCX)

Click here for additional data file.

Changes in physical health prior to, during, and after the community lockdown in Victoria compared to the Rest of Australia.

(DOCX)

Click here for additional data file.

Summary of restrictions leading to, during, and following the 2020 extended community lockdown in Victoria, Australia.

(DOCX)

Click here for additional data file.

Comparisons of participants included in the analyses and those excluded.

(DOCX)

Click here for additional data file.

Comparisons of participants residing in Victoria compared to participants in the Rest of Australia.

(DOCX)

Click here for additional data file.

Impacts of the lockdown in Victoria on determinants of health including work, social interactions, and finance.

(DOCX)

Click here for additional data file.

Media statements and Victorian Government publications on restrictions.

(DOCX)

Click here for additional data file.

11 Oct 2021

Reviewer 1 and reviewer 5 provide more substantial critique of the paper and its methods. Please review these comments and respond to each point raised in a neatly set out table (with verbatim extract of the relevant reviewer comment) appended in a separate document (your rebuttal letter) with a brief comment against each query raised with reference to the areas (or main areas) where you have made edits in response. Please also ensure that any edits to the manuscript are provided in mark up for review.

Please submit your revised manuscript by Nov 25 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

David James Carter, PhD, LLM (Res), LLB (Hon I), BA

Academic Editor

PLOS ONE

Journal requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please provide additional details regarding participant consent. In the ethics statement in the Methods and online submission information, please ensure that you have specified what type you obtained (for instance, written or verbal, and if verbal, how it was documented and witnessed). If your study included minors, state whether you obtained consent from parents or guardians. If the need for consent was waived by the ethics committee, please include this information.

3. Please state whether you validated the questionnaire prior to testing on study participants. Please provide details regarding the validation group within the methods section.

4. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For more information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

5. Your ethics statement should only appear in the Methods section of your manuscript. If your ethics statement is written in any section besides the Methods, please move it to the Methods section and delete it from any other section. Please ensure that your ethics statement is included in your manuscript, as the ethics statement entered into the online submission form will not be published alongside your manuscript.

6. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

Reviewer #3: Partly

Reviewer #4: Yes

Reviewer #5: No

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: Yes

Reviewer #3: I Don't Know

Reviewer #4: I Don't Know

Reviewer #5: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

Reviewer #5: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer #4: Yes

Reviewer #5: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Summary:

This paper summarized a survey to find the association of lockdown with psychological distress in Victorian community in Australia. N=898 subjects who finished the survey in all time points was included in the study. The paper contributes to understanding people’s psychological distress influenced by the lockdown policy, but recovered after the lockdown. This paper is practical and meaningful.

The topic of this study is interesting. It helped us to understand how the lockdown policy’s influence on people’s mental health. The survey monitored people’s status longitudinally across the time points. This paper was well written and structured logically. It clearly introduce the background and findings in the discussion.

Concerns:

There are quite a few concerns on methods and analysis and unclear points in this paper. There are some major issues:

1. The data collection background was not clearly explained. The survey Reponses were from the working age group, so the population is the entire working-age group in corresponding areas/community, but what is the population of the study subjects? What is the exact definition of working-age group?

From the supplementary table S1, the retention group N=898 contains subjects with age 65 and more. So, more clarification and discussion are necessary.

2. How was the survey sample size decided? There is no power analysis discussion in the paper. What is the power or effect size from the current sample?

3. The data collection process was confusing. How the survey participants were recruited? Randomly selection? Volunteer response? Response rate? More explanation is needed.

4. There is only a rough descriptive table between sample and subjects excluded from the study. But, there is no descriptive tables to compare exposure group (N=305) and Comparison group (N=593) on age, gender, etc. it is concern that if there were unbalance between these two groups.

5. From table S2, the descriptive table between Retention group and attribution group (excluded subjects), it seems the retention group is mainly telephone interview (80%), and attribution group were more taking online survey (60%). I think there is bias. Revision and More discussion is needed

6. In terms of the main finding, for example the figure of mental health, the conference intervals are needed to show the variation of each group across different time points (before, lockdown and After). What is the effect size of this differences claimed in the paper?

7. Also, sensitivity analysis needs to be added to show the robustness of the finding.

Reviewer #2: Thank you for the opportunity to review such a beautifully written paper on a topic of great national importance.

I would recommend acceptance if one issue can be remedied.

The study evaluates mental health at different time points and finds no lasting effects from the lockdown. To reach this conclusion it used the Kessler 6 scale and the mental component of the SF-12.

My concern is that the scales used may not capture the full range of mental health impacts of lockdown. For instance, this range includes disorders that have a complicated aetiology, such as post-traumatic stress disorder (PTSD), where symptoms may emerge over time. In fact, in the literature on disaster response suggests there may be a 'honeymoon phase' characterised by increased community cohesion in the immediate aftermath of a mass trauma event (see e.g. DeWolfe, 2000, p.5). Post-traumatic stress is one of the most likely sequelae of the crisis, yet I can find no research on the validity of the SF-12 or the K-6 for discerning PTSD symptoms in a study population.

I would request the manuscript be amended to account for these limitations. It would be more accurate to say, hewing closer to the wording of the scales themselves, that day-to-day stress and distress returned to pre-lockdown levels, but impacts on the longer-term mental health of participants cannot be assessed using the methodology adopted in the study.

