Differences in Pertussis Incidence by Race and Ethnicity in the United States, 2010–2017

Abstract

Background

An increased pertussis burden has been demonstrated among Hispanic or Latino and American Indian or Alaska Native (AI/AN) infants. However, data on potential disparities among other age and racial groups are limited.

Methods

We analyzed pertussis cases reported through Enhanced Pertussis Surveillance from 2010 to 2017. Pertussis and severe pertussis incidence were calculated by race (White, Black or African American, AI/AN, and Asian or Pacific Islanders), ethnicity (Hispanic or Latino and non-Hispanic or non-Latino), and age.

Results

Compared with White persons, overall incidence was lower among Black or African American (incidence rate ratio [IRR], .57; 95% confidence interval [CI], .53-.61), AI/AN (IRR, 0.65; 95% CI, .58-.72), and Asian or Pacific Islander persons (IRR, 0.39; 95% CI, .35-.43). Overall incidence of pertussis was higher (1.5-fold; 95% CI, 1.37-1.60) among Hispanic or Latino compared with non-Hispanic or non-Latino adults, potentially related to household size or lower pertussis vaccine uptake among adult Hispanic or Latino cases. Severe pertussis incidence was similar among Black or African American and AI/AN persons compared with White persons. Among infants, severe pertussis incidence was 1.4-fold higher (95% CI, 1.03-1.82) among Black or African American infants than among White infants, and 2.1-fold higher (95% CI, 1.67-2.57) among Hispanic or Latino infants than non-Hispanic or non-Latino infants.

Conclusions

The contrast between lower reported incidence but similar or higher severe pertussis incidence among Black or African American and AI/AN persons compared with White persons warrants further investigation and may reflect underdiagnosis or underreporting of mild disease.

Pertussis, or whooping cough, is a highly contagious vaccine-preventable disease caused by the bacterium Bordetella pertussis. Clinical manifestations of pertussis can vary from mild cough illness to severe disease and death in infants, especially during the early months of life [1]. Reported pertussis incidence in the United States had been increasing since the late 1980s until the start of the COVID-19 pandemic in 2020, with some large peaks in reported disease, despite stable or increasing pertussis vaccination coverage. In 2019, vaccine coverage was 90% with tetanus, diphtheria, and acellular pertussis vaccine among adolescents [2], whereas vaccine coverage with 4 doses of diphtheria, tetanus, and acellular pertussis-containing vaccine (DTaP) was 81% among 2 year olds [3].

Several studies have suggested differences in pertussis epidemiology by race and ethnicity in the United States. An analysis of national pertussis surveillance data demonstrated higher incidence among Hispanic or Latino persons and lower incidence among Black or African American persons compared with non-Hispanic or non-Latino or White persons, respectively [4]. Black or African American and Hispanic or Latino persons have also accounted for a disproportionate number of cases among infants [4] and among hospitalized pertussis patients [5]. Several additional studies have demonstrated an increased incidence of pertussis disease, hospitalization, and/or deaths among Hispanic or Latino and American Indian or Alaska Native (AI/AN) infants [6–9]. Although these findings suggest that there may be important differences in the epidemiology or reporting of U.S. pertussis cases by race and ethnicity, there remain important gaps in these findings particularly for noninfant age groups.

We analyzed data on U.S. pertussis cases collected through Enhanced Pertussis Surveillance (EPS)/Emerging Infections Program network during 2010 through 2017. Pertussis incidence and disease severity were compared among different racial and ethnic groups. Individual characteristics of persons and their households, including differences in vaccine coverage were examined to better understand disparities in pertussis incidence and disease severity.

METHODS

Pertussis cases reported through EPS with cough onset between 1 January 2010 and 31 December 2017 were included in this analysis [10]. EPS collects data and isolates from pertussis cases reported from 7 states: statewide from Connecticut (beginning in 2011), Minnesota (2010), and New Mexico (2011); and selected counties from Colorado (5 Denver counties) (2011), Georgia (20 metro Atlanta counties) (2014), New York (15 Rochester and Albany counties) (2011), and Oregon (3 Portland counties) (2010). Information on sex, age, race, ethnicity, geographic location, pertussis vaccination history, antibiotic treatment timing, and clinical symptoms is collected. Sources of information include medical records, state immunization registries, patient-held vaccination cards, school vaccine records, and patient self-report. Persons had the option to identify their ethnicity as Hispanic or Latino or non-Hispanic or non-Latino. Persons had the option to identify their race as AI/AN, Asian or Pacific Islander (API), Black or African American, Multiracial, Other, or White. Race and ethnicity variables were used to categorize cases corresponding to race and ethnicity groups reported in the 2010 U.S. census. Cases with multiple races reported or self-reporting as multiracial were categorized as multiracial.

