Authors: Shu-Yu Tai, Ying-Chen Chi, Joh-Jong Huang, Ying-Yeh Chen, Ichiro Kawachi, Tsung-Hsueh Lu
Categories: Regular Article, Adolescent, Leading cause of death, Mortality rates, Race and ethnicity, Accidents, Injuries, COVID-19
Source: SSM - Population Health
Authors: Shu-Yu Tai, Ying-Chen Chi, Joh-Jong Huang, Ying-Yeh Chen, Ichiro Kawachi, Tsung-Hsueh Lu
The World Health Organization (WHO) classification scheme for leading causes of death (LCODs) provides greater granularity than the National Center for Health Statistics (NCHS) scheme by subdividing broad categories such as cancer, heart disease, and accidents into 17, 8, and 7 specific groups, respectively. This study compared LCOD rankings among U.S. adolescents aged 10–19 years across racial and ethnic groups using CDC WONDER mortality data. Across 48 sex–age–race/ethnicity–period strata, top-ranked causes differed between schemes in 31 strata. For example, policy priorities for non-Hispanic White boys aged 15–19 years would differ by accidents ranked first under the NCHS scheme, whereas suicide or transportation accidents ranked highest under the WHO scheme. Substantial changes in ranks of top five LCODs across years were observed in some racial/ethnic groups. Among boys aged 10–14 years, transportation accidents rose from third in 2018–2019 to first in 2020–2021 and remained first in 2022–2023 among non-Hispanic Asians, while remaining consistently second among non-Hispanic Black and White boys and first among Hispanic boys. The WHO and NCHS schemes offer complementary LCOD profiles that can guide adolescent mortality surveillance and health policy decision-making.
Leading causes of death (LCOD) rankings are widely used to assess the relative burden of diseases and injuries and to inform resource allocation and prevention priorities (Anderson, 2011; Gerzoff & Williamson, 2001). The National Center for Health Statistics (NCHS) publishes annual LCOD rankings by sex, age, and race/ethnicity in the United States (Curtin et al., 2024). Multiple studies have shown that LCOD rankings differ across racial and ethnic groups (Chang et al., 2016; Espey et al., 2014; Faust et al., 2022; Hastings et al., 2015; Luck et al., 2022; Simon et al., 2021).
However, a key limitation of most studies is their reliance on the NCHS classification scheme, which includes 51 categories (including a recently added COVID-19 category) encompassing broad groupings such as cancer, heart disease, and accidents (Curtin et al., 2024). As noted by Tai and Lu (2021) and Tai et al. (2022), such broad categories can mask important differences in the relative burden of specific causes. Because prevention strategies often differ by disease or injury subtype, more granular classification is needed to support targeted policy and resource allocation.
To address this limitation, the World Health Organization (WHO) developed a more detailed classification scheme comprising 66 categories (including a newly added COVID-19 category) that subdivide cancer, heart disease, and accidents into 17, 8, and 7 categories, respectively (Becker et al., 2006). Compared with the NCHS scheme, the WHO system offers substantially greater granularity. Several studies have compared LCOD rankings using the NCHS and WHO schemes for general populations (Gjertsen et al., 2018; Griffiths et al., 2005; Hsiao et al., 2015; Tai, Chi, Chien, et al., 2023, Tai, Chi, Lo, et al., 2023), but none have examined how these classifications perform for U.S. adolescents.
Applying the WHO scheme is particularly important for adolescent mortality, as injuries account for a large proportion of deaths in this age group. The WHO scheme disaggregates the broad NCHS category of “accidents” (unintentional injuries) into seven specific mechanisms—transportation accidents, falls, firearm-related injuries, drowning, suffocation, poisoning, and events of undetermined intent—providing more actionable insights for prevention. Similarly, cancer profiles among adolescents differ from those in adults and older populations, making site-specific rankings more relevant for guiding prevention and treatment priorities. Therefore, this study aimed to compare LCOD rankings by sex, age, and race/ethnicity among U.S. adolescents across time using both the WHO and NCHS classification schemes.
In this descriptive temporal trends study, we queried “2018–2023: Underlying Cause of Death by Single-Race Categories” dataset from CDC WONDER (Centers for Disease Control and Prevention, 2024). In query 3 “select demographics” we selected sex, five-age groups, Hispanic origin, and single race 6. In query 4 “select year” we selected 2018–2019, 2020–2021, and 2022–2023. In query 6 “select cause of death” we selected the International Classification of Disease Tenth Revision (ICD-10) codes for each category in the WHO (Becker et al., 2006) and NCHS scheme (Curtin et al., 2024).