----

DeWolfe 2000 https://www.hsdl.org/?view&did=4017 accessed 13 Sept 2021.

Reviewer #3: Your paper is both timely and adds to our knowledge of the impact of lockdowns on the mental health of the population. I believe your paper could be strengthened with more detail - at the moment, yes it substantiates the fact that people in lockdown in Victoria suffered from worse mental health, but we don't know much about the nuances therein. I note that you have not analysed the experiences of your subjects based on whether they are Aboriginal or Torres Strait Islanders or from CALD backgrounds, or for example where they lived specifically, and even though you note that you looked a gender, you did not discuss any differences in detail.

Reviewer #4: This well-written and timely paper investigates the short- and long-term negative effects of lockdown in the State of Victoria, Australia. The authors found that the Victorian lockdown led to increased psychological distress during the lockdown, but these effects resolved after restrictions are lifted.

The study raises important points regarding the negative effects during and following lockdowns and contributes to existing works on this important topic.

However, the persistence of the negative health effects needed to be evaluated more carefully. Only showing that negative mental health consequences resolve over time is too simplistic. Since the Covid-19 pandemic began, we have learned that lockdown and social isolation measures have different effects for different populations. Some communities are more vulnerable than others. For example, disadvantaged low-income communities and people with mental health and addiction problems are more susceptible to negative impacts from interventions such as lockdowns. It would be useful for the author to provide information about the impacts of these measures on vulnerable individuals during and following lockdowns. If the authors don’t have the data regarding the effects on these populations, they should acknowledge and discuss this limitation.

The same applies to the discussion around the Australian government’s efforts to mitigate the harm caused by public health measures against Covid-19. Discussion is needed on the strategies for mitigating the negative psychological consequences of lockdown specifically for populations more susceptible to public health measures. Can the author also discuss strategies to assist vulnerable communities during lockdowns and relate these to the situation in Victoria?

Last, there would be value in the author adding a paragraph comparing their findings to similar studies conducted in other countries (e.g., the short-/long-term negative effects of lockdowns) and discuss variation in outcomes (if such exists).

Reviewer #5: Comments to Author:

Ms. Ref. No. PONE-D-21-24391

Title: The health impacts of a 4-month long community-wide COVID-19 lockdown: Findings from a prospective longitudinal study in the state of Victoria, Australia.

• Overall objective and the extent to which this was achieved:

This paper attempts to address a very important and, to date, a sparsely researched topic: When lockdown ceases, to what extent do the established negative health effects of lockdown persist and what is the speed of recovery of individuals and communities to these negative effects?

The authors state that they took advantage of a natural experiment where, contemporaneously, there was a comparison group outside the lockdown area in Victoria, Australia.

Their stated aim was to examine (lines 83-87):

1. the impact on mental and physical health of the extended lockdown.

2. the impact of the extended lockdown on established determinants of health: unemployment, social interaction, and finances.

3. Effects persisting after lockdown was lifted.

Of these three aims, the literature is flooded with the impact of lockdowns on health, mental, physical, and psychological. This is also mentioned in their introduction. However, it is the third aspect that is under-researched as again, they explain (lines 70 -73) even before the three objectives are listed.

Main takeaway: Unfortunately, it is in this third objective that the paper falls short. In their conclusion, they mainly talk about the negative impact of the extended lockdown (not new ground) and the mitigating effect of employment and social interactions which also has been shown. They end by saying “longer-term follow-up is required to identify any persistent health effects of community lockdowns.” They actually could have done this had they exploited the 2 months of post-lockdown data. But that was not done.

• Clarification needed: The paper seems remarkably similar to this already published:

Griffiths, D., Sheehan, L., Vreden, C.V., Petrie, D., Sim, M. and Collie, A., 2021. 1360 Health impacts of a 4 month community-wide lockdown: a prospective longitudinal study in Victoria, Australia. International Journal of Epidemiology, 50(Supplement_1), pp.dyab168-240.

Though this appears in Google Scholar, do I assume that this is just the publication of an abstract of a conference presentation? I was unable to get hold of this paper. The authors can clarify.

• The Model:

1. I failed to find out what is the model that they are estimating. Lines 170 to 189 give a general description of what was done. The exact specification of the model/models needs to be spelled out, and the control variables indicated. This is the heart of the paper and a general description does not provide the information needed to assess the underpinnings of the analysis

2. In the estimation, they also need to add a fixed time effect.

• Variables:

3. What were all the variables and how were they defined and measured? For instance, I could find no definition of the two key variables they have used and even shown graphs with them: psychological distress and mental health.

4. Of the variables they have discussed, I have the following comments:

i) Financial stress is important to capture. However, the authors have taken only one out of the 3 indicators that the GSS Summary Results for Australia have considered. The GSS considers three elements:

(https://www.abs.gov.au/statistics/people/people-and-communities/general-social-survey-summary-results-australia/2014)

a. A cash flow problem in the last year in terms of either inability to pay bills on time or having to get help to pay.

b. The ability to raise $2000 within a week for something important.

c. Taking a dissaving action such as “drawing on savings, increasing a credit card balance by $1,000 or more and taking out a personal loan”.

Please see Graph 2 in https://www.abs.gov.au/statistics/people/people-and-communities/general-social-survey-summary-results-australia/2014. It can be seen that of all the 3 stressors, cash flow and dissaving have higher percentages in general than the ability to raise $2000 dollars. In fact, it is the unemployed that have the most problems.

