Interactions Between Early-Life Adversity, Pandemic Stress, and Social Support on Psychiatric Disorders in a Nationally Representative Sample of Canadian Adults

doi:10.21203/rs.3.rs-4614109/v1

Download PDF

Research Article

Interactions Between Early-Life Adversity, Pandemic Stress, and Social Support on Psychiatric Disorders in a Nationally Representative Sample of Canadian Adults

https://doi.org/10.21203/rs.3.rs-4614109/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

PURPOSE

Mental health problems increased during the COVID-19 pandemic, but the interaction between pandemic stress, early-life adversity, and social support in predicting mental health is not well understood. This study examined independent and interactive associations between pandemic stress, early-life adversity, and social support on psychiatric disorders in a nationally-representative sample of Canadian adults.

METHODS

Participants (n = 9,409) were from the Mental Health and Access to Care Survey, a cross-sectional survey of Canadian adults during the pandemic. Measures included pandemic stressors (Statistics Canada), early-life adversity (Childhood Experiences of Violence Questionnaire), social support (Social Provisions Scale), and past 12-month psychiatric problems (World Mental Health-Composite International Diagnostic Interview). Statistical analyses included two-step logistic regression models adjusted for covariates and weighted for complex survey design.

RESULTS

Early-life adversity (aOR = 1.24 [1.15–1.35]-aOR = 1.53 [1.39–1.69]) and pandemic stress (aOR = 1.18 [1.12–1.25]-aOR = 1.32 [1.26–1.39]) were associated with higher odds of psychiatric problems. Significant interactions between early-life adversity and pandemic stress for depression (aOR = 0.96 [0.93–0.98]) suggested an attenuated effect of pandemic stress at higher levels of early-life adversity. Social support was negatively associated with psychiatric problems (aOR = 0.88 [0.86–0.91]-aOR = 0.97 [0.94–0.99]), while pandemic stress was positively associated (aOR = 1.20 [1.15–1.26]-aOR = 1.33 [1.27–1.40]). An interaction between social support and pandemic stress for suicidality (aOR = 1.02 [1.01–1.03]) suggested higher odds of pandemic stress at higher levels of social support.

CONCLUSION

Early-life adversity and pandemic stress increased psychiatric disorder likelihood, while social support was protective. However, complex interactions indicate nuanced relationships in mental health risk during the pandemic.

'COVID-19 pandemic'

'Pandemic stress'

'Early-life adversity'

'Social support'

'Psychiatric disorders'

‘Epidemiology’

During the COVID-19 pandemic, numerous financial, health, and social stressors were shown to significantly increase the risk of mental health problems [1]. During the early pandemic period, approximately one-quarter of Canadians screened positive for an affective disorder, which increased the risk for substance abuse problems [2]. These elevated levels of distress led to a 50% increase in the average spending rates on mental health services [3]. Psychiatric problems are associated with both long-term disability and early mortality, and thus remain a critical concern even in the post-pandemic period [4, 5]. There is also a need to understand how pandemic stressors interact with preexisting vulnerability factors in predicting mental health outcomes in the general population, and how certain environmental factors may buffer against these stressors in large-scale representative surveys in order to inform preventative public health strategies for managing current and future crises.

Early-Life Adversity and Pandemic Stress

Early-life adversity describes a range of adverse experiences that occur in childhood and adolescence, including direct forms of maltreatment, with anticipated negative effects on psychosocial and physical health. These include direct exposure to threat (e.g., physical and sexual abuse), observing violence (e.g., intimate partner violence and community violence), and deprivation of development-enhancing experiences (e.g., emotional/physical neglect). Approximately 60% of individuals report at least one type of adversity in their lifetime, with about one-quarter of the population experiencing three or more adversities [6]. Early-life adversity increases the risk of many negative health outcomes, including psychopathology, physical disability, and premature mortality [7–9]. It has been estimated that nearly one-third of all mental health disorders across the life course are attributable to early-life adversity [10–12]. Beyond the significant impact on individuals, there is also a considerable social burden associated with early-life adversity. For example, it has been estimated that a 10% reduction in early-life adversity would generate annual savings of 3 million Disability-adjusted life years (~$105 billion USD) [13].

While the long-term risks of early-life adversity are well documented, it is unclear how these experiences intersect with stressors imbued by the COVID-19 pandemic in predicting mental health problems. Understanding these associations at the population level is crucial given significant individual differences in stress responsiveness that, in part, generate and sustain mental health inequities. The pandemic context provides an opportunity to capture greater variability in recent stress exposure across individuals and ultimately determine how early-life adversity may moderate the effect of pandemic stress on mental health problems in the general population. To date, findings on the joint effects of early-life adversity and recent stress have been mixed. Some evidence suggests a stress sensitizing effect of early-life adversity [14], whereby exposure to early-life adversity increases the effect of later stress, thus potentiating risk for poor psychiatric outcomes [14–16]. There is also evidence for stress accumulation, whereby the effects of early-life adversity and later stress independently impair functioning and increase risk for poor psychiatric outcomes [17]. There are also documented sex differences in these effects, where a smaller study conducted during the pandemic identified evidence of stress accumulation in female caregivers and stress sensitization in male caregivers, which was contingent on the mental health outcomes examined [18]. These associations have yet to be examined in nationally-representative studies during the pandemic.

