Association of Global Cognitive Function With Psychological Distress and Adherence to Public Health Recommendations During the Coronavirus Disease 2019 Pandemic: The Women’s Health Initiative Free

Abstract

Background

The association of cognitive function with symptoms of psychological distress during the coronavirus disease 2019 (COVID-19) pandemic or adherence to COVID-19 protective health behaviors is not well-understood.

Methods

We examined 2 890 older women from the Women’s Health Initiative cohort. Prepandemic (ie, within 12 months prior to pandemic onset) and peripandemic global cognitive function scores were assessed with the modified Telephone Interview for Cognitive Status (TICS-m). Anxiety, stress, and depressive symptom severity during the pandemic were assessed using validated questionnaires. We examined adherence to protective behaviors that included safe hygiene, social distancing, mask wearing, and staying home. Multivariable models were adjusted for age, race, ethnicity, education, region of residence, alcohol intake, and comorbidities.

Results

Every 5-point lower prepandemic TICS-m score was associated with 0.33-point mean higher (95% confidence interval [CI], 0.20, 0.45) perceived stress and 0.20-point mean higher (95% CI, 0.07, 0.32) depressive symptom severity during the pandemic. Higher depressive symptom severity, but not anxiety or perceived stress, was associated with a 0.69-point (95% CI, −1.13, −0.25) mean decline in TICS-m from the prepandemic to peripandemic period. Every 5-point lower peripandemic TICS-m score was associated with 12% lower odds ratio (OR, 0.88; 95% CI, 0.80, 0.97) of practicing safe hygiene.

Conclusions

Among older women, we observed that: (a) lower prepandemic global cognitive function was associated with higher stress and depressive symptom severity during the pandemic; (b) higher depressive symptom severity during the pandemic was associated with cognitive decline; and (c) lower global cognitive function during the pandemic was associated with lower odds of practicing safe hygiene.

Throughout the course of the coronavirus disease 2019 (COVID-19) pandemic, lockdowns, mask mandates, and social distancing measures brought unprecedented disruptions to our daily lives. Emerging evidence shows potentially negative psychosocial consequences from the response to the COVID-19 pandemic (1–4). COVID-19 pandemic-related stressors that may be associated with psychological distress include persistent fear of infection, social isolation, negative economic consequences, and uncertainty about the future (1).

In a nationally representative sample of 1 337 U.S. adults, 13.0% reported serious psychological distress due to the COVID-19 pandemic in April 2020 (2). The most common pandemic-related stressors included concerns about contracting COVID-19 (65.9%) and the pandemic’s adverse effects on employment (65.1%) and personal finances (60.6%). Because psychosocial distress is associated with risk of future cognitive decline, psychological distress during the COVID-19 pandemic represents an important public health issue (5–12). For example, depression and anxiety, 2 common symptoms of psychological distress, are associated with increased risk of cognitive decline, cognitive impairment, and dementia among older adults (5–9). In the Health and Retirement Study, participants who had higher levels of psychological distress (including anxiety, negative emotions, hostility, pessimism, and hopelessness) had 20%–30% higher risk of dementia (10).

Few studies have examined the association of cognitive function with psychological distress during the pandemic. The relationship between cognitive function and adherence to COVID-19 protective health behaviors (eg, wearing masks and practicing safe hygiene) is also not well-understood and may represent added risk among older adults with low cognitive function. In this analysis, we aimed to determine whether: (a) lower prepandemic cognitive function was associated with higher likelihood of experiencing symptoms of psychological distress during the pandemic; (b) symptoms of psychological distress during the pandemic were associated with cognitive decline from the prepandemic to peripandemic period; and (c) lower peripandemic cognitive function was associated with lower likelihood of adhering to key COVID-19 protective health behaviors. To address these questions, we leveraged data from 2 ongoing observational ancillary studies and 1 intervention ancillary study from the Women’s Health Initiative (WHI) cohort of older women. Relative to men, women have reported greater stress and poorer mental health during the COVID-19 pandemic (13). Thus, from a public health perspective, continued monitoring and reporting of women’s mental health outcomes during the pandemic is important.