The authors need to justify why the other two were not considered and why focusing on the selected item was appropriate.

ii) I do not understand how, (if the responses to that one question was taken as “yes” or “no” or “don’t know”), financial stress could be measured (for each person) on a 1 to 10 scale. Is this an overall (average) measure for the county? This needs clarification.

5. Social Interaction – two binary measures are taken:

Was any time spent with anyone not a household member and

Whether contact was made with 7 or fewer people via speech, phone or online communication.

The original 35 item Duke Social Support Index was abbreviated to a 23-item index and also to a 11-item index for use on elderly patients who may get exhausted by the use of the longer questionnaire. However, studies have generally used either the 23 item or the 11 item DSSI. Usually these have been for older people. In this paper (not confined only to the elderly), only two items have been picked and the justification of these two items has not been given. In other words, why would we believe in the validity of a 2-item DSSI?

The authors need to justify and discuss this.

6. What is the need for the variable “engagement in work” since “employment” at time of survey is also included? Is there any extra information being captured?

• Other comments regarding results:

Turning to the results with regard to the stated objectives:

7. The negative health findings (they focus on psychological distress and mental health --- again, how were these measured?) during lockdown for both exposed and comparison group is in line with findings elsewhere in the COVID literature and so do not break new ground.

• Unobserved heterogeneities as the two groups do not appear similar

8. During the lockdown period the mental health of the comparison group actually improved significantly from pre-lockdown. There is no further exploration of this counter-intuitive result. However, this pattern did not show up in the exposure group thus leading one to suspect that there are other factors at play or unobserved heterogeneities that have not been captured and which may, thus, invalidate this “natural experiment”.

9. Right now, looking at the mental health and psychological graphs, it can be seen that the two groups did not begin at the same level. Are these differences statistically significant? This is important as otherwise, this further shows that the two groups were quite different to begin with (another reason to suspect other unaccounted for factors at play).

10. What accounts for the different trajectories of psychological distress? Again, the two groups appear different.

• The graphs

11. The graphs are not easy to understand.

i. Since there is no discussion of “health score” in the cases of mental health and psychological distress, one does not know how to interpret the y-axes. The x-axis is not labelled.

ii. How were the health scores obtained?

iii. As I understand, the first dot is for data collected PRIOR to lockdown. The second dot is the data collected DURING lockdown. The third dot is the data collected POST lockdown. It seems a valuable opportunity to explore the POST lockdown period was not taken by averaging the October to December data. This would have shown the speed of recovery and might have added some new information to the literature.

• Precision of estimates and statistical power:

12. Since I am not clear on the models estimated to find the mental and physical effects of a lockdown, I wanted to make sure that if any of the explanatory variables were correlated, that was clearly brought out and the implications mentioned. Employment, social interaction, and financial stress are likely to be highly correlated. Though there are methods to deal with this problem, at least the impact on the precision of the estimates needs to be mentioned.

13. The state of Victoria has 6.68 million people (line 94) accounting for over a quarter of the 6.68 million population of Australia (line 95). I find it worrying that the analysis has only 305 people in Victoria (exposed group) and 593 in the rest of Australia (the comparison group). The sample sizes seem too small for the statistical tests to have much power. At the very least, this should be mentioned.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

Reviewer #4: No

Reviewer #5: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

14 Dec 2021

The authors have addressed each of the comments from the 5 reviewers in full. This documented in a Table, and also repeated here below.

Responses to reviewer comments

Manuscript ID: PONE-D-21-24391

Title: The health impacts of a 4-month long community-wide COVID-19 lockdown: Findings from a prospective longitudinal study in the state of Victoria, Australia.

We thank the reviewers for their consideration of this paper and comments. Our responses are listed in the table below.

REVIEWER 1:

This paper summarized a survey to find the association of lockdown with psychological distress in Victorian community in Australia. N=898 subjects who finished the survey in all time points was included in the study. The paper contributes to understanding people’s psychological distress influenced by the lockdown policy, but recovered after the lockdown. This paper is practical and meaningful.

The topic of this study is interesting. It helped us to understand how the lockdown policy’s influence on people’s mental health. The survey monitored people’s status longitudinally across the time points. This paper was well written and structured logically. It clearly introduce the background and findings in the discussion.

REVIEWER COMMENT:

Concerns:

There are quite a few concerns on methods and analysis and unclear points in this paper. There are some major issues:

1. The data collection background was not clearly explained. The survey Reponses were from the working age group, so the population is the entire working-age group in corresponding areas/community, but what is the population of the study subjects? What is the exact definition of working-age group?

From the supplementary table S1, the retention group N=898 contains subjects with age 65 and more. So, more clarification and discussion are necessary.

AUTHOR RESPONSE:

Text has been added outlining the inclusion criteria for the study which includes being aged 18 years or more, and were working before the pandemic. The recruitment process is now outlined in more detail.

Paragraph one has been added to the section:

Methods>Participants, data and context

REVIEWER COMMENT:

2. How was the survey sample size decided? There is no power analysis discussion in the paper. What is the power or effect size from the current sample?

AUTHOR RESPONSE:

A power analysis was included for primary study outcomes for the longitudinal cohort study. However, this paper describes an unplanned experiment looking at the impacts of a lockdown that took place within the study period, thus we had no opportunity to choose (i.e. increase) the sample size at this stage and a power analysis was not conducted to observe the health impacts during, and after, an unanticipated community lockdown.

