Sage Journals: Discover world-class research

Abstract

Objectives:

Black, Indigenous, and People of Color have borne a disproportionate incidence of COVID-19 cases in the United States. However, few studies have documented the completeness of race and ethnicity reporting in national COVID-19 surveillance data. The objective of this study was to describe the completeness of race and ethnicity ascertainment in person-level data received by the Centers for Disease Control and Prevention (CDC) through national COVID-19 case surveillance.

Methods:

We compared COVID-19 cases with “complete” (ie, per Office of Management and Budget 1997 revised criteria) data on race and ethnicity from CDC person-level surveillance data with CDC-reported aggregate counts of COVID-19 from April 5, 2020, through December 1, 2021, in aggregate and by state.

Results:

National person-level COVID-19 case surveillance data received by CDC during the study period included 18 881 379 COVID-19 cases with complete ascertainment of race and ethnicity, representing 39.4% of all cases reported to CDC in aggregate (N = 47 898 497). Five states (Georgia, Hawaii, Nebraska, New Jersey, and West Virginia) did not report any COVID-19 person-level cases with multiple racial identities to CDC.

Conclusion:

Our findings highlight a high degree of missing data on race and ethnicity in national COVID-19 case surveillance, enhancing our understanding of current challenges in using these data to understand the impact of COVID-19 on Black, Indigenous, and People of Color. Streamlining surveillance processes to decrease reporting incidence and align reporting requirements with an Office of Management and Budget–compliant collection of data on race and ethnicity would improve the completeness of data on race and ethnicity for national COVID-19 case surveillance.

Keywords

data surveillance COVID-19 data quality race and ethnicity SARS-CoV-2

To date, several studies have examined the unequal incidence of disease caused by COVID-19 on Black, Indigenous, and People of Color,^1
-4 but little research has focused on the completeness of reporting on race and ethnicity in COVID-19 data.^5
-7 Several studies on racial and ethnic disparities in COVID-19 incidence and prevalence note that missingness of complete data on race and ethnicity occurs at both the state and national levels.^5
-9 There is a lack of attention in the literature to the tracking and completeness of case, hospitalization, and mortality data for Hispanic people and people of multiple races, as well as a lack of critical review of completeness of Centers for Disease Control and Prevention (CDC) data on race and ethnicity for cases, hospitalizations, and deaths.

In 1997, the Office of Management and Budget (OMB) revised its standards on race and ethnicity to include the 5 racial categories we use today—American Indian/Alaska Native (AI/AN), Asian, Native Hawaiian/Pacific Islander, Black, and White—with the 2 ethnic categories being Hispanic origin or non-Hispanic origin.¹⁰ OMB criteria also hold that data on race and ethnicity should be collected separately and that individuals should be allowed to select more than 1 race when self-identifying their racial identity or identities. For the remainder of this article, we refer to racial and ethnic data that include these 5 racial categories, collect data on race and ethnicity separately, allow for individuals to be represented in data with more than 1 racial identity, and collect ethnicity as “complete.”

While this OMB directive applies to federal agencies reporting on demographic characteristics, the transmission of state surveillance data on COVID-19 to CDC is voluntary. Furthermore, while the Coronavirus Aid, Relief, and Economic Security (CARES) Act Section 18115¹¹ requires data on race and ethnicity to be recorded by laboratories when they report on COVID-19 test results, there is no enforcement mechanism. As a result, broad variation exists by state as to which racial and ethnic categories are recorded and how they are reported.¹²

The impact of these variations on the recording and reporting of race and ethnicity directly decreases the ability of local and national governments to provide accurate surveillance for the most racially and ethnically diverse communities. Per 2020 postcensal estimates, 39.4% of AI/AN people identify as multiracial, and 37.6% of AI/AN people identify as being of Hispanic origin.¹³ The collection of only 1 racial identity or the collection of only race or only ethnicity in COVID-19 data likely disproportionately underestimates the incidence of COVID-19 in AI/AN communities relative to more homogenous demographic communities, such as White people (3.3% multiracial, 21.3% Hispanic).¹³ Without an appropriate understanding of the quality of data used in the surveillance of these communities, interventions or funding aimed at addressing the disproportionate impacts of COVID-19 may be misguided.

This study describes several elements of data quality on race and ethnicity in person-level, national COVID-19 case surveillance data. First, we wanted to understand the proportion of all COVID-19 cases that were sent to the person-level COVID-19 case surveillance dataset with complete information on race and ethnicity. Second, we examined the percentage of COVID-19 cases with complete data on race and ethnicity compared with the CDC aggregate COVID-19 case counts to better understand the generalizability of the person-level COVID-19 dataset to all known COVID-19 cases.¹⁴ CDC aggregate COVID-19 case counts are collected from jurisdictions and provide the most up-to-date numbers on cases and deaths but do not contain specific information on demographic characteristics of individual cases, which we contrast in this study with case numbers from person-level case surveillance with detailed demographic information. Third, we examined the proportion of all COVID-19 cases that were identified as having multiple racial identities as compared with a state’s underlying percentage of its population that identified as having multiple racial identities. Finally, we examined state-by-state discrepancies in the collection of data on race and ethnicity.

