Mental Health Struggles Among Older Adults During the Pandemic

The Household Pulse Survey has provided regular insight into respondents’ mental health and well-being throughout the pandemic.

Since April 2020, the HPS has asked two questions about anxiety symptoms and two questions about depression symptoms. This analysis relies on pooled data from two distinct collection phases of the survey:

• Phase 1 (April 23-July 21, 2020) – 14.2 million invitations, 1.1 million responses, 7.7% response rate.
• Phase 3.10 (August 23-October 30, 2023) – 3.2 million invitations sent, 220,664 responses, 7.0% response rate.

These two collection phases enabled us to analyze data over roughly a three-year period, providing a snapshot of experiences during the pandemic’s early days and after the pandemic emergency ended.

From the beginning, the HPS has included four mental health questions, two relating to reported symptoms of anxiety and two about reported symptoms of depression. 

These questions were developed in partnership with the National Center for Health Statistics (NCHS). This analysis follows an approach outlined by NCHS, which categorizes individuals based on how frequently they reported feeling either anxious or depressed, consistent with diagnoses of generalized anxiety disorder or major depressive disorder, although it is important to highlight that these are self-reported outcomes to a survey, and are not intended to reflect an official diagnosis of clinical anxiety or depression. 

For this analysis, respondents are categorized as experiencing mental health struggle if they reported symptoms of either anxiety or depression at a frequency meeting the thresholds outlined by NCHS.

The degree to which adults 65 and older struggled with mental health varied by sociodemographic factors.

During Phase 1, older women (26.5%) were more likely than older men (18.6%) to struggle with mental health. This disparity was also present during Phase 3.10, with more older women (21.5%) than men (16.9%) reporting symptoms in the latest cycle.

Reported struggles with anxiety and depression, however, decreased among both groups during the cycle, indicating mental health outcomes broadly improved as the pandemic wore on (Figure 2). 

Results were not as straightforward by race and ethnicity (Figure 3).

During Phase 1, for example, both non-Hispanic White and Asian adults 65 and older reported rates of struggle less than other groups. (The rates for non-Hispanic Whites and Asians were not statistically different.)

But most race and Hispanic origin differences were no longer statistically significant during Phase 3.10, although both older non-Hispanic White and Black adults reported lower levels of struggle than those who were non-Hispanic of Some Other Race or Hispanic of any race. In this latter phase, rates for older non-Hispanic White, Asian and Black adults were not statistically different.

Rates of mental health struggle improved between the two phases for both non-Hispanic White and Black adults 65 and older but not among older non-Hispanic Asian and non-Hispanic of Some Other Race individuals. 

In line with past research on the population as a whole, living arrangements of older people continued to make a difference in reported symptoms of mental health struggle (Figure 4).

In Phase 3.10, for example, older adults living with other adults but without children in the home (17.2%) fared better than those living by themselves (22.5%) or with others including children (27.3%).

Other characteristics associated with older adults’ mental health:

• Education.  The percentage of the older population struggling with mental health decreased when educational attainment was higher: Nearly a third (32.6%) of those with less than a high school diploma experienced mental health struggles compared to 12.9% of those with a bachelor’s degree or higher.
• Health insurance. The vast majority of older adults receive health insurance through Medicare. Those who were not insured (40.7%) were more likely to experience mental health struggles than those who had health insurance (19.1%). Roughly 3 in 10 older adults insured via Medicaid struggled with mental health, a number that was not statistically different from those without any insurance.
• Disability. Older adults with a disability (44.9%) were much more likely than those without one (13.6%) to struggle with mental health, a finding that reflects the relationship between physical and mental health outcomes for the population in general and for older adults in particular. 

In addition to sociodemographic differences, older adults’ struggles with mental health also varied based on their economic circumstances and experiences (Figure 5).

For example, older adults who reported food insufficiency were also more likely to report symptoms of mental health struggle (34.2%) than those who had enough food (11.2%).

Similarly, those who struggled with paying basic household expenses were more likely to experience mental health challenges (28.2%) than those who did not struggle (7.8%).

As the nation grappled with price increases and the fear of possible future increases, older adults stressed by current inflation (22.5%) or concerned about future inflation (20.3%) were more likely to struggle with mental health than those not experiencing inflation-related stress (4.8%) or worried about future prices going up (7.8%).

The HPS is designed to provide near real-time data on how the pandemic affected people’s lives. Information on the methodology and reliability of these estimates can be found in the Source and Accuracy statements for each data release. These statements also include information on the survey invitations and response rates.

Data users interested in state-level sample sizes, the number of respondents, weighted response rates and occupied housing unit coverage ratios can review the quality measures file available on the Household Pulse Survey Technical Documentation webpage. 

In comparison to other Census Bureau surveys, HPS response rates are low and data users should exercise caution when interpreting estimates from the survey, especially with regard to the impact of potential nonresponse bias [PDF <1.0 MB]. Responding adults may differ from the general population with respect to both mental health struggle and demographic characteristic prevalence.

All comparative statements in this analysis have undergone statistical testing, and, unless otherwise noted, all comparisons are statistically significant at the 90 percent significance level.

As part of the Census Bureau’s experimental data series, the HPS was designed to have low respondent burden, provide quick turnaround on product releases and produce estimates that meet urgent public needs. All estimates discussed here were calculated from public-use microdata files.

This analysis was supported by the National Institute on Aging’s Division of Behavioral and Social Research.