Racial/Ethnic Differences in the Application and Receipt of Services to Address Social Needs: Impact of COVID-19.

Black and Hispanic Americans are disproportionately affected by COVID-19, which impacts their social needs. The objective of this study was to examine differences in white, Black, and Hispanic adults' application and receipt of services to address their social needs during COVID-19. Utilizing weeks 1, 2, and 3 from the COVID Impact Survey, descriptive statistics analyzed covariates and the 12 social services by participants' race/ethnicity. Unweighted frequencies and weighted percentages were computed for the services score by race/ethnicity. Forward stepwise binary logistic regression analyses examined the relationship between services needed and race/ethnicity. All analyses were conducted using STATA MP 14. Among 20 533 participants, unemployment insurance was a common service participants reported applying for or trying to apply for. Compared with white participants, Hispanic participants had higher adjusted odds of needing unemployment insurance services (adjusted odds ratio [AOR] = 1.58; 95% confidence interval [CI], 1.18-2.11). Black (AOR = 3.25; 95% CI, 2.49-4.25) and Hispanic (AOR = 1.55; 95% CI, 1.14-2.10) participants had higher adjusted odds of needing Supplemental Nutrition Assistance Program services than white participants. It is important for Black and Hispanic Americans to have access to these social services. Research and evaluation studies are warranted to inform policies that sustain/modify social services for future use. These studies must include a representative sample of Black and Hispanic Americans.

ALTHOUGH coronavirus disease 2019 (COVID-19) was first recognized in December 2019,1 it was not declared a global health emergency until January 20202 and a national emergency in the United States (US) until March 2020.3 COVID-19 has become a leading cause of death,4 claiming more than 574 679 lives in the US as of May 3, 2021.5 Some minorities are disproportionately affected by this condition. Compared with white Americans, Black, Hispanic, and American Indian or Alaska Native Americans have been identified to have a higher risk for COVID-19 infection, hospitalization, and mortality.6

A disproportionately higher risk of contracting COVID-19 in minorities is influenced by social determinants of health, such as race/ethnicity, neighborhood environment, and employment level.7 Research has consistently shown that predominantly Black7 and Hispanic8 communities are more densely populated and increased residents' risk of contracting COVID-19. In addition to their geographical residence, Black9 and Hispanic8 Americans are more likely to hold job positions such as frontline workers, increasing their COVID-19 risk.

In addition to managing their higher risk of contracting COVID-19, Black and Hispanic Americans may have higher social needs, such as financial10 and food insecurity,11,12 than their counterparts. The rapid and widespread transmission of this disease has resulted in state-level COVID-19 restrictions, such as stay-at-home orders.13 Although evidence suggests the restrictions decreased population movement, which aided in managing transmission, they also led to the closing of businesses and schools.10 Consequently, COVID-19 has drastically increased unemployment, with more than 40 million workers filing for unemployment in May 2020.14 Unfortunately, Black and Hispanic Americans' employment status was inordinately affected, with higher unemployment and a decrease in the number of hours worked by both groups.15

Numerous social safety net services, such as food pantries, Social Security, and Temporary Assistance for Needy Families, are available and may have reduced this racial/ethnic inequity. Moreover, national-level policies such as the Coronavirus Aid, Relief, and Economic Security Act extended benefits (unemployment insurance) that may aid in lessening Black and Hispanic Americans' social needs.14 However, there is limited understanding of these racial/ethnic groups' application and receipt of these services. Understanding their use of these services can advance the science on racial and ethnic health inequities experienced by families and communities. Moreover, it can inform policies to aid Black and Hispanic Americans in accessing and utilizing these services. It also can inform initiatives to diminish underlying racial/ethnic inequities in the social determinants of health. Therefore, the objective of this study was to examine differences in white, Black, and Hispanic adults' application and receipt of services to address their social needs during COVID-19.

