The association of social, economic, and health-related variables with suicidal and/or self-harm thoughts in individuals admitted to a psychiatric inpatient unit during the COVID-19 pandemic: An exploratory study

Abstract

Individuals with pre-existing psychiatric diagnoses appear to be vulnerable to worsening mental health symptoms during the coronavirus disease 2019 (COVID-19) pandemic. Furthermore, psychiatric hospitalizations during the pandemic may be complicated by increased risk of SARS-Cov-2 infection and limited social engagement due to changes in hospital policies. The objective of our exploratory study was to determine whether social, economic, and health-related variables were associated with thoughts of suicide and/or self-harm since March 2020 in individuals admitted to a psychiatric inpatient unit during the COVID-19 pandemic. Chi-square tests revealed four variables were significantly associated with thoughts of suicide and/or self-harm: 1) difficulty with cancellation of important events, 2) some form of loneliness, 3) decreased time spent in green spaces, and 4) increased time spent using devices with screens. The logistic regression model showed a significant association between suicidal and/or self-harm thoughts and cancellation of important events. Further investigation of the loneliness variable components revealed a significant association between suicidal and/or self-harm thoughts and feeling a lack of companionship, feeling isolated, and feeling alone. These results suggest that social challenges experienced during the pandemic were associated with negative mental health symptoms of individuals admitted to a psychiatric inpatient unit.

1. Introduction

The coronavirus disease 2019 (COVID-19) pandemic has led to mental health impacts of social isolation, lockdowns, fear of infection, and SARS-Cov-2 infection/COVID-19 disease itself. Individuals with pre-existing mental illnesses were more likely to experience a decline in their mental health during the pandemic (O'Connor et al., 2021; Pierce et al., 2021). Socio-demographic, economic, and health determinants such as income, housing, social isolation, substance use, and sexual orientation are independent risk factors for suicidal and/or self-harm thoughts (Mulholland et al., 2021; Näher et al., 2020). Before the pandemic, previous studies found that living alone and feelings of loneliness were associated with suicidal outcomes including suicidal ideation in the general adult population (Calati et al., 2019). Other factors, such as engaging in physical activity, meditation, and sleep duration and quality, may be protective against experiencing suicidal ideation (Littlewood et al., 2019; Vancampfort et al., 2018; Wu et al., 2021). After the onset of the pandemic, variables such as COVID-19 stress, loneliness, financial strain, alcohol misuse, and testing positive for COVID-19 were associated with suicidal and/or self-harm thoughts in the general adult population (Elbogen et al., 2021). Moreover, a negative perception of social distancing was bidirectionally associated with suicidal ideation (Ammerman et al., 2021). In Ontario, suicide rates increased at the beginning of the pandemic (Pirkis et al., 2022). Given this finding, more studies are needed to understand what has driven thoughts of suicide and/or self-harm during the pandemic.

Despite the surge in research related to the COVID-19 pandemic, relatively few studies have examined the effects of the pandemic on psychiatric inpatient populations. At the beginning of the pandemic in March 2020, psychiatric hospitalizations decreased in Canada, Germany, and France (Jollant et al., 2021; Kim et al., 2021; Zielasek et al., 2021). Examining the effects of the COVID-19 pandemic on inpatient populations is important because these patients are potentially the most severely ill, have an increased SARS-Cov-2 exposure risk, and face challenges related to being hospitalized within an acute care setting during a pandemic (Vigo et al., 2020; Xiang et al., 2020; Zhu et al., 2020). At a time when clinical research may have been limited by COVID-19 restrictions, the current study provides insight into a particularly vulnerable group of psychiatric inpatients at risk of suicide and/or self-harm. In this exploratory study, we investigate whether social, economic, and health-related variables were associated with thoughts of suicide and/or self-harm in individuals admitted to a psychiatric inpatient unit during the COVID-19 pandemic.

2. Methods

Participants were recruited from the Sunnybrook Health Sciences Centre Inpatient Psychiatry Unit in Toronto, Canada from August 2020 to May 2021. The average age of participants was 38.4 years (range 19–77 years, standard deviation 13.5), with 50.8% female, 44.3% male and 4.9% other. Race/ethnicity of participants included 59% White, 13.1% Asian, 4.9% Black, 4.9% Indian-Caribbean, 4.9% Middle Eastern and 13% other. Primary psychiatric diagnoses leading to hospitalization included mood disorders (58.9%), substance-use disorders (17.6%), personality disorders (9.8%), psychotic disorders (9.7%), anxiety disorders (1.6%), and stressor-related disorders (1.6%). The majority of participants were single (62.3%), with 18% divorced or separated, 17.9% married or living common law, and 1.6% widowed. The majority of participants identified as heterosexual (73.8%), with 9.8% identifying as bisexual, 4.9% gay, 4.9% lesbian, 1.6% queer, and 4.9% chose not to disclose.