Methods

For this analysis, we assessed case surveillance data from CDC’s national person-level COVID-19 case surveillance data from April 5, 2020, through December 1, 2021, from all 50 states. Jurisdictions voluntarily submit deidentified, standardized information electronically for individual COVID-19 cases to CDC. This information contains fields on age, race and ethnicity, sex, underlying comorbidities, and health outcomes.¹⁵ Data on race and ethnicity could be collected in a method compliant with the 1997 OMB revised guidelines with the dataset including 1 field for each of the 5 OMB racial groups (AI/AN, Asian, Black, Native Hawaiian/Pacific Islander, and White), as well as a field for other race, other specified race, and race unknown. Individuals could be identified with 1 or more racial identities. Data on race were kept separate from data on ethnicity. Ethnicity was captured as either Hispanic/Latino or non-Hispanic/Latino. More information on the standardization and collection of these data is available elsewhere.¹⁶ We compared these data with CDC national aggregate COVID-19 case surveillance data from April 5, 2020, through December 1, 2021, from all 50 states.

Both datasets obtained from CDC by the Urban Indian Health Institute were downloaded initially via the DCIPHER (Data Collation and Integration for Public Health Emergency Response) outbreak response database and subsequently from the CDC protect outbreak response database. Analysis was restricted to confirmed COVID-19 cases with an identified county of residence and complete CDC case notification that indicates the first day CDC was notified of the case. Analysis was restricted to after April 5, 2020, to reflect the period when the Council of State and Territorial Epidemiologists published the first interim case surveillance definition, making COVID-19 a nationally notifiable condition.¹⁷

We analyzed the completeness of the collection of data on racial and ethnic identity via several steps. First, we determined the percentage of COVID-19 cases in the CDC database with complete ascertainment of race and ethnicity. For the purposes of this article, complete ascertainment of race and ethnicity is met only when a COVID-19 case was assigned to 1 or more of the 5 OMB race categories or had any response included in the “other specified race” field and ethnicity was identified as either non-Hispanic/Latino or Hispanic/Latino. Second, we used a Pearson correlation coefficient to determine whether there was a significant association in the number of person-level COVID-19 cases reported to CDC with complete race and ethnicity and time since April 2020 by day when divided by the number of aggregate COVID-19 cases on that same day. Third, we divided the number of COVID-19 cases with complete race and ethnicity in the CDC person-level database by the number of COVID-19 cases reported by aggregate counts of COVID-19 cases. Fourth, to examine the representation of individuals with multiple racial identities, we examined the percentage of COVID-19 cases with at least 1 identified racial identity and that indicated 2 or more racial identities. We then compared this number with the percentage of individuals in a state identified as having 2 or more racial identities per the 2020 postcensal estimates. To identify states whose proportions of multiracial individuals in COVID-19 case surveillance differed from state population estimates, we used an exact binomial test with a Bonferroni adjustment.¹⁸ Fifth, to examine the collection of both race and ethnicity, we examined the percentage of individuals who had both race and ethnicity assessed among those with either race or ethnicity assessed. We performed these analyses for the aggregate of all 50 states and individually by state. Findings were considered significant at α = .05.

We used R version 4.0.2 (R Foundation for Statistical Computing) to conduct all analyses. We conducted analyses as part of public health surveillance activities, and the Seattle Indian Health Board determined this project to be exempt from human subjects approval.

Results

As of December 1, 2021, a total of 35 579 801 laboratory-confirmed COVID-19 cases were in CDC’s COVID-19 person-level case surveillance dataset, of which 33 815 355 (95.0%) had complete information on county of residence and CDC notification date.

Of these 33 815 355 COVID-19 cases, 22 347 629 (66.1%) had complete ascertainment of race, 23 562 878 (69.7%) had complete ascertainment of ethnicity, and 18 881 379 (55.8%) had complete ascertainment of both race and ethnicity (Table 1). On the same date, there were a total of 47 898 497 COVID-19 cases for all 50 states per aggregate counts. Therefore, 70.6% of COVID-19 cases are included in the CDC person-level case surveillance dataset. Proportional complete ascertainment of both race and ethnicity in CDC case surveillance relative to the aggregate number of COVID-19 cases was lowest in periods when the number of COVID-19 cases increased (eFigure 1 in Supplemental Material). We observed a significant negative correlation between the percentage of aggregate COVID-19 cases reported to the person-level case surveillance dataset with complete race and ethnicity information and days since April 5, 2020 (R = −0.18, P < .001) (Figure 1). We observed a significant positive correlation between the percentage of person-level COVID-19 cases with complete race and ethnicity information and days since April 5, 2020 (R = 0.71, P < .001) (eFigure 2 in Supplemental Material).

Table 1.

