Insomnia symptoms among older adults during the first year of the COVID-19 pandemic: A longitudinal study

Abstract

Objective

To identify sociodemographic, psychological, and health factors related to trajectories of insomnia symptoms in older adults during the COVID-19 pandemic.

Methods

From May 2020 to May 2021, 644 older adults (mean age = 78.73, SD = 5.60) completed telephone-administered self-reported measures (ie, Insomnia Severity Index, consensus sleep diaries, UCLA Loneliness Scale, Kessler Psychological Distress Scale, Post-Traumatic Checklist, perceived health threat, and International Physical Activity Questionnaire) and provided sociodemographic data at 4 timepoints. Using the Insomnia Severity Index score at each timepoint, group-based trajectory modeling was conducted to identify groups with distinct insomnia trajectories.

Results

On average, there was no significant change in insomnia symptoms over time. Three groups with distinct sleep trajectories were identified: clinical (11.8%), subthreshold (25.3%), and good sleepers (62.9%). Older adults who were younger, male, had elevated psychological distress and posttraumatic stress disorder symptoms, perceived more SARS-CoV-2 health threat, spent more time in bed, and had shorter sleep duration during the first wave of the pandemic were more likely to belong to the clinical than to the good sleepers group. Those who were younger, female, had elevated psychological distress and PTSD symptoms, greater loneliness, spent more time in bed, and had reduced sleep duration during the first wave were more likely to belong to the subthreshold than to the good sleepers group.

Conclusions

Over 1 in 3 older adults experienced persistent subthreshold or clinically significant insomnia symptoms. Both sleep-related behaviors as well as general and COVID-19-related psychological factors were associated with insomnia trajectories.

Introduction

At the beginning of the coronavirus 19 (COVID-19) pandemic in March 2020, strict lockdown and physical distancing directives were imposed around the globe to mitigate the spread of the virus, protect vulnerable populations, and reduce pressure on healthcare systems. While infection from the SARS-CoV-2 virus was widespread across all demographics, older adults were at greater risk for COVID-19-related morbidity and mortality. Although stringent confinement measures can help reduce the spread of the virus within the population, they can also impact the psychological well-being and sleep of the general population, especially older adults.¹

Meta-analyses of cross-sectional studies during the first wave of the COVID-19 pandemic suggest that 18.0%-36.7% of the general population reported significant sleep disturbances.2, 3 Using retrospective self-report data, older adults reported greater insomnia severity, poorer sleep quality, longer sleep onset latency, increased sleep duration, greater time in bed, lower sleep efficiency, more daytime impairment, and more frequent use of sleep medications during the pandemic compared to before the pandemic.4, 5 However, longitudinal studies have found mixed results when comparing sleep quality and insomnia severity between the prepandemic period and the first wave of the pandemic. A longitudinal study using prepandemic data from a population-based cohort of 594 middle-aged adults found that the prevalence of insomnia symptoms increased from 25.4% to 32.2% and the prevalence of insomnia disorder increased from 16.8% to 19% during the first wave of the pandemic.⁶ Another longitudinal study indicated that 10.2% of healthcare providers with insomnia prior to the COVID-19 pandemic no longer had the disorder during the first wave of the pandemic, whereas 43.4% without insomnia prior to the pandemic subsequently developed insomnia.⁷ However, these results conflict with other longitudinal findings reflecting no change in sleep quality or prevalence of insomnia in older adults during the first wave of the pandemic compared to the prepandemic period.8, 9, 10 Thus, the change in sleep quality and insomnia in older adults during the first wave of the COVID-19 pandemic remains unclear.

Although the most stringent public health mitigation measures were generally put in place during the first wave of the pandemic, changing yet protracted sanitary measures were maintained in most countries during subsequent waves of the pandemic. Yet, far less information is available on insomnia symptoms during the later waves of the pandemic. Two Italian longitudinal studies of the general population examined sleep changes over the first and second waves of the pandemic, up until autumn 2020. One study found later bed and rise times, and longer time in bed and sleep latency among adults between the ages of 18 and 79 years in the first wave compared to the second wave, with poor sleep quality in both waves compared to retrospective data reflecting the period prior to each subsequent lockdown.¹¹ The other study found similar results within the same age group, including earlier bed and rise times, decreased sleep latency, improved subjective sleep quality, and decreased insomnia severity during the second wave of the pandemic compared to the first wave.¹² In addition, 1 German longitudinal study examined insomnia severity in 267 adults with a mean age of 31 over the course of 6 months, beginning in summer 2020 and ending in early winter 2021. These findings suggested an initial prevalence of clinically significant insomnia of 10.1% and subthreshold insomnia of 27.0%, with no statistically significant change in insomnia over the 6-month period.¹³ How insomnia symptoms changed over the first year of the pandemic remains unclear, with no research specifically examining older adults.