Participants were recruited as a part of a Caring Contacts randomized controlled trial (RCT). Once recruited for the RCT, they were given the option to complete an additional COVID-19-related questionnaire adapted from standardized questions used by the International COVID-19 Suicide Prevention Research Collaboration (ICSPRC) (https://www.iasp.info/research-collaboration-icsprc/). Participants completed this questionnaire prior to receiving any of the RCT treatments. The questionnaire had 60 questions and 14 sections (Box 1).

• •Physical/Social Space
• •Violence
• •Family Relationships
• •Cancellation of Important Events
• •Children in School
• •Isolation
• •Change in Subjective Financial Status
• •Concerned About Future Finances
• •Employment Status Changes
• •Adverse Financial Outcomes During the Pandemic
• •COVID-19 Infection Related Questions, Changes in Health-Related Behaviours
• •Subjective Mental Health
• •Experience of Care by People in a Suicidal Crisis

The pre-specified aim of this study was to identify associations between thoughts of suicide and/or self-harm in psychiatric inpatients and other variables on the COVID-19-related questionnaire. Given the exploratory nature of this study, we did not generate a priori hypotheses. After data collection, from the COVID-19 related self-reported questionnaire we identified descriptive variables of interest that included employment status, SARS-CoV-2 infection, high risk category for COVID-19 illness severity, COVID-19 information seeking, self-rating of mental health before lockdown, how participants sought help, and rating of mental health care interaction. We then identified a dependent variable that was suicidal and/or self-harm thoughts, which was derived from the question since the official lockdown was announced on March 14 2020, have you had thoughts that you would be better off dead or of hurting yourself in some way? The questions on the survey were complex and many had multiple choice answers. Given the modest sample size, both in terms of participants and responses on the questionnaire, and the number of items on the questionnaire, we combined some related questions into a single variable for analysis and all variables were dichotomized. For example, in the questionnaire, there were separate questions about whether participants had experienced psychological harm and physical harm since the start of lockdown. For the purposes of our analyses, we combined the two questions to be a single variable of experience of physical or psychological harm since lockdown. There was a total of 23 dichotomized variables for analysis (Table 1 ). We chose the direction of the dichotomization based on the available literature but we did not have sufficient power to test this as a hypothesis. Other examples include the some form of loneliness since lockdown variable that was combined from multiple source questions and dichotomized. The four component questions related to loneliness were do you (1) feel you lack companionship (2) feel left out (3) feel isolated and (4) feel alone? We combined these questions into a single variable of some form of loneliness since lockdown. Similarly, the question pertaining to difficulty with cancellation of important events included life events such as weddings, funerals, graduations, and proms. A chi-square test of independence was conducted for each of the 23 variables. We conducted a false discovery rate (FDR; Benjamini and Hochberg, 1995) correction on these 23 variables with the p-values that were generated from the chi-squares. If any variables representing a combination of questions were found to be significant, further post-hoc chi-square tests of independence were conducted on the individual source questions to determine which drove significance. Variables with a p-value less than or equal to 0.10 were included in a backwards stepwise logistic regression that controlled for age and gender. We also compared the demographic data for participants in the RCT who completed the questionnaire compared to participants who chose not to complete the questionnaire using a t-test and chi-squares.

Table 1.

Variables for analysis.

Live alone	Y/N
Living arrangement change due to COVID-19 pandemic	Y/N
Type of place you live in	House/Other
Home with a safe outdoor space	Y/N
Experience of physical or psychological harm since lockdown	Y/N
Worsening in quality of family relationships since lockdown	Y/N
Difficulty with cancellation of important events in your life	Y/N
Some form of loneliness since lockdown	Y/N
Difficulty with cancellation of holidays	Y/N
Financial deterioration since lockdown	Y/N
Worried about future finances	Y/N
Major life stressor since lockdown	Y/N
COVID-19 illness stressor for self, family or friends	Y/N
Essential worker	Y/N
Change in the amount that you sleep	Y/N
Decreased physical activity	Y/N
Decreased time spent in green spaces	Y/N
Increase in amount you smoke or consume alcohol	Y/N
Decreased time spent practicing relaxation	Y/N
Increased time spent listening to the news	Y/N
Increased time spent using devices with a screen	Y/N
Decreased time spent doing hobbies	Y/N
Difficulty accessing mental health care	Y/N