METHODS

Study population

Data from the COVID Impact Survey were utilized to meet this study's objective. The COVID Impact Survey, conducted by NORC at the University of Chicago and funded by the Data Foundation, examined physical and mental health, economic status, and social systems of US adults 18 years and older.16 A 2-prong sampling approach was undertaken. The sample consisted of a nationally representative sample of adults (AmeriSpeak sample).16 The second sample incorporated adults from 18 regional areas, including 10 states (California, Colorado, Florida, Louisiana, Minnesota, Missouri, Montana, New York, Oregon, and Texas) and 8 metropolitan statistical areas (Atlanta, Baltimore, Birmingham, Chicago, Cleveland, Columbus, Phoenix, and Pittsburgh) (address-based sample).16 AmeriSpeak sample was selected from the AmeriSpeak Panel using a 48 sampling strata sampling.17 The strata were based on age, race/Hispanic ethnicity, education, and sex.17 The address-based sample utilized the US Postal Service delivery-sequence file, covering approximately 97% of US households.17

Interviews were conducted in English and Spanish, and participants received a monetary incentive for participating in the study.16 The survey was administered over 3 weeks (April 20-26, 2020; May 4-10, 2020; and May 30-June 8, 2020) to obtain national and regional estimates.16 Each of the 3 weeks is independent cross-sectional waves of data that were merged for this study. Additional details on the methodological approach of the COVID Impact Survey can be found on the Data Foundation's Web site.18 The pooled 3 weeks of data totaled 25 269 participants. Participants who self-identified as white, Black, or Hispanic were included in the study (n = 20 533). All other racial categories were excluded because of limited responses.

Measures

Dependent variable

The dependent variables of interest were the 12 forms of services participants received, applied for, tried to apply for, or did not receive/apply for in the past 7 days. The services included in the survey were unemployment insurance; Supplemental Nutrition Assistance Program (SNAP); Temporary Assistance for Needy Families; other aid from the government; assistance from a union/other association; service from a community organization; other types of assistance; Social Security; help from a church/religious organization; assistance from a food pantry; any kind of government health insurance or health coverage (including Medicaid Medical Assistance or Medicare); and Supplemental Social Security.

Each of the 12 service variables was recategorized because of limited responses to the needed service (need vs not needed service) variables. Specifically, 3 responses, “received,” “applied for,” or “tried to apply for,” were combined for each of the 12 services to create binary variables, with no services selected (did not receive/apply for) defined as “0” versus selected services (received, applied for, or tried to apply for) denoted as “1.” The services score variable was developed from these binary variables (need vs not needed variables). The services score variable was formed to understand the number of services participants needed. The 12 binary variables (need vs not needed variables) were summated to create the score variable. The services score variable ranged from 0, defined as zero services selected, to a score of 12, which indicated all 12 services were selected.

Independent variables

Participants' self-identified race/ethnicity (white, Black, or Hispanic) was the independent variable.

Covariates

Covariates included were age in years (18-29, 30-44, 45-59, and 60+), which was recategorized to 18-29, 30-44, 45+ years due to limited responses, and sex (male/female). Education level (less than high school, high school graduate, some college, associate, or bachelor, master's, or doctorate degree) was also incorporated but recategorized to address limited responses (high school graduate or less, some college, or bachelor degree or higher). Participants reported their ability to pay an unexpected $400 expense (yes/no). Household income was included but recategorized from $10 000-$19 999, $20 000-$29 999, $30 000-$39 999, $40 000-$49 999, $50 000-$74 999, $75 000-$99 999, $100 000-$149 999, and $150 000+ to $10 000-$19 999, $20 000-$29 999, $30 000-$49 999, and $50 000+ to address low responses.

Additional covariates included were have worked for pay in the past 7 days (yes, for someone else; yes, self-employed; or no), population density (rural, suburban, or urban), household size (1, 2, 3, 4, 5, or 6+ persons), and self-reported having health conditions (hypertension, cardiovascular disease, asthma, chronic lung disease/chronic obstructive pulmonary disease, bronchitis/emphysema, allergies, mental health, cystic fibrosis, liver disease/end-stage liver disease, cancer, compromised immune system, overweight/obesity), which was categorized into yes/no.