The sizes of effects are described in Table 2 in the forms of adjusted estimates on the scales of either differences in scores on the Kessler-6 scale or the SF-12 mental or physical component summary scale. Effect sizes range from medium to low depending on the outcome measure and the confidence intervals demonstrate, given our sample size, the statistical power we have for inference.

REVIEWER COMMENT:

3. The data collection process was confusing. How the survey participants were recruited? Randomly selection? Volunteer response? Response rate? More explanation is needed.

AUTHOR RESPONSE:

We have added details within this paper explaining the recruitment process. This is also described in reference 9.

Paragraph one has been added to the section:

Methods>Participants, data and context

REVIEWER COMMENT:

4. There is only a rough descriptive table between sample and subjects excluded from the study. But, there is no descriptive tables to compare exposure group (N=305) and Comparison group (N=593) on age, gender, etc. it is concern that if there were unbalance between these two groups.

AUTHOR RESPONSE:

Supplementary Table S3 has been added comparing descriptive statistics of the exposure group and comparison group. The groups are largely similar and the reported variables in this Table are incorporated in the multivariate regression models to account for any imbalance between groups.

REVIEWER COMMENT:

5. From table S2, the descriptive table between Retention group and attribution group (excluded subjects), it seems the retention group is mainly telephone interview (80%), and attribution group were more taking online survey (60%). I think there is bias. Revision and More discussion is needed.

AUTHOR RESPONSE:

This comment now refers to Table S3 since the addition of an extra supplementary table.

The comparison throughout this paper is between the exposure group and control group which have very similar distributions for the survey mode. The new Supplementary Table S2 has been added following a reviewer comment to illustrate this.

Whilst there is a clear difference in attrition based on the survey mode, both groups share the same level of attrition. Furthermore, survey mode is included as a covariate in all models.

The following sentence has been added to the Discussion:

“the use of a comparison group assisted to account for other forms of bias, such as the differences in cohort retention based on survey mode”

REVIEWER COMMENT:

6. In terms of the main finding, for example the figure of mental health, the conference intervals are needed to show the variation of each group across different time points (before, lockdown and After). What is the effect size of this differences claimed in the paper?

AUTHOR RESPONSE:

Figures on health outcomes, and those on determinants of health have been adapted to include an estimate of the uncertainty within each group at each time point using 95% confidence intervals.

Effect sizes (and respective 95% CIs) are explicitly stated in Table 1, (and Supplementary Table S4) in the form of changes in scores to either the Kessler-6 scale or the SF-12 component summary score, (and group percentage changes in determinants of health). Note that the uncertainty in the average change in scores (tables) may be smaller than the uncertainty at each point in time (Figures) due to the within individual correlation in scores over time.

REVIEWER COMMENT:

7. Also, sensitivity analysis needs to be added to show the robustness of the finding.

AUTHOR RESPONSE:

While we do not refer to an explicit ‘sensitivity analysis’, the figures represent the unadjusted estimates and the regression results the adjusted estimated (and both illustrate the same story). In addition, the regression models have been replicated on 9 occasions across 4 themes (health, work, social interactions and finance).

For example, the robustness of the findings regarding aspects of mental health can be observed by similar patterns in mental health SF-12 scores and (inverted) Kessler-6 psychological distress scores. Changes observed in mental health and distress do not extend to health more generally such as physical health (based on the measures used).

Similarly, the regression model for social isolation demonstrates a significant effect of lockdown, and recovery thereafter, whereas the lack of an observed significant effect for virtual interactions acts to demonstrate an element of robustness.

Analogous considerations can be inferred by comparing the outcomes for engagement in work vs. employment, and financial stress vs. financial resources.

Taken together, this approach demonstrates where key differences are observed during lockdown across the aforementioned four themes, and their recovery thereafter. It also demonstrates outcomes where significant effects are not observed.

REVIEWER 2:

Thank you for the opportunity to review such a beautifully written paper on a topic of great national importance.

I would recommend acceptance if one issue can be remedied.

REVIEWER COMMENT:

The study evaluates mental health at different time points and finds no lasting effects from the lockdown. To reach this conclusion it used the Kessler 6 scale and the mental component of the SF-12.

My concern is that the scales used may not capture the full range of mental health impacts of lockdown. For instance, this range includes disorders that have a complicated aetiology, such as post-traumatic stress disorder (PTSD), where symptoms may emerge over time. In fact, in the literature on disaster response suggests there may be a 'honeymoon phase' characterised by increased community cohesion in the immediate aftermath of a mass trauma event (see e.g. DeWolfe, 2000, p.5). Post-traumatic stress is one of the most likely sequelae of the crisis, yet I can find no research on the validity of the SF-12 or the K-6 for discerning PTSD symptoms in a study population.

I would request the manuscript be amended to account for these limitations. It would be more accurate to say, hewing closer to the wording of the scales themselves, that day-to-day stress and distress returned to pre-lockdown levels, but impacts on the longer-term mental health of participants cannot be assessed using the methodology adopted in the study.

DeWolfe 2000 https://www.hsdl.org/?view&did=4017 accessed 13 Sept 2021.

AUTHOR RESPONSE:

The reviewer raises an important point on capturing the full range of mental health impacts and the scope of the mental health measures used in the study.

A penultimate paragraph has been added to the Discussion section addressing the points raised by the reviewer.