Method

Study Population

The WHI is a large, prospective study investigating major determinants of chronic diseases among postmenopausal women. The WHI recruited 161 808 postmenopausal women aged 50–79 years during 1993–1998 to participate in an observational study (N = 93 676), or 1 or more of 3 clinical trials (N = 68 133). The clinical trials included 2 hormone therapy (HT) trials (which ended in 2002 and 2004), a dietary modification trial (which ended in 2005), and a calcium and vitamin-D supplementation trial (which ended in 2005). In 2005, 76.9% of 150 075 eligible WHI women consented for an additional 5 years of follow-up in the first Extension Study. In 2010, 86.8% of 107 706 women consented to further follow-up through 2015 in the second Extension Study. The third Extension Study is currently ongoing. The WHI study design and methods are described in detail elsewhere (14).

The present study included WHI participants from 3 ancillary studies with measures of global cognitive function: Women’s Health Initiative Memory Study Epidemiology of Cognitive Health Outcomes (WHIMS-ECHO), WHI Sleep Hypoxia Effects on Resilience (WHISPER), and Cocoa Supplement and Multivitamin Outcomes Study for the Mind (COSMOS-Mind). Participants for these ancillary studies were drawn from the WHI cohort, and there is no overlap in participation. All participants provided written informed consent, and these studies obtained Institutional Review Board approval at the Fred Hutchinson Cancer Research Center.

WHIMS-ECHO (N = 2 893) is an ancillary, observational follow-up study to the larger WHIMS, a randomized clinical trial that aimed to determine the effects of hormone therapy on global cognitive function and incidence of dementia among a subset of women aged 65–79 years recruited into the WHI HT trials, which ended in 2002 and 2004 but continued to follow women through 2007 (15). Follow-up in the observational WHIMS-ECHO study for cognitive outcomes started in 2008 and continued through November 2021. WHISPER (N = 5 000) examines the relationship between sleep-disordered breathing and other sleep exposures, and cardiovascular disease (CVD), cancer, and cognitive trajectory. WHI women aged 65 and older were enrolled into WHISPER between 2017 and 2018, and followed for up to 3 years. COSMOS-Mind (N = 2 262), an ancillary study to the larger COSMOS randomized clinical trial, assessed the effects of cocoa extract and a multivitamin on cognitive function in adults aged 65 years and older; the trial ended on December 31, 2020 (16). COSMOS-Mind recruited a group of women from the WHI who were not participants of WHIMS-ECHO. Women in COSMOS-Mind were followed for 3 years.

Global Cognitive Function Measure

The modified Telephone Interview for Cognitive Status (TICS-m) is designed to assess global cognitive function and was administered annually by trained and certified interviewers as part of the WHIMS-ECHO, WHISPER, and COSMOS-Mind protocols. TICS-m is a 16-item instrument that assesses orientation, attention and concentration, short- and long-delay free recall, mental calculation, naming, repetition, social knowledge, and praxis (17). TICS-m has good reliability (ie, Cochran’s alpha = 0.78) and validity (ie, sensitivity of 94% and specificity of 100% for discriminating cognitively normal from demented individuals) (18,19).

For the present study, we designated January 30, 2020, the day when the Director-General of the World Health Organization declared COVID-19 a public health emergency of international concern, as the start of the pandemic. The participant sample included women who completed the TICS-m within 12 months prior to the onset of the COVID-19 pandemic (ie, January 30, 2019 through January 29, 2020), and the WHI COVID-19 questionnaire that was administered to all WHI participants between June 2020 and October 2020. The most recent TICS-m score was used. For analyses in the subset of women with cognitive assessments completed during the pandemic, peripandemic TICS-m scores were collected between January 30, 2020 and February 23, 2021 (ie, the last date through which TICS-m scores were available at the time of these analyses).