Person-level COVID-19 case surveillance data, aggregate national COVID-19 case counts, and number of person-level COVID-19 cases with complete data on race and ethnicity,^a overall and by state, United States, April 5, 2020–December 1, 2021

State	Person-level surveillance cases	No. of CDC person-level COVID-19 cases with complete^a race and ethnicity	CDC aggregate COVID-19 counts, no.	Percentage of person-level cases with complete^a race and ethnicity	Cases in person-level surveillance divided by CDC aggregate COVID-19 counts, %^b	Cases in person-level surveillance with complete^a race and ethnicity divided by CDC aggregate COVID-19 counts, %^b
United States	33 813 254	18 881 379	47 898 497	55.8	70.6	39.4
Alabama	387 638	212 467	843 866	54.8	45.9	25.2
Alaska	122 765	53 351	152 716	43.5	80.4	34.9
Arizona	1 163 595	720 891	1 267 511	62.0	91.8	56.9
Arkansas	372 121	304 672	529 739	81.9	70.2	57.5
California	4 817 603	1 970 689	5 049 589	40.9	95.4	39.0
Colorado	755 129	341 415	836 731	45.2	90.2	40.8
Connecticut	343 499	95 690	416 017	27.9	82.6	23.0
Delaware	130 588	35 100	150 911	26.9	86.5	23.3
Florida	2 634 281	1 883 536	3 720 972	71.5	70.8	50.6
Georgia	923 663	397 224	1 657 813	43.0	55.7	24.0
Hawaii	62 058	26 767	87 293	43.1	71.1	30.7
Idaho	243 097	124 534	305 770	51.2	79.5	40.7
Illinois	1 624 830	815 095	1 798 617	50.2	90.3	45.3
Indiana	862 333	478 971	1 096 769	55.5	78.6	43.7
Iowa	402 910	230 551	517 988	57.2	77.8	44.5
Kansas	367 834	273 457	465 367	74.3	79.0	58.8
Kentucky	336 112	250 227	791 632	74.4	42.5	31.6
Louisiana	11 660	3085	757 868	26.5	1.5	0.4
Maine	83 711	61 831	120 130	73.9	69.7	51.5
Maryland	576 838	306 715	582 526	53.2	99.0	52.7
Massachusetts	849 742	451 868	902 925	53.2	94.1	50.0
Michigan	885 246	643 867	1 458 786	72.7	60.7	44.1
Minnesota	834 406	605 938	911 437	72.6	91.5	66.5
Mississippi	44 607	32 683	512 533	73.3	8.7	6.4
Missouri	464 355	313 710	911 300	67.6	51.0	34.4
Montana	158 939	101 556	190 727	63.9	83.3	53.2
Nebraska	149 748	87 659	314 381	58.5	47.6	27.9
Nevada	426 524	227 559	455 500	53.4	93.6	50.0
New Hampshire	75 110	41 152	160 623	54.8	46.8	25.6
New Jersey	1 020 730	472 454	1 215 200	46.3	84.0	38.9
New Mexico	308 100	182 872	313 640	59.4	98.2	58.3
New York	2 305 342	909 867	2 596 895	39.5	88.8	35.0
North Carolina	1 296 709	738 984	1 528 247	57.0	84.8	48.4
North Dakota	136 278	22	162 087	0	84.1	0
Ohio	1 347 917	890 844	1 686 170	66.1	79.9	52.8
Oklahoma	497 559	339 069	665 672	68.1	74.7	50.9
Oregon	381 087	193 260	390 031	50.7	97.7	49.5
Pennsylvania	1 434 641	830 423	1 726 370	57.9	83.1	48.1
Rhode Island	125 546	48 158	182 946	38.4	68.6	26.3
South Carolina	743 843	432 021	916 270	58.1	81.2	47.1
South Dakota	163 305	121 852	165 967	74.6	98.4	73.4
Tennessee	994 910	675 313	1 314 225	67.9	75.7	51.4
Texas	95 451	51 595	4 304 156	54.1	2.2	1.2
Utah	570 976	477 533	594 167	83.6	96.1	80.4
Vermont	46 381	40 437	50 015	87.2	92.7	80.8
Virginia	704 500	437 036	966 484	62.0	72.9	45.2
Washington	677 512	273 289	768 097	40.3	88.2	35.6
West Virginia	7623	6397	294 478	83.9	2.6	2.2
Wisconsin	755 510	611 248	978 460	80.9	77.2	62.5
Wyoming	88 392	56 445	110 883	63.9	79.7	50.9

Abbreviations: CDC, Centers for Disease Control and Prevention; HHS, US Department of Health and Human Services.

“Complete” race and ethnicity refers to a person-level COVID-19 case in which 1 of 5 Office of Management and Budget 1997 fields for race¹⁰ (American Indian/Alaska Native, Asian, Black, Native Hawaiian/Pacific Islander, White) was listed as yes or another race was specified, and ethnicity was collected.

Centers for Disease Control and Prevention.¹⁴

Figure 1.