Age-appropriate pertussis vaccines received by the patient were ascertained using multiple sources, including vaccination cards, medical records, and state immunization registries. Vaccination status was classified according to Advisory Committee on Immunization Practices guidelines [11]. Vaccine doses received after or <2 weeks before cough onset were not considered for vaccination status. Cases with a missing or incomplete pertussis vaccine history were classified as “unknown.” Cases in infants <2 months old with no recorded vaccine administered were classified as “too young.” Cases in persons ≥2 months old that were verified by investigators as having not received pertussis vaccine, or who were verified as receiving less than the age-appropriate recommended number of pertussis vaccine doses were classified as “unvaccinated” and “insufficiently vaccinated,” respectively. All other cases were considered “up-to-date” on pertussis vaccination.

Cases that met the Council of State and Territorial Epidemiologists’ case definitions for probable or confirmed cases were included [12]. The case definition for 2010 through 2014 defined probable cases as cough illness lasting ≥2 weeks with at least 1 of the following symptoms: paroxysms, inspiratory whoop, or posttussive vomiting, absence of laboratory confirmation, and no epidemiologic linkage to a laboratory confirmed case. Confirmed cases included (1) any probable case with positive polymerase chain reaction or an epidemiologic link to a laboratory-confirmed case; or (2) a case with cough of any duration with isolation of Bordetella pertussis by culture. The case definition was revised in 2014 for infants aged <1 year to include apnea with or without cyanosis as 1 of the clinical symptoms. Severe pertussis disease was defined as a case with 1 or more of the following: seizure, encephalopathy, positive x-ray for pneumonia, hospitalization, or death.

Multiple imputation was conducted to fill in missing data on race and ethnicity using fully conditional specification. Ten imputations were performed with the following variables to inform imputation.

• Demographic information: state, age, sex, employed at/attended school, employed at/attended daycare, number of residents in household, race (to impute ethnicity), and ethnicity (to impute race).

• Clinical and epidemiological information: presence of paroxysmal cough, vomiting, inspiratory whoop, an x-ray positive for pneumonia, or apnea; any laboratory testing of clinical specimens; number of health care visits to diagnosis; case classification; hospitalization; prescription of antibiotics; receipt of any pertussis-containing vaccine; source of infection; association with known pertussis outbreak; and epidemiologic link to laboratory-confirmed case.

• ZIP Code Tabulation Areas–level race and ethnicity data from the 2010 U.S. census were merged with EPS data and used to inform the imputation [13].

Pertussis and severe pertussis incidence were calculated using U.S. census data for the EPS catchment area. Incidence rate ratios (IRRs) and corresponding 95% confidence intervals (CIs) were calculated overall and stratified by age using Poisson regression with White race and non-Hispanic or non-Latino ethnicity as the reference categories and adjusting for site. Odds ratios were calculated using bivariate analyses to identify case characteristics that differed by race and ethnicity. Kolmogorov-Smirnov significance testing was performed to compare overall and site-specific distributions of count data between race and ethnicity groups. All data were analyzed in SAS 9.4 (Cary, NC). This study was reviewed in accordance with human research protection procedures and was determined to be nonresearch, public health surveillance by the Centers for Disease Control and Prevention.

RESULTS

Descriptive Epidemiology of Cases

During 2010 through 2017, 20,759 pertussis cases were reported in EPS (Table 1). Minnesota accounted for almost half of cases (n = 9863, 48%) (Figure 1). Of the cases, 16 293 (79%) were in persons aged <20 years and 1495 (7%) were in infants aged <1 year. A total of 1072 (5%) of cases were classified as severe; 4 deaths were reported (Table 2).

Table 1.