Psychological Distress Measures

Anxiety

Anxiety was assessed using the Patient-Reported Outcomes Measurement Information System (PROMIS) Anxiety Short Form 4a questionnaire, which was included as part of the WHI COVID-19 questionnaire (20). This questionnaire has excellent reliability (ie, Cronbach’s alpha = 0.89) and good construct validity, with higher PROMIS anxiety scores associated with greater number of anxiety disorders in a prior study (21). The questionnaire contains the following 4 items rated on a 5-point Likert-type scale (where response choices vary from 1 = “never” to 5 = “always”) that assess anxiety levels in the past 7 days: (a) “I felt fearful”; (b) “I found it hard to focus on anything other than my anxiety”; (c) “My worries overwhelmed me”; and (c) “I felt uneasy.” Raw scores were summed and then converted to standardized T-scores according to the PROMIS Anxiety Short Form 4a Conversion Table, resulting in a mean score of 50 and a standard deviation (SD) of 10. Higher scores indicate greater levels of perceived anxiety. Scores of ≥60 are 1 SD above the population mean and are indicative of clinically significant anxiety symptoms (22–24).

Stress

Stress was assessed with the Perceived Stress Scale (PSS-4), which was included on the WHI COVID-19 questionnaire (25). PSS-4 has demonstrated good reliability (Cronbach’s alpha = 0.74) and validity, showing positive correlations with the Symptom Checklist-90-Revised anxiety and depression scales (r = 0.51 and r = 0.69, respectively) (26). PSS-4 asks: “In the past 4 weeks, how often have you felt: (a) that you were unable to control the important things in your life; (b) confident about your ability to handle your personal problems; (c) that things were going your way; and (c) that difficulties were piling up so high that you could not overcome them?” Each item was rated on a 5-point Likert-type scale (with response choices varying from 1 = “never” to 5 = “very often”). Total scores were obtained by summing across the 4 items, with higher scores representing greater levels of perceived stress.

Depressive symptom severity

Depressive symptom severity was assessed with the 15-item Geriatric Depression Scale (GDS), which was collected from January 30, 2020 to February 23, 2021 as part of the WHIMS-ECHO, WHISPER, and COSMOS-Mind annual cognitive assessments (27). The GDS is a self-report screening instrument assessing various symptoms of depression, including satisfaction with life, boredom, level of happiness, and feeling helpless, with responses rated as “yes” or “no.” The GDS has excellent reliability (with an average reliability across studies of 0.85) and validity (with sensitivity and specificity rates of 92.7% and 65.2%, respectively, and positive and negative predictive values of 82.6% and 83.3%, respectively, when measured against-International Classification of Diseases (ICD)-10 diagnostic criteria for major depressive episode). (28,29). Total scores are summed, with higher scores indicating greater depressive symptom severity. A score of ≥5 is indicative of mild to severe depressive symptom severity (27).

Adherence to COVID-19 Protective Health Behaviors

The WHI COVID-19 questionnaire assessed adherence to COVID-19 protective health behaviors using the following question: “Since March 2020, what steps have you taken to reduce your risk of being infected by COVID-19?” Response options included protective health behaviors recommended by the Centers for Disease Control and Prevention at the time of the questionnaire (June 2020–October 2020), which were analyzed as 4 separate binary variables (fully adherent or inconsistently/not adherent) for analysis, similar to a recent study (30): (a) practices safe hygiene (women were considered to have practiced safe hygiene if they responded “yes” to all of the following options: washes hands frequently, tries not to touch face, and disinfects surfaces frequently); (b) maintains social distancing (women were considered to maintain social distancing if they responded “yes” to all of the following options: maintains physical distance from people outside household, avoids in-person social or religious activities, avoids or limits in-person shopping, and avoids shaking hands); (c) wears a face mask in public; and (d) stays home.

