Depression and anxiety during the first and second waves of the COVID‐19 pandemic in two large, prospective, aging cohorts in rural and urban India

Abstract Introduction The COVID‐19 pandemic resulted in a wide variety of adverse consequences, including disruption of long‐term, human research studies globally. Two long‐term, prospective, aging cohort studies, namely, Srinivaspura Aging, Neurosenescence and COGnition (SANSCOG) study and Tata Longitudinal Study of Aging (TLSA), conducted in rural and urban India, respectively, had to be suspended during first and second waves of COVID‐19. Methods We conducted telephonic assessments to screen for depression and anxiety in the above two cohorts comprising of adults ≥45 years, during the first wave (2020) and second wave (2021) lockdown periods in India. Further, we included depression assessments data from two additional time periods—pre‐COVID (2019) and the “inter‐wave” period (between the first and second waves) to compare proportions of depression in these cohorts, during four distinct time periods—(i) pre‐COVID, (ii) COVID first wave lockdown, (iii) inter‐wave period, and (iv) COVID second wave lockdown (rural: 684, 733, 458, 611 and urban: 317, 297, 204, 305 respectively). Results During COVID first wave, 28.8% and 5.5% had depression and anxiety, respectively in the rural cohort. Corresponding figures in the urban cohort were 6.5% and 1.7%. During second wave, 28.8% of rural subjects had depression and 3.9% had anxiety, whereas corresponding figures in urban subjects were 13.1% and 0.66%. During the above‐mentioned four time periods, proportions of depression were: rural—8.3%, 28.8%, 16.6%, 28.8%; urban—12%, 6.1%, 8.8%, 13.1%. Conclusions Multi‐fold increase in depression among aging, rural Indians during first and second waves, with high depression among subjects ≥65 years and those with comorbidities during the first wave, is concerning. Urgent public health measures are needed to address this added mental health burden and thereby, prevent further potential adverse consequences.


| INTRODUCTION
The COVID-19 pandemic resulted in an enormous adverse impact on humankind and caused tremendous suffering all over the world. India was substantially impacted by both the first and second waves of the pandemic in 2020 and 2021, respectively. The first wave of the pandemic resulted in stringent nationwide lockdowns from late March 2020 till the end of May 2020. 1 Further, phased easing of restrictions continued for several months after this. Just as the country was on the road to recovery from the first wave, the second wave hit in April 2021. The impact of the second wave was far more detrimental than the first wave, not only because of the higher magnitude of cases and deaths but also due to its precipitous onset. [2][3][4] The dimensions of the adverse impact of this pandemic have extended far beyond the direct or acute effects of the infection or disease manifestations. In particular, the psychological impact has been prominent, though understudied and not well understood.
There could be several possible reasons for substantial, adverse mental health consequences due to COVID-19 in developing countries such as India. These include widespread fear 5,6 and stigma 7,8 regarding the disease, disruption in accessing healthcare during the lockdowns, 9,10 economic impact, 11 overburdened, and disparate healthcare systems including weak primary healthcare infrastructure 12,13 and the already high burden of mental health disorders in the country. 14,15 In this scenario, it is crucial to understand the extent and nature of the psychological impact of COVID-19 in India, so that appropriate preventive or mitigative interventions can be developed to counter or mitigate the adverse consequences.
Older adults are not only more vulnerable to COVID-19 infection along with poorer disease outcomes, 16 they are likely to have suffered considerably due to the psychosocial and economic impact of the pandemic and its associated restrictions. [17][18][19][20][21] Though studies [22][23][24][25][26][27][28][29] from other parts of the world have attempted to study the psychological impact of the pandemic in various population groups, such studies from India are limited, [30][31][32][33][34] particularly in the aging population. [35][36][37][38] We are conducting two large-scale longitudinal cohort studies-Srinivaspura Aging, Neurosenescence and COGnition (SANSCOG) study in rural India (projected n = 10,000) since 2018 and Tata Longitudinal Study of Aging (TLSA) in urban India (projected n = 1000) since 2015. These studies are aimed at studying the diverse cognitive trajectories of aging and thereby, identifying risk and protective factors for aging-related, neurodegenerative disorders such as dementia. Both these studies had to be halted during the first and second wave COVID-19 lockdowns in India. However, during both these lockdown periods, we conducted telephonic assessments on consenting subjects from both cohorts to screen for depression and anxiety. In this paper, we present the proportions of depression and anxiety in both the above cohorts during the first and second waves of the pandemic. Further, we utilized data from depression assessments that were carried out as part of the periodically scheduled clinical assessments in these two cohorts to establish two additional COVID-related time periods-"pre-COVID period" (assessments done in 2019) and "inter-wave period" (assessments done between the first wave and second waves, when the cohort studies had briefly resumed). With this additional data, we compared proportions of depression in the rural and urban cohorts, during four distinct time periods-(i) pre-COVID period, (ii) first-wave lockdown period, (iii) inter-wave period, and (iv) second-wave lockdown period.
We hypothesized that there would be significant rise in the proportion of depression during the first wave and second wave of the pandemic as compared with either the pre-COVID or the interwave periods in subjects from both cohorts.