Regarding analysis, we classified the age groups into 10–14 years and 15–19 years as in the NCHS annual report (Curtin et al., 2024). We specifically identified non-Hispanic Asian group as one category because Asian Americans are the fastest-growing racial or ethnic group in the United States (Budiman & Ruiz, 2021). We calculated sex–age–race/ethnicity–specific mortality rates (deaths per 100 000) for each category in both the NCHS and WHO schemes and used these rates to determine LCOD rankings.
A total of 92 812 adolescents who died in the United States between 2018 and 2023 were included in the analysis (69.4 % boys and 30.6 % girls; 77.9 % aged 15–19 years and 22.1 % aged 10–14 years; 2.6 % non-Hispanic Asian, 25.4 % non-Hispanic Black, 44.8 % non-Hispanic White, 22.7 % Hispanic, and 4.5 % other races/ethnicities).
Table 1 presents the top-ranked causes of death across 48 strata defined by sex (2 categories), age (2 categories), race/ethnicity (4 categories), and period (3 categories). In 31 of the 48 strata, the top-ranked causes differed between the WHO and NCHS schemes. Under the WHO scheme, the most frequent top-ranked causes were suicide (19 strata), transportation accidents (19 strata), and homicide (10 strata). In contrast, under the NCHS scheme, accidents ranked first in 33 strata, followed by suicide in 7 strata, cancer in 4 strata, and homicide in 4 strata.Table 1Top-ranked cause of death among U.S. adolescents by sex, age, and race/ethnicity according to the World Health Organization (WHO) and the National Center for Health Statistics (NCHS) classification schemes, before (2018–2019), during (2020–2021), and after (2022–2023) COVID-19 pandemic.Table 1Non-Hispanic AsianNon-Hispanic BlackNon-Hispanic WhiteHispanicSchemeBeforeDuringAfterBeforeDuringAfterBeforeDuringAfterBeforeDuringAfterBoys aged 15–19 yearsWHOSuicSuicSuicHomiHomiHomiSuicSuicTranTranHomiTranNCHSSuicSuicSuicHomiHomiHomiAcciAcciAcciAcciAcciAcciGirls aged 15–19 yearsWHOSuicSuicSuicHomiHomiHomiTranTranTranTranTranTranNCHSSuicSuicSuicAcciAcciAcciAcciAcciAcciAcciAcciAcciBoys aged 10–14 yearsWHOSuicSuicTranSuicHomiHomiSuicSuicSuicTranTranTranNCHSCaSuicAcciAcciAcciHomiAcciAcciAcciAcciAcciAcciGirls aged 10–14 yearsWHOTranSuicSuicTranHomiTranSuicSuicSuicTranTranTranNCHSAcciCaCaAcciAcciAcciAcciAcciAcciCaAcciAcciAbbreviations: Acci, accidents; Ca, cancer; Homi, homicide; Suic, suicide; Tran, transportation accident.
Fig. 1A–D presents the top ten LCODs according to the WHO and NCHS schemes in 2018–2019, 2020–2021, and 2022–2023 by race/ethnicity for boys aged 15–19 years (Fig. 1A), girls aged 15–19 years (Fig. 1B), boys aged 10–14 years (Fig. 1C), and girls aged 10–14 years (Fig. 1D). Of new specific categories created in the WHO scheme, five—transportation accidents, drowning, poisoning, hematopoietic cancer, brain cancer, and epilepsy—were ranked among the top ten in nearly every sex–age–race/ethnicity group for both adolescents aged 10–14 and 15–19 years; however, poisoning was among the top ten in every racial/ethnic group in adolescents aged 15–19 years only.Fig. 1Ranks and mortality rates (deaths per 100 000) for the 10 leading causes of death (LCODs) among U.S. adolescent boys aged 15–19 years (A), girls aged 15–19 years (B), boys aged 10–14 years (C), and girls aged 10–14 years (D) by race and ethnicity according to two classification schemes in 2018-19, 2020–21, and 2022–23 (gray solid line indicates no change in ranks, red solid arrow line indicates increase in ranks, and blue dotted arrow line indicates decline in ranks for top five LCODs in 2018-19).Fig. 1
By contrast, several WHO-specific categories appeared among the top ten only in particular sex, age, and race/ethnicity groups. For example, unintentional firearm injury was among the top ten for non-Hispanic Black and Hispanic boys aged 15–19 years (Fig. 1A) and for non-Hispanic White boys aged 10–14 years (Fig. 1C). Suffocation (accidental threats to breathing) appeared only in non-Hispanic Asian and White boys aged 10–14 years (Fig. 1C).