Demographic Characteristics of Pertussis Cases in Enhanced Pertussis Surveillance, 2010-2017 (N = 20 759)

Characteristic	N	% EPS Cases	% EPS Catchmenta
State
Colorado	4025	19.4	12.6
Connecticut	657	3.2	17.1
Georgia	426	2.1	10.9
Minnesota	9863	47.5	29.6
New Mexico	2756	13.3	9.9
New York	1280	6.2	10.4
Oregon	1752	8.4	9.4
Sex
Male	9464	45.7	49.1
Female	11 254	54.3	50.9
Unknown	41	0.4	N/A
Age (y)
<1	1495	7.2	1.2
1-6	3774	18.2	7.5
7-10	3762	18.1	5.2
11-19	7262	35	11.8
20+	4466	21.5	74.3
Race
White persons	16 994	81.9	81.4
Black or African American persons	829	4.0	11.4
American Indian or Alaskan Native persons	370	1.8	2.1
Asian/Pacific Islander	364	1.8	5.1
Multiracial	267	1.3	N/A
Other	560	2.7	N/A
Unknown	1375	6.6	N/A
Ethnicity
Hispanic or Latino persons	3793	18.3	14.8
Non-Hispanic or non-Latino persons	15 741	75.8	85.2
Unknown	1225	5.9	N/A
case classificationb
Confirmed	18 230	87.8	…
Probable	2529	12.2	…

Abbreviations: EPS, Enhanced Pertussis Surveillance; N/A, not available.

^a2010 US Census estimates.

^bCases before 2014 and from 2014 through 2017 classified according to 2010 and 2014 Council of State and Territorial Epidemiologists (CSTE) case definitions, respectively.

Figure 1.

Annual incidence of pertussis (A) and severe pertussis (B) reported by EPS site from 2010 through 2017. Data available for Colorado, Connecticut, New Mexico, and New York since 2011 and Georgia since 2014. Rates are per 100 000 population. *A large peak in pertussis disease was reported nationally in 2012 [4]. All EPS sites reported higher pertussis incidence in 2012, with the highest incidence reported in Minnesota.

Table 2.