Covariates

Covariates included: age; education; region of residence; history of CVD (defined as myocardial infarction, coronary artery bypass graft, percutaneous transluminal coronary angioplasty, or stroke), diabetes, hypertension, and cancer; and alcohol use, which were collected as part of ongoing WHI follow-up. These covariates were selected as they may affect prepandemic cognitive or mental health status, or affect adherence to health behaviors in older adults, according to prior literature (5–7,30–32). We included covariates for race and ethnicity due to differences in the levels of exposure and outcomes to the COVID-19 pandemic across racial and ethnic groups (33). We also examined number of people living in household and concerns about the pandemic, which were captured on the WHI COVID-19 questionnaire.

Statistical Analysis

Descriptive statistics of study characteristics in the overall sample are presented by prepandemic TICS-m score categorized using an established cutpoint, with >31 indicating normal cognition and ≤31 indicating impaired cognition (34). Continuous variables are presented as means (SD), and categorical variables are presented as frequencies and proportions. Differences in participant characteristics by TICS-m-grouping were evaluated using 2-sample t tests for continuous variables and chi-square tests for categorical variables.

Multiple linear regression models were used to determine associations of prepandemic global cognitive function with anxiety, stress, or depressive symptoms during the pandemic, which were examined separately as the dependent variables. The independent variable in each model was prepandemic TICS-m score, which was examined both as a dichotomous variable using cutpoints (≤31 vs >31) and a continuous variable to evaluate linear associations. Models were adjusted for: age; study source (WHIMS-ECHO, WHISPER, and COSMOS-Mind); race; ethnicity; education; history of CVD, diabetes, hypertension, and cancer; alcohol intake; region of residence; number of people living in household; total number of prior TICS-m screens to control for practice effects; and number of months from pandemic start (January 30, 2020) to date of psychological distress symptom collection (ie, on the COVID-19 questionnaire) during the pandemic. Analyses are presented as a 5-point change in TICS-m score, which represents an approximate 1 SD change in our data.

Associations of stress, anxiety, or depressive symptoms during the pandemic with change in prepandemic to peripandemic global cognitive function measured using the TICS-m were examined using multiple linear regression models. Separate models were fit with anxiety, perceived stress, and depressive symptom severity as the independent variables of interest. The dependent variable in the models was calculated as the difference between the prepandemic and peripandemic TICS-m scores. Models were adjusted for similar confounders as described previously, with the addition of number of months between prepandemic and peripandemic TICS-m measurements.

Associations of peripandemic global cognitive function with adherence to COVID-19 protective health behaviors were evaluated using logistic regression models. The following outcomes were examined in separate logistic regression models: “practices safe hygiene (yes/no),” “maintains social distancing (yes/no),” “wears a face mask in public (yes/no),” “stays home (yes/no),” or “adherence to all measures (yes/no),” as defined previously. Models were adjusted for age; study source; race; ethnicity; education; history of CVD, diabetes, hypertension, and cancer; alcohol intake; region of residence; number of people living in household; total number of prior TICS-m tests; and number of months from pandemic start to date of psychological distress symptom collection.

Sensitivity analyses were conducted restricting peripandemic TICS-m screens to those occurring after March 13, 2020, the day when COVID-19 was declared a national emergency in the United States and after which most social and other changes were implemented throughout the country. Sensitivity analyses were performed using cutpoints for depressive symptom severity (ie, mild to severe depressive symptoms [GDS score ≥5] relative to not depressed (GDS score 0–4) (27) and clinically significant anxiety symptoms (ie, PROMIS anxiety ≥60 vs <60) (22–24) to examine: (a) the association of prepandemic cognitive function with any depressive symptoms or clinically significant anxiety symptoms during the pandemic using logistic regression models, adjusted for similar confounders as previously and (b) the association of any depressive symptoms or clinically significant anxiety symptoms during the pandemic with changes in prepandemic to peripandemic cognitive function, using similar analyses as previously (27). To determine whether cognitive impairment affected participant response to questions on adherence to COVID-19 protective health behaviors, sensitivity analyses were performed excluding women with TICS-m ≤31 from the analyses.