Rolling 14-day average percentage volume of COVID-19 cases as reported by aggregate counts to the Centers for Disease Control and Prevention (CDC) and COVID-19 cases reported by person-level national COVID-19 case surveillance with complete race and ethnicity information, 50 states, April 5, 2020–December 1, 2021. Solid points indicate daily percentage of aggregate COVID-19 cases reported in the person-level CDC dataset with complete race and ethnicity information. Dashed line indicates the least-squares best-fit line between date of case reported to CDC and percentage of aggregate COVID-19 cases reported in the person-level CDC dataset with complete race and ethnicity information. Shaded regions indicate 95% CIs. “Complete” race and ethnicity refers to a person-level COVID-19 case in which 1 of 5 Office of Management and Budget 1997 fields for race¹⁰ (American Indian/Alaska Native, Asian, Black, Native Hawaiian/Pacific Islander, White) was listed as yes or another race was specified and ethnicity was collected. Data source: Centers for Disease Control and Prevention.¹⁴

The states with the highest and lowest percentage of COVID-19 cases with complete race and ethnicity information relative to the aggregate count of COVID-19 cases were Vermont (n = 40 437/50 015; 80.8%) and North Dakota (22/162 087, 0%), respectively (Table 1). Eight states (Connecticut, Delaware, Georgia, Louisiana, Mississippi, North Dakota, Texas, and West Virginia) had <25% of aggregate COVID-19 cases sent to CDC with complete race and ethnicity information (Figure 2).

Figure 2.

Percentage of COVID-19 cases as reported by aggregate counts to the Centers for Disease Control and Prevention and COVID-19 cases reported by person-level national COVID-19 case surveillance with complete information on race and ethnicity, 50 states, April 5, 2020–December 1, 2021. “Complete” race and ethnicity refers to a person-level COVID-19 case in which 1 of 5 Office of Management and Budget 1997 fields for race¹⁰ (American Indian/Alaska Native, Asian, Black, Native Hawaiian/Pacific Islander, White) was listed as yes or another race was specified and ethnicity was collected. Data source: Centers for Disease Control and Prevention.¹⁴

Of a total of 22 347 629 COVID-19 cases that ascertained at least 1 racial identity, 222 260 (1.0%) had 2 or more racial identities. When stratified by state, 5 states (Georgia, Hawaii, Nebraska, New Jersey, and West Virginia) had no COVID-19 cases that were identified as having 2 or more racial identities (Figure 3). The greatest percentage discrepancy between the percentage of COVID-19 cases identified as having 2 or more racial identities per state and the postcensal estimates of COVID-19 cases with 2 or more racial identities per state was in Hawaii (0% vs 24.6%).

Figure 3.

Percentage of individuals listed as having 2 or more racial identities in the Centers for Disease Control and Prevention’s national COVID-19 case surveillance system versus state population data, 50 states, April 5, 2020–December 1, 2021. Hawaii’s 2020 postcensal estimates of its population that identified as multiracial was 24.6%.¹³ Open circles indicate the percentage of person-level COVID-19 cases with at least 1 racial identity that have 2 or more racial identities. Closed circles indicate the 2020 postcensal estimates of a given state’s percentage population that identity themselves as having 2 or more racial identities. Horizontal lines indicate the difference between the state’s percentage of person-level COVID-19 cases with 2 or more racial identities and the state’s percentage population that identify themselves as having 2 or more racial identities. Error bars indicate 95% CIs for the estimate of the proportion of states’ COVID-19 cases with 1 or more racial identities identified that have 2 or more racial identities using exact binomial test with an associated Bonferroni adjustment for 50 tests. Data sources: US Census Bureau,¹³ Centers for Disease Control and Prevention.¹⁴

Among 33 815 355 COVID-19 cases with complete information on county of residence and CDC notification, 27 029 128 (79.9%) had either race or ethnicity information collected, and 18 881 379 (55.8%) had both race and ethnicity information collected. Of the 27 029 128 cases with either race or ethnicity information collected, 3 466 250 (12.8%) had only race information collected and 4 681 499 (17.3%) had only ethnicity information collected. When stratified by state, among cases for which race or ethnicity was ascertained, states differed in the percentage of cases with complete information on race, ethnicity, or both (Table 2).

Table 2.

Percentage of data by missingness type among COVID-19 cases with data on either race or ethnicity, 50 states, April 5, 2020–December 1, 2021