Statistical analysis

The data set was weighted according to the recommended weighting procedure to obtain national estimates.16 Chi-square tests were conducted to examine the association between race/ethnicity and participants' sociodemographic characteristics. Chi-square tests were also conducted to assess the relationship between participants' race/ethnicity and the 12 forms of services (received, applied for, try to apply for, or did not receive/apply). Unweighted frequencies and weighted percentages were computed for the services score by race/ethnicity. Forward stepwise binary logistic regression analyses examined the relationship between services needed variables and race/ethnicity. Covariates were included and adjusted for in the regression models. Sex and education level were included in the models because the literature shows both to be affected by COVID-19.12,15 Additional covariates included in the regression models had a P value of less than .05 in the univariate analysis. All models' fit was subsequently assessed using the Hosmer-Lemeshow goodness-of-fit test.19 For all statistical tests, a value of P < .05 was considered the significance level. All analyses were conducted using STATA MP 14.20

RESULTS

The pooled 3 weeks of data totaled 25 269 participants, with 20 533 participants meeting the inclusion criteria. Table 1 depicts the unweighted frequencies and weighted percentages of respondents' sociodemographic characteristics by race/ethnicity. More than half of the participants were female; among female participants, the largest percentage self-identified as Black (60.1%), followed by Hispanic (51.6%), and then white (51.5%). Education levels varied across the 3 racial/ethnic groups, with the highest proportion of those who self-identified as white reported having a bachelor degree or higher (36.0%). However, for Black and Hispanic respondents, a high school degree or less had the greatest percentage of responses, 43.3% and 48.8%, respectively.

TABLE 1.

Unweighted Frequencies and Weighted Percentage of Participants' Sociodemographic Characteristics, by Race/Ethnicity

	White (n = 15985), n (%)	Black (n = 2290), n (%)	Hispanic (n = 2258), n (%)	Total	P
Age, y					<.001
18-29	1 595 (17.4)	319 (17.2)	597 (32.9)	2 511 (20.20)
30-44	3 494 (23.4)	636 (26.2)	772 (28.4)	4 902 (24.70)
45+	10 896 (59.2)	1 335 (56.6)	889 (38.7)	13 120 (55.10)
Sex					.01
Male	7 216 (48.6)	739 (39.9)	920 (48.4)	8 875 (47.38)
Female	8 765 (51.6)	1 551 (60.1)	1 337 (51.6)	11 653 (52.62)
Education level					<.001
High school graduate or less	2 294 (35.6)	589 (43.3)	688 (48.8)	3 571 (39.06)
Some college	4 760 (28.4)	915 (28.3)	853 (28.3)	6 528 (28.38)
Bachelor's degree or higher	8 915 (35.1)	781 (28.4)	717 (22.9)	10 413 (32.56)
Income					<.001
$10 000-$19 999	946 (7.3)	295 (14.8)	247 (12.6)	1 488 (9.11)
$20 000-$29 999	1 288 (11.4)	304 (17.9)	355 (19.3)	1 947 (13.57)
$30 000-$49 999	2 534 (19.1)	460 (25.0)	457 (19.8)	3 451 (20.56)
$50 000+	10 295 (61.4)	849 (42.2)	992 (48.3)	12 136 (56.76)
Employment status					.22
Yes, I worked for salary	6 652 (41.5)	954 (40.2)	961 (42.4)	8 567 (41.49)
Yes, self-employed	1 284 (8.0)	97 (5.3)	149 (5.1)	1 530 (7.30)
No, I did not work	7 939 (50.5)	1 211 (54.6)	1 130 (51.7)	10 280 (51.22)
Population density					<.001
Rural	899 (11.7)	48 (5.9)	42 (3.2)	989 (9.36)
Suburban	2 747 (23.1)	162 (10.5)	187 (9.7)	3 096 (19.60)
Urban	12 334 (64.4)	2 080 (83.6)	2 029 (87.1)	16 443 (71.04)
Household size					<.001
1 person	4 978 (35.3)	844 (39.5)	469 (24.4)	6 291 (33.86)
2 persons	6 042 (27.6)	555 (19.9)	540 (19.4)	7 137 (25.10)
3 persons	2 130 (13.1)	361 (15.6)	383 (14.6)	2 874 (13.68)
4 persons	1 555 (9.6)	218 (8.2)	339 (14.1)	2 112 (10.25)
5 persons	675 (5.1)	144 (5.1)	235 (11.6)	1 054 (6.42)
≥6 persons	581 (9.3)	167 (10.1)	290 (15.9)	1 038 (10.69)
Reported having health conditions					<.001
No	3 403 (21.2)	493 (22.1)	707 (30.8)	4 603 (23.21)
Yes	12 582 (78.8)	1 797 (77.1)	1 551 (69.2)	15 930 (76.79)
Ability to pay unexpected $400 expense
No	14 543 (87.2)	1 563 (72.5)	1 778 (78.4)	17 884 (83.67)	<.001
Yes	1 442 (12.8)	727 (27.6)	480 (21.6)	2 649 (16.33)