Case Characteristics by Race and Ethnicity, 2010-2017 (Nonimputed Data) ^a

Characteristic	Total	White Persons			Black or African American Persons			American Indian or Alaskan Native Persons			Asian/Pacific Islander			Non-Hispanic or Non-Latino Persons			Hispanic Or Latino Persons
	N	N	(%)	P	N	(%)	P	N	(%)	P	N	(%)	P	N	(%)	P	N	(%)	P
Age (y)																		…
<1	1495	1028	(6.0)	Ref	163	(19.7)	<.0001	46	(12.4)	<.0001	42	(11.5)	.0003	976	(6.2)	Ref	450	(11.9)	<.0001
1-6	3774	3054	(18.0)		193	(23.3)		62	(16.8)		67	(18.4)		2728	(17.3)		865	(22.8)
7-10	3762	3151	(18.5)		119	(14.4)		64	(17.3)		66	(18.1)		2852	(18.1)		714	(18.8)
11-19	7262	6093	(35.9)		183	(22.1)		105	(28.4)		109	(29.9)		5734	(36.4)		1004	(26.5)
20+	4466	3668	(21.6)		171	(20.6)		93	(25.1)		80	(22.0)		3451	(21.9)		760	(20.0)
Vaccination status
Up-to-dateb	12 539	10 358	(71.8)	Ref	446	(62.9)	<.0001	223	(72.9)	.007	230	(74.7)	.0025	9437	(70.8)	Ref	2370	(73.4)	<.0001
Insufficiently vaccinated	2680	2149	(14.9)		110	(15.5)		52	(17.0)		43	(14.0)		2029	(15.2)		464	(14.4)
Unvaccinated	1930	1658	(11.5)		83	(11.7)		20	(6.5)		22	(7.1)		1595	(12.0)		248	(7.7)
Too young	421	269	(1.9)		70	(9.9)		11	(3.6)		13	(4.2)		263	(2.0)		145	(4.5)
Unknown	3189	2560	…	-	120	…	-	64	…	-	56	…	…	2417	…	-	566	…	-
Symptom severity
Not severe	19 687	16 217	(95.4)	Ref	737	(88.9)	<.0001	336	(90.6)	<.0001	340	(93.4)	.069	15 009	(95.3)	Ref	3518	(92.7)	<.0001
Severe	1072	777	(4.6)		92	(11.1)		35	(9.4)		24	(6.6)		732	(4.7)		275	(7.3)
Seizures	46	33	(0.2)	-	1	(0.1)	-	4	(1.1)	-	3	(0.8)	-	31	(0.2)	-	12	(0.3)	-
Encephalopathy	24	18	(0.1)	-	1	(0.1)	-	1	(0.3)	-	1	(0.3)	-	17	(0.1)	-	6	(0.2)	-
Pneumonia (x-ray)	467	354	(2.1)	-	27	(3.3)	-	13	(3.5)	-	11	(3.0)	-	354	(2.2)	-	73	(1.9)	-
Hospitalized	678	473	(2.8)	-	73	(8.8)	-	23	(6.2)	-	17	(4.7)	-	430	(2.7)	-	220	(5.8)	-
Death	4	3	(0.0)	-	0	(0.0)	-	0	(0.3)	-	0	(0.0)	-	4	(0.0)	-	0	(0.0)	-
Time to antibiotic treatmentc
<1 wk	3646	3043	(19.3)	Ref	127	(17.0)	.0502	52	(15.3)	.1507	64	(19.1)	.4622	2745	(18.8)	Ref	696	(19.9)
1-<2 wk	5850	4836	(30.6)		214	(28.7)		103	(30.3)		102	(30.4)		4553	(31.2)		971	(27.8)
2-<3 wk	4417	3601	(22.8)		179	(24.0)		78	(22.9)		77	(23.0)		3347	(22.9)		802	(22.9)	.0015
3 + wk	4783	3914	(24.8)		195	(26.2)		93	(27.4)		78	(23.3)		3584	(24.5)		926	(26.5)
Never	506	405	(2.6)		30	(4.0)		14	(4.1)		14	(4.2)		386	(2.6)		102	(2.9)
Unknown	1557	1195	…	-	84	…	-	30	…	-	29	…	-	1126	…	-	296	…	-
Total	20 759	16 994	…		829	…		370	…		364	…		15 741	…		3793	…

^aComparison of characteristic distribution performed for each race and ethnicity group using White persons and non-Hispanic or non-Latino persons, respectively, as reference groups.

^bAvailable vaccination information was used to determine whether persons were up-to-date for pertussis vaccination per current Advisory Committee on Immunization Practices recommendations.

^cTime between cough onset and initiation of antibiotic treatment.

Before multiple imputation, most cases (n = 16 994, 82%) were in White persons, followed by 829 (4%) in Black or African American, 370 (2%) in AI/AN, 364 (2%) in API, and 267 (1%) in multiracial persons; 560 (3%) were in persons classified as other race, and 1375 (7%) in persons with unknown race. There were 3793 (18%) cases in Hispanic or Latino persons, 15 741 (76%) in non-Hispanic or non-Latino persons, and 1225 (6%) in persons with unknown ethnicity (Table 1). The age distribution of cases varied by race and ethnicity. Infants made up 6% of White, 20% of Black or African American, and 12% each of AI/AN and API persons; 6% of non-Hispanic or non-Latino persons were infants compared with 12% of Hispanic or Latino persons (Table 2). The proportion of cases that was severe also differed by race and ethnicity, ranging from 5% among White and non-Hispanic or non-Latino to 11% among Black or African American persons (Table 2).

Overall, vaccination status differed significantly by race and ethnicity, with Hispanic or Latino (73%) and API (75%) persons more likely to be up-to-date on vaccinations than non-Hispanic or non-Latino (71%) and White (72%) persons, respectively, whereas Black or African American (10%), API (4%), and Hispanic (5%) persons were more likely to be too young to be vaccinated than White (2%) or non-Hispanic or non-Latino (2%) persons, respectively (Table 2). However, when stratified by age, Hispanic or Latino adults were less likely to be up-to-date than non-Hispanic or non-Latino adults, whereas AI/AN adults were more likely to be up-to-date than White adults (Supplementary Table 1).