Social Support and Pandemic Stress

Whereas early-life adversity may magnify the association between pandemic stress and psychopathology, certain positive experiences may buffer against the effect of these stressors. Social support has been suggested to play such a role, and may have an especially important benefit in times where social capital is essential for promoting well-being (i.e., during the pandemic). Evidence supports the promotive (i.e., main) effects of social support in reducing risk for mental health problems during the pandemic. One study during the first year of the pandemic identified social support to be associated with a 46% decrease in the risk of depression [19]. This study also suggested that social support may buffer (i.e., interact with) pandemic-related financial stressors on depression. A smaller study of college students made a similar observation, whereby social support buffered the effect of social isolation on psychological health during the pandemic [20]. Again, however, evidence of either promotive or buffering effects of social support against pandemic stress on psychiatric problems in nationally-representative surveys is limited.

The Present Study

This study examined associations between pandemic stressors during the COVID-19 pandemic and psychiatric disorders and associated outcomes in a nationally-representative sample of Canadian adults. The first objective was to examine independent and interactive associations between pandemic stress and early-life adversity with psychiatric outcomes. We hypothesized that both pandemic stress and early-life adversity would have significant positive associations with a range of psychiatric outcomes, and that associations between pandemic stress and psychiatric problems would be exacerbated by higher levels of early-life adversity. The second objective was to examine independent and interactive associations between pandemic stress and social support with psychiatric outcomes. We hypothesized that pandemic stress and social support would have significant but opposing effects on a range of psychiatric outcomes, and that social support would attenuate the association between pandemic stress on psychiatric problems. We tested these associations across a range of psychiatric problems, but made no disorder- or domain-specific hypotheses given their shared etiology and transdiagnostic overlap [21]. Finally, we explored whether the abovementioned associations were moderated by sex [22, 23]. While we hypothesized that these effects would be observed in both sexes, we speculated that early-life adversity and pandemic stressors would have a stronger effect on suicidality and substance use disorders in males [24], and a stronger effect on depression and anxiety in females [25].

Study Population

Participants were selected from the Mental Health and Access to Care Survey (MHACS), a survey conducted by Statistics Canada with the broader goal of collecting information about the mental health of Canadians [26]. MHACS is a nationally-representative, cross-sectional survey that was completed over the period March to July 2022. The sampling frame was developed from the 2021 Census and represented all provinces of Canada (excluding territories); however, individuals from minoritized groups were oversampled. Individuals living on reserves and other Indigenous settlements, full-time members of the Canadian Forces, and those in collective dwellings (e.g., institutional residences) were not sampled in MHACS (approximately 3% of the Canadian population meets one of these criteria). Participants were 15 years or older and, of the 39,485 individuals sampled, a response was obtained from 9,861 individuals (response rate of 25%). Non-response adjustment weighting was implemented to reduce bias, and the population was determined to be well-represented when compared with the 2021 Census. Data were stored and accessed at a secure Research Data Centre (RDC) facility at the University of Toronto. Individuals under 18 years of age (n = 452) did not report on early-life adversity and were therefore excluded, leaving an eligible sample for analyses of n = 9,409.

Measures

Pandemic Stress. A total score of pandemic stressors was derived from a set of 14-items asking the respondent, “have you experienced any of the following impacts due to the COVID-19 pandemic?”. Items included stressors relating to finances (e.g., loss of job/income, inability to meet financial obligations or essential needs), service access (e.g., difficulty accessing required health care services), physical health (e.g., diagnosis of COVID-19, hospitalization), and distress/isolation. These items were responded to with a binary response option (i.e., “Yes” [1] or “No” [0]). Items were summed to produce a total score of different stressors experienced, whereby a higher score indicated more stressors (range = 0–14). Acceptable internal consistency was observed across items (α = .76). Due to a possible overlap in the qualities of a subset of these items (i.e., “Feelings of loneliness or isolation”, “Emotional distress”, and “Physical health problems”) with our psychiatric outcomes (e.g., depression symptoms), a sensitivity analysis was conducted which omitted these items and re-summed the overall scale (range = 0–11).

Early-Life Adversity. A modified, 6-item short version of the Childhood Experiences of Violence Questionnaire (CEVQ-SF) was used to assess early-life adversity. Respondents were asked about, “things that may have happened to you before you were 16 in your school, in your neighbourhood, or in your family”, with items related to observing intimate partner violence, direct physical abuse, and direct sexual abuse. Respondents specify how many times the events occurred: “Never”, “1 to 2 times”, “3 to 5 times”, “6 to 10 times”, and “More than 10 times”. Each of these items were dichotomized (“Never” [0] vs. “1 or more” [1]) and a summed scale of early-life adversity was produced (range = 0–6), with higher scores indicating more types of early-life adversity. The CEVQ-SF has been demonstrated to have strong psychometric properties, as well as good internal consistency (α = .85) [27].

Social Support. The 10-item version of the Social Provisions Scale (SPS-10) was used to measure current social supports. The SPS-10 is derived from a larger 24-item scale (SPS-24) and comprises ten items where individuals report “to what extent each statement describes your current relationships with other people” (e.g., people they can depend on, people that provide emotional security, people that provide advice). Items are responded to on a four-point Likert scale ranging from “Strongly agree” to “Strongly disagree”. Items are summed to produce a total social support score (range = 10–40), with higher scores indicating higher levels of perceived social support. The SPS-10 has been demonstrated to have strong psychometric properties, with strong concurrent validity with the SPS-24 and similar predictive power, as well as good internal consistency (α = .88) [28].

Psychiatric Outcomes. The following psychiatric outcomes were assessed in the current study: major depressive episode, bipolar disorder, generalized anxiety disorder, social anxiety disorder, substance use disorder, and suicidality. All psychiatric outcomes (except suicidality) were assessed with an adapted version of the World Mental Health-Composite International Diagnostic Interview (CIDI) [29], based on criteria drawn from the Diagnostics and Statistics Manual, Fourth Edition (DSM-IV). Evaluations of reliability and validity have consistently demonstrated the effectiveness of the CIDI in discerning the presence or absence of psychiatric disorder across various clinical and non-clinical contexts [30]. Suicidality was measured using three items in which the respondent specified whether they had thought about, planned, or attempted suicide. This variable was dichotomized, such that endorsement of any suicidality was treated as “Present suicidality” and endorsing “No” for all three items was treated as “Absent suicidality”. To capture outcomes particular to the pandemic period, all outcomes were based on the previous twelve-month period and were reported as “Present” [1] or “Absent” [0].