In all analyses, we explored interactions with age (categorized according to 80 years, as the majority [57.9%] of the sample was above this age), number of people living in household, history of CVD, and history of diabetes. p Values were considered statistically significant at p < .017 (ie, 0.05/3 using Bonferroni correction) for analyses examining stress, anxiety, and depressive symptoms, and p < .01 (ie, 0.05/5) for analyses examining the 5 outcomes for COVID-19 protective behaviors. Statistical analyses were conducted using Statistical Analyses Software (SAS), Version 9.4 (SAS Institute, Cary, NC).

Results

Analytic Sample

In total, 3 283 women completed TICS-m within the 12 months prior to the start of the COVID-19 pandemic and completed the WHI COVID-19 questionnaire administered between June 2020 and October 2020. After excluding first those who self-reported having previously tested positive for COVID-19 (n = 22) and then those missing data on stress or anxiety symptoms during the pandemic (n = 371), there remained 2 890 women with prepandemic global cognitive function data (n = 2 211 from WHISPER; n = 434 from WHIMS-ECHO; and n = 245 from COSMOS-Mind). A subset (n = 1 690) of these women had peripandemic depressive symptom data.

Overall, 3 040 women completed TICS-m on or after January 30, 2020 and completed the WHI COVID-19 questionnaire. After excluding first those who tested positive for COVID-19 (n = 16) and then those missing data on stress or anxiety (n = 324), there remained 2 700 women with peripandemic global cognitive function data (n = 2 240 from WHISPER; n = 374 from WHIMS-ECHO; and n = 86 from COSMOS-Mind). A subset of these women (n = 1 690) also had global cognitive function data collected within 12 months prior to pandemic onset.

Descriptive Characteristics of Sample

Among the 2 890 women with prepandemic global cognitive function data, mean age was 82.8 (SD 5.8) years; 78.8% were White; 16.8% were African American; 1.0% were Asian; 1.1% were more than 1 race; 0.2% were American Indian/Alaskan Native; 2.1% were of unknown race; and 6.7% were of Hispanic ethnicity. Overall, 52.1% were college graduates; 9.2% had history of CVD; 22.1% had history of diabetes; and 71.5% had history of hypertension. Approximately half (50.4%) reported living with at least 1 person in the same household; 6.6% had a family member or close friend die from COVID-19; and 92.7% were somewhat or very concerned about the pandemic. The most common pandemic-related concerns included the health, safety, and infection of family and friends (79.8%); the ability to be with family and friends (63.4%); and perceived risk of COVID-19 infection (61.9%). Regarding adherence to COVID-19 protective health behaviors, 96.7% wore a face mask; 81.5% stayed home; 60.3% maintained social distancing; and 43.9% practiced safe hygiene. The mean (SD) most recent prepandemic TICS-m score was 35.2 (4.6). The mean (SD) most recent peripandemic TICS-m score was 35.3 (SD 4.7). There was a mean of 16.0 (SD 2.4) months between prepandemic and peripandemic TICS-m screens.

Relative to women with prepandemic TICS-m scores >31, women with scores ≤31 were older, less likely to be college graduates, less likely to report >1 alcoholic drink per week, and more likely to have history of diabetes or hypertension (Table 1). Women with TICS-m scores ≤31 were less likely to be: concerned about the risk of COVID-19 infection, the health and safety of family and friends, the ability to be with family or friends, and the nation and economy (Supplementary Table S1). They were more likely to be concerned about having enough money, personal safety, and financial security (Supplementary Table S1).