Overall, 90% of cases were known to have received antibiotic treatment. The most common antibiotics prescribed were clarithromycin and azithromycin (96.2%). Hispanic or Latino persons were less likely to have received antibiotic treatment 1 to <2 weeks after cough onset compared with non-Hispanic or non-Latino persons (Table 2). Significant differences in the distributions of reported household residents and doctors’ visits between different racial and ethnic groups were observed, with more cases having higher numbers of household residents and doctors’ visits among Black or African American and Hispanic or Latino persons as well as more cases with relatively larger household size among AI/AN persons (Supplementary Figure 1). After multiple imputations, the proportions of cases by racial and ethnic group remained similar.

Pertussis and Severe Pertussis Incidence by Race and Ethnicity

Overall pertussis incidence was 14.1/100 000 person-years, whereas severe pertussis incidence was 0.7/100 000 person-years. Patterns in incidence of overall and severe pertussis were largely similar when stratified by year and jurisdiction (Figure 1); however, in Minnesota in 2012, the ratio of severe to total pertussis cases was relatively lower than for other years or sites. Pertussis incidence was significantly lower among Black or African American (IRR, 0.57; 95% CI, .53-.61), AI/AN (IRR, 0.65; 95% CI, .58-.72), and API persons (IRR, 0.39; 95% CI, 0. 35-.43) than White persons (Table 3). Severe pertussis incidence was lower among API persons (IRR, 0.54, 95% CI, .36–.80) than White persons. By ethnicity, incidence of both pertussis (IRR, 1.33; 95% CI, 1.28-1.38) and severe pertussis (IRR, 1.89; 95% CI, 1.62-2.21) was significantly higher among Hispanic or Latino than non-Hispanic or non-Latino persons.

Table 3.

Racial and Ethnic Differences in Pertussis Incidence and Severity, Adjusted by EPS Site

Characteristic	All Cases				Severe Cases
	Incidencea	IRRb	95% CI	P Value	Incidencea	IRRb	95% CI	P Value
Race
White	10.0	REF			0.63	REF
Black or African American	5.7	0.57	0.53-0.61	<.001	0.76	1.21	0.98-1.51	.081
American Indian or Alaskan Native persons	6.5	0.65	0.58-0.72	<.001	0.79	1.26	0.90-1.78	.184
Asian/Pacific Islander	3.9	0.39	0.35-0.43	<.001	0.34	0.54	0.36-0.80	.002
ethnicity
Non-Hispanic or non-Latino persons	9.0	REF			0.58	REF
Hispanic or Latino persons	12.0	1.33	1.28-1.38	<.001	1.09	1.89	1.62-2.21	<.001

Abbreviations: CI, confidence interval; IRR, incident rate ratio; REF, reference.

^aIncidence per 100,000 population.

^bIRR and 95% CI calculated from Poisson regression on post multiple imputation dataset.

Pertussis incidence was highest among infants (84.0 cases/100 000 person years) and lowest among adults aged >20 years (4.1 cases/100 000 person-years). Pertussis incidence was similar among AI/AN (IRR, 0.96; 95% CI, .71-1.28) and Black or African American infants (IRR, 1.14; 95% CI, .96-1.34) but significantly lower among API infants (IRR, 0.60; 95% CI, .44-.81) compared with White infants (Figure 2). Severe pertussis incidence was significantly higher among Black or African American (IRR, 1.37; 95% CI, 1.03-1.82) and significantly lower among API (IRR, 0.53; 95% CI, .29-.96) compared with White infants. Among other age groups, pertussis incidence was lower among Black or African American, AI/AN, and API persons compared with White persons. In contrast, severe pertussis incidence in all other age categories among Black or African American, AI/AN, and API persons was similar to severe pertussis incidence in White persons.

Figure 2.

EPS site-adjusted, age-stratified incidence of overall and severe cases by race (A, B) and ethnicity (C, D), respectively, after multiple imputations. Incidence rate calculated by Poisson regression per 100 000 population, *P < .05, **P < .001. Data source: EPS, 2010-2017.