Covariates. Multiple covariates were considered, where variables that could confound the association between pandemic stress, early-life adversity, social support, and psychiatric outcomes were controlled. Covariates included sex at birth (“Female” [0] vs. “Male” [1]), age (continuous), household income (we dichotomized the income variable to reflect Canada’s household poverty line, “>=$40k” [0] vs. <$40k), minoritized group (“No” [0] vs. “Yes” [1]), and marital status (dichotomized as “Married/Living common law” [0] vs. “Not married/Not living common law” [1]). Sex was explored as a moderator in the primary analyses.

Statistical Analyses

In order to adjust for unequal selection probabilities due to the complex survey design, the MHACS design effect weights (adjusted for non-response) were incorporated into to all descriptive and inferential modelling. In addition, variance estimates were produced using bootstrap weights (BSW1-BSW1000) to incorporate variability coming from MHACS design. Weighted prevalence estimates were calculated for all psychiatric outcomes, as well as weighted means and frequencies for all predictors and covariates, stratified across subgroups with and without a psychiatric disorder. Unweighted frequency counts are presented, whereas proportions were weighted. Separate two-step logistic regression models were conducted for each psychiatric outcome. The first set of models assessed the main effects of early-life adversity and pandemic stress on psychiatric outcomes (Model 1). The second step then added the interaction between early-life adversity and pandemic stress (Model 2). The second set of models assessed the main effects of social support and pandemic stress on psychiatric outcomes (Model 1). The second step then added the interaction between social support and pandemic stress (Model 2). All models were then replicated with sex as a moderator, with main and interactive effects included (i.e., main effects, two-way interactions, and three-way interactions involving sex). To address the issue of multiple comparisons, we applied a conservative Bonferroni correction (α = 0.05/6 psychiatric outcomes in each model = .0083).

For each outcome of interest, cases were excluded when outcome data was missing (substance use disorder [n = 679 missing]; generalized anxiety disorder [n = 499]; major depression [n = 390]; social anxiety disorder [n = 239]; bipolar disorder [n = 182]; suicidality [n = 102]) but were maintained in other outcome analyses if data was available. Fractional Hotdeck Imputation (FHDI) [31] was used to impute missing data for predictors and covariates, with the highest amount of data missing for social support (n = 715) and early-life adversity (n = 579). FHDI is a robust method for imputation with complex survey designs, where one missing recipient can be imputed with multiple fractional donors (10 donors used in the present study). Each donor contributes a fraction of the original weight of the recipient, whereby the sum of fractional donor weights is equivalent to the single recipient. This analysis then recalculates study survey weights and bootstrapped weights to use in subsequent analyses. Full details of missing data are presented in supplemental material (Table S1). Continuous variables were mean centered in all models to reduce multicollinearity between predictors involved in interactions and to assist in interpreting model intercepts. All analyses were conducted using SAS Systems for Windows (version 9.4) [32]. SAS procedures accommodating complex survey design were used for all imputation, descriptive, and inferential analyses (i.e., PROC SURVEYIMPUTE, PROC SURVEYFREQ, PROC SURVEYMEANS, PROC SURVEY LOGISTIC) [32].

Descriptive Statistics

Table 1 shows the characteristics of the 9,409 participants included in the study, with a mean age 48.66 years, of whom 51% were female. Descriptive statistics are reported on participants with complete data (missing data reported in Supplemental materials, Table S1). Overall prevalence estimates were 8.21% for major depression, 5.11% for generalized anxiety disorder, 1.97% for bipolar disorder, 6.63% for social anxiety, 3.75% for substance use disorder, and 3.58% for suicidality. The results of independent samples t-tests and Rao-Scott chi-square tests of independence demonstrated that, for all psychiatric outcomes, individuals with a present disorder/suicidality were younger and not married/cohabitating compared to those without. Individuals from minoritized groups had lower prevalence of generalized anxiety disorder, social anxiety disorder, and substance use disorder as compared to those from non-minoritized groups. Females had significantly higher prevalence of major depression, generalized anxiety disorder, and social anxiety disorders than males, while males had significantly higher prevalence of substance use disorders. Higher prevalence of major depression, generalized anxiety disorder, bipolar disorder, and suicidality were reported among those in low-income households compared to those in higher-income households.