During the first wave of the COVID-19 pandemic, cross-sectional research has identified certain sociodemographic factors, such as being female, lower education level, and greater financial strain, to be related to greater insomnia severity and poorer subjective sleep quality among older adults.14, 15, 16 Yet, evidence for the relationship between age itself and these adverse sleep outcomes in the context of the pandemic is still mixed.14, 16 Beyond age-related biological changes that can lead to earlier shifted sleep-wake cycles, more sleep fragmentation, and increased napping behavior among older adults,¹⁷ the COVID-19 pandemic has introduced additional factors that may negatively impact sleep in these individuals.¹⁴ In the early phase of the pandemic, older adults also reported disruptions in their social rhythm, described as a loss of regularity in the timing of daily behaviors, such as bed or wake-up times, socialization, eating, physical activity, and entertainment.¹⁸ Many older adults engaged in less physical activity during the pandemic compared to the prepandemic period, which may negatively impact their sleep quality.¹⁹ Furthermore, the COVID-19 confinement measures may have created a context of increased loneliness. Accordingly, older adults who reported greater feelings of loneliness during the first wave of the pandemic also experienced worse sleep quality and increased insomnia symptoms compared to the prepandemic period.15, 18 In addition, given that older adults are at greater risk for severe COVID-19 outcomes,¹ the pandemic was found to be associated with worse anxiety, depression, and posttraumatic stress symptoms, which may also be associated with a higher risk for insomnia symptoms.20, 21 Thus, the current research includes these potential predictors of insomnia throughout the later stages of the pandemic to shed light on the effect of these factors over time.

Objectives and hypotheses

The present study examined changes in insomnia symptoms during the first year of the COVID-19 pandemic among older adults. In addition, sociodemographic, psychological, and health factors were examined to identify potential risk or protective factors for insomnia symptoms during the first year of the pandemic.

Participants and methods

To be eligible for this study, participants had to be at least 60 years old and had to speak French or English. They also needed to be able to complete telephone-based semistructured interviews. They were excluded if they had hearing impairment making them unable to provide oral informed consent, or severe cognitive impairment, as assessed by a cutoff score of 14 points on the telephone-based Mini-Mental State Examination.²² While 1377 older adults were contacted to participate in this study, 644 participants were enrolled in the study (participation rate: 46.8%). Participants were recruited from ongoing cohort studies and participant pools from the Centre de Recherche de l’Institut de Gériatrie de Montréal (CRIUGM) or l’Étude sur la Santé des Aînés (ESA)-Service studies. At the second timepoint, 63 additional participants were recruited through newspaper and radio advertisements.

Study design and procedure

A 4-wave longitudinal study was carried out from May 6, 2020 to May 24, 2021, which covered 3 waves of the COVID-19 pandemic. Data were collected at 4 timepoints: Spring 2020 (Time 1), Summer 2020 (Time 2), Fall 2020 (Time 3), and Spring 2021 (Time 4). Participants were undergoing stricter confinement and physical distancing measures during the Time 1 and Time 4 assessments. Confinement measures were relaxed during the Time 2 assessment and gradually increased during the Time 3 assessment. The local infection rate increased over time in Quebec (roughly 50 new cases/day at T2 to around 320 new cases/day at T4). The study received ethical approval from the Centre intégré universitaire de santé et de services sociaux (CIUSSS) du Centre-Sud-de-l’Île-de-Montréal Ethics Review Board. All participants provided verbal informed consent prior to the start of the study.

Measures

Sleep variables

The Insomnia Severity Index (ISI) measures the severity of insomnia symptoms within the last 2 weeks. It is composed of 7 items on a 5-point Likert scale (0 = no problem to 4 = very severe problem). A higher overall score signifies greater insomnia symptoms.²³ The internal consistency of the ISI scores within the sample was .86.

Retrospective sleep diaries adapted from the consensus sleep diary²⁴ assessed the average duration of nap time, total sleep time, and total time in bed in the last 2 weeks. To examine the stability of social rhythms, participants provided information about the earliest and latest wake-up times in the past 2 weeks. The difference between the earliest and latest wake-up times was computed to measure the stability of social rhythm, with greater scores indicating a more unstable rhythm.²⁵

Sociodemographic variables

Sociodemographic characteristics consisted of age, gender, education level, and income level. Participants with an annual income below $25 000 were considered to live in poverty.²⁶ Participants also indicated if they lived alone or with others.