Prepandemic Global Cognitive Function and Symptoms of Psychological Distress During the Pandemic

Multivariable associations of prepandemic global cognitive function with stress, anxiety, and depressive symptom severity during the pandemic are shown in Table 2. Prepandemic cognitive function was not significantly associated with peripandemic anxiety. Every 5-point lower prepandemic TICS-m score was associated with 0.33-point mean higher perceived stress during the pandemic (95% CI, 0.20, 0.45; p < .001). Women with TICS-m scores ≤31 compared with >31 had 0.62-point mean higher perceived stress (95% CI, 0.35, 0.90; p < .001). Every 5-point lower prepandemic TICS-m score was associated with 0.20-point mean higher depressive symptom severity score during the pandemic (95% CI, 0.07, 0.32; p = .003). Findings for anxiety, perceived stress, or depressive symptom severity did not vary by age, number living in household, history of CVD, or history of diabetes (Supplementary Table S2). In sensitivity analyses using cutpoints for depressive symptom severity, 223/1 690 women (13.2%) had any depressive symptoms during the pandemic, whereas 1 467/1 690 (86.8%) were not depressed. Every 5-point lower prepandemic TICS-m score was associated with 21% higher odds of any depressive symptoms during the pandemic (OR 1.21; 95% CI, 1.02–1.45; Supplementary Table S3). Overall, 268/2 890 (9.3%) women had clinically significant anxiety symptoms (ie, a PROMIS anxiety score ≥60) during the pandemic. Prepandemic TICS-m was not significantly associated with peripandemic anxiety defined using this cutpoint (Supplementary Table S3).

Psychological Distress and Prepandemic to Peripandemic Change in Global Cognitive Function

Multivariable associations of stress, anxiety, and depressive symptom severity during the pandemic with prepandemic to peripandemic change in global cognitive function are presented in Table 3. Anxiety or perceived stress scores during the pandemic were not significantly associated with prepandemic to peripandemic change in global cognitive function (Table 3). Every 5-point higher depressive symptom severity during the pandemic was associated with 0.69-point mean decline in TICS-m score from the prepandemic to peripandemic period (95% CI, −1.13, −0.25; p = 0.002; Table 3). Findings for anxiety, perceived stress, or depressive symptom severity did not vary by age, number living in household, history of CVD, or history of diabetes (Supplementary Table S4). In sensitivity analyses limited to participants with peripandemic TICS-m scores collected after March 13, 2020, findings were similar (Supplementary Table S5). Having any depressive symptoms during the pandemic was associated with a nonsignificant 0.41-point mean decline in global cognitive function (95% CI, −0.99, 0.16) from the prepandemic to peripandemic period (Supplementary Table S6). Having clinically significant anxiety symptoms during the pandemic was not significantly associated with change in global cognitive function (Supplementary Table S6).

Peripandemic Global Cognitive Function and Adherence to COVID-19 Protective Health Behaviors

Multivariable associations of peripandemic global cognitive function with adherence to COVID-19 protective health behaviors are presented in Table 4. Every 5-point lower peripandemic TICS-m score was associated with 12% lower odds (OR, 0.88; 95% CI, 0.80–0.97; p = 0.007) of practicing safe hygiene. There were no significant associations of peripandemic global cognitive function with other COVID-19 protective behaviors. The odds of maintaining social distancing were 17% lower for every 5-point lower TICS-m score among women older than 80 years (OR, 0.83; 95% CI, 0.74–0.94) but not among those 80 years or younger in age-stratified analyses. Other findings did not vary by age, number living in household, history of CVD, or history of diabetes (Supplementary Table S7). Findings were similar when excluding women with TICS-m ≤31 from the analyses (Supplementary Table S8).

Discussion

In a sample of older WHI women, we observed that lower prepandemic global cognitive function was associated with higher perceived stress and depressive symptom severity during the COVID-19 pandemic, independent of age, race, ethnicity, education, region of residence, alcohol intake, and comorbidities. We also observed that greater depressive symptom severity during the pandemic was associated with a decline in prepandemic to peripandemic global cognitive function. Women with lower peripandemic global cognitive function were less likely to report practicing safe hygiene during the pandemic.