Table 1. Descriptive statistics for predictors variables and covariates, stratified by presence or absence of major depressive episode, generalized anxiety disorder, bipolar disorder, social anxiety disorder, substance use disorder, and suicidality (n=9,409)
	Major Depressive Episode			Generalized Anxiety Disorder			Bipolar Disorder
	Yes	No	p-value*	Yes	No	p-value*	Yes	No	p-value*
	Mean / %	Mean / %	p-value*	Mean / %	Mean / %	p-value*	Mean / %	Mean / %	p-value*
Age	39.22 (0.72)	49.27 (0.12)	<.0001	40.77 (0.88)	48.91 (0.11)	<.0001	36.41 (1.5)	48.88 (0.10)	<.0001
Early-life adversity	1.64 (0.08)	0.92 (0.02)	<.0001	1.70 (0.1)	0.94 (0.02)	<.0001	1.99 (0.16)	0.95 (0.02)	<.0001
Social support	32.55 (0.31)	35.46 (0.06)	<.0001	31.96 (0.39)	35.41 (0.06)	<.0001	32.40 (0.61)	35.29 (0.06)	<.0001
Pandemic stressors	4.55 (0.13)	2.34 (0.03)	<.0001	4.71 (0.16)	2.38 (0.03)	<.0001	4.81 (0.25)	2.45 (0.03)	<.0001
Sex			<.0001			<.0001			.9617
Female	8.99	91.01		6.59	93.41		1.98	98.02
Male	5.77	94.23		3.58	96.42		1.96	98.04
Visible Minority			.0564			<.0001			.3197
Yes	6.39	93.61		3.34	96.66		1.73	98.27
No	7.69	92.31		5.74	94.26		2.05	97.95
Low Household Income (<$40k)			.0145			.0478			.0160
Yes	9.94	90.06		6.91	93.09		3.34	96.66
No	7.12	92.88		4.91	95.09		1.80	98.20
Marital Status			<.0001			<.0001			<.0001
Yes	11.10	88.90		7.40	92.60		3.21	96.79
No	4.88	95.12		3.53	96.47		1.12	98.88

Table 1. cont.
	Social Anxiety Disorder			Substance Use Disorder			Suicidality
	Yes	No	p-value*	Yes	No	p-value*	Yes	No	p-value*
	Mean / %	Mean / %	p-value*	Mean / %	Mean / %	p-value*	Mean / %	Mean / %	p-value*
Age	37.09 (0.75)	49.34 (0.11)	<.0001	36.03 (0.90)	48.98 (0.12)	<.0001	38.49 (1.17)	49.07 (0.1)	<.0001
Early-life adversity	1.43 (0.08)	0.94 (0.02)	<.0001	1.54 (0.10)	0.95 (0.02)	<.0001	2.00 (0.11)	0.94 (0.02)	<.0001
Social support	32.97 (0.29)	35.40 (0.06)	<.0001	34.20 (0.39)	35.26 (0.06)	0.0078	31.77 (0.5)	35.36 (0.06)	<.0001
Pandemic stressors	4.08 (0.14)	2.40 (0.03)	<.0001	3.92 (0.18)	2.44 (0.03)	<.0001	4.30 (0.2)	2.44 (0.03)	<.0001
Sex			<.0001			0.0005			0.3818
Female	8.57	91.43		2.86	97.14		3.79	96.21
Male	4.62	95.38		4.69	95.31		3.36	96.64
Visible Minority			<.0001			<.0001			0.4682
Yes	4.81	95.19		2.02	97.98		3.24	96.76
No	7.41	92.59		4.31	95.69		3.60	96.40
Low Household Income (<$40k)			.3089			0.2868			0.002
Yes	7.64	92.36		4.57	95.43		5.91	94.09
No	6.53	93.47		3.66	96.34		3.29	96.71
Marital Status			<.0001			<.0001			<.0001
Yes	9.65	90.35		2.00	98.00		2.16	97.84
No	4.56	95.44		6.30	93.70		5.64	94.36
* = Bivariate analyses (Rao-Scott Chi-Square for binary variables & independent samples t-test for continuous variables; tested at alpha level of .05) Standard errors displayed in parentheses All values weighted to account for complex survey design

Table 2

Main and interactive effects between pandemic stressors, early-life adversity, and social support on major depressive episode and generalized anxiety disorder
	Major Depressive Episode (n = 9,019)						Generalized Anxiety Disorder (n = 8,910)
	Model 1: Main Effects			Model 2: Interactions			Model 1: Main Effects			Model 2: Interactions
	aOR	Lower CI	Upper CI	aOR	Lower CI	Upper CI	aOR	Lower CI	Upper CI	aOR	Lower CI	Upper CI
Early-life adversity and Pandemic Stress
Early-life Adversity	1.32	1.22	1.42	1.42	1.34	1.50	1.31	1.20	1.43	1.34	1.24	1.45
Pandemic Stress	1.31	1.25	1.37	1.31	1.27	1.35	1.32	1.26	1.39	1.32	1.27	1.37
Early-life adversity x Pandemic Stress	--	--	--	0.96	0.93	0.98	--	--	--	0.99	0.95	1.02
Social Support and Pandemic Stress
Social Support	0.90	0.88	0.92	0.89	0.86	0.91	0.88	0.86	0.91	0.88	0.84	0.91
Pandemic Stress	1.33	1.27	1.38	1.33	1.29	1.37	1.33	1.27	1.40	1.33	1.29	1.38
Social Support x Pandemic Stress	--	--	--	1.00	1.00	1.01	--	--	--	1.00	0.99	1.01
All models computed with complex survey weights and adjusted for age, sex, minority status, marital status, and household income. Model 1—main effects only. Model 2—main effects and interaction. Both sets of models adjusted for covariates. aOR = Adjusted Odds Ratio CI = Confidence Interval (95%)