Psychological and health variables

A 3-item UCLA Loneliness Scale evaluated perceived loneliness within the last 2 weeks. The items are rated on a 3-point Likert scale (1 = hardly ever to 3 = often). A higher total score reflects greater loneliness.²⁷ The internal consistency of the UCLA Loneliness Scale scores within the sample was .77.

Kessler Psychological Distress Scale 6-item version (K6) assessed psychological distress in the past 2 weeks. Each item is rated on a 5-point Likert scale (1 = none of the time to 5 = all of the time). A higher total score indicates more psychological distress.²⁸ The internal consistency of the K6 scores within the sample was .84.

The Post-Traumatic Checklist – Civilian Version (PCL-C) measured PTSD symptoms associated with the COVID-19 pandemic in the last 2 weeks. This scale consists of 17 items on a 5-point Likert scale (0 = not at all to 4 = very often). A higher total score reflects greater severity of PTSD symptoms, with a cutoff score of 30 or above suggesting the presence of probable PTSD.²⁹ The internal consistency of the PCL-C scores within the sample was .83.

The perceived health threat of COVID-19 was assessed based on the health belief model using 4 items evaluating the extent to which they believed that themselves or their close others were susceptible to getting infected by the virus and how dangerous the virus could be to them or their closer others on a 5-point Likert scale (0 = not at all to 4 = extremely). A higher total score indicated a greater perceived threat. The internal consistency of the score was .70.

International Physical Activity Questionnaire (IPAQ) – short form assessed physical activity in the past week. Walking, moderate-intensity, and vigorous activity times were evaluated. The metabolic equivalent of task for the combination of all of these activities was calculated.³⁰

Statistical analysis

Hierarchical linear modeling (HLM) evaluated whether the severity of insomnia symptoms changed, on average, across the 4 timepoints within our sample. Group-based trajectory modeling (GBTM) was then conducted to identify groups of older adults who exhibited distinct patterns of changes in insomnia symptoms throughout the pandemic. Unlike growth-curve modeling techniques that assume that each individual’s trajectory evolves around 1 population mean, GBTM assumes that the population is made up of multiple groups of individuals who commonly evolve together over time. The purpose of GBTM is to decrease the within-group variability in trajectories while also increasing the between-group variability such that the trajectory for each group will be distinct.³¹ To select the number of groups, we examined the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), cross-validation error (CVE), and the overall group size (at least 50 individuals per group). The model with lower BIC and CVE values, reflecting better model fit, was selected. Maximum likelihood was used to estimate the probability that each individual belonged to each subgroup and to allocate each individual to the most probable subgroup. The final model included 3 subgroups with a quadratic trajectory. All participants with at least 1 timepoint were included in the modeling of the group-based trajectories. GBTM was conducted using the software package, “crimCV,” in R. Subsequently, multinomial logistic regression was performed to evaluate which sociodemographic, psychological, and behavioral factors at T1 independently predicted group membership using the “nnet” package in R. All variables were entered simultaneously in the model in order to examine their independent effects adjusting for all other variables. The odds ratio (OR) of belonging to a particular trajectory group relative to a reference group, the good sleepers group, was estimated for each predictor. An alpha was set to .05 for these analyses.

Results

Overall, the longitudinal study consisted of 644 older adults with a mean age of 78.73 (SD = 5.60 and age range: 64.20-97.05). In total, 537 participants (83.39%) completed T1, 454 participants (70.50%) completed T2, 429 participants (66.61%) completed T3, and 395 participants (61.34%) completed T4. In addition, 259 (40.22%) participants had data for all 4 timepoints. Sociodemographic, psychological, and health characteristics of the sample at the start of the pandemic are presented in Table 1. Out of all participants who shared their COVID-19 infection status at T1, none of them obtained a positive COVID-19 test result. However, 0.65%, 0.24%, and 6.27% of participants self-reported that they tested positive for COVID-19 at T2, T3, and T4, respectively.

Table 1.

