Association of Severe COVID-19 and Persistent COVID-19 Symptoms With Economic Hardship Among US Families

Key Points Question Are COVID-19 outcomes, including severe COVID-19 and persistent COVID-19 symptoms, associated with economic hardship among US families? Findings In this cohort study of 6932 families, the odds of reporting economic hardship were higher for families headed by an adult with persistent COVID-19 symptoms and, to a lesser extent, families headed by an adult with previous severe COVID-19 compared with families with no history of COVID-19. Families with lower income before the pandemic were more vulnerable to employment disruptions and earnings losses associated with an adult family member’s COVID-19 illness. Meaning Policy actions to mitigate the household financial hardship of post–COVID-19 conditions merit continued discussion.


Introduction
More than 5.9 million COVID-19-associated hospital admissions have been recorded among US adults. 1 It is estimated that at least 3 million to 5 million US adults are currently living with activitylimiting post-COVID-19 conditions (PCCs) and that approximately 15% of the US adult population (approximately 39 million adults) have, at some point, experienced PCCs. 2,3Beyond significant personal health effects, there is accumulating evidence that patients with severe COVID-19 and PCCs face considerable economic consequences.Symptoms of PCCs can adversely affect daily functioning and have been associated with a lower likelihood of working full time and a higher likelihood of being unemployed. 4,5COVID-19-associated hospitalizations have been similarly associated with job loss. 67][8][9] The consequences may be even more severe for families with low income who have fewer resources available to buffer against COVID-19-related financial shocks but appear to be more likely to experience severe or long-term health effects from COVID-19. 10,11is study used data from the Panel Study of Income Dynamics (PSID), a long-running, nationally representative household panel survey with interviews conducted both before and during the COVID-19 pandemic, to examine associations between COVID-19 exposure and economic hardship at the family level and, furthermore, whether these associations vary according to a family's prepandemic socioeconomic status.We hypothesized (1) that COVID-19 exposure would be positively associated with experiences of economic hardship, (2) that the strength of these associations would vary according to the duration and severity of COVID-19 symptoms, and (3) that the economic sequelae of COVID-19 would be most pronounced among families with lower income before the pandemic.

Methods
This cohort study was deemed exempt from review by the Office of Research Compliance, an administrative office that supports the University of South Carolina Institutional Review Board, because it involved secondary analysis of publicly available, deidentified data.We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies. 12

Data Source
We used unrestricted public use data from the PSID, a nationally representative, longitudinal study of families in the US. 13 The core PSID survey, fielded biennially, collects detailed information on income, employment, wealth, and family composition.The PSID collects detailed information for the family's reference person (RP) and, if present, their spouse or partner (SP).

Measures Economic Hardship
We examined 3 indicators of family-level economic hardship reported by the RP in the PSID-2021 survey.They were (1) resident family member laid off or furloughed due to the pandemic, (2) resident family member lost earnings due to the pandemic, and (3) resident family member had financial difficulties due to the pandemic.

COVID-19 Exposure
Information related to COVID-19 infection and severity and duration of COVID-19 symptoms was collected for the RP and SP (if present).The RP was asked if they or their SP had "talked to a doctor or other health care professional about whether [they] may have had COVID-19."An affirmative response was followed by, "Did they say that [you/your SP] definitely had COVID-19, probably had it, may have had it, probably did not have it, or definitely did not have COVID-19?"If the RP reported that they or their SP were told that they definitely or probably had COVID-19, they were considered to have a positive COVID-19 diagnosis.Follow-up questions ascertained whether the RP or their SP was admitted to a hospital because of COVID-19 (yes or no); had experienced any COVID-19 symptoms (yes or no) and, if so, how bad or bothersome these symptoms were at their worst (mild, moderate, severe, or very severe); or were currently experiencing lingering physical or mental health effects from COVID-19 (yes or no).Families headed by at least 1 adult (RP and/or SP) with a COVID-19 diagnosis were assigned to 1 of 3 ordinal exposure categories:

Covariates
We included the following sociodemographic characteristics collected in the PSID-2019: any children younger than 16 years in the family (yes or no), age of the RP, self-reported race and ethnicity of the RP (Hispanic, non-Hispanic Black, non-Hispanic White, or other [including non-Hispanic Alaska Native, American Indian, Asian, or Pacific Islander persons or individuals who reported other race]), highest level of education completed by the RP or SP (less than college, some college, or college

JAMA Network Open | Public Health
COVID-19 Symptoms and Economic Hardship Among US Families graduate), total family income relative to US Census Bureau poverty thresholds 14 (income below or above 200% of the poverty threshold), whether any family member went without health insurance in the past 2 years (yes or no), geographic region (northeast, north central, south, or west), and residence in a nonmetropolitan area (yes or no).For each outcome of interest, we also identified a closely related indicator of economic hardship collected in the PSID-2019: whether the RP or SP missed work because they were temporarily laid off (yes or no), total family labor income, or whether any family member had credit card or store card debt (yes or no).