More than 60% of white participants reported having an income of $50 000 or higher, whereas approximately 42% of Black participants and approximately 48% of Hispanic participants reported having an income of $50 000 or higher. It should also be noted that those who self-identified as Black had the greatest percentage of participants who reported not working (54.6%) than working for pay (40.2%) in the last 7 days, followed by those who self-identified as Hispanic (51.7% vs 42.4%), and then those who identified as white (50.5% vs 41.5%).

The assessment of the services score by race/ethnicity depicts more than 47% of all racial and ethnic groups needed zero services, with those who self-identified as white at 50% (Figure). Hispanic participants were more likely than white and Black participants to need 1 or 2 services; however, Black participants were more likely to need 3, 4, or 5 services.

Number of services participants selected by race/ethnicity.

Table 2 depicts the unweighted frequencies and weighted percentages of the 12 forms of social services by respondents' race/ethnicity. Social Security and any kind of government health insurance (eg, Medicaid) were 2 predominant services all racial/ethnic groups received. In comparison, unemployment insurance was the most common service participants reported applying for or trying to apply for. There were a substantial number of services participants reported they applied for. SNAP was one of these services, with 1.4% of white, 4.6% of Black, and 5.0% of Hispanic respondents reporting they applied for this service. The most common service participants tried to apply for varied by race and ethnicity. Per racial/ethnic group, unemployment insurance among white and Hispanic participants (2.7% vs 4.6) and SNAP among Black participants (4.5%) were the leading services for which the participants applied.

TABLE 2.

Services Participants Received, Applied for, Tried to Apply for, or Did not Receive/Apply, by Race/Ethnicity