Table 3

Main and interactive effects between pandemic stressors, early-life adversity, and social support on bipolar disorder and social anxiety disorder
	Bipolar Disorder (n = 9,227)						Social Anxiety Disorder (n = 9,170)
	Model 1: Main Effects			Model 2: Interactions			Model 1: Main Effects			Model 2: Interactions
	aOR	Lower CI	Upper CI	aOR	Lower CI	Upper CI	aOR	Lower CI	Upper CI	aOR	Lower CI	Upper CI
Early-life adversity and Pandemic Stress
Early-life Adversity	1.47	1.28	1.68	1.56	1.40	1.72	1.24	1.15	1.35	1.30	1.21	1.38
Pandemic Stress	1.29	1.21	1.39	1.30	1.24	1.37	1.20	1.14	1.25	1.19	1.15	1.24
Early-life adversity x Pandemic Stress	--	--	--	0.97	0.93	1.02	--	--	--	0.97	0.95	0.99
Social Support and Pandemic Stress
Social Support	0.92	0.89	0.96	0.90	0.85	0.95	0.90	0.88	0.92	0.90	0.87	0.92
Pandemic Stress	1.33	1.24	1.42	1.34	1.27	1.41	1.20	1.15	1.26	1.20	1.16	1.25
Social Support x Pandemic Stress	--	--	--	1.01	1.00	1.02	--	--	--	1.00	1.00	1.01
All models computed with complex survey weights and adjusted for age, sex, minority status, marital status, and household income. Model 1—main effects only. Model 2—main effects and interaction. Both sets of models adjusted for covariates. aOR = Adjusted Odds Ratio CI = Confidence Interval (95%)

Table 4

Main and interactive effects between pandemic stressors, early-life adversity, and social support on substance use disorder and suicidality
	Substance Use Disorder (n = 8,730)						Suicidality (n = 9,307)
	Model 1: Main Effects			Model 2: Interactions			Model 1: Main Effects			Model 2: Interactions
	aOR	Lower CI	Upper CI	aOR	Lower CI	Upper CI	aOR	Lower CI	Upper CI	aOR	Lower CI	Upper CI
Early-life adversity and Pandemic Stress
Early-life Adversity	1.33	1.20	1.48	1.38	1.27	1.48	1.53	1.39	1.69	1.54	1.43	1.65
Pandemic Stress	1.18	1.12	1.25	1.18	1.13	1.24	1.20	1.13	1.27	1.20	1.14	1.26
Early-life adversity x Pandemic Stress	--	--	--	0.98	0.95	1.01	--	--	--	1.00	0.96	1.03
Social Support and Pandemic Stress
Social Support	0.97	0.94	0.99	0.97	0.93	1.00	0.89	0.86	0.92	0.87	0.83	0.90
Pandemic Stress	1.22	1.15	1.28	1.22	1.16	1.27	1.23	1.16	1.30	1.26	1.20	1.32
Social Support x Pandemic Stress	--	--	--	1.00	0.99	1.01	--	--	--	1.02	1.01	1.03
All models computed with complex survey weights and adjusted for age, sex, minority status, marital status, and household income. Model 1—main effects only. Model 2—main effects and interaction. Both sets of models adjusted for covariates. aOR = Adjusted Odds Ratio CI = Confidence Interval (95%)

Early-Life Adversity and Pandemic Stress

In Model 1 (main effects only), controlling for pandemic stress, early-life adversity was significantly associated with all psychiatric outcomes (adjusting for multiple comparisons), where a one-unit increase in the number of adversities was associated with a 24–53% increased odds of experiencing a psychiatric problem. The strongest association was observed for suicidality (aOR = 1.53 [1.39, 1.69]) and the weakest association was observed for social anxiety (aOR = 1.24 [1.15, 1.35]). Similarly, controlling for early-life adversity, pandemic stress was significantly associated with all psychiatric outcomes, where a one-unit increase in the number of pandemic stressors was associated with an 18–32% increased odds of experiencing a psychiatric problem. The strongest association was observed for generalized anxiety disorder (aOR = 1.32 [1.26, 1.39]), and the weakest was observed for substance use disorder (aOR = 1.18 [1.12, 1.25]). However, for major depression, these effects need to be interpreted cautiously in light of a significant interaction between early-life adversity and pandemic stress (described below).

In Model 2 (main effects and interaction), a significant interaction between early-life adversity and pandemic stress was observed for major depression only (aOR = 0.96 [0.93, 0.98]) (adjusting for multiple comparisons). The pattern of this interaction suggested that, when pandemic stressors were low, the probability of depression was lower when early-life adversity was low (i.e., lower intercept) compared to when early-life adversity was high (Supplement Figure S1). However, when early-life adversity was low, there was a sharper acceleration in the probability of depression as pandemic stress increased. In contrast, when early-life adversity was high, although the intercept was higher than when early-life adversity was lower, a more gradual and linear increase in the probability of depression was observed. At the highest levels of pandemic stress, those with fewer adversities had higher odds of depression compared to those with more adversities.

Social Support and Pandemic Stress

In Model 1, social support was a significant promotive factor for all psychiatric outcomes (adjusting for multiple comparisons), where a one-unit increase in the number of social supports was associated with a 3–12% decreased odds of experience a psychiatric problem. The strongest association was observed for generalized anxiety disorder (aOR = 0.88 [0.86, 0.91]), and the weakest association was observed for substance use disorder (aOR = 0.97 [0.94, 0.99]). Controlling for social support, pandemic stress was significantly associated with all psychiatric outcomes, where a one-unit increase in the number of pandemic stressors was associated with a 20–33% increased odds of experiencing a psychiatric problem. The strongest association was observed for major depression (aOR = 1.33 [1.27, 1.38]), generalized anxiety disorder (aOR = 1.33 [1.27, 1.40]), and bipolar disorder (aOR = 1.33 [1.24, 1.42]), and the weakest was observed for social anxiety (aOR = 1.20 [1.15, 1.26]).