Statistical Analysis
First, we estimated a multinomial logistic regression model to assess correlations between families' sociodemographic characteristics and experiences of economic hardship at baseline and COVID-19 exposure.The no COVID-19 exposure category served as the reference outcome.Results are presented as adjusted relative risk ratios (ARRRs) with 95% CIs.
Next, we estimated logistic regression models to examine associations between COVID-19 exposure (persistent; severe; or moderate, mild, or asymptomatic COVID-19 vs no COVID-19) and 3 indicators of pandemic-related economic hardship (laid off or furloughed, lost earnings, and financial difficulties) among US families.Separate models were fit for each economic outcome.Results are presented as unadjusted odds ratios (ORs) or adjusted ORs (AORs) with 95% CIs.Adjusted models included a range of potentially confounding sociodemographic characteristics and, to account for the possibility that COVID-19 exposure may be associated with economic hardship at baseline, a closely related economic indicator collected in the PSID-2019 (Table 1).We further examined whether associations between COVID-19 exposure and economic hardship varied by prepandemic socioeconomic status.Differences across subgroups were evaluated by fitting separate regression models stratified by prepandemic family income (below vs above 200% of the poverty threshold).
All analyses were conducted using the survey commands in Stata, version 16 (StataCorp LLC) to apply family weights provided by the PSID.Estimates were considered statistically significant if 95% CIs did not include 1 (equivalent to 2-sided P < .05).
a Sample includes 6932 families.
b All adjusted models include controls for resident children, age of the reference person (RP), race and ethnicity of the RP, highest level of education completed by the RP or spouse or partner (SP), total family income relative to poverty thresholds, uninsurance, geographic region, and residence in a nonmetropolitan area.In addition, to account for potential differences in baseline economic hardship, each adjusted model includes a similar economic indicator collected in the Panel Study of Income Dynamics-2019 survey as a covariate: whether the RP or SP missed work because they were temporarily laid off, total family labor income, or whether anyone in the family had credit card or store card debt.
Among families with higher income, the odds of a resident member being laid off or furloughed  On balance, families with lower income before the pandemic (ie, families who had fewer resources available to buffer against COVID-19-related financial shocks) were more vulnerable to employment disruptions and earnings losses associated with an adult family member's COVID-19 illness.
Individuals living in economically vulnerable households are more likely to hold essential jobs, which have been associated with increased employment-related exposure and risk of severe COVID-19 among workers and their household members. 15,16Low-wage essential workers, who typically do not have the ability to work from home, are also least likely to have access to paid leave. 17,18Our results highlight the extent to which the economic sequelae of COVID-19 vary according to socioeconomic status and, consequently, may compound preexisting inequalities.
A total of 15.4% of households in our sample were headed by an adult with a prior COVID-19 diagnosis.Of those, close to 1 in 4 (weighted 28.4%; 4.4% of the overall sample) reported persistent symptoms consistent with PCCs.[24]

Limitations
This study has some limitations.

Conclusions
This cohort study suggests that persistent COVID-19 symptoms and, to a lesser extent, severe COVID-19 were associated with increased odds of pandemic-related economic hardship among a cohort of US families.The economic consequences of COVID-19 varied according to socioeconomic status; families with lower income before the pandemic were more vulnerable to employment disruptions and earnings losses associated with an adult family member's COVID-19 illness.