State	No. of COVID-19 cases with data on race or ethnicity in CDC person-level national COVID-19 case surveillance^a	No. (%) lacking data on complete^b race	No. (%) lacking data on ethnicity	No. (%) with complete^b data on race and ethnicity
United States	27 029 128	4 681 499 (17.3)	3 466 250 (12.8)	18 881 379 (69.9)
Alabama	282 505	14 828 (5.2)	55 210 (19.5)	212 467 (75.2)
Alaska	82 491	4018 (4.9)	25 122 (30.5)	53 351 (64.7)
Arizona	966 618	174 598 (18.1)	71 129 (7.4)	720 891 (74.6)
Arkansas	342 724	16 076 (4.7)	21 976 (6.4)	304 672 (88.9)
California	3 792 249	1 630 666 (43.0)	190 894 (5.0)	1 970 689 (52.0)
Colorado	560 474	200 056 (35.7)	19 003 (3.4)	341 415 (60.9)
Connecticut	179 143	30 962 (17.3)	52 491 (29.3)	95 690 (53.4)
Delaware	105 266	68 166 (64.8)	2000 (1.9)	35 100 (33.3)
Florida	2 366 994	384 368 (16.2)	99 090 (4.2)	1 883 536 (79.6)
Georgia	550 145	68 914 (12.5)	84 007 (15.3)	397 224 (72.2)
Hawaii	51 406	19 336 (37.6)	5303 (10.3)	26 767 (52.1)
Idaho	167 618	13 235 (7.9)	29 849 (17.8)	124 534 (74.3)
Illinois	1 218 408	214 350 (17.6)	188 963 (15.5)	815 095 (66.9)
Indiana	726 408	41 393 (5.7)	206 044 (28.4)	478 971 (65.9)
Iowa	274 447	14 486 (5.3)	29 410 (10.7)	230 551 (84.0)
Kansas	327 843	20 828 (6.4)	33 558 (10.2)	273 457 (83.4)
Kentucky	294 353	16 419 (5.6)	27 707 (9.4)	250 227 (85.0)
Louisiana	7274	109 (1.5)	4080 (56.1)	3085 (42.4)
Maine	75 619	1352 (1.8)	12 436 (16.4)	61 831 (81.8)
Maryland	436 150	52 490 (12.0)	76 945 (17.6)	306 715 (70.3)
Massachusetts	737 383	216 016 (29.3)	69 499 (9.4)	451 868 (61.3)
Michigan	741 312	51 886 (7.0)	45 559 (6.1)	643 867 (86.9)
Minnesota	771 279	29 756 (3.9)	135 585 (17.6)	605 938 (78.6)
Mississippi	39 215	1631 (4.2)	4901 (12.5)	32 683 (83.3)
Missouri	386 907	27 971 (7.2)	45 226 (11.7)	313 710 (81.1)
Montana	128 361	6251 (4.9)	20 554 (16.0)	101 556 (79.1)
Nebraska	116 875	12 051 (10.3)	17 165 (14.7)	87 659 (75.0)
Nevada	342 968	79 630 (23.2)	35 779 (10.4)	227 559 (66.3)
New Hampshire	52 619	5410 (10.3)	6057 (11.5)	41 152 (78.2)
New Jersey	752 233	144 709 (19.2)	135 070 (18.0)	472 454 (62.8)
New Mexico	258 678	61 504 (23.8)	14 302 (5.5)	182 872 (70.7)
New York	1 552 026	297 532 (19.2)	344 627 (22.2)	909 867 (58.6)
North Carolina	1 058 127	147 573 (13.9)	171 570 (16.2)	738 984 (69.8)
North Dakota	99 292	Suppressed^c	99 269 (<100.0)	22 (0)
Ohio	1 105 173	59 675 (5.4)	154 654 (14.0)	890 844 (80.6)
Oklahoma	408 194	18 569 (4.5)	50 556 (12.4)	339 069 (83.1)
Oregon	299 255	61 331 (20.5)	44 664 (14.9)	193 260 (64.6)
Pennsylvania	1 310 743	52 943 (4.0)	427 377 (32.6)	830 423 (63.4)
Rhode Island	95 928	38 559 (40.2)	9211 (9.6)	48 158 (50.2)
South Carolina	572 321	61 760 (10.8)	78 540 (13.7)	432 021 (75.5)
South Dakota	152 375	4527 (3.0)	25 996 (17.1)	121 852 (80.0)
Tennessee	840 573	66 303 (7.9)	98 957 (11.8)	675 313 (80.3)
Texas	79 471	13 257 (16.7)	14 619 (18.4)	51 595 (64.9)
Utah	535 020	43 959 (8.2)	13 528 (2.5)	477 533 (89.3)
Vermont	44 318	956 (2.2)	2925 (6.6)	40 437 (91.2)
Virginia	580 715	67 845 (11.7)	75 834 (13.1)	437 036 (75.3)
Washington	369 461	74 596 (20.2)	21 576 (5.8)	273 289 (74.0)
West Virginia	7191	391 (5.4)	403 (5.6)	6397 (89.0)
Wisconsin	712 926	40 602 (5.7)	61 076 (8.6)	611 248 (85.7)
Wyoming	70 054	7655 (10.9)	5954 (8.5)	56 445 (80.6)

Abbreviation: CDC, Centers for Disease Control and Prevention.

Data source: Centers for Disease Control and Prevention.¹⁴

Complete race refers to a person-level COVID-19 case in which 1 of the 5 Office of Management and Budget fields for race¹⁰ (American Indian/Alaska Native, Asian, Black, Native Hawaiian/Pacific Islander, White) was listed as yes or another race was specified. Complete ethnicity refers to COVID-19 cases in which ethnicity was listed as Hispanic/Latino or non-Hispanic/Latino.

Data were suppressed when the total number of individuals was <10.

Discussion

Our findings provide considerations for using national COVID-19 case surveillance information. First, 33 states had <50% of all COVID-19 cases reported in aggregate sent to CDC’s person-level COVID-19 case surveillance database with complete race and ethnicity information. Second, many states either are not collecting information on multiple racial identities or are collecting this information only rarely. Third, many individuals had information on only race or ethnicity captured.