The phases of disaster described by the DeWolfe Training manual is relevant and has been cited.

REVIEWER 3

Your paper is both timely and adds to our knowledge of the impact of lockdowns on the mental health of the population.

REVIEWER COMMENT:

I believe your paper could be strengthened with more detail - at the moment, yes it substantiates the fact that people in lockdown in Victoria suffered from worse mental health, but we don't know much about the nuances therein. I note that you have not analysed the experiences of your subjects based on whether they are Aboriginal or Torres Strait Islanders or from CALD backgrounds, or for example where they lived specifically, and even though you note that you looked a gender, you did not discuss any differences in detail.

AUTHOR RESPONSE:

While we would have loved to explore further in terms of how the lockdowns impacted different groups in society we were constrained by our sample size to be able to provide robust estimates for these different groups..

We now more clearly acknowledge that the nuances of changes in mental health across different groups of people are not described in this paper.

We reference a paper looking at the risk of poor mental health based on several determinants of health and gender-differences [9].

A penultimate paragraph has been added to the Discussion section addressing the points raised by the reviewer, including our limited sample of Aboriginal and Torres Strat Islanders and CALD communities.

REVIEWER 4:

This well-written and timely paper investigates the short- and long-term negative effects of lockdown in the State of Victoria, Australia. The authors found that the Victorian lockdown led to increased psychological distress during the lockdown, but these effects resolved after restrictions are lifted.

The study raises important points regarding the negative effects during and following lockdowns and contributes to existing works on this important topic.

REVIEWER COMMENT:

However, the persistence of the negative health effects needed to be evaluated more carefully. Only showing that negative mental health consequences resolve over time is too simplistic. Since the Covid-19 pandemic began, we have learned that lockdown and social isolation measures have different effects for different populations. Some communities are more vulnerable than others. For example, disadvantaged low-income communities and people with mental health and addiction problems are more susceptible to negative impacts from interventions such as lockdowns. It would be useful for the author to provide information about the impacts of these measures on vulnerable individuals during and following lockdowns. If the authors don’t have the data regarding the effects on these populations, they should acknowledge and discuss this limitation.

AUTHOR RESPONSE:

A similar point has also been described by reviewer 3.

A penultimate paragraph has been added to the Discussion section addressing the points raised by the reviewer.

We did not collect data on addiction, but reference a paper describing the role of pre-existing mental health conditions on changes in mental health upon loss of work during the pandemic [9].

REVIEWER COMMENT:

The same applies to the discussion around the Australian government’s efforts to mitigate the harm caused by public health measures against Covid-19. Discussion is needed on the strategies for mitigating the negative psychological consequences of lockdown specifically for populations more susceptible to public health measures. Can the author also discuss strategies to assist vulnerable communities during lockdowns and relate these to the situation in Victoria?

AUTHOR RESPONSE:

These are important points, although we have little information on this resulting from the reported analysis in this paper for the cohort.

A penultimate paragraph has been added to the Discussion section briefly covering the points raised by the reviewer

REVIEWER COMMENT:

Last, there would be value in the author adding a paragraph comparing their findings to similar studies conducted in other countries (e.g., the short-/long-term negative effects of lockdowns) and discuss variation in outcomes (if such exists).

AUTHOR RESPONSE:

There have been a lot of variation between lockdowns in terms of the measures imposed, and also in the study designs to describe the impacts of lockdown. In the second paragraph of the discussion we outline some of these differences described by a meta-analysis of psychological health in COVID-19 induced lockdowns (reference 19).

A penultimate paragraph has been added to the Discussion section briefly covering the points raised by the reviewer including context of a paper [32] describing changes in physical health post-lockdown for a specific group of people experiencing anxiety post-lockdown.

REVIEWER 5:

REVIEWER COMMENT:

Overall objective and the extent to which this was achieved:

This paper attempts to address a very important and, to date, a sparsely researched topic: When lockdown ceases, to what extent do the established negative health effects of lockdown persist and what is the speed of recovery of individuals and communities to these negative effects?

The authors state that they took advantage of a natural experiment where, contemporaneously, there was a comparison group outside the lockdown area in Victoria, Australia. Their stated aim was to examine (lines 83-87):

1. the impact on mental and physical health of the extended lockdown.

2. the impact of the extended lockdown on established determinants of health: unemployment, social interaction, and finances.

3. Effects persisting after lockdown was lifted.

Of these three aims, the literature is flooded with the impact of lockdowns on health, mental, physical, and psychological. This is also mentioned in their introduction. However, it is the third aspect that is under-researched as again, they explain (lines 70 -73) even before the three objectives are listed.

Main takeaway: Unfortunately, it is in this third objective that the paper falls short. In their conclusion, they mainly talk about the negative impact of the extended lockdown (not new ground) and the mitigating effect of employment and social interactions which also has been shown. They end by saying “longer-term follow-up is required to identify any persistent health effects of community lockdowns.” They actually could have done this had they exploited the 2 months of post-lockdown data. But that was not done.

AUTHOR RESPONSE:

This manuscript does describe outcomes post-lockdown and we have now made this clearer in the paper.

For example, Table 2 refers to outcomes during a ‘post-lockdown’ period, and a test labelled as ‘After lockdown’ describes differences between the exposure group and control group post-lockdown whilst accounting for differences between group pre-lockdown.