In Model 2 (main effects and interaction), a significant interaction between social support and pandemic stress was observed for suicidality (aOR = 1.02 [1.01, 1.03]). The pattern of this interaction suggested that, when pandemic stressors were low, the probability of suicidality was higher when social support was lower (i.e., higher intercept) compared to when social support was higher (Supplement Figure S2). However, there was a reversal in the direction of association from an apparent decrease in suicidality as pandemic stress increased at low levels of social support, to an increase in suicidality as pandemic stress increased at high levels of social support. However, it is important to note that the confidence bands were quite wide at low levels of social support, and the average level of social support was quite high overall (mean of 35.22 on a scale from 10–40). The reliability of the negative association between pandemic stress and suicidality at lower levels of social support is therefore questionable. Moreover, even at relatively high levels of pandemic stress, the odds of suicidality were not appreciably different at low compared to high levels of social support, with the mean odds of suicidality across levels of pandemic stressors generally higher at low compared to high levels of social support. Thus, caution should be exercised in interpreting this interaction effect, which was only observed for suicidality.

Sex Effects

No significant effects of sex were observed in any models for depression, bipolar disorder, generalized anxiety, social anxiety, bipolar, or suicidality after adjustment for multiple comparisons. For substance use disorder, social support was a promotive factor for females (aOR = 0.89 [0.84, 0.94]), but not for males (aOR = 1.01 [0.97, 1.05]).

Sensitivity Analysis

We tested the same models wherein items from the pandemic stress scale were removed to reduce overlap with psychiatric outcomes. The results were substantively similar to the main primary findings. In order to further investigate the reliability of the observed interactions in light of low numbers of participants in the tail of the distribution for high early-life adversity and low social support, sensitivity analyses were conducted wherein early-life adversity and social support were recoded into categorical variables. The results are presented in the Supplement (Text S1). Both interactions reported above were no longer observed when using categorical predictors, suggesting caution should be exercised when interpreting the interactions.

The present study examined independent and joint associations of early-life adversity, social support, and pandemic stress on psychiatric outcomes in a nationally-representative sample of Canadian adults during the COVID-19 pandemic. To our knowledge, this is the first demonstration of how pandemic stress interacts with pre-existing adversities and current social support in predicting a range of psychiatric outcomes within a large nationally-representative survey. The results add to a large body of evidence, mostly from smaller-scale studies with selected samples, in showing the complex associations between pandemic stress, early-life adversity and social support on the well-being of Canadian adults.

Consistent with our hypotheses and prior literature, early-life adversity and pandemic stress were independently associated with higher odds of psychiatric outcomes. These effects were observed across psychiatric outcomes. In contrast, our hypotheses pertaining to the interactive effects of early-life adversity and pandemic stress were not supported. It was hypothesized that early-life adversity would exacerbate the association between pandemic stress and psychiatric outcomes. Instead, results showed that the positive association between pandemic stress and depression was weaker at higher levels of early-life adversity. At low to moderate levels of pandemic stress, those with more early-life adversity had a higher likelihood of experiencing depression that those with less early-life adversity, while at higher levels of pandemic stress, there was lower odds of depression among those experiencing more early-life adversity. Overall, while this interaction effect was significant, it was relatively weak. Moreover, it is important to note that the confidence bands increased (reflecting greater variation) at higher levels of pandemic stress—likely due to fewer scores in this range—and that the findings at higher levels of pandemic stress may thus be less reliable than at low-to-medium levels of pandemic stress. When interpreting this finding it is also important to consider the possibility for survivor bias [33], where the sample in question may represent individuals who, on average, have shown stronger resilience in the face of early-life adversity than those not surveyed. Finally, there may be factors that were unaccounted for (e.g., at family and community levels) which explain why some individuals with higher levels of early-life adversity are able to cope with the negative impact of pandemic stress with respect to their emotional well-being [34].

Additionally, while it was hypothesized that social support would buffer against the impacts of pandemic stress on psychiatric outcomes, results showed that the positive association between pandemic stress and suicidality was stronger at higher levels of social support. Importantly, at low to moderate levels of pandemic stress, the odds of suicidality were lower among those with higher social support, yet at high levels of pandemic stress, there appeared to be an increased odds of suicidality among those with higher social support. This latter finding, while unexpected, may reflect a stress proliferation effect in which stressors in one domain create stressors in other domains [35]. In this case, individuals with greater social networks may be more attuned to the suffering of those in their network and the broad-scale risks posed by the pandemic, thus increasing the impact of pandemic stress on their own mental health. It may also be that the added demands of having to maintain social networks and support others in a time of ecological crises compromises their ability to handle increasing numbers of stressors, ultimately precipitating odds for suicidality. Again, however, interpretation of these results should consider the relatively modest size of the effect and wide confidence bands at higher levels of pandemic stress. Further research is needed to identify the nuances of social support and the potential upside and downside to managing the demands of maintaining social networks during periods of elevated stress periods. Moreover, we did not distinguish between suicidal ideation, plans, and attempts, yet there may be important differences between these outcomes with respect to the ways in which social support and recent stress interact. Overall, as the size of all observed interactions was small, there is a need for replication of other national cohorts to explore the robustness of the findings.

Finally, our hypotheses pertaining to sex differences were not supported, and instead, we observed a significant association between social support and substance use disorders in females as compared to males. This suggests that males benefitted less from social support than females with respect to substance use. While typically social support has been demonstrated to decrease the odds of psychiatric disorder for both sexes, consistent with the present findings, evidence has shown this effect to be stronger in females as compared to males [36].

Strengths and Limitations

This study has several strengths, foremost being that it is the first study to investigate the association between pandemic stress, early-life adversity, and social support in relation to psychiatric outcomes in a nationally-representative sample of Canadian adults during the pandemic. Second, the analytic approach was robust, considering important covariates and effect modifiers, as well as incorporating study survey weights and imputation-adjusted survey weights to ensure the findings are generalizable. The assessment of multiple psychiatric outcomes was made possible by the large study sample, which prior studies have often been unable to disentangle. However, there are important limitations to consider in interpreting the findings. First, the cross-sectional nature of the survey design precludes the ability to comment on the directionality of the findings. Relatedly, while many studies use retrospective reporting of early-life adversity, some evidence suggests concurrent psychiatric factors may influence the reporting of traumatic childhood experiences, potentially biasing results [37]. Finally, while Statistics Canada adjusted the weighting of the sample to reflect the Canadian population, the response rate for the sample was low (approximately 25% of those surveyed). Thus, it is possible that individuals most significantly impacted by the pandemic (and less likely to complete the survey) may be underrepresented in the sample.