Our findings indicate that the total number of COVID-19 cases with complete data on race and ethnicity in COVID-19 case surveillance is <50% of the volume of COVID-19 cases reported in aggregate. To our knowledge, this analysis is the first to publish this information. Previous research focused separately on the percentage of COVID-19 cases sent to CDC and the percentage of COVID-19 cases in person-level surveillance with complete racial and ethnic information, while our study focused on the percentage of COVID-19 cases counted in aggregate that were sent to CDC with complete racial and ethnic information.^4,16 We present the percentage of COVID-19 cases reported for each state with complete race and ethnicity information by comparing person-level surveillance data reported by CDC with CDC aggregate counts of cases. We found significant differences by state, with many states sending no information on race or ethnicity to CDC. This finding is important because COVID-19 datasets with race and ethnicity information aggregated by county of origin are now publicly available.¹⁹ Understanding which states have low overall availability of data on race and ethnicity provides insight into the generalizability of studies examining race and ethnicity using these data.²⁰

According to the 2020 postcensal estimates, individuals with multiple racial identities compose 10.2% of the US population and at least 1.4% of every state’s population.¹³ However, our study found that 49 states have significantly less than expected representation of individuals with multiple racial identities in COVID-19 data. Furthermore, 9 states had <10 individuals with multiple racial identities, and 5 states had no individuals with multiple racial identities. These findings indicate that several states are not able to store or transmit case surveillance data on multiple racial identities and suggest that most states are struggling to collect, store, or transmit case surveillance data on multiple racial identities. Our findings indicate that individuals with multiple racial identities are currently poorly represented in national COVID-19 case surveillance data. As has been observed in other studies, individuals with multiple racial identities have unique health risks and experiences compared with single-race individuals, such as lack of access to health care, increased risk of respiratory diseases, and increased risk of depression.^21,22 As some states have struggled to collect even a single racial identity for COVID-19 cases, the capacity to assess and collect multiple racial identities poses clear logistical issues. However, without the collection of full and complete data on racial identities for COVID-19 cases, our understanding of the disease burden of COVID-19 on certain racial and ethnic communities (eg, AI/AN people, who are 39.4% multiracial) is incomplete.¹³

Our study identified substantial differences in the collection of data on race and ethnicity by states, which undermine efforts to understand the burden of disease for all populations.²³ In cases in which data on race or ethnicity were obtained, states differed in the proportions of cases for which they recorded race or ethnicity. Prior research found that the collection of data on race and ethnic identity can result in the omission of race among those who identify as Hispanic/Latino, many of whom consider their Hispanic identity to be a part of their racial identity.²³ Furthermore, compared with non-Hispanic individuals, Hispanic individuals may have higher levels of discrepancies in how they see their racial identity and how their race is recorded by health care providers.²³ This is especially relevant to the AI/AN population, among whom 37.5% identify as Hispanic/Latino. As a result, AI/AN individuals may be more likely than non-Hispanic White individuals to have their race and ethnicity inaccurately recorded. The inconsistent collection of data on both race and ethnicity highlights the need for self-reporting of race and standardization in organizations on the methods used to capture data on race and ethnicity.

Limitations

Our analysis had several limitations. First, our analysis of person-level COVID-19 surveillance data includes only laboratory-confirmed COVID-19 cases, as all states report on this category at a minimum. As such, estimates may overestimate missingness for states that report aggregate numbers for both confirmed and probable COVID-19 cases. Notably, data obtained by CDC are not one-to-one copies of data collected by individual states. Interoperability issues limit the abilities of states to send their information to CDC. Third, the cases in the aggregate COVID-19 dataset do not represent all cases in an area because of COVID-19 cases not reported to the state or transmitted in aggregate. Fourth, lags from when cases are reported to CDC in aggregate versus when cases are transmitted to the person-level dataset might cause poor correlation between the 2 datasets. Fifth, the numbers in this study from state aggregate COVID-19 counts from CDC and from person-level COVID-19 case surveillance are aggregated at the state level. Therefore, there may be additional nuances of reporting practices at the county level that were not analyzed in our study. Finally, COVID-19 infection resulting in severe complications or death may make it impossible to obtain self-reported data on race and ethnicity, which may have resulted in less reliable assessments of these factors.

Conclusions

While state-specific insights into why variation in racial and ethnic data quality among states is occurring are generally unavailable, a report published by the Council of State and Territorial Epidemiologists identified several factors limiting the transmission of race and ethnicity data to CDC: information system limitations at the point of collection and public health agency, patient hesitation to indicate race or ethnicity, limited resources or staffing, and state laws or agency policies that limit or prohibit further sharing of race and ethnicity data for individuals with COVID-19.²⁴ Data modernization initiatives such as data storage and interoperability of systems offer promising results for streamlining the processes that are bottlenecking state capacity to fully and accurately document the race and ethnicity of individuals with COVID-19.^25,26 Data modernization initiatives may especially benefit health departments in case surges when automated, standardized, and streamlined systems would be able to handle sudden increases in workload that might overwhelm the less flexible capacity of public health workers.²⁷ However, data modernization initiatives should be coupled with appropriate incentives for state and local health departments to reward adherence to already existing laws mandating the collection of data on race and ethnicity in COVID-19 case surveillance.¹⁹