Pertussis incidence was lower among Hispanic or Latino persons aged 7 to 10 years and similar among those aged 1 to 6 years compared with non-Hispanic or non-Latino persons of similar age (Figure 2). Pertussis incidence was 1.2-fold higher among Hispanic or Latino adults aged 20+ years (95% CI, 1.08-1.29) and 1.53-fold higher among Hispanic or Latino infants (95% CI, 1.35-1.73) compared with non-Hispanic or non-Latino adults and infants. Among Hispanic or Latino infants, severe pertussis incidence was 2.07-fold higher than among non-Hispanic or non-Latino infants (95% CI, 1.67-2.57).

The disparities in severe pertussis incidence across racial and ethnic groups fluctuated during the study period, with the greatest differences reported in 2012 and smaller differences in more recent years (Figure 3). Nevertheless, incidence of pertussis and severe pertussis remained elevated in Hispanic or Latino persons throughout the analytic period.

Figure 3.

Annual pertussis incidence and severity by race (A, B) and ethnicity (C, D), respectively, after multiple imputations. Rates per 100 000 population. Data source: EPS, 2010-2017. *A large peak in pertussis disease was reported nationally in 2012 [4]. All EPS sites reported higher pertussis incidence in 2012, with the highest incidence reported in Minnesota.

DISCUSSION

This analysis of U.S. pertussis cases reported through EPS during 2010 through 2017 highlights and reaffirms previously identified disparities in pertussis incidence among Hispanic or Latino persons and severe pertussis disease among Hispanic or Latino and AI/AN infants. The magnitude of increased risk for pertussis and severe pertussis disease that we identified among Hispanic or Latino and AI/AN infants was comparable to that observed in prior studies [6–8]. However, we also found an elevated risk of pertussis, but not of severe pertussis, among Hispanic or Latino adults. This increased risk may be related to our finding that Hispanic or Latino adults were less likely to be up-do-date on vaccinations than non-Hispanic or non-Latino adults. We also observed larger household sizes among Hispanic and Latino persons, in alignment with prior studies [6], as well as among Black or African American and AI/AN persons. Larger households and fewer up-to-date vaccinated persons represent increased opportunities for pertussis exposure and infection. Hispanic or Latino and Black or African American cases reported a higher number of doctors’ visits and Hispanic or Latino persons had delayed receipt of antibiotics after cough onset. We could not determine if delay in receipt of antibiotics was due to the number of doctor visits needed for a diagnosis or a delay in seeking care. Further evaluation to better understand these differences in pertussis epidemiology, vaccination status, and treatment by race and ethnicity can help inform effective prevention strategies targeted for disproportionately affected populations.

Although pertussis incidence was similar among AI/AN and Black or African American infants compared with White infants, this trend did not hold true among other age groups, in which a substantially lower pertussis incidence was observed among AI/AN and Black or African American persons compared with White persons. In contrast, we found overall severe pertussis incidence was similar among Black or African American persons and among AI/AN compared with White persons. Although our findings could suggest true differences in pertussis incidence and severity of disease by race and age group, the findings may more likely indicate differences in pertussis diagnosis or reporting by healthcare professionals by race and ethnicity. The comparable or higher rates of pertussis and severe pertussis among infants paired with lower rates of nonsevere disease in older persons among Black or African American and AI/AN persons suggests underdiagnosis and/or underreporting of mild pertussis cases in these populations, especially among older persons. The underdiagnosis or underreporting of pertussis could reflect differential health care seeking behavior or access to health care among minority populations as indicated by higher number of doctors’ visits required for diagnosis among Hispanic or Latino and Black or African American persons. Poor access to health care has been well documented to disproportionately affect racial and ethnic minority populations in the United States [14–16]. Underdiagnosis of pertussis in adults increases the potential risk of disease transmission to unvaccinated infants and increases the risk of missed diagnosis among exposed infants.

In contrast to AI/AN and Black or African American persons, incidence of both overall and severe pertussis was lower among API persons than among White persons across all age groups. The consistency of this finding across age groups and disease severity is more suggestive of a real difference in disease incidence rather than underreporting. However, 1 limitation to this finding is that we could not separate Asian and Pacific Islander/Native Hawaiian Islander populations in this analysis. Given that these populations have different characteristics and might have differing access to health care or health care–seeking behaviors [17], our analysis might be masking important disparities by combining these 2 populations.