This is the first nationally-representative study of Canadians to explore the interplay between pandemic stress, early-life adversity, and social support on psychiatric disorders and suicidality. Pandemic stressors and early-life adversity were associated with increased odds for psychiatric outcomes, whereas social support was a promotive factor, reducing odds. Contrary to hypotheses, high levels of early-life adversity did not have a sensitizing effect by exacerbating the effect of pandemic stress on psychiatric outcomes, and social support did not buffer against pandemic stress in the expected way. As these interactions deviated from our expected hypotheses, future research should attempt to replicate these findings with other national cohorts to examine the robustness of these observations.

Funding:

The authors did not receive support from any organization for the submitted work.

No funding was received to assist with the preparation of this manuscript.

No funding was received for conducting this study.

No funds, grants, or other support was received.

Conflicts of interest/Competing interests:

The authors have no relevant financial or non-financial interests to disclose.

The authors have no conflicts of interest to declare that are relevant to the content of this article.

All authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript.

The authors have no financial or proprietary interests in any material discussed in this article.

Author Contribution

All authors contributed to the study conception and design. Material preparation and analysis were performed by D.J. The first draft of the manuscript was written by D.J. and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Manchia M, Gathier AW, Yapici-Eser H et al (2022) The impact of the prolonged COVID-19 pandemic on stress resilience and mental health: A critical review across waves. Eur Neuropsychopharmacol 55:22–83. https://doi.org/10.1016/J.EURONEURO.2021.10.864
Public Health Agency of Canada (2022) Cycle 2: Symptoms of anxiety and depression during the pandemic. American Psychiatric Association Publishing, Ottawa
Cantor JH, McBain RK, Ho PC et al (2023) Telehealth and In-Person Mental Health Service Utilization and Spending, 2019 to 2022. JAMA Heal Forum 4:e232645–e232645. https://doi.org/10.1001/JAMAHEALTHFORUM.2023.2645
Turecki G, Brent DA, Gunnell D et al (2019) Suicide and suicide risk. Nat Rev Dis Prim 2019 51 5:1–22. https://doi.org/10.1038/s41572-019-0121-0
Johnson D, Dupuis G, Piche J et al (2018) Adult mental health outcomes of adolescent depression: A systematic review. Depress Anxiety 35:700–716. https://doi.org/10.1002/DA.22777
Merrick MT, Ford DC, Ports KA, Guinn AS (2018) Prevalence of Adverse Childhood Experiences From the 2011–2014 Behavioral Risk Factor Surveillance System in 23 States. JAMA Pediatr 172:1038–1044. https://doi.org/10.1001/JAMAPEDIATRICS.2018.2537
Calegaro G, Soares PSM, Colman I et al (2023) Adverse childhood experiences (ACEs) and suicidal behaviors in emerging adulthood: The 1993 Pelotas birth cohort. Child Abuse Negl 146:106517. https://doi.org/10.1016/J.CHIABU.2023.106517
Hammond NG, Colman I, Orr SL (2022) Adverse childhood experiences and onset of migraine in Canadian adolescents: A cohort study. Headache 62:319–328. https://doi.org/10.1111/HEAD.14256
Wade M, Wright L, Finegold KE (2022) The effects of early life adversity on children’s mental health and cognitive functioning. Transl Psychiatry 12. https://doi.org/10.1038/S41398-022-02001-0
Moore SE, Scott JG, Ferrari AJ et al (2015) Burden attributable to child maltreatment in Australia. Child Abuse Negl 48:208–220. https://doi.org/10.1016/J.CHIABU.2015.05.006
Huang Y, Li P, Lai Z et al (2018) Association between sexual minority status and suicidal behavior among Chinese adolescents: A moderated mediation model. J Affect Disord 239:85–92. https://doi.org/https://dx.doi.org/10.1016/j.jad.2018.07.004
Gloger S, Martínez P, Behn A et al (2021) Population-attributable risk of adverse childhood experiences for high suicide risk, psychiatric admissions, and recurrent depression, in depressed outpatients. Eur J Psychotraumatol 12. https://doi.org/10.1080/20008198.2021.1874600
Bellis MA, Hughes K, Ford K et al (2019) Life course health consequences and associated annual costs of adverse childhood experiences across Europe and North America: a systematic review and meta-analysis. Lancet Public Heal 4:e517–e528. https://doi.org/10.1016/S2468-2667(19)30145-8
McLaughlin KA, Conron KJ, Koenen KC, Gilman SE (2010) Childhood adversity, adult stressful life events, and risk of past-year psychiatric disorder: a test of the stress sensitization hypothesis in a population-based sample of adults. Psychol Med 40:1647–1658. https://doi.org/10.1017/S0033291709992121
Wade M, Zeanah CH, Fox NA et al (2019) Stress sensitization among severely neglected children and protection by social enrichment. Nat Commun 10. https://doi.org/10.1038/S41467-019-13622-3
Russo JE, Dhruve DM, Oliveros AD (2022) Coping with COVID-19: Testing the stress sensitization hypothesis among adults with and without a history of adverse childhood experiences. J Affect Disord Rep 10:100379. https://doi.org/10.1016/J.JADR.2022.100379