Notably, some categories are present only in the NCHS scheme and therefore provide distinct information not captured by the WHO scheme. For example, legal intervention (ICD-10 code Y35 including injuries inflicted by the police or other law-enforcing agents, including military on duty, in the course of arresting or attempting to arrest lawbreakers, suppressing disturbances, maintaining order, and other legal action) ranked among the top ten LCODs for non-Hispanic Black and Hispanic boys aged 15–19 years (Fig. 1A), and anemias appeared among the top ten for non-Hispanic Black boys and girls aged 15–19 years, and girls aged 10–14 years (Fig. 1A, B, and 1D).
Among the 16 strata defined by sex (2 categories), age (2 categories), and race/ethnicity (4 categories) in 2020–2021, COVID-19 ranked among the top ten LCODs in 11 and 13 strata according to the WHO and NCHS schemes, respectively. In 2022–2023, the number declined to 5 under the WHO scheme but remained at 13 under the NCHS scheme.
Changes in the top-ranked causes across periods were observed only among certain racial and ethnic groups among adolescents aged 10–14 years (Table 1). For instance, among non-Hispanic Asian girls, the top-ranked cause changed from transportation accidents in 2018–2019 to suicide in 2020–2021 under the WHO scheme; under the NCHS scheme the top-ranked cause changed from accidents to cancer during the same periods. No comparable change was observed among non-Hispanic White girls.
Notable changes in ranks of top five LCODs across study periods were observed among non-Hispanic Asian boys and girls aged 10–14 years, whereas the ranks remained stable among other racial and ethnic groups. Among boys aged 10–14 years, the rank of transportation accidents increased from third in 2018–2019 to second in 2020–2021 and first in 2022–2023 among non-Hispanic Asians, while remaining consistently second among non-Hispanic Black and White boys and first among Hispanic boys (Fig. 1C). Among girls aged 10–14 years, brain cancer increased in rank from fifth in 2018–2019 to second in both 2020–2021 and 2022–2023 among non-Hispanic Asians, whereas ranks for other racial and ethnic groups showed no substantial change (Fig. 1D).
The findings demonstrate that the ranking profiles of LCODs among U.S. adolescents vary depending on whether the WHO or NCHS scheme is applied. These differences yield distinct implications for public health policy decision-making. For example, when using the NCHS scheme, policy makers might prioritize accidents as most important health issue among non-Hispanic White boys aged 15–19 years, whereas the WHO scheme would suggest suicide or transportation accidents as top priorities. Similarly, for non-Hispanic Black girls aged 15–19 years, the NCHS scheme identifies accidents as the leading cause, while the WHO scheme highlights homicide as the primary concern.
Compared with the NCHS scheme, the WHO classification provides greater granularity for informing policy decisions. Nonetheless, only a subset of the WHO-defined categories had sufficient deaths to be ranked among the top ten LCODs. For instance, while the WHO scheme disaggregates the broad “accidents” category into seven specific injury types, only two—transportation accidents and drowning—consistently appeared among the top ten across most sex–age–race/ethnicity groups. Poisoning was among the top ten for adolescents aged 15–19 years in all racial and ethnic groups. Similarly, although the WHO scheme separates cancer and heart disease into 17 cancer sites and 8 specific heart conditions, respectively, only two cancer sites (hematopoietic and brain cancers) ranked among the top ten across most groups, and only two heart conditions (pulmonary hypertensive heart disease and cardiomyopathy) appeared in the top ten for selected groups. Thus, information about the relative importance of specific injuries and disease types becomes accessible only when the WHO classification is used.
However, certain NCHS categories offer greater specificity than their WHO counterparts. For example, the NCHS scheme distinguishes five digestive system disease categories, compared with only two under the WHO scheme. Because substantial numbers of deaths are concentrated in broad NCHS categories such as cancer, heart disease, and accidents, subcategories with relatively small counts may still rank among the top ten LCODs. This pattern helps explain why causes such as anemia, tuberculosis, cholelithiasis, legal intervention, and complications of medical and surgical care appeared among the top ten in certain sex–age–race/ethnicity groups under the NCHS classification.
Furthermore, we noted that more strata had COVID-19 ranked among the top ten LCODs in 2022–2023 when the NCHS scheme was applied, compared with the WHO scheme. This finding suggests that the NCHS scheme may be more sensitive in identifying emerging health concerns associated with causes of death that account for relatively small numbers of deaths. Overall, the LCOD profiles based on the NCHS scheme provide information that is distinct from—yet complementary to—that derived from the WHO scheme.