| Study design
Mixed methods study with cross-sectional assessments as well as utilizing previously acquired longitudinal data for comparison.

| Setting
Community setting in the villages of Srinivaspura (SANSCOG study) and urban Bangalore (TLSA). Both study sites are approximately 60 miles apart, located in the southern Indian state of Karnataka.

| Cohort characteristics
Rural (SANSCOG) cohort participants are village-dwelling, predominantly from a low/lower-middle income (annual household income <200,000 Indian Rupees) agricultural background, with limited levels of literacy (ability to both read and write with understanding, in any language) and have not been exposed to substantial lifestyle changes or modernization. However, the urban (TLSA) cohort subjects are city-dwelling, highly educated, belong to the working middle/highincome (annual household income >200,000 Indian Rupees), and have seen significant lifestyle changes due to urban living in the last few decades. Details of the two cohorts, sampling frames, and recruitment strategies are given in Appendix I.

| Study sample
The study sample includes cognitively healthy, aging individuals (45 years and above), who were enrolled in two, prospective, aging cohorts, namely, SANSCOG and TLSA (described earlier). In the rural cohort, 733 and 611 subjects consented and underwent telephonic assessments for depression and anxiety during the first wave (June

| Ethics and informed consent
Both SANSCOG and TLSA studies have been cleared by the Institutional Ethics Committee of the Centre for Brain Research, Indian Institute of Science, and the collaborating institutions. Written, informed consent was obtained from all participants before recruitment into the two studies. During the lockdown periods of the first and second waves of the COVID-19 pandemic, we initially attempted to telephonically contact all participants in both rural (SANSCOG) and urban (TLSA) cohorts to enquire about their well-being and to provide them awareness regarding COVID-19-related safety precautions as well as medical guidance, when necessary. Further details of these initial rounds of calls have been presented in another manuscript. 39 During these calls, participants from both cohorts were informed about the telephonic assessments for depression and anxiety and willing participants provided voluntary, telephonic consent for these assessments.

| Assessments
To avoid bias, we used the same assessment instruments in both cohorts during both waves of the pandemic. Depression was assessed by using the 7-item version of the Geriatric Depression Scale (GDS-7) and anxiety symptoms were assessed using the 7-item version of the Generalized Anxiety Disorder Questionnaire (GAD-7).
Details of the scales, administration, scoring, and interpretation are given in Appendix II.
Even before the onset of the pandemic, regular, in-person, clinical assessments of both our cohorts included depression assessments (GDS-30) as well as collecting information on selfreported history of depression, alcohol use, smoking, and medical comorbidities. Similarly, between the first and second wave lockdowns, when we briefly resumed both studies (October 8, 2020-April 21, 2021), all the above-mentioned information was collected during the clinical assessments done in-person (as part of the regular SANSCOG & TLSA study schedules). Thus, we had two more time periods, during which depression assessments were performed in these cohorts-one, before the onset of the COVID-19 pandemic (referred to in this paper as "pre-COVID" period) and another, when assessments were conducted after the first wave and until the onset of the second wave (referred to in this paper as "interwave" period). The four distinct time periods when depression assessments were carried out in the rural and urban cohorts are represented in Figure 1. Proportions of subjects screened to have depression during the above four time periods were compared in both the cohorts.
Further, in the rural cohort, we identified a longitudinal subgroup of subjects (n = 261), who had GDS data for pre-COVID, COVID firstwave, and second-wave lockdown periods. This enabled us to observe how the proportion of depression differed in the same subgroup of subjects during these three distinct time periods. We could not include the inter-wave period in this analysis due to the parent studies' follow-up time frames (once in 3 years for subjects <65 years and once in 2 years for subjects ≥65 years). We analyzed how factors such as age, gender, literacy, presence of co-morbidities, past history of depression, alcohol use, and tobacco use played out F I G U R E 1 The flow diagram represents the number of rural (SANSCOG) and urban (TLSA) cohort participants at different stages of recruitment for undergoing telephonic assessments for depression and anxiety during the first and second wave lockdowns periods of COVID-19 in India. Also depicted below are number of rural (SANSCOG) and urban (TLSA) cohort participants for whom depression assessment data were available during four distinct COVID-related time periods: (i) Pre-COVID period, (ii) COVID first wave lockdown period, (iii) COVID interwave period, and (iv) COVID second wave lockdown period. SANSCOG, Srinivaspura Aging, Neurosenescence and COGnition; TLSA, Tata Longitudinal Study of Aging. SUNDARAKUMAR ET AL. | 3 of 10 during these different time periods-before the pandemic and during the first and second wave lockdown periods.

| Statistical analysis
Data was entered in excel sheets and were subsequently exported to the JASP open-source software (version 0.14.1). Differences in overall proportions of depression during different assessment periods in the rural as well as urban cohort were analyzed. Further analysis of the longitudinal subgroup of SANCOG cohort subjects based on age, gender, presence of comorbidity, past of depression, and substance use was done using Pearson's χ² test. A p-value of <0.05 was considered statistically significant. Cramér's V was also reported wherever the p-value was significant and interpreted as the effect size measure (Cramér's V <0.2-weakly associated, 0.2-0.6moderately associated, and >0.6-strongly associated). There was no missing data for depression and anxiety assessments in both the cohorts and hence, no imputation analysis was done.