Addressing point 3, our findings describe that after the conclusion of lockdown there were no persistent effects based upon the measures described in this paper. As other reviewers have pointed out, our findings do not evaluate some forms of mental health disorders such as PTSD which may develop after longer follow-up periods, or upon repeated exposure to lockdowns.

Further discussion has been added on this point in the penultimate paragraph of the Discussion section.

The following text has been added to the Conclusion section :” After the removal of lockdown restrictions, the mental health of those exposed to the extended lockdown returned to be equivalent to those in the non-exposed comparison group”.

REVIEWER COMMENT:

Clarification needed: The paper seems remarkably similar to this already published:

Griffiths, D., Sheehan, L., Vreden, C.V., Petrie, D., Sim, M. and Collie, A., 2021. 1360 Health impacts of a 4 month community-wide lockdown: a prospective longitudinal study in Victoria, Australia. International Journal of Epidemiology, 50(Supplement_1), pp.dyab168-240. Though this appears in Google Scholar, do I assume that this is just the publication of an abstract of a conference presentation? I was unable to get hold of this paper. The authors can clarify.

AUTHOR RESPONSE:

The editor has been informed this is an abstract for an oral conference presentation.

REVIEWER COMMENT:

• The Model:

1. I failed to find out what is the model that they are estimating. Lines 170 to 189 give a general description of what was done. The exact specification of the model/models needs to be spelled out, and the control variables indicated. This is the heart of the paper and a general description does not provide the information needed to assess the underpinnings of the analysis

AUTHOR RESPONSE:

Changes have been made to the text to clarify the analytical approach.

Figures 1 and 2 have updated y-axis labels describing the scales used to measure health outcomes.

Fixed effects for time and location have been emphasised for mixed linear regression models (for health outcomes) in addition to describing the exposure.

Methods>Analytical approach:

“Fixed effects were included for gender, age group, survey mode, residential location and survey timepoint.”

Main effects for time and location have been emphasised for generalised estimating equations (for outcomes describing known determinants of health)

Methods>Analytical approach:

“Main effects were included for gender, age group, survey mode, residential location and survey timepoint.

An additional supplementary table S2 has been added describing the exposure group and control group across a set of covariates.

REVIEWER COMMENT:

2. In the estimation, they also need to add a fixed time effect.

AUTHOR RESPONSE:

We have included a fixed time effect in each model as the exposure and have added text to the Analytic Approach section to emphasise this point. This is described in the previous response.

REVIEWER COMMENT:

Variables:

3. What were all the variables and how were they defined and measured? For instance, I could find no definition of the two key variables they have used and even shown graphs with them: psychological distress and mental health.

AUTHOR RESPONSE:

The y-axis for Figures 1 and 2 have been renamed to reflect the description of the Kessler 6 score and 2-item Short Form health survey (SF-12) mental, or physical, component scores.

This matches the descriptions listed in the Methods>Health outcomes subsection.

REVIEWER COMMENT:

4. Of the variables they have discussed, I have the following comments:

i) Financial stress is important to capture. However, the authors have taken only one out of the 3 indicators that the GSS Summary Results for Australia have considered. The GSS considers three elements:

(https://www.abs.gov.au/statistics/people/people-and-communities/general-social-survey-summary-results-australia/2014) a. A cash flow problem in the last year in terms of either inability to pay bills on time or having to get help to pay.

b. The ability to raise $2000 within a week for something important.

c. Taking a dissaving action such as “drawing on savings, increasing a credit card balance by $1,000 or more and taking out a personal loan”.

Please see Graph 2 in https://www.abs.gov.au/statistics/people/people-and-communities/general-social-survey-summary-results-australia/2014. It can be seen that of all the 3 stressors, cash flow and dissaving have higher percentages in general than the ability to raise $2000 dollars. In fact, it is the unemployed that have the most problems. The authors need to justify why the other two were not considered and why focusing on the selected item was appropriate.

AUTHOR RESPONSE:

We acknowledge that the analysis includes the question listed by the reviewer as point b without the other two questions listed.

There are two indicators described by the reviewer. Both are in reference to a 12-month period of time, which are not suitable for shorter-term changes in financial stress. Our approach measures changes across three time-points within a 6 month period. Thus, the question on “The ability to raise $2000 within a week for something important” is most relevant for our purpose.

The approach of taking point b alone has been published previously (below).

Example publications using the financial resources question without the others from the General Social Survey:

Pollock D, Shepherd CC, Adane AA, Foord C, Farrant BM, Warland J. Knowing your audience: Investigating stillbirth knowledge and perceptions in the general population to inform future public health campaigns. Women and Birth. 2021 Jul 30.

Thomas DP, Briggs V, Anderson IP, Cunningham J. The social determinants of being an Indigenous non‐smoker. Australian and New Zealand journal of public health. 2008 Apr;32(2):110-6.

Griffiths D, Sheehan L, van Vreden C et al. The Impact of Work Loss on Mental and Physical Health During the COVID-19 Pandemic: Baseline Findings from a Prospective Cohort Study. J Occup Rehabil. 2021. doi:10.1007/s10926-021-09958-7

REVIEWER COMMENT:

ii) I do not understand how, (if the responses to that one question was taken as “yes” or “no” or “don’t know”), financial stress could be measured (for each person) on a 1 to 10 scale. Is this an overall (average) measure for the county? This needs clarification.