Sociodemographic, psychological, and health characteristics of the sample at T1

Characteristic		M (SD)/%
Mean age (years)		78.73 (5.60)
	<65	0.33
	65-74	28.78
	75-84	57.07
	85>	13.82
Gender	Male	26.93
	Female	73.07
Highest education level	None/Primary school	7.51
	Secondary school	24.21
	College	22.54
	University	45.74
Annual income	<$25,000	25.26
	$25,000+	74.74
Living situation	Lives alone	54.25
	Lives with others	45.75
Total UCLA Loneliness Scale score		4.60 (1.73)
	Not lonely (3-5)	71.40
	Lonely (6-9)	28.60
Total Kessler Psychological Distress Scale score		9.39 (3.25)
	Likely to be well (10-19)	79.07
	Likely to have a mild disorder (20-24)	14.53
	Likely to have a moderate disorder (25-29)	3.98
	Likely to have a severe disorder (30-50)	2.42
Total Post-Traumatic Checklist score		7.79 (7.35)
	No PTSD (0-29)	98.53
	PTSD (30-68)	1.47
Perceived health threat of COVID-19		6.83 (2.90)
Total International Physical Activity Questionnaire score (MET-minutes/week)		2089.04 (1936.00)
	Low ($\le$ 599)	36.59
	Moderate (600-2999)	42.79
	High ($\ge$3000)	20.62
Total Insomnia Severity Index score		4.17 (4.57)
	No clinically significant insomnia (0-7)	79.52
	Subthreshold insomnia (8-14)	16.76
	Moderate-severe insomnia (15-28)	3.73
Nap time (minutes/day)		23.13 (37.78)
	<15	55.35
	15-29.99	12.95
	30-44.99	11.07
	45-59.99	3.00
	$\ge$ 60	17.64
Sleep duration (hours/night)		7.42 (1.41)
	<5	3.55
	5-5.99	8.22
	6-6.99	13.08
	$\ge$ 7	75.14
Time in bed (hours/night)		8.34 (1.36)
	<5	0.19
	5-6.99	12.38
	7-8.99	49.34
	$\ge$ 9	38.09
Social rhythm (minutes)a		114.39 (80.87)
	<30	8.08
	30-59.99	10.90
	60-89.99	17.11
	90-119.99	15.98
	$\ge$ 120	47.93

Note. % represents the valid percentage.

^aSocial rhythm scores were computed as the difference between the earliest and latest wake-up times within the past 2 weeks that the questionnaire was administered.

The HLM model indicated that there was no statistically significant change in the severity of insomnia symptoms across time, β ₁ = 0.10, SD = 0.07, t = 1.42, and P = .16. Furthermore, GBTM identified 3 groups with distinct sleep trajectories: clinical (G1), subthreshold (G2), and good sleepers (G3) groups. As shown in Fig. 1, the clinical group (11.78% of the sample) had stably higher total ISI scores than the other 2 groups between T1 and T4. For the subthreshold group (25.27% of the sample), the total ISI scores were fairly stable and resided within the subthreshold range of clinical insomnia between T1 and T3 but exhibited a slight decrease at T4. Finally, the good sleepers group (62.95% of the sample) had stably low ISI scores in total across time.

Fig. 1.

Insomnia severity group trajectories throughout the COVID-19 pandemic. Notes: N = 644

Sociodemographic, psychological, and health factors associated with group membership were identified ( Table 2). Older adults who were younger, male, had elevated psychological distress and PTSD symptoms, perceived more SARS-CoV-2 health threats, spent more time in bed, and had shorter sleep duration at Time 1 were more likely to belong to the clinical group than to the good sleepers group. Those who were younger, female, had elevated psychological distress and PTSD symptoms, greater loneliness, spent more time in bed, and had reduced sleep duration at Time 1 were more likely to be a part of the subthreshold group as opposed to the good sleepers group. Education level, annual income, living situation, physical activity, nap duration, and variability in wake time were not independently associated with group membership in these multivariable analyses.

Table 2.

Multinomial logistic regression

	Clinical group		Subthreshold group
Predictor	OR	95% CI	OR	95% CI
Age	0.65a	[0.45, 0.86]	0.70b	[0.56, 0.84]
Gender (ref = male)	0.20c	[0.08, 0.32]	2.10b	[1.25, 2.95]
Education (ref = no university)	1.66	[0.76, 2.56]	1.04	[0.69, 1.39]
Annual income (ref = no poverty)	2.31	[0.98, 3.65]	0.95	[0.58, 1.32]
Living situation (ref = not living alone)	0.63	[0.28, 0.98]	1.24	[0.81, 1.68]
UCLA Loneliness Scale score at T1	1.36	[0.93, 1.79]	1.44c	[1.17, 1.71]
Kessler Psychological Distress Scale score at T1	1.46a	[1.03, 1.90]	1.40c	[1.15, 1.66]
Post-Traumatic Checklist score at T1	3.67c	[2.44, 4.89]	2.51c	[1.96, 3.07]
Perceived health threat of COVID-19 at T1	1.59a	[1.08, 2.10]	1.18	[0.95, 1.42]
International Physical Activity Questionnaire score at T1	0.83	[0.56, 1.10]	0.98	[0.80, 1.15]
Nap time at T1	1.13	[0.82, 1.44]	1.16	[0.96, 1.37]
Sleep duration at T1	0.02c	[0.01, 0.03]	0.10c	[0.07, 0.12]
Time in bed at T1	6.04c	[3.96, 8.12]	3.89c	[2.85, 4.93]
Social rhythm at T1	1.15	[0.82, 1.49]	1.08	[0.89, 1.26]