Scorza P, Corbeil T, Wall M et al (2022) Adverse Childhood Experiences and Perceived Stress in Early Adulthood in the Context of Disadvantage. Child Abuse Negl 131:105687. https://doi.org/10.1016/J.CHIABU.2022.105687
Wade M, Prime H, Johnson D et al (2021) The disparate impact of COVID-19 on the mental health of female and male caregivers. Soc Sci Med 275:113801. https://doi.org/10.1016/J.SOCSCIMED.2021.113801
Choi KW, Lee YH, Liu Z et al (2023) Social support and depression during a global crisis. Nat Ment Heal 2023 16 1:428–435. https://doi.org/10.1038/s44220-023-00078-0
Szkody E, Stearns M, Stanhope L, McKinney C (2021) Stress-Buffering Role of Social Support during COVID‐19. Fam Process 60:1002. https://doi.org/10.1111/FAMP.12618
Goerigk S, Reinhard MA, Barton BB et al (2023) Transdiagnostic analysis of adverse childhood experiences and their association with psychopathology-A TRANS-D conform study. Psychiatry Res 329. https://doi.org/10.1016/J.PSYCHRES.2023.115545
Devries KM, Mak JYT, Child JC et al (2014) Childhood sexual abuse and suicidal behavior: a meta-analysis. Pediatrics 133. https://doi.org/10.1542/PEDS.2013-2166
Jones MS, Pierce H, Shafer K (2022) Gender differences in early adverse childhood experiences and youth psychological distress. J Crim Justice 83:101925. https://doi.org/10.1016/J.JCRIMJUS.2022.101925
Leban L, Gibson CL (2020) The role of gender in the relationship between adverse childhood experiences and delinquency and substance use in adolescence. J Crim Justice 66:101637. https://doi.org/10.1016/J.JCRIMJUS.2019.101637
Whitaker RC, Dearth-Wesley T, Herman AN et al (2021) The interaction of adverse childhood experiences and gender as risk factors for depression and anxiety disorders in US adults: a cross-sectional study. BMC Public Health 21. https://doi.org/10.1186/S12889-021-12058-Z
Statistics Canada (2023) Mental Health and Access to Care Survey (MHACS). In: Stat. Canada. https://www.statcan.gc.ca/en/survey/household/5015. Accessed 6 Jul 2023
Tanaka M, Wekerle C, Leung E et al (2011) Preliminary Evaluation of the Childhood Experiences of Violence Questionnaire Short Form. J Interpers Violence 27:396–407. https://doi.org/10.1177/0886260511416462
Caron J (2013) Une validation de la forme abrégée de l’Échelle de provisions sociales : l’ÉPS-10 items. Sante Ment Que 38:297. https://doi.org/10.7202/1019198AR
Robins LN, Wing J, Wittchen HU et al (1988) The Composite International Diagnostic Interview: An Epidemiologic Instrument Suitable for Use in Conjunction With Different Diagnostic Systems and in Different Cultures. Arch Gen Psychiatry 45:1069–1077. https://doi.org/10.1001/ARCHPSYC.1988.01800360017003
Wittchen HU (1994) Reliability and validity studies of the WHO–Composite International Diagnostic Interview (CIDI): a critical review. J Psychiatr Res 28:57–84. https://doi.org/10.1016/0022-3956(94)90036-1
Kim J, Fuller W (2004) Fractional hot deck imputation. Biometrika 91:559–578. https://doi.org/10.1093/biomet/91.3.559
SAS Institute Inc (2023) SAS/STAT 15.3 User’s Guide. Cary, NC
Czeisler M, Wiley JF, Czeisler CA et al (2021) Uncovering survivorship bias in longitudinal mental health surveys during the COVID-19 pandemic. Epidemiol Psychiatr Sci 30. https://doi.org/10.1017/S204579602100038X
Racine N, Eirich R, Madigan S (2022) Fostering Resilience in Children Who Have Been Maltreated: A Review and Call for Translational Research. Can Psychol 63:203–213. https://doi.org/10.1037/CAP0000312
Arpawong TE, Mekli K, Lee J et al (2022) A longitudinal study shows stress proliferation effects from early childhood adversity and recent stress on risk for depressive symptoms among older adults. Aging Ment Health 26:870. https://doi.org/10.1080/13607863.2021.1904379
Milner A, Krnjacki L, LaMontagne AD (2016) Age and gender differences in the influence of social support on mental health: a longitudinal fixed-effects analysis using 13 annual waves of the HILDA cohort. Public Health 140:172–178. https://doi.org/10.1016/J.PUHE.2016.06.029
Colman I, Kingsbury M, Garad Y et al (2016) Consistency in adult reporting of adverse childhood experiences. Psychol Med 46:543–549. https://doi.org/10.1017/S0033291715002032

No competing interests reported.

Supplement.pdf

Download PDF

Reviewers invited by journal
04 Jul, 2024
Editor assigned by journal
28 Jun, 2024
Submission checks completed at journal
20 Jun, 2024
First submitted to journal
20 Jun, 2024

You are reading this latest preprint version

Interactions Between Early-Life Adversity, Pandemic Stress, and Social Support on Psychiatric Disorders in a Nationally Representative Sample of Canadian Adults

Status:

Version 1

Abstract

PURPOSE

METHODS

RESULTS

CONCLUSION

Introduction

Early-Life Adversity and Pandemic Stress

Social Support and Pandemic Stress

The Present Study

Materials and Methods

Study Population

Measures

Statistical Analyses

Results

Descriptive Statistics

Early-Life Adversity and Pandemic Stress

Social Support and Pandemic Stress

Sex Effects

Sensitivity Analysis

Discussion

Strengths and Limitations

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1