	White, n (%)	Black, n (%)	Hispanic, n (%)	Total	P
Unemployment insurance					<.001
Received	773 (5.69)	131 (5.4)	145 (9.2)	1049 (6.30)
Applied for	625 (4.48)	166 (5.8)	153 (9.0)	944 (5.48)
Try to apply	287 (2.68)	91 (3.4)	106 (4.6)	484 (3.12)
Did not receive/apply	14 051 (87.15)	1 850 (85.4)	1 824 (77.2)	17 725 (85.10)
Supplemental Nutrition Assistance Program					<.001
Received	906 (8.7)	480 (21.3)	337 (12.8)	1723 (11.07)
Applied for	137 (1.4)	117 (4.6)	74 (4.1)	328 (2.48)
Try to apply	122 (1.2)	96 (4.5)	73 (3.3)	291 (2.01)
Did not receive/apply	14 536 (88.7)	1 555 (69.6)	1 737 (78.9)	17 828 (84.44)
Temporary Assistance for Needy Families					<.001
Received	54 (0.6)	39 (2.1)	23 (1.1)	116 (1.01)
Applied for	30 (0.5)	25 (1.5)	22 (1.6)	77 (0.87)
Try to apply	54 (0.5)	36 (2.0)	53 (2.0)	143 (0.99)
Did not receive/apply	15 518 (98.3)	2 125 (93.5)	2 108 (95.2)	19 751 (97.14)
Other aid from government					.04
Received	1 542 (9.1)	175 (7.3)	223 (9.5)	1 940 (8.97)
Applied for	262 (2.9)	56 (4.4)	78 (3.6)	396 (3.19)
Try to apply	162 (0.9)	40 (2.9)	56 (2.4)	258 (1.40)
Did not receive/apply	13 764 (87.1)	1 973 (85.8)	1 863 (84.5)	17 600 (86.44)
Union or other association					<.001
Received	170 (1.1)	30 (1.4)	27 (0.8)	227 (1.06)
Applied for	41 (0.7)	18 (1.9)	20 (0.8)	79 (0.85)
Try to apply	36 (0.2)	17 (1.1)	30 (1.1)	83 (0.64)
Did not receive/apply	15 489 (98.1)	2 177 (95.6)	2 144 (96.5)	19 810 (97.45)
Community organization					<.001
Received	133 (1.0)	66 (2.9)	64 (2.1)	263 (1.45)
Applied for	38 (0.3)	20 (1.5)	24 (1.3)	82 (0.61)
Try to apply	60 (0.7)	34 (2.3)	32 (1.6)	126 (1.04)
Did not receive/apply	15 629 (98.1)	2 133 (93.3)	2 114 (95.1)	19 876 (96.89)
Other assistance					.01
Received	287 (2.1)	64 (2.5)	84 (3.7)	435 (2.44)
Applied for	73 (0.8)	32 (2.2)	44 (1.2)	149 (1.06)
Try to apply	78 (0.5)	33 (0.9)	33 (1.7)	144 (0.78)
Did not receive/apply	15 413 (96.6)	2 120 (94.5)	2 067 (93.4)	19 600 (95.72)
Social Security					<.001
Received	4 304 (24.2)	483 (23.5)	302 (13.2)	5 089 (22.10)
Applied for	102 (0.9)	33 (1.3)	27 (1.5)	162 (1.08)
Try to apply	64 (0.7)	25 (0.7)	28 (1.6)	117 (0.89)
Did not receive/apply	11 332 (74.1)	1 704 (74.5)	1 864 (83.7)	14 900 (75.93)
Church or religious organization					<.001
Received	128 (1.2)	79 (3.5)	80 (2.5)	287 (1.76)
Applied for	34 (0.7)	15 (1.1)	26 (1.2)	75 (0.88)
Try to apply	38 (0.3)	29 (1.1)	33 (2.5)	100 (0.93)
Did not receive/apply	15 672 (97.7)	2 142 (93.2)	2 103 (93.8)	19 917 (96.43)
Food pantry					<.001
Received	526 (5.4)	303 (12.8)	239 (8.3)	1 068 (6.85)
Applied for	40 (0.8)	39 (1.4)	36 (1.5)	115 (1.01)
Try to apply	69 (0.4)	46 (2.4)	40 (2.2)	155 (0.96)
Did not receive/apply	15 218 (93.5)	1 868 (83.4)	1 915 (88.1)	19 001 (91.18)
Any government health insurance					<.001
Received	3 289 (21.3)	566 (25.1)	432 (18.5)	4 287 (21.39)
Applied for	117 (1.2)	49 (3.2)	54 (2.6)	220 (1.71)
Try to apply	77 (0.1)	39 (2.4)	39 (2.1)	155 (1.34)
Did not receive/apply	12 326 (76.6)	1 600 (68.4)	1 701 (76.9)	15 627 (75.56)
Supplemental Social Security					<.001
Received	440 (3.5)	192 (7.7)	106 (5.8)	738 (4.49)
Applied for	47 (0.8)	29 (1.5)	21 (1.3)	97 (0.99)
Try to apply	51 (0.3)	29 (1.4)	29 (2.5)	109 (0.85)
Did not receive/apply	15 214 (95.4)	2 005 (89.4)	2 067 (90.4)	19 286 (93.67)