Note. N = 644; ref refers to reference categories for the respective predictors; T1 refers to the first wave of the pandemic. All variables were entered simultaneously in the statistical models. For the calculation of the odd ratios, the good sleepers group served as the reference group.

^aP < .05;

^bP < .01;

^cP < .001.

Discussion

The current study investigated changes in insomnia symptoms in older individuals during the COVID-19 pandemic. The average insomnia severity within this sample remained stable during the first year of the COVID-19 pandemic. Group-based trajectory identified 3 trajectory groups with varying degrees of insomnia symptoms. Membership to the clinical group was associated with younger age, male gender, elevated psychological distress, more posttraumatic stress symptoms, greater perceived health threat, shorter sleep duration, and longer time spent in bed at Time 1 compared to the good sleepers group. Membership to the subthreshold group was predicted by younger age, female gender, elevated psychological distress, more posttraumatic stress symptoms, more loneliness, shorter sleep duration, and more time spent in bed at T1 compared to the good sleepers. Thus, both sleep and psychological factors were associated with group membership.

In the present sample, insomnia symptoms of older adults were stable during the first year of the COVID-19 pandemic, consistent with results from a longitudinal study in middle-aged individuals.³² The clinical group, comprising about 11.8% of the sample, exhibited higher levels of insomnia symptoms throughout the follow-up compared to the rest of the sample. The proportion of individuals meeting the clinical threshold for clinical insomnia was somewhat lower than the 15.6% prevalence of clinical insomnia symptoms observed in a meta-analysis of cross-sectional studies with older adults in China during the pandemic.³³ Furthermore, it is important to note that not all individuals in the clinical group reached the clinical cutoff for insomnia (ISI ≥15), and our sample had a relatively low percentage of severe insomnia (3.73% prevalence) compared to the 13.0% prevalence of severe insomnia in older adults from Sweden prior to the pandemic,³⁴ suggesting possible sampling biases. Similarly, our “subthreshold” group did not reach the ISI>8 cutoff at all timepoints. The subthreshold group accounted for an additional 25.3% of older adults, suggesting that a total of 37.1% of the participants reported some level of insomnia symptoms that may be negatively impacting their well-being. This is consistent with the 36.9% prevalence of subthreshold and clinical insomnia reported before the pandemic in older adults from Sweden but higher than the 22.2% prevalence among those from Canada.^{34 , 35}.

The factors associated with the largest odds ratio of membership to the clinical and subthreshold groups were time in bed and sleep durations at Time 1. During the first wave of the COVID-19 pandemic, these older adults spent more time lying in bed while reporting shorter sleep duration than the good sleepers. This greater time in bed combined with lower sleep duration reflects poor sleep efficiency. In one study, older adults maintained their usual sleep schedule during the pandemic. Here, it could be that, although their sleep duration remained unchanged from before the pandemic, their sleep efficiency decreased by spending more time in bed than usual.³⁶ However, given the lack of prepandemic data, it is not clear if the time spent in bed increased during the pandemic. Regardless, the extensive time in bed observed in both the clinical and subthreshold groups has been proposed to be an important factor perpetuating insomnia symptoms. Remaining in bed while awake for long periods may promote conditioned arousal, where the bed environment loses its learned association with sleep and becomes associated with wake or negative feelings.³⁷

The second largest odds ratio of membership to the clinical and subthreshold groups consisted of elevated COVID-19-related PTSD symptoms at Time 1. This is in line with previous findings that greater PTSD symptoms were associated with poorer sleep quality and increased frequency of early awakenings during the early phase of the pandemic.²¹ Hyperarousal, the heightened emotional, cognitive, and physiological activation associated with acute stress symptoms, may interfere with sleep initiation or maintenance, or lead to restless sleep.³⁸ Similarly, higher psychological distress was a significant predictor of membership to the clinical and subthreshold groups.