Our findings highlight a high degree of missingness of data on race and ethnicity in the US case surveillance system, which can enhance our understanding of current challenges of using these data to understand COVID-19’s impact on Black, Indigenous, and People of Color. First, our study identified substantial variation in the rate at which states are providing case surveillance with complete racial information to CDC. Second, our study highlights states that have no representation of individuals with multiple racial identities because of the methods by which these states collect data on racial identity. Finally, our study highlights ongoing issues in the simultaneous collection of data on race and ethnicity. These elements weaken effective surveillance of COVID-19 outcomes among Black, Indigenous, and People of Color, especially among multiracial and Hispanic communities. Without access to high-quality data on race and ethnicity, local, state, and federal health departments are unable to adequately allocate scarce public resources and address health disparities. These shortcomings in completeness of data on race and ethnicity require investments into public health infrastructure, including data modernization initiatives, adequate funding of public health staff, and education on the importance of collecting race and ethnicity data to build public health systems that provide a clear picture of disease burden in all communities.

Supplemental Material

sj-docx-1-phr-10.1177_00333549231154577 – Supplemental material for Completeness of Race and Ethnicity Reporting in Person-Level COVID-19 Surveillance Data, 50 States, April 2020–December 2021

Supplemental material, sj-docx-1-phr-10.1177_00333549231154577 for Completeness of Race and Ethnicity Reporting in Person-Level COVID-19 Surveillance Data, 50 States, April 2020–December 2021 by Scott Erickson, Rachael Bokota, Christine Doroshenko, Kate Lewandowski, Kojo Osei, Kaeli Flannery and Adrian Dominguez in Public Health Reports

Footnotes

Acknowledgements

The authors acknowledge Shannen Keene, MS, for assisting in editing and proofreading this article.

Authors’ Note

Coauthor Adrian Dominguez, MPH, passed away January 14, 2023. We dedicate this work to his memory as an epidemiologist, mentor, and champion for public health.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Research was supported by the US Department of Health and Human Services, Indian Health Service, Epidemiology Program for American Indian/Alaska Native Tribes and Urban Indian Communities, grant number U1B1IHS0006-21-00.

ORCID iDs

Scott Erickson, MPH

Christine Doroshenko, MPH

Supplemental Material

Supplementary figures for this article can be found at . The authors have provided these supplemental materials to give readers additional information about their work. These materials have not been edited or formatted by Public Health Reports’s scientific editors and, thus, may not conform to the guidelines of the AMA Manual of Style, 11th Edition.

References

Price-Haywood

Burton

Fort

Seoane

. Hospitalization and mortality among Black patients and White patients with COVID-19. N Engl J Med. 2020;382(26):2534-2543. doi:10.1056/NEJMsa2011686

Moore

Ricaldi

Rose

, et al. Disparities in incidence of COVID-19 among underrepresented racial/ethnic groups in counties identified as hotspots during June 5-18, 2020—22 states, February–June 2020. MMWR Morb Mortal Wkly Rep. 2020;69(33):1122-1126. doi:10.15585/mmwr.mm6933e1

Stokes

Zambrano

Anderson

, et al. Coronavirus disease 2019 case surveillance—United States, January 22–May 30, 2020. MMWR Morb Mortal Wkly Rep. 2020;69(24):759-765. doi:10.15585/mmwr.mm6924e2

Hatcher

Agnew-Brune

Anderson

, et al. COVID-19 among American Indian and Alaska Native persons—23 states, January 31–July 3, 2020. MMWR Morb Mortal Wkly Rep. 2020;69(34):1166-1169. doi:10.15585/mmwr.mm6934e1

Karaca-Mandic

Georgiou

Sen

. Assessment of COVID-19 hospitalizations by race/ethnicity in 12 states. JAMA Intern Med. 2021;181(1):131-134. doi:10.1001/jamainternmed.2020.3857

Tirupathi

Muradova

Shekhar

Salim

Al-Tawfiq

Palabindala

. COVID-19 disparity among racial and ethnic minorities in the US: a cross sectional analysis. Travel Med Infect Dis. 2020;38:101904. doi:10.1016/j.tmaid.2020.101904

Gross

Essien

Pasha

Gross

Wang

Nunez-Smith

. Racial and ethnic disparities in population-level COVID-19 mortality. J Gen Intern Med. 2020;35(10):3097-3099. doi:10.1007/s11606-020-06081-w

Mahajan

Larkins-Pettigrew

. Racial demographics and COVID-19 confirmed cases and deaths: a correlational analysis of 2886 US counties. J Public Health (Oxf). 2020;42(3):445-447. doi:10.1093/pubmed/fdaa070

Centers for Disease Control and Prevention. Office of Management and Budget directive no. 15: race and ethnic standards for federal statistics and administrative reporting. As adopted on May 12, 1977. Accessed November 2, 2021. https://wonder.cdc.gov/wonder/help/populations/bridged-race/directive15.html

10.

Office of Management and Budget. Revisions to the standards for the classification of federal data on race and ethnicity. October 30, 1997. Accessed November 24, 2021. https://www.federalregister.gov/documents/1997/10/30/97-28653/revisions-to-the-standards-for-the-classification-of-federal-data-on-race-and-ethnicity

11.