JAMA Network Open | Public Health
COVID-19 Symptoms and Economic Hardship Among US Families Fewer details are collected for other family members.In 2021, the PSID added questions to assess the outcome of the JAMA Network Open | Public Health COVID-19 Symptoms and Economic Hardship Among US Families COVID-19 pandemic on study families, including COVID-19-related health outcomes and pandemicrelated financial difficulties.The PSID-2021 survey was fielded from March 19 to December 31, 2021.The PSID-2021 Early Release file includes data collected from a balanced panel of 8468 families who were active in the PSID in both 2019 and 2021.We merged postpandemic measures of family economic hardship and COVID-19-related health outcomes collected in the PSID-2021 survey with prepandemic sociodemographic characteristics collected in the PSID-2019 survey.We excluded families with incomplete sociodemographic or economic information (n = 259) or indeterminate COVID-19 exposure (n = 1277).The study cohort comprised 6932 families.
First, we cannot definitively rule out the possibility that the increased odds of economic hardship we observed among families headed by an adult with previous severe COVID-19 or persistent COVID-19 symptoms were associated with factors other than COVID-19 illness.However, leveraging the longitudinal nature of PSID data, we were able to adjust for a range of potentially confounding factors measured prior to the pandemic.The absence of a strong association between economic hardship and previous moderate, mild, or asymptomatic COVID-19 further strengthens our interpretation of the study findings.Second, measures of economic hardship, COVID-19 diagnosis, and severity and duration of COVID-19 symptoms were self-reported and thus subject to recall and misclassification bias.Third, we excluded those with possible symptoms of COVID-19 who did not talk to a health care professional.Fourth, the PSID-2021 survey was fielded from March to December 2021, so our results may not reflect later phases of the pandemic.

Table 2
presents results from a multinomial logistic regression in which the reference outcome is no COVID-19 exposure.Accounting for baseline family characteristics, prepandemic experiences of economic hardship were associated with COVID-19 exposure.Families headed by an adult who missed work because they were temporarily laid off or furloughed were at higher risk of persistent COVID-19 symptoms (ARRR, 2.82 [95% CI, 1.34-5.93]),while families with credit card or store card debt (an indicator of financial difficulties) were at higher risk of moderate, mild, or asymptomatic COVID-19 infection (ARRR, 1.58 [95% CI, 1.25-2.00]).

Table 3
presents results from unadjusted and adjusted logistic regression models.Compared with families with no history of COVID-19, families headed by an adult with persistent COVID-19

Table 3 .
Unadjusted ORs and AORs of Family Economic Hardship by COVID-19 Exposure a (AOR, 1.81 [95% CI, 1.19-2.75]),losingearnings(AOR,2.37 [95% CI, 1.62-3.46]),orhavingfinancialdifficulties(AOR, 3.74 [95% CI, 2.48-5.63])due to the pandemic were 1.8 to 3.7 times greater for families headed by an adult with persistent COVID-19 symptoms than for families with no COVID-19 (Table4).Positive associations between previous severe COVID-19 and pandemic-related economicDiscussionThis cohort study contributes timely, policy-relevant information on how COVID-19 and PCCs may be associated with families' economic well-being.Using data from a long-running, nationally representative household panel survey with interviews conducted both before and during the COVID-19 pandemic, we found that persistent COVID-19 symptoms and, to a lesser extent, previous (resolved) severe COVID-19 were associated with increased odds of pandemic-related economic hardship among US families.In regression models adjusted for prepandemic sociodemographic characteristics and experiences of economic hardship, the odds of a resident family member being laid off or furloughed, losing earnings, or having financial difficulties were 2.0 to 3.7 times higher among families headed by an adult with persistent COVID-19 symptoms and 1.7 to 2.0 times higher among families headed by an adult with previous severe COVID-19 compared with families with no

Table 4 .
AORs of Family Economic Hardship by COVID-19 Exposure and Prepandemic Income a All adjusted models include controls for resident children, age of the reference person (RP), race and ethnicity of the RP, highest level of education completed by the RP or spouse or partner (SP), total family income relative to poverty thresholds, uninsurance, geographic region, and residence in a nonmetropolitan area.In addition, to account for potential differences in baseline economic hardship, each adjusted model includes a similar economic indicator collected in the Panel Study of Income Dynamics-2019 survey as a covariate: whether the RP or SP missed work because they were temporarily laid off, total family labor income, or whether anyone in the family had credit card or store card debt.-19Symptoms and Economic Hardship Among US Families history of COVID-19.We found no statistically significant association in the overall sample between moderate, mild, or asymptomatic COVID-19 and indicators of family economic hardship.Compared with similar families with no history of COVID-19, families headed by an adult with persistent COVID-19 symptoms had increased odds of experiencing economic hardship regardless of their prepandemic financial status.In contrast, an adult family member's severe COVID-19 illness was more strongly associated with indicators of economic hardship among families with lower income.
b c Sample includes 2222 families with total family income below 200% of the US Census Bureau poverty threshold.dSampleincludes 4710 families with total family income above 200% of the US Census Bureau poverty threshold.JAMA Network Open | Public HealthCOVID