AUTHOR RESPONSE:

We report two measures described in the final 3 sentences of the Methods>Determinants of health subsection

(1) Financial resources

Question:‘If all of a sudden you had to get $2000 for something important, could the money be obtained within a week?

Response: Yes, No or Don’t know

(2) Financial stress (current level)

Response: a 10-point scale from 1 (not at all stressed) to 10 (as stressed as can be)

These are two different measures, which are described in the final 3 sentences of the Methods>Determinants of health subsection

REVIEWER COMMENT:

5. Social Interaction – two binary measures are taken:

Was any time spent with anyone not a household member and

Whether contact was made with 7 or fewer people via speech, phone or online communication.

The original 35 item Duke Social Support Index was abbreviated to a 23-item index and also to a 11-item index for use on elderly patients who may get exhausted by the use of the longer questionnaire. However, studies have generally used either the 23 item or the 11 item DSSI. Usually these have been for older people. In this paper (not confined only to the elderly), only two items have been picked and the justification of these two items has not been given. In other words, why would we believe in the validity of a 2-item DSSI?

The authors need to justify and discuss this.

AUTHOR RESPONSE:

We have added justification for the inclusion of 2 questions that are used to derive the Social Interaction sub-scale of the Duke Social Support Index.

The following text has been added in Methods>Determinants of health:

“The social interaction sub-scale itself was not derived following participant feedback on the question regarding taking part in group meetings (such as sports or other clubs). Many of these activities were not legally permitted under lockdown restrictions and therefore if included the overall interaction sub-scale would instead partially reflect compliance with lockdown restrictions “

REVIEWER COMMENT:

6. What is the need for the variable “engagement in work” since “employment” at time of survey is also included? Is there any extra information being captured?

AUTHOR RESPONSE:

Public health measures to reduce viral transmission during the pandemic resulted in a large number of people either losing their job or being temporarily stood-down from work (but remaining employed).

The two variables referred to as ‘employment’ and ‘engagement in work’ reflect this difference, which is also reflected in the study results. Many people were employed but temporarily not engaged in work.

REVIEWER COMMENT:

• Other comments regarding results:

Turning to the results with regard to the stated objectives:

7. The negative health findings (they focus on psychological distress and mental health --- again, how were these measured?) during lockdown for both exposed and comparison group is in line with findings elsewhere in the COVID literature and so do not break new ground..

AUTHOR RESPONSE:

There is a section Methods>Health Outcomes that describes how health was measured using the Kessler-6 psychological distress scale and scores of mental and physical health from the 12-item short form health survey.

The y-axes for Figures 1 and 2 have been updated to assist the reader identify the scales upon which health outcomes are presented.

We also examine this post-lockdown.

REVIEWER COMMENT:

Unobserved heterogeneities as the two groups do not appear similar

8. During the lockdown period the mental health of the comparison group actually improved significantly from pre-lockdown. There is no further exploration of this counter-intuitive result. However, this pattern did not show up in the exposure group thus leading one to suspect that there are other factors at play or unobserved heterogeneities that have not been captured and which may, thus, invalidate this “natural experiment”.

AUTHOR RESPONSE:

The pre-lockdown setting took place during the early stages of the pandemic. We have now made this clearer in the paper. Also in the final sentence the subsection Methods>Participants date and context we describe significant shifts in health from pre-pandemic levels.

Given the significant rapid declines in mental health at the start of the pandemic (pre-lockdown), it is perhaps unsurprising that mental health would recover over time in the rest of Australia (having relatively few restrictions).

We didn’t observe any significant differences post-lockdown between the group exposed to lockdown and the control group, suggesting a similar pattern of recovery in mental health, which was delayed for those experiencing lockdown.

This is described in: Discussion>paragraph 1

“Following an abrupt national decline in mental health observed during the early stages of the pandemic, people exposed to an extended lockdown experienced a delay in the recovery of mental health observed in the rest of Australia. Following the easing of lockdown measures, we then observe a resolution of the negative mental health impacts of lockdown within a two-month period, bringing the mental health of Victorian residents back in line with that of the rest of Australia.”

REVIEWER COMMENT:

9. Right now, looking at the mental health and psychological graphs, it can be seen that the two groups did not begin at the same level. Are these differences statistically significant? This is important as otherwise, this further shows that the two groups were quite different to begin with (another reason to suspect other unaccounted for factors at play).

AUTHOR RESPONSE:

Firstly, we graphically present averages for the exposure group and comparison group across three time-points which are unadjusted, or raw data.

However, in the regression models we account for differences between the study groups by incorporating covariates into the models. After adjusting for these covariates there are no significant differences between the two groups at the first survey time-point (Table 2).

The part of Table 2 labelled ‘The Lockdown Impact’ involves another form of adjustment. It models differences at one time point whilst controlling for differences at another time point.

In summary the analytical approach accounts for the concerns of the reviewer observed from the raw averages.

REVIEWER COMMENT:

10. What accounts for the different trajectories of psychological distress? Again, the two groups appear different..

AUTHOR RESPONSE:

This paper models changes in health over time starting just after the first wave for an exposure group that experienced an extended community lockdown due to a second wave, and a comparison group that did not experience a community lockdown.

We conclude that some differences in psychological distress coincide with a community lockdown but these appear to resolve quickly after the lockdown ended.

Additionally, there are trends describing a recovery in mental health over time in general towards pre-pandemic population levels.