Multiple studies have reported a negative effect of the COVID-19 pandemic on psychological health among younger and older adults (1–4,35–40). A systematic review and meta-analysis of 55 studies found a 16.0% and 15.2% prevalence of depression and anxiety, respectively, during the COVID-19 pandemic (36). In a representative sample of 1 441 U.S. adults aged 18–39 years, depressive symptoms were more than 3-fold higher during COVID-19 compared with before the pandemic, increasing from 8.5% to 27.8% (39). Furthermore, women were more likely to report depressive symptoms than men (39). Prior to COVID-19, studies had identified an 8.8% and 7.3% prevalence of depression and anxiety, respectively, in the general population (41,42).

Because maintaining good mental health is critical to overall well-being and quality of life in older adults, identifying subgroups who may be particularly vulnerable to pandemic-related psychological distress is important in the public health response to the pandemic. Older adults may have unique risk factors rendering them more vulnerable to psychological distress. For example, a study among 2 785 older adults with a median age of 70 years observed that frailty and multimorbidity, 2 common geriatric syndromes among older adults, were associated with greater psychological distress during the COVID-19 pandemic (35). Our findings extend this prior study by suggesting that low cognitive function, which is prevalent especially among the oldest of older adults, may be a risk factor for pandemic-related psychological distress. Our findings also suggest that depressive symptom severity during the pandemic may be associated with cognitive decline; however, given the short period of observation and the fact that affective state can impact cognitive test performance, this finding should be interpreted with caution.

The mechanisms underlying the association between psychological distress and poor cognitive health have not been fully elucidated. Prior studies have shown that depression is associated with higher risk of cognitive decline and cognitive impairment in older adults (7,43). It is possible that late-life depression is a symptom in the progression to future cognitive impairment, which may partly explain the association of lower prepandemic global cognitive function with depressive symptom severity during the pandemic in our study (7,44). Our study was unique in that it examined symptoms of psychological distress during an unprecedented pandemic that severely disrupted people’s daily lives. Future studies are needed to determine the long-term effect of COVID-19 pandemic-related psychological distress on cognitive outcomes among older women, including the extent to which psychological distress during the pandemic may accelerate cognitive decline and the development of mild cognitive impairment or dementia.

Key public health measures recommended by the Centers for Disease Control and Prevention to mitigate the spread of COVID-19 have included wearing masks, washing hands, maintaining 6 feet of distance with others who do not live in your household, avoiding crowds and indoor spaces, and cleaning and disinfecting frequently touched surfaces (45). However, the efficacy of these recommendations depends in large part on the public’s willingness to adhere to them. Prior studies identified male sex, younger age, high psychological distress, low perceived risk of COVID-19, and limited exposure to and perceived efficacy of public health recommendations, as predicting nonadherence to COVID-19 protective behaviors (46,47). Female sex and having a chronic disease have been identified as predicting adherence to these behaviors among adults (46,48). Our findings suggest that low cognitive function during the pandemic may be associated with lower adherence to practicing safe hygiene. This finding is supported by a recent analysis from the Health and Retirement Study, which showed that better memory performance was associated with greater adherence to COVID-19 protective behaviors (eg, mask usage, hand washing, social distancing, and use of hand sanitizers) (30). Memory and executive function are essential to the initiation and maintenance of health behaviors and the development of habits (27). Thus, individuals with impaired cognition may have difficulty adhering to COVID-19-related protective behaviors, as such behaviors must be planned, decisions must be made about situations in which to engage in such behaviors (eg, large crowds), and failure to protect oneself (eg, through mask wearing) must be monitored and detected (27). Thus, older adults with impaired cognition should be prioritized in public health messaging to promote COVID-19-related protective health behaviors.