The forward stepwise binary logistic regression models for each of the 12 social services can be found in Supplemental Digital Content Appendix, Tables A-C (available at: http://links.lww.com/FCH/A36, http://links.lww.com/FCH/A37, and http://links.lww.com/FCH/A38, respectively). Supplemental Digital Content Appendix, Table A (available at: http://links.lww.com/FCH/A36) depicts the models with no covariates included. Supplemental Digital Content Appendix, Table B (available at: http://links.lww.com/FCH/A37) shows the inclusion of age and sex. Supplemental Digital Content Appendix, Table C (available at: http://links.lww.com/FCH/A38) portrays the final model in which education, ability to pay $400 expense, income, work for pay, population density, reported chronic conditions, and household size were incorporated in the respective models as appropriate. A comparison of Supplemental Digital Content Appendix, Tables A-C (available at: http://links.lww.com/FCH/A36, http://links.lww.com/FCH/A37, and http://links.lww.com/FCH/A38, respectively) demonstrates substantial changes in the adjusted odds ratios (AORs) among Black and Hispanic Americans for the 12 social services.

Table 3 (available at: http://links.lww.com/FCH/A39) depicts services needed by racial and ethnic groups, with the models adjusted for covariates. Compared with participants who self-identified as white, those who self-identified as Hispanic had higher adjusted odds of needing unemployment insurance (AOR = 1.54; 95% confidence interval [CI], 1.15-2.06). Hispanic participants were also more likely to have needed other assistance than white participants (AOR = 1.70; 95% CI, 1.03-2.81). In contrast, participants who identified as Black were more likely to need assistance from a union or other association (AOR = 2.12; 95% CI, 1.05-4.27), service from a church or religious organization (AOR = 3.24; 95% CI, 1.87-5.62), and any kind of government health insurance (AOR = 1.38; 95% CI, 1.06-1.79) than white participants.

Nonetheless, there were similar services needed by both Black and Hispanic participants. Participants who identified as Black (AOR = 3.25; 95% CI, 2.49-4.25) or Hispanic (AOR = 1.55; 95% CI, 1.14-2.10) had higher adjusted odds of needing SNAP services than respondents who identified as white. Likewise, Black participants (AOR = 3.75; 95% CI, 2.08-6.77) had higher adjusted odds, and Hispanic participants (AOR = 2.02; 95% CI, 1.03-3.95) had higher adjusted odds of needing Temporary Assistance for Needy Families services than white participants. In comparison with participants who identified as white, Black and Hispanic participants had higher adjusted odds of needing assistance from a community organization (AOR = 3.21; 95% CI, 1.84-5.60; and AOR = 1.97; 95% CI, 1.07-3.64), food pantry assistance (AOR = 2.20; 95% CI, 1.54-3.14; and AOR = 1.54; 95% CI, 1.08-2.21), and Supplemental Social Security (AOR = 2.39; 95% CI, 1.61-3.54; and AOR = 2.19; 95% CI, 1.35-3.54).

It should also be noted that the adjusted odds of having the inability to pay an unexpected $400 expense was higher among participants who needed unemployment insurance (AOR = 1.60; 95% CI, 1.19-2.16), other government aid (AOR = 1.44; 95% CI, 1.07-1.95), other assistance (AOR = 1.62; 95% CI, 1.02-2.58), food pantry (AOR = 3.45; 95% CI, 2.56-4.64), and any government health insurance (AOR = 1.97; 95% CI, 1.56-2.49) than those who did not need these services.