REVIEWER COMMENT:

• The graphs

11. The graphs are not easy to understand.

i. Since there is no discussion of “health score” in the cases of mental health and psychological distress, one does not know how to interpret the y-axes. The x-axis is not labelled.

AUTHOR RESPONSE:

Figure 1 and Figure 2 have been updated.

The y-axes are labelled according to the specific scale from which the values are taken.

The x-axis is now labelled as time.

REVIEWER COMMENT:

ii. How were the health scores obtained?

AUTHOR RESPONSE:

Health scores are described in the short subsection Methods>Health outcomes. The updated y-axis on Figures also act to emphasise the nature of the scales.

REVIEWER COMMENT:

iii. As I understand, the first dot is for data collected PRIOR to lockdown. The second dot is the data collected DURING lockdown. The third dot is the data collected POST lockdown. It seems a valuable opportunity to explore the POST lockdown period was not taken by averaging the October to December data. This would have shown the speed of recovery and might have added some new information to the literature.

AUTHOR RESPONSE:

We agree that exploring time-dependent differences in the post-lockdown time-point would be interesting for the exposed group, but this was beyond the scope of this paper.

REVIEWER COMMENT:

• Precision of estimates and statistical power:

12. Since I am not clear on the models estimated to find the mental and physical effects of a lockdown, I wanted to make sure that if any of the explanatory variables were correlated, that was clearly brought out and the implications mentioned. Employment, social interaction, and financial stress are likely to be highly correlated. Though there are methods to deal with this problem, at least the impact on the precision of the estimates needs to be mentioned.

AUTHOR RESPONSE:

Employment, social interaction and financial stress (along with mental and physical health) are defined as outcomes in the regression models and are not treated as explanatory variables.

This information can be found in

Methods>Analytical approach:

“Generalised estimating equations were used with binary logistic models for binary outcomes describing work status, social isolation, and financial resources”

Whilst we observe that changes in mental health coincide with social isolation and loss of work during lockdown, these are each treated as separate outcomes for separate models. This approach is aligned to the research questions.

REVIEWER COMMENT:

13. The state of Victoria has 6.68 million people (line 94) accounting for over a quarter of the 6.68 million population of Australia (line 95). I find it worrying that the analysis has only 305 people in Victoria (exposed group) and 593 in the rest of Australia (the comparison group). The sample sizes seem too small for the statistical tests to have much power. At the very least, this should be mentioned.

AUTHOR RESPONSE:

The sample sizes were sufficient to identify significant differences in outcomes during the lockdown (adjusted estimates reported in Table 2). And if these effects had persisted at the same level after the lockdown we would have had enough statistical power to detect these. Following the same individuals over time also provides us with more statistical power when estimating the change in outcomes as outcomes are correlated within individuals.

Within the discussion, we acknowledge the limitation of using a cohort sample, compared with population level data, and have added a remark in alignment with the reviewers comment in regard to financial stress:

Discussion:

“Furthermore, we did not identify any significant differences in financial stress due to the lockdown within our sample”

Submitted filename: Responses to reviewer comments_dec_sub.docx

Click here for additional data file.

14 Mar 2022

18 Mar 2022

Reviewer 2 did not respond provide a response on the question of whether comments were addressed, and provided a comment that our analysis does not describe PTSD.

AUTHOR RESPONSE TO REVIEWER 2 COMMENT:

The study objectives do not propose to measure post-traumatic stress disorder (PTSD) but rather it is designed to measure mental health generally – Mental Health as a person’s emotional, cognitive and intellectual status. It could be considered positive or negative at a given time.

Mental illness, or psychological disorders such as PTSD, is a different construct which can affect thinking, feeling, mood, behaviour and ability to function over time – we don’t propose to describe this outcome in our study objectives.

There is a difference between mental health and mental illness/conditions/disorders, although these can often be used interchangeably.

We claim to describe mental health using the SF-12 score (on a numeric scale) – an approach which is appropriate for this construct, validated, and well-defined in the Methods section.

We also describe a measure of overall physical health - analogously this does not describe specific types of physical impairments, or types of injury for example.

The reviewer states that the methodology does not directly capture post-traumatic stress disorders, yet the reviewer also describes that this is already acknowledged in the study limitations of the paper, citing part of the following sentence:

From Line 385:

“Whilst our findings describe a recovery in aspects of mental health such as symptoms of anxiety and psychological distress, mental disorders such as post-traumatic stress disorder may re-emerge over longer periods of time [33], or upon repeated exposure to lockdowns or crises.”

We do not propose to make claims about PTSD and we also acknowledge the limitation of the measures used in not being able to provide information about specific mental illnesses such as PTSD, Depression or Anxiety.

No changes to the text were made.

Submitted filename: Responses to reviewer comments_2022march_sub.docx

Click here for additional data file.

25 Mar 2022

The health impacts of a 4-month long community-wide COVID-19 lockdown: Findings from a prospective longitudinal study in the state of Victoria, Australia.

PONE-D-21-24391R2

Dear Dr. Griffiths,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Giuseppe Carrà, PhD

Academic Editor

PLOS ONE

31 Mar 2022

PONE-D-21-24391R2

The health impacts of a 4-month long community-wide COVID-19 lockdown: Findings from a prospective longitudinal study in the state of Victoria, Australia.

Dear Dr. Griffiths:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Giuseppe Carrà

Academic Editor

PLOS ONE