Greater perceived health threat was a significant predictor of membership to the clinical group, whereas loneliness was a significant predictor of membership to the subthreshold group. These results are consistent with previous findings, suggesting that psychological distress²⁰ and perceived loneliness15, 18 during the early stages of the COVID-19 pandemic were associated with higher insomnia severity and worse sleep quality. The lack of social relationships may also enhance the association between psychological distress and insomnia among older adults.³⁹ In addition, it was previously shown that older adults who were more concerned about themselves or their kin contracting COVID-19 reported more insomnia symptoms.¹⁵

Within this cohort of older adults, individuals who were younger were more likely to have subthreshold or clinical levels of insomnia symptoms throughout the pandemic compared to the good sleepers group. One study found that increased age was associated with poorer sleep quality and greater insomnia symptoms.¹⁶ However, other studies suggest that older individuals presented with better psychological adaptation and less insomnia during the pandemic than younger individuals.^{14, 40} We also found that females were more likely to have subthreshold levels of insomnia symptoms. However, males were more likely to belong to the clinical group. This contradicts what has previously been reported in past literature.12, 14, 36 One explanation could be that older females who tend to have more social support than males were better equipped to cope with confinement-related stress.⁴¹

Although nap duration and annual income were positively associated with insomnia severity in bivariate analyses, they were not independent predictors in the multivariable models. They might then be associated with insomnia severity through their shared associations with other variables included in the model. In the present study, education level, living situation, variability in wake-up time, and physical activity were not associated with insomnia severity. The association between these factors and insomnia may be attenuated among older adults in the context of a pandemic with protracted confinement measures. Alternatively, sampling bias may account for some of the discrepancies with other findings observed in the literature.

The most notable strength of this study is its longitudinal design, which enabled us to examine patterns of sleep changes among older adults across a 12-month period that covered 3 waves of the COVID-19 pandemic. Interpretation of these findings must be done in light of the study limitations. First, without prepandemic insomnia data, it is impossible to ascertain whether the current insomnia symptoms were triggered or intensified by the pandemic or represented preexisting issues. Furthermore, the generalizability of our findings is limited by the fact that our convenience sample consisted predominantly of white and female participants who were well-educated. Likely due to these sampling biases, the prevalence of clinical insomnia at Time 1 was lower than in other population-based studies, highlighting the importance of replicating these findings in more representative samples.⁴² Moreover, our sleep diary assessment was retrospective and, therefore, prone to recall biases. Finally, we did not account for any medication use in our analyses because precise information on this matter was not collected.

Despite these limitations, the present results highlight that more than a third of older adults displayed persistent insomnia symptoms at a clinical or subclinical level throughout the first year of the pandemic. Many of these older adults would likely benefit from sleep-focused intervention. Cognitive-behavioral therapy for insomnia (CBTi) has been shown to be effective for older adults with insomnia and comorbid disorders.⁴³ Accordingly, nearly all factors associated with greater insomnia severity in the current study can be addressed with different components of CBTi. For instance, the cognitive component of CBTi can target PTSD symptoms and psychological distress, whereby excessive worrying can be addressed through Socratic questioning and behavioral experiments. Notably, digital CBTi has been shown to be effective in improving both insomnia and depressive symptoms at a 1-year follow-up in older adults, providing a scalable intervention in the context of a pandemic.⁴⁴ However, other forms of CBT may be needed to treat distress associated with the pandemic, for example, PTSD-related symptoms.

In conclusion, this study revealed that nearly 12% of older adults maintained higher insomnia symptoms than the rest of the sample over the first year of the COVID-19 pandemic with an additional 25% experiencing less, yet subthreshold, symptoms of insomnia that may still be negatively impacting their well-being and daily functioning. These results can help identify a subgroup of older adults with subthreshold or clinically significant levels of insomnia symptoms that would benefit from psychological intervention targeting sleep behaviors and cognitions. Future research should examine the trajectories of insomnia severity in the (post)pandemic era and investigate the effects of sleep-related interventions to improve sleep health among older adults and improve their postpandemic quality of life and functioning.

Declaration of conflicts of interest

There are no conflicts of interest to declare.

Funding

This study was supported by the Centre de recherche de l’Institut universitaire de gériatrie de Montréal and the Quebec Network for Research on Aging, a thematic network supported by the Fonds de Recherche du Québec – Santé, and a Canadian Institutes of Health Research Project Grant.