As noted earlier, vaccine uptake among Hispanic or Latino than in non-Hispanic or non-Latino adults was lower, which may have contributed not only to the increased incidence of pertussis among Hispanic or Latino adults, but also among Hispanic or Latino infants because of the potential for increased transmission from infected adults to infants. In contrast, the remaining observed differences in pertussis and severe pertussis incidence by racial and ethnic group appear unlikely to be driven primarily by differential pertussis vaccine uptake. Among other age groups, the primary differences in vaccination status were that Hispanic or Latino persons aged 1 to 19 years and AI/AN persons aged 1 to 6 and 20+ years were more often up-to-date on vaccinations than non-Hispanic or non-Latino or White persons, respectively, and that more cases in Black or African American or Hispanic or Latino infants were in persons too young to be vaccinated. Our findings align with other studies showing that vaccination coverage with tetanus, diphtheria, and acellular pertussis vaccine in adolescents and with at least 3 doses of DTaP in children is similar across racial and ethnic groups [3, 18]. However, prior studies have also shown that coverage with 4 doses of DTaP is slightly lower among Black or African American and Hispanic or Latino children than in non-Hispanic or non-Latino White children [3], which differs from our findings.

Although EPS provided us with more complete data on race, ethnicity, and other variables compared with national pertussis surveillance data, our analysis still has limitations. The generalizability of our analysis is limited by the EPS catchment area, which includes approximately 7% of the U.S. population and only covers major metropolitan areas in Colorado, Georgia, New York, and Oregon; furthermore, during our analysis period nearly half of cases were reported from 1 site (Minnesota). Although we did have missing race or ethnicity information for 7% and 6% of cases, respectively, we used multiple imputation to increase completeness and more accurately estimate pertussis incidence by race and ethnicity. Finally, our analysis relied on surveillance data and therefore included only reported pertussis cases. Because there is generally a lack of awareness and underdiagnosis of pertussis in the adult population, we are likely not capturing all adult cases of pertussis in the study area. In addition, although the majority of the cases did occur in White persons, we found that the number of cases in people of other races and was still large enough to support calculation of incidences and risk ratios. In some cases, the precision of our estimates was affected as indicated by wide confidence intervals, which did limit our inferences. As noted previously, it is likely that some of the differences in pertussis incidence we observed reflect differences in health care seeking behavior or access and/or testing or reporting by health care professionals rather than true differences in pertussis incidence.

Our analyses showed significant differences in pertussis incidence among racial and ethnic groups that varied by age group. Our findings reaffirm previously identified disparities in severe pertussis disease among Hispanic or Latino infants and highlight disparities in pertussis incidence among Hispanic or Latino adults, which may be related to lower up-to-date vaccination status, larger household sizes, and higher number of doctors’ visit for a diagnosis in this group. Among AI/AN, API, and Black or African American adult populations, lower incidence of total cases but not of severe pertussis among older age groups might suggest underdiagnosis/underreporting of mild disease. Further investigation into this potential underdiagnosis/underreporting is warranted to inform interpretation of surveillance data, identify potential disparities in health care access or pertussis diagnosis, and enhance pertussis control efforts. In addition, the study findings would be enhanced by national validation. In the wake of the COVID-19 pandemic, which has brought renewed attention to the health disparities among U.S. racial and ethnic groups, these findings can be leveraged to continue reducing disparities and ultimately improve access to health care and health outcomes among all people in the United States.

Patient Consent Statement: This study does not include factors necessitating patient consent.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. The authors thank Christine Miner and Amanda E. Faulkner, Centers for Disease Control and Prevention, for their contributions to managing and compiling the data used in this analysis. We thank the EPS surveillance staff in Colorado, Connecticut, Minnesota, New Mexico, New York, and Oregon for collecting the data on pertussis cases used for this analysis. The authors also thank Roxanne Ryan, Kristen Gerard, Jillian Karr, Glenda Smith, Eva Pradhan, Suzanne McGuire, and Rachel Ester for their efforts on this study.

Disclaimer. The findings and conclusions in this article are those of the authors and do not necessarily represent the official position of the U.S. Centers for Disease Control and Prevention.

Potential conflicts of interest. All authors: No reported conflicts.