DISCUSSION

The results of this study show that US adults have substantial social needs during the COVID-19 pandemic. Across racial/ethnic groups, Social Security was the most commonly needed service. Although no other studies have examined Social Security benefits during the pandemic, the use of this service may be associated with individuals' decrease in earnings due to the stay-at-home orders. The stay-at-home orders and older adults' severe risk of developing COVID-19 may have resulted in eligible participants, such as those who have sought retirement due to the pandemic, receiving Social Security services. Older and disabled individuals are vulnerable populations who may qualify for Social Security, and COVID-19 may have impacted their long-term plans (eg, delaying retirement). Therefore, further research is needed to better understand the impact of COVID-19 on these populations to ensure all their needs are met.

Among all participants, unemployment insurance was the most frequent service for which participants applied. The high number of applications for this service was expected because of the record number of US adults who lost their jobs or were furloughed.10,21 Projections of the unemployment rate for years 2021 and 2022 have been conducted, with unemployment levels continuing to be impacted by COVID-19 in 2022.22 With the known unpredictability of COVID-19 and COVID-19 vaccine uptake, it may be beneficial to continue to conduct research to ensure unemployment insurance extended benefits are not removed before the economy reaches full employment. The latest statistics indicate a decrease in unemployment in the US;23 however, the inequities among US adults must be considered in these analyses to prevent the exacerbation of health disparities.

Unemployment insurance was a service a substantial number of participants tried to apply for. The high number of attempts may be an indicator of difficulties in applying for this service, which has been reported.10 An earlier study stated more than 54% of US adults aged 18 to 64 years who applied for but did not receive unemployment insurance reported the process as difficult (21.6%) or very difficult (33.0%).10 This finding highlights the potential benefit of reforming the application process, such as modification to the software or assistance in completing the application, to reduce barriers to receiving this service. Moreover, streamlining the application process to allow eligible individuals to apply for more than 1 service at a time (eg, unemployment insurance and SNAP) may be beneficial. This benefit is portrayed in the results of this study, which showed more than 32% of participants reported receiving, applying, or trying to apply for more than 1 service. Applying for multiple services can be time-consuming and may have intensified participants' mental stress. Combining the application process for some services may increase participants' access to services and improve their behavioral health. Research should be conducted to determine the feasibility and effectiveness of combining multiple services.

The forward stepwise binary logistic regression outcomes in which covariates were controlled showed participants' service needs also varied by race and ethnicity. Hispanic respondents had significantly higher adjusted odds of needing unemployment insurance than their white counterparts. Research has suggested COVID-19 had a more severe impact on Hispanic workers.15,24 Their disproportionate needs for unemployment insurance may stem from underlying structural determinants such as education and income that can influence the type and number of job opportunities.15 This population may benefit from targeted community-based workforce training and employment assistance initiatives. The literature discusses the effectiveness of targeted interventions,25 and this strategy may aid Hispanic Americans in obtaining employment in a timely manner.

Needed services also differed for Black participants. Black participants had higher adjusted odds of needing any kind of government health insurance (eg, Medicaid) than their white counterparts. This was not surprising, as the literature is replete on the challenges faced by this group in accessing medical care and discussed health insurance as a prominent indicator.26–28 In addition, Black participants' service needs were associated with church/religious and community-based organizations. This association may stem from accessibility and/or the cultural aspects of the Black community. Research has previously discussed the critical role religion plays among Black Americans.29 Ultimately, church/religious and community-based organizations may be addressing structural factors to Black Americans' receipt of services. As a result, these organizations must receive the appropriate funding to ensure they are sustained to meet the needs of their community members.