3. Results

Out of 103 eligible inpatient participants who were approached to complete the COVID-19 related questionnaire, a total of 61 participants agreed to do so (59.2%). Forty-nine of the 61 participants (80.3%) indicated that since the lockdown was announced in March 2020, they have had thoughts that they would be better off dead or thoughts of hurting themselves in some way (suicidal and/or self-harm thoughts).

3.1. Variables associated with suicidal and/or self-harm thoughts

The chi-square tests of independence that tested for an association with suicidal and/or self-harm thoughts found four out of 23 variables to have a p-value less than or equal to 0.10, with two of the variables having a p-value less than 0.05. When we ran the FDR correction, the significance level became p = 0.003. The four variables were: (1) difficulty with cancellation of important events in your life (χ²(1, N = 61) = 4.11, p = 0.042, FDR corrected p = 0.007); (2) some form of loneliness since lockdown (χ² (1, N = 61) = 7.03, p = 0.023, FDR corrected p = 0.003); (3) decreased time spent in green spaces such as parks or gardens (χ² (1, N = 60) = 3.33, p = 0.068, FDR corrected p = 0.010); and (4) increased time spent using devices with a screen (χ² (1, N = 61) = 3.02, p = 0.086, FDR corrected p = 0.0133). With the FDR correction one variable remained significant. Since this was an exploratory study, we chose to include all four variables in the logistic regression analysis.

Fewer than half of participants (42.6%) reported having difficulty with cancellation of important events, while the majority of participants (88.5%) reported experiencing some form of loneliness since lockdown. Despite 68.3% of participants reporting having a safe outdoor space (e.g., a garden or a yard) to exercise or play, most participants (56.7%) indicated that their time spent in green spaces such as parks or gardens decreased, and 70.5% of participants indicated that their time spent using devices with a screen increased.

3.2. Logistic regression model

The logistic regression model controlling for age and gender was statistically significant, (χ²(5) = 18.53, p = 0.002) (Table 2 ). The model explained 43.2% (Nagelkerke R ²) of the variance in suicidal and/or self-harm thoughts and correctly classified 86% of cases. Individuals who found the cancellation of important events in their life difficult were 13.2 times more likely to experience suicidal and/or self-harm thoughts.

Table 2.

Variables in the Equation

		B	S.E.	Wald	df	Sig.	Exp(B)	95% C.I.for EXP(B)
								Lower	Upper
Step 1a	Age	0.044	0.037	1.478	1	0.224	1.045	0.973	1.123
	Gender(1)	1.316	0.884	2.214	1	0.137	3.727	0.659	21.088
	Difficulty with cancellation of important events in your life(1)	2.504	1.124	4.960	1	0.026	12.236	1.350	110.860
	Some form of loneliness since lockdown(1)	2.034	1.249	2.653	1	0.103	7.641	0.661	88.297
	Decreased time spent in green spaces such as parks or gardens(1)	1.550	0.915	2.867	1	0.090	4.712	0.783	28.340
	Increased time spent using devices with a screen(1)	0.928	0.882	1.108	1	0.292	2.530	0.449	14.239
	Constant	-4.818	2.133	5.102	1	0.024	0.008
Step 2a	Age	0.033	0.033	0.988	1	0.320	1.033	0.969	1.101
	Gender(1)	1.332	0.863	2.384	1	0.123	3.789	0.699	20.550
	Difficulty with cancellation of important events in your life(1)	2.579	1.101	5.490	1	0.019	13.186	1.525	114.037
	Some form of loneliness since lockdown(1)	2.216	1.213	3.340	1	0.068	9.174	0.852	98.824
	Decreased time spent in green spaces such as parks or gardens(1)	1.502	0.898	2.800	1	0.094	4.491	0.773	26.092
	Constant	-3.990	1.904	4.391	1	0.036	0.018

^aVariable(s) entered on step 1: Difficulty with cancellation of important events in your life, Some form of loneliness since lockdown, Decreased time spent in green spaces such as parks or gardens, Increased time spent using devices with a screen.