One plausible explanation for the limited number of notable rank changes in LCODs across periods among certain racial and ethnic groups of older adolescents is the substantial mortality gap between the top- and second-ranked causes. When mortality from the leading cause is markedly higher, short-term rank reversals become unlikely. For example, homicide mortality rates—the leading cause—were three to four times higher than those for transportation accidents—the second-ranked cause—among non-Hispanic Black boys aged 15–19 years across all three study periods, making it improbable for transportation accident mortality to surpass homicide between 2018–2019 and 2020–2021. In contrast, the smaller gap between the first and second LCODs among non-Hispanic White and Hispanic boys aged 15–19 years was associated with observed rank changes between 2020–2021 and 2022–2023 (Fig. 1A).
With respect to poisoning mortality, several studies have documented increases in drug-overdose deaths—particularly those involving fentanyl—among U.S. adolescents since the latter half of 2019 (Friedman et al., 2022; Friedman & Shover, 2023; Gaither, 2023; Tanz et al., 2022; Wolf et al., 2024; Zhao et al., 2024; Friedman and Hadland, 2024). However, these studies did not examine changes in mortality through an intersectional lens that considers sex, age, and race/ethnicity simultaneously. Our findings reveal a notable increase in the relative rank of poisoning as a LCOD among non-Hispanic Asian and Hispanic girls aged 15–19 years and among non-Hispanic Asian boys aged 15–19 years. Further research is warranted to elucidate how non-Hispanic Asian and Hispanic older adolescents gain access to counterfeit pills containing fentanyl, as this information is critical for designing culturally and contextually tailored prevention strategies.
Another important policy implication is that persistent increases in the rank of specific LCODs within certain sex–age–race/ethnicity groups—reflecting their growing relative importance—may serve as sentinel indicators of emerging health concerns. Particular attention should be directed toward deaths of undetermined intent among non-Hispanic Black boys aged 15–19 years, transportation accidents among non-Hispanic Asian boys aged 10–14 years, and homicide among Hispanic girls aged 10–14 years. We also observed an atypical rise-and-fall pattern in hematopoietic cancer among non-Hispanic Asian and Hispanic boys aged 10–14 years. Further research is warranted to identify the mechanisms and contextual factors over the six-year study period that contributed to these rank shifts and to inform the design of targeted prevention strategies.
This study has several limitations. First, both the WHO and NCHS schemes consider intent (unintentional, suicide, homicide, and undetermined intent) before mechanism (transportation accidents, falls, firearm-related injuries, suffocation, drowning, and poisoning) for injury classification. As a result, many unintentional injuries may be misclassified as events of undetermined intent, and some suicides may be misclassified as unintentional injuries. To address such misclassification, several studies have adopted broader, mechanism-based definitions that include all intents—such as firearm-related injuries (Lee et al., 2025; Roberts et al., 2023; Wolf et al., 2024) or drug overdose (Friedman et al., 2022; Gaither, 2023)—to more accurately capture the overall injury burden.
Second, the race and ethnicity categories in this study were those provided by CDC WONDER and were not further disaggregated into single-race versus multiple-race groups. Third, some sex–age–race/ethnicity strata contained small numbers of deaths, leading to the exclusion of certain racial groups with low counts (e.g., American Indian or Alaska Native and Native Hawaiian or Other Pacific Islander adolescents), despite their disproportionately high mortality rates. Future research should consider disaggregating national data by state or region to enable more granular analyses. Comparing LCOD ranking profiles across regions could provide valuable insights into emerging health problems that may be shaped by local contextual factors.
The findings illustrate different LCOD ranking profiles among U.S. adolescents depending on whether the WHO or NCHS classification scheme is used. Examining changes in LCOD rankings (relative importance compared with other LCODs) is a preliminary step to identify emerging health issues in specific sex–age–race/ethnicity groups and to guide more detailed investigations. Compared with the NCHS scheme, the WHO scheme generally provides more granular information to guide policy decisions. However, in some instances LCOD profiles derived from the NCHS scheme capture details that the WHO scheme does not. We therefore recommend using both schemes to obtain complementary information to better inform health policy decisions.
Shu-Yu Tai: Writing – original draft, Formal analysis, Data curation, Conceptualization. Ying-Chen Chi: Writing – original draft, Formal analysis, Data curation, Conceptualization. Joh-Jong Huang: Writing – original draft, Formal analysis, Data curation, Conceptualization. Ying-Yeh Chen: Writing – review & editing, Supervision, Project administration. Ichiro Kawachi: Writing – review & editing, Supervision. Tsung-Hsueh Lu: Writing – review & editing, Supervision, Formal analysis, Data curation, Conceptualization.
This study used public available data from CDC WONDER and was approved by the Institutional Review Board of National Cheng Kung University Hospital, Taiwan (approval no. B-EX-112-016).
The authors declare that they have no known competing financial interests or personal relationships that could have influenced the work reported in this paper.