| RESULTS
The flow diagram depicting number of rural and urban cohort participants who underwent telephonic assessments for depression and anxiety during the first-and second-wave lockdowns are displayed in Figure 1. Mean age, gender, and literacy status of rural and urban participants belonging to the 4 studied groups (pre-COVID, COVID first wave lockdown period, COVID inter-wave period, and COVID second wave lockdown period) are presented in Table 1. Among the longitudinal subgroup participants in the rural cohort (n = 261), who had depression assessments in the pre-COVID as well as COVID first wave and second wave, 56% (147/261) were males, mean age was 55.6 ± 8.8 years (males: 55.9 ± 8.9, females: 55.2 ± 8.6) and literacy rate was 70%; 183/261 (males: 129/147;88%, females: 54/114; 47%).    Table 2.
Rural subjects with one or more comorbidities were found to have a significantly higher proportion of depression than those without comorbidities (32%; 49/153 vs. 15.7%; 17/108, p = 0.003) during the COVID first wave, as represented in Figure 4. However, no significant difference was observed between these two groups in the pre-COVID and COVID second wave periods. The proportion of depression in subjects with past history (self-reported) of depression was significantly higher than those without past history of depression during all three time periods ( Figure 4)  However, we did not have pre-COVID anxiety assessments to make a T A B L E 2 Comparison of proportions of depression between groups stratified by literacy status, tobacco use, and alcohol use interesting to note that the proportions of subjects with anxiety as compared to depression in both cohorts were substantially lower.
One of the possible reasons could be that older adults in this part of the world could consider openly admitting or expressing anxiety or fear as a sign of weakness, whereas expressing distress or sadness could be more socially acceptable. It has also been previously shown that older persons are less likely to identify anxiety symptoms than depressive symptoms in themselves. 40 On sub-analyzing the longitudinal subgroup of 261 rural subjects who had undergone assessments in the pre-COVID as well as COVID first and second waves, we found that subjects in the age group ≥65 years were significantly more depressed than those in the age group <65 years only during the first wave of the pandemic. Similarly, those subjects with one or more comorbidities were significantly more depressed than those without comorbidities only during the first wave. social isolation has been associated with depression in elderly 43 and is known to predict worse disease outcomes. 44 On the other hand, the second wave's social restrictions in India were at a substantially lesser scale than the first wave's strictly enforced, nationwide lockdowns, which is potentially the reason why the trend discussed above was not seen during the second wave period. COVID period as compared to telephonic assessments using GDS 7item version during the COVID first and second wave lockdown periods). Also, we did not have adequate numbers to conduct subanalysis in the same subset of urban cohort subjects as we did for the rural subjects. Finally, these results may not be generalizable due to non-probabilistic sampling method that was used in this study. We propose to conduct follow-up assessments for depression and anxiety in our subjects in the coming months, with larger sample sizes, to further observe how trends in depression and anxiety play out in the future.
Geriatric depression is a mounting challenge in India, especially considering the sharp rise in its proportion of older persons in recent years. 45 Wide variations have been reported in the prevalence of elderly depression in India depending on study settings, geographical regions, assessment tools used, and so on. 46,47 Depression in the elderly is known to be associated with higher risk for cognitive decline, 48 physical disability, 49,50 frailty, 51 and mortality. 52 The current COVID-19 crisis has further exposed the challenge of mental health disorders in the elderly and brought to light the challenges of delivering adequate mental healthcare to the elderly population of India. 53 Prompt assessment of the impact of COVID-19 on mental health of the elderly is essential to put in place mitigative measures or F I G U R E 4 This figure highlights the differences in proportions of depression between subjects with/without comorbidities and with/ without self-reported history of depression among a longitudinal subgroup of rural cohort subjects (n = 261) who had depression assessments pre-COVID as well as during the first wave and second waves of COVID.
develop preventive strategies for the furture. 54

ACKNOWLEDGMENTS
We would like to acknowledge the TLSA and SANSCOG study teams for staying connected with our cohorts and carrying out telephonic assessments during both lockdown periods. We would like to place our heartfelt thanks to all our study subjects for their participation and kind co-operation. SANSCOG study is funded by Mr. Kris Gopalakrishnan (Cofounder, Infosys) through the Centre for Brain Research and TLSA is funded by Tata Trusts. The funding source did not have any role in the study design, data collection, analysis, and interpretation, writing of the manuscript or the decision to submit the manuscript for publication.

CONFLICT OF INTEREST
The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT
Data will be shared in conformity with statutory requirements of the government of India, through the Alzheimer's Disease Data Initiative (ADDI) platform.

ETHICS STATEMENT
This study involves human participants and was approved by the

TRANSPARENCY STATEMENT
The lead author Jonas S. Sundarakumar affirms that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been explained.