Our study has several limitations. Our findings may generalize only to older women. Further, our study included women from 3 WHI ancillary studies, and we did not include men from COSMOS-Mind, given our focus on the women in WHI; selection bias may be possible. The period of follow-up to determine prepandemic to peripandemic change in cognitive function was relatively short, and we assessed psychological distress measures at a single time point. This may partly explain the modest associations of psychological distress measures with cognitive function observed in our study. Greater follow-up is needed to determine the extent of the effect of the COVID-19 pandemic on cognitive and mental health among older women, including assessment of social and lifestyle changes that occurred due to the pandemic and whether these are associated with cognitive decline in the long term. All measures of psychological distress were self-reported using screening instruments, and we lacked information on diagnosis of depression and anxiety. We also lacked information on psychiatric treatments that could have influenced cognitive function. Adherence to protective behaviors was assessed via self-report and thus may not fully reflect actual adherence for those with cognitive impairment. However, our findings were similar when we excluded women with TICS-m ≤31, suggesting this is an unlikely explanation of our findings. Strengths of this study include a large sample, information on a broad range of covariates, examination of several key psychological distress measures, and availability of well-validated global cognitive function measures both preceding and during the pandemic.

In conclusion, in a sample of older WHI women, lower prepandemic global cognitive function was associated with higher perceived stress and depressive symptom severity during the COVID-19 pandemic. Furthermore, depressive symptom severity during the pandemic was associated with a decline in global cognitive function from the prepandemic to peripandemic periods. We also observed that older women with lower global cognitive function during the pandemic were less likely to practice safe hygiene. The long-term effect of the COVID-19 pandemic on cognitive and mental health outcomes among older women requires further study.

Acknowledgments

Program office: (National Heart, Lung, and Blood Institute, Bethesda, MA) Jacques Rossouw, Shari Ludlam, Joan McGowan, Leslie Ford, and Nancy Geller. Clinical coordinating center: (Fred Hutchinson Cancer Research Center, Seattle, WA) Garnet Anderson, Ross Prentice, Andrea LaCroix, and Charles Kooperberg. Investigators and academic centers: (Brigham and Women’s Hospital, Harvard Medical School, Boston, MA) JoAnn E. Manson; (MedStar Health Research Institute/Howard University, Washington, DC) Barbara V. Howard; (Stanford Prevention Research Center, Stanford, CA) Marcia L. Stefanick; (The Ohio State University, Columbus, OH) Rebecca Jackson; (University of Arizona, Tucson/Phoenix, AZ) Cynthia A. Thomson; (University at Buffalo, Buffalo, NY) Jean Wactawski-Wende; (University of Florida, Gainesville/Jacksonville, FL) Marian Limacher; (University of Iowa, Iowa City/Davenport, IA) Jennifer Robinson; (University of Pittsburgh, Pittsburgh, PA) Lewis Kuller; (Wake Forest University School of Medicine, Winston-Salem, NC) Sally Shumaker; (University of Nevada, Reno, NV) Robert Brunner. Women’s Health Initiative Memory Study: (Wake Forest University School of Medicine, Winston-Salem, NC) Mark Espeland.

Funding

We gratefully acknowledge the 161,808 women who enrolled in the Women’s Health Initiative almost 30 years ago, and those who are able to remain active and engaged partners in this journey to advance our understanding of the health and well-being of women. We are particularly grateful to the 49,695 women who participated in the WHI COVID-19 Survey – 1, launched during a disruptive period in their lives. Their generosity adds critical information on the impact of the pandemic on one of our most vulnerable populations. We also thank the WHI Clinical Coordinating Center at Fred Hutchinson Cancer Research Center for conducting the COVID survey and providing financial and statistical support for this effort, as well as the four WHI Regional Centers who assisted in conducting the COVID-19 Survey - 1 (Stanford School of Medicine, The Ohio State University College of Medicine, University at Buffalo School of Public Health & Health Professions, and Wake Forest University Health Sciences). Finally, we thank Wake Forest University Health Sciences for providing the financial support for publication costs that made this supplement possible.

Conflict of Interest

None declared.

Author Contributions

Conception or design of the work: A.H.S., S.R.R., S.A.S., and L.D.B.; Acquisition and analysis of data: S.R.R., S.A.S., L.D.B., and J.C.L.; Preparation of manuscript: A.H.S.; Interpretation of data: all authors; Critical revision of the work for important intellectual content: all authors; Final approval of the version to be published: all authors.

References