3.3. Post-hoc tests on loneliness variable

Seventy-two percent of participants reported feeling a lack of companionship, 80.3% reported feeling isolated, 82% reported feeling alone, and 60.7% reported feeling left out. Post-hoc chi-square tests of independence conducted on the four component questions of the some form of loneliness variable showed a significant association between suicidal and/or self-harm thoughts and feeling a lack of companionship (χ²(1, N = 61) = 6.90, p = 0.026), feeling isolated (χ²(1, N = 61) = 4.57, p = 0.047), and feeling alone (χ²(1, N = 61) = 10.33, p = 0.005). The was no significant association between thoughts of suicide or self-harm and feeling left out (χ²(1, N = 61) = 0.71, p = 0.513).

3.4. Population characteristics

Before the lockdown in March 2020, 47.5% of participants were employed or self-employed, 19.7% were in educational settings, 9.8% unemployed, 16.4% permanently sick or disabled, 4.9% in an unpaid work/apprenticeship, and 1.6% retired. At the time they completed the survey, 32.8% were unemployed/unable to work/on unpaid leave, 14.8% employed and currently working, 13.2% employed but working reduced hours or on unpaid leave, 16.4% permanently sick or disabled, 13.1% in education, 4.9% retired, and 4.9% in unpaid work/apprenticeship. Only one participant was known to have been previously infected with SARS-CoV-2 and 13.1% of participants were deemed a high risk category for COVID-19 illness severity (e.g., aged 70+, underlying health condition). Most participants sought out COVID-19 related information less than once a day (52.5%), 42.6% sought out COVID-19 information 1–5 times a day, and 4.9% sought out COVID-19 information 5 times a day. Participants rated their mental health before March 2020 as poor (41%), fair (32.8%), and good (26.2%). Of the 49 participants who indicated they had suicidal and/or self-harm thoughts since the start of lockdown 65.5% of people went to a health care professional, 45.9% of people went to family or friends, 27.8% used a crisis hotline, and 14.8% used an online service. Participants rated their mental health care interaction as good (67.2%), fair (21.3%), and poor (1.6%).

In our comparison of the demographic data between participants in the RCT who completed the questionnaire compared to participants who chose not to complete the questionnaire we found that there were no significant differences in age (p = 0.713), gender (p = 0.417), and psychiatric disorder (p = 0.289). There was a significant difference in race (p = 0.042), marital status (p = 0.035), and sexual orientation (p = 0.033). Our sample had a large portion of white, single and divorced/separated/widowed, and gay/lesbian/queer participants.

4. Discussion

Our study adds to the current literature by exploring the relationship between social, economic, and health-related variables and suicidal and/or self-harm thoughts in individuals admitted to a psychiatric inpatient unit during the COVID-19 pandemic. We found that difficulty with cancellation of important events was associated with thoughts of suicide and/or self-harm. There was also an indication that loneliness, and decreased time spent in green spaces may have been associated with suicidal and/or self-harm thoughts, but to a lesser degree.

Although less than half of participants experienced difficulty with cancellation of important events, it was the variable with the strongest association with suicidal and/or self-harm thoughts. Weddings, graduations and funerals are all rituals that mark a significant occasion in life and, in the context of the COVID-19 pandemic, all had to be adapted (Imber-Black, 2020). Individuals dealing with the loss of a loved one are negatively affected by the absence of funeral rituals during the COVID-19 pandemic and as a result of an impacted grieving process may even be at risk of developing complicated grief (Cardoso et al., 2020; Mortazavi et al., 2020).

When the components of the loneliness variable were tested individually, feeling a lack of companionship, feeling isolated, and feeling alone were all associated with suicidal and/or self-harm thoughts. This is in line with previous studies that have shown that feelings of loneliness have increased during the COVID-19 pandemic and have been associated with suicidal thoughts (Allan et al., 2021; Elbogen et al., 2021; Killgore et al., 2020a, 2020b). Our results provide further support for this finding and suggest that an association between loneliness and suicidal thoughts is also present and may be amplified in psychiatric inpatient populations.