SNAP and food pantry assistance were 2 services Black and Hispanic participants had higher odds of needing. These findings show both groups are experiencing food and financial insecurity and support studies on the subject matter.30,31 Their receipt of SNAP benefits aligns with a nationally representative study conducted by Morales and colleagues,12 which showed Black and Hispanic households were more likely to report not having the financial means to purchase food. A short-term objective may be to evaluate the impact of the modification to SNAP to expand services. This may assist in determining how long the modifications should continue and their implications on SNAP services postpandemic. In addition, research is needed to determine whether the extended benefits are effectively addressing food insecurity. Previous research suggests that some SNAP beneficiaries before COVID-19 remain food insecure.32 Thus, benefits obtained during the pandemic may not effectively meet enrollees' needs. A long-term objective may be to examine policies (eg, minimum wage) that can address structural determinants such as education and income level, which may influence Black and Hispanic Americans' use of this service.

Temporary Assistance for Needy Families was a service Black and Hispanic participants had higher odds of needing. Temporary Assistance for Needy Families addresses family and childhood hardship determinants by disseminating block grants to states, territories, and American Indian tribes to fund benefits and services (eg, economic assistance and work support).33 This finding was unique in the sense that little research has been conducted on this relationship. However, with the known increase in unemployment due to COVID-19 and the substantial increase in food and financial insecurity among Black and Hispanic Americans, the need for the Temporary Assistance for Needy Families services is anticipated. The Temporary Assistance for Needy Families modified its policies, such as its income thresholds or increased benefit amounts for kinship caregivers.34 The changes in policies varied by state, which suggests the need for additional research to understand their impact on residents' access and utilization of this service.

It is well known that Americans experience financial hardship.35,36 However, the results of this study are novel because it demonstrates the relationship between experiencing financial hardship and the need for social service assistance. It showed that participants who could not pay an unexpected $400 expense had higher adjusted odds of needing unemployment insurance, other government aid, other assistance, food pantry assistance, and any government health insurance. The need for these services has direct implications on participants' health outcomes, diet/nutrition, and access to health care services. These findings provide supporting evidence for the need for health and social policies to address structural barriers to the receipt of these services.

This study had several limitations and strengths. First, this study utilized a cross-sectional design, which prevents the determination of a causal relationship. Second, the survey questions do not distinguish between receiving services before COVID-19 and services they started receiving during COVID-19; therefore, the impact of COVID-19 on the need for these services may be under- or overestimated. Although the study consists of a nationally representative sample, participants are skewed toward urban than rural residents. This may have under- or overestimated services needed among the geographical groups. Finally, the responses were self-reported by participants and were subjected to recall bias. Nonetheless, this study consisted of 3 waves of cross-sectional data collected during 3 months (April, May, and June), which may give accurate insight into the need and receipt of services. This study also discusses the receipt, application, and attempt to apply for a broad range of services, which can inform policies on delivering these services.

CONCLUSION

The results of this study depict racial and ethnic differences in services needed (ie, received, applied for, or tried to apply for) during the COVID-19 pandemic. Several of these services were unemployment insurance, SNAP, food pantry, and Temporary Assistance for Needy Families. This racial and ethnic difference in needed services has implications on current national- and state-level policies that safeguard the availability of such services. The findings of this study can inform federal- and state-level policies to protect access to these services for Black and Hispanic Americans. Specifically, when and how long the social services are extended varies by state, which has health implications for states with a higher minority population.

These services address underlying social determinants such as education and income that impact health and mitigate outcomes of the economic crisis. In addition, health and financial crises, such as COVID-19 and economic downturns, are imminent. Thus, these social services should be sustained and modified, as needed, to protect Black and Hispanic Americans' access to these services for future use. Research and evaluation studies are warranted to understand the effectiveness of these services during the COVID-19 pandemic. The outcome of this study highlights the importance of incorporating a representative sample of Black and Hispanic Americans in the research and evaluation studies to inform the development of population-level interventions (eg, tailored employment assistance initiatives for Hispanic Americans) to have the greatest impact.

Further research is warranted to understand racial and ethnic groups' decision to apply or not to apply for the 12 social services. These findings may identify additional structural barriers and/or inadequacy in services as deterrents to applying for social services, and they can inform health policies that can improve the health of vulnerable populations.