COVID-19 pandemic response, Laboratory data reporting: CARES Act Section 18115 (2022).

12.

Douglas

Respress

Gaglioti

, et al. Variation in reporting of the race and ethnicity of COVID-19 cases and deaths across US states: April 12, 2020, and November 9, 2020. Am J Public Health. 2021;111(6):1141-1148. doi:10.2105/AJPH.2021.306167

13.

US Census Bureau. County population by characteristics: 2010-2020. Accessed December 16, 2021. https://www.census.gov/programs-surveys/popest/technical-documentation/research/evaluation-estimates/2020-evaluation-estimates/2010s-county-detail.html

14.

Centers for Disease Control and Prevention. FAQ: COVID-19 data and surveillance. Updated October 5, 2022. Accessed October 20, 2022. https://www.cdc.gov/coronavirus/2019-ncov/covid-data/faq-surveillance.html

15.

Centers for Disease Control and Prevention. Human infection with 2019 novel coronavirus case report form. April 2020. Accessed August 9, 2022. https://www.cdc.gov/coronavirus/2019-ncov/downloads/pui-form.pdf

16.

Gold

JAW

DeCuir

Coyle

, et al. COVID-19 case surveillance: trends in person-level case data completeness, United States, April 5–September 30, 2020. Public Health Rep. 2021;136(4):466-474. doi:10.1177/00333549211006973

17.

Council of State and Territorial Epidemiologists. Update to the Standardized Surveillance Case Definition and National Notification for 2019 Novel Coronavirus Disease (COVID-19): Position Paper. Updated 2021. Accessed December 16, 2021. https://asprtracie.hhs.gov/technical-resources/resource/8322/standardized-surveillance-case-definition-and-national-notification-for-2019-novel-coronavirus-disease-covid-19

18.

Bland

Altman

. Multiple significance tests: the Bonferroni method. BMJ. 1995;310(6973):170. doi:10.1136/bmj.310.6973.170

19.

Centers for Disease Control and Prevention. COVID-19 case surveillance public use data with geography. Accessed December 16, 2021. https://data.cdc.gov/Case-Surveillance/COVID-19-Case-Surveillance-Public-Use-Data-with-Ge/n8mc-b4w4

20.

Beaulieu

. Addressing Gaps in Public Health Reporting of Race and Ethnicity Data for COVID-19: Findings and Recommendations Among 45 State & Local Health Departments. Council of State and Territorial Epidemiologists; 2022. Accessed May 31, 2022. https://preparedness.cste.org/wp-content/uploads/2022/04/RaceEthnicityData_FINAL.pdf

21.

Udry

Hendrickson-Smith

. Health and behavior risks of adolescents with mixed-race identity. Am J Public Health. 2003;93(11):1865-1870. doi:10.2105/ajph.93.11.1865

22.

Running Bear

Asdigian

Beals

Manson

Kaufman

. Health outcomes in a national sample of American Indian and Alaska Native adults: differences between multiple-race and single-race subgroups. PLoS One. 2020;15(12):e0242934. doi:10.1371/journal.pone.0242934

23.

Magaña López

Bevans

Wehrlen

Yang

Wallen

. Discrepancies in race and ethnicity documentation: a potential barrier in identifying racial and ethnic disparities. J Racial Ethn Health Disparities. 2017;4(5):812-818. doi:10.1007/s40615-016-0283-3

24.

Beaulieu

. Addressing Gaps in Public Health Reporting of Race and Ethnicity Data for COVID-19: Findings and Recommendations Among 45 State & Local Health Departments. Council of State and Territorial Epidemiologists; 2022. Accessed October 20, 2022. https://preparedness.cste.org/wp-content/uploads/2022/04/RaceEthnicityData_FINAL.pdf

25.

Labgold

Hamid

Shah

, et al. Estimating the unknown: greater racial and ethnic disparities in COVID-19 burden after accounting for missing race and ethnicity data. Epidemiology. 2021;32(2):157-161. doi:10.1097/EDE.0000000000001314

26.

MacKenzie

Davidson

Wiesenthal

, et al. The promise of electronic case reporting. Public Health Rep. 2016;131(6):742-746. doi:10.1177/0033354916670871

27.

Centers for Disease Control and Prevention. Why is eCR important? December 3, 2021. Accessed December 17, 2021. https://www.cdc.gov/ecr/why-is-eCR-important.html

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.13 MB

Completeness of Race and Ethnicity Reporting in Person-Level COVID-19 Surveillance Data,50 States,April 2020–December 2021

Abstract

Objectives:

Methods:

Results:

Conclusion:

Keywords

Methods

Results

Discussion

Limitations

Conclusions

Supplemental Material

sj-docx-1-phr-10.1177_00333549231154577 – Supplemental material for Completeness of Race and Ethnicity Reporting in Person-Level COVID-19 Surveillance Data, 50 States, April 2020–December 2021

Footnotes

Acknowledgements

Authors’ Note

Declaration of Conflicting Interests

Funding

ORCID iDs

Supplemental Material

References

Supplementary Material