There was a weaker association between suicidal and/or self-harm thoughts and decreased time spent in green spaces. Prior studies have identified a relationship between mental health symptoms and access to green spaces during the COVID-19 pandemic (Pouso et al., 2020; Soga et al., 2021), with one study finding that, depending on the strictness of the lockdown, individuals with a nature view and those with access to private outdoor space had fewer symptoms of depression and anxiety (Pouso et al., 2020). Given this finding, we may have expected our variable home with a safe outdoor space to have an inverse association with suicidal and/or self-harm thoughts however it did not. This could be because having a home with a safe outdoor space does not necessarily translate to spending time in that outdoor space. Soga et al. (2021) found that people who use green space frequently during the COVID-19 pandemic reported lower levels of depression and anxiety, and the frequency of green space use was positively related to self-esteem, life satisfaction, and happiness. Our data showed that although 68.3% of participants had a safe outdoor space, more than half of participants indicated that their time spent in green spaces decreased. This supports the notion that time spent outdoors during the COVID-19 pandemic can be associated with mental health symptoms, and adds that a decrease in time spent outdoors may be associated with suicidal and/or self-harm thoughts in individuals who have experienced psychiatric hospitalization during the pandemic.

The finding of 70.5% of our sample reporting increased screen time during the COVID-19 pandemic is in line with previous epidemiological studies (Colley et al., 2020; Wagner et al., 2021). With physical distancing measures in place, the increase in screen time is likely a combination of many aspects of day-to-day life becoming remote including work, virtual health care appointments, and an increase in recreational device use including for the purposes of socializing. While there are many benefits for the use of devices during the COVID-19 pandemic (Riva et al., 2020), studies have shown that greater time spent using devices has been linked to poorer mental health during the COVID-19 pandemic (Colley et al., 2020; Meyer et al., 2020; Smith et al., 2020). Our findings provide further support for the association between increased screen time and worsening mental health in our clinical sample.

Our study has some notable limitations. First, all participants were hospitalized for an acute illness (including a high proportion with a mood disorder). While we believe this is a strength given the paucity of clinical research on severely ill clinical samples during the pandemic, it must also be acknowledged that we do not know the degree to which these findings generalize to others outside of our clinical sample, including clinical populations with different distributions of mental illness. Second, our study has a small sample size, again limiting the available power to identify findings that otherwise may have been significant. Third, participants responded to the study questionnaire during their stay in the hospital when they were experiencing an acute episode of illness which may have impacted the accuracy of recall. Additionally, the dependent variable captured thoughts of suicide and self-harm in a single measure, not separately, so we were not able to differentiate thoughts of suicide from thoughts of self-harm. Furthermore, we did not have a measure for symptom severity or coping ability which may have provided more insight into the participants’ experience and ability. Finally, our finding of a significant difference in race, marital status, and sexual orientation between the participants in the RCT who completed this questionnaire compared to the non-completers, shows that we were not able to capture a fully representative sample.

In sum, our results suggest a few social variables including difficulty with cancellation of important events, feeling a lack of companionship, feeling isolated, and feeling alone, were all associated with suicidal and/or self-harm thoughts in individuals admitted to a psychiatric inpatient unit during the COVID-19 pandemic. These results suggest that the social challenges individuals experienced during the COVID-19 pandemic were substantial and were associated with negative mental health symptoms for individuals admitted to a psychiatric inpatient unit during the pandemic. The variables found to be associated with suicidal and/or self-harm thoughts may be important targets of mitigation strategies in the later stages of the pandemic and in future pandemics.

Funding

This work was supported by the COVID-19 specific Innovation Fund of the Alternative Funding Plan from the Academic Health Sciences Centres of Ontario. The sponsor had no involvement in study design, collection, analysis and interpretation, writing of the report, nor in the decision to submit the article for publication.

CRediT authorship contribution statement

Sarah Holman: Project administration, Investigation, Formal analysis, Visualization, Writing – original draft, Writing – review & editing. Rosalie Steinberg: Funding acquisition, Conceptualization, Methodology, Supervision, Writing – review & editing. Ayal Schaffer: Funding acquisition, Conceptualization, Methodology, Supervision, Writing – review & editing. Lisa Fiksenbaum: Conceptualization, Formal analysis, Writing – review & editing. Mark Sinyor: Funding acquisition, Conceptualization, Methodology, Supervision, Writing – review & editing.

Declaration of Competing Interest

Declaration of interest: The authors declare no conflicts of interest. Dr. Steinberg and Dr. Schaffer receive salary support through academic scholars’ awards from the Department of Psychiatry at Sunnybrook Health Sciences Centre. Dr. Sinyor receives salary support through academic scholars’ awards from the Departments of Psychiatry at Sunnybrook Health Sciences Centre and the University of Toronto.