Open Forum Infectious Diseases MAJOR ARTICLE 1,2 2 2 2,3 2,4 Evette Cordoba, Gil Maduro, Mary Huynh, Jay K. Varma, and Neil M. Vora 1 2 Department of Epidemiology, University of North Carolina Gillings School of Global Public Health, Chapel Hill, North Carolina; New York City Department of Health and Mental Hygiene, New 3 4 York, New York; National Center for Emerging and Zoonotic Infectious Diseases and Career Epidemiology Field Officer Program, Office of Public Health Preparedness and Response, Centers for Disease Control and Prevention, Atlanta, Georgia Background. “Pneumonia and influenza” are the third leading cause of death in New York City. Since 2012, pneumonia and influenza have been the only infectious diseases listed among the 10 leading causes of death in NYC. Most pneumonia and influ- enza deaths in NYC list pneumonia as the underlying cause of death, not influenza. We therefore analyzed death certificate data for pneumonia in NYC during 1999–2015. Methods. We calculated annualized pneumonia death rates (overall and by sociodemographic subgroup) and examined the etiologic agent listed. Results. er Th e were 41 400 pneumonia deaths during the study period, corresponding to an annualized age-adjusted death rate of 29.7 per 100 000 population. Approximately 17.5% of pneumonia deaths specified an etiologic agent. Age-adjusted pneumonia death rate declined over the study period and across each borough. Males had an annualized age-adjusted pneumonia death rate 1.5 (95% confidence interval [CI], 1.5–1.5) times that of females. Non-Hispanic blacks had an annualized age-adjusted pneumonia death rate 1.2 (95% CI, 1.2–1.2) times that of non-Hispanic whites. The annualized pneumonia death rate increased with age group above 5–24 years and neighborhood-level poverty. Staten Island had an annualized age-adjusted pneumonia death rate 1.3 (95% CI, 1.2–1.3) times that of Manhattan. In the multivariable analysis, pneumonia deaths were more likely to occur among males, non-His- panic blacks, persons aged ≥65 years, residents of neighborhoods with higher poverty levels, and in Staten Island. Conclusions. While the accuracy of death certificates is unknown, investigation is needed to understand why certain popula- tions are disproportionately recorded as dying from pneumonia in NYC. Keywords. mortality; New York City; pneumonia. Pneumonia is a clinical syndrome characterized by pulmo- definition that is also used nationally, but these statistics have nary infection. There are numerous infectious etiologies of 2 notable limitations for understanding pneumonia burden pneumonia, though an etiology is not identified in the major - in NYC. ity of cases even aer ext ft ensive diagnostic evaluation . An First, the International Classicfi ation of Diseases, Tenth important contributor to pneumonia is infection with influenza Revision (ICD-10), codes that the standard definition uses to virus, a common seasonal pathogen that can cause pneumonia identify pneumonia and influenza (ICD-10 codes J09–J18) are through primary pulmonary infection or secondary bacter- not inclusive of all forms of pneumonia that might be contrib- ial complications . Because of their association, pneumonia uting to the overall burden of pneumonia in NYC. For example, and influenza are oen g ft rouped together when analyzing death newborns can experience pneumonia, yet congenital pneumo- certificate data. nias (ICD-10 codes P23.1–P23.6) are not part of the standard “Pneumonia and influenza” have ranked as the third lead- definition, nor are pneumonias following aspiration events (ICD-10 code J69.0) [4, 5]. Furthermore, this standard defin- ing cause of death in New York City since 1998 . Since 2012, ition does not fully reflect local epidemiologic circumstances pneumonia and influenza have been the only infectious diseases for pneumonia. The pneumonia and influenza diagnoses (ICD- listed among the 10 leading causes of death in NYC; HIV, which 10 codes J09–J18) used by this standard definition notably do was previously among the 10 leading causes of death, has now not include Legionnaires’ disease (ICD-10 code A48.1), a severe fallen from the list . These statistics are generated annually pneumonia caused by Legionella bacteria whose incidence is in NYC using death certificate data processed via a standard rising in NYC and which has been responsible for severe out- breaks in NYC [4, 6–8]. Second, while grouping pneumonia and influenza together Received 14 September 2017; editorial decision 9 January 2018; accepted 15 January 2018. is accurate in some cases, the 2 conditions are not always asso- Correspondence: N. Vora, BA, MD, NYC Department of Health and Mental Hygiene, Division ciated, and combining them might obscure differences in epi- of Disease Control, 42-09 28th St, Long Island City, NY 11101 (firstname.lastname@example.org). Open Forum Infectious Diseases demiology. Within the category pneumonia and influenza Published by Oxford University Press on behalf of Infectious Diseases Society of America 2018. (ICD-10 codes J09–J18), there are substantially more deaths This work is written by (a) US Government employee(s) and is in the public domain in the US. recorded for pneumonia (ICD-10 codes J12–J18) than deaths DOI: 10.1093/ofid/ofy020 Deaths From Pneumonia—New York City, 1999–2015 • OFID • 1 Downloaded from https://academic.oup.com/ofid/article-abstract/5/2/ofy020/4810679 by Ed 'DeepDyve' Gillespie user on 16 March 2018 recorded for influenza (ICD-10 codes J09–J11). For example, number of pneumonia and pneumonia-related deaths, respec- in NYC during 1999–2015, pneumonia was listed as the under- tively, by the corresponding population using intercensal popu- lying cause of death in 99.3% of all pneumonia and influenza lation estimates from the US Census for each year . Second, deaths, whereas influenza was listed as the underlying cause of we multiplied the monthly counts for pneumonia and pneumo- death in only 0.7%. nia-related deaths by 12 to annualize the rates. In both cases, the er Th e are limited published data on the epidemiology of death rates were expressed per 100 000 population. Pneumonia pneumonia in NYC . Understanding the aggregate burden deaths were also categorized into 7 age groups (<5, 5–24, 25–44, of pneumonia in NYC, as well as burden for various socio- 45–64, 65–74, 75–84, ≥85 years). Where applicable, we age-ad- demographic groups, could inform clinical and public health justed the annualized death rates to the 2000 projected US pop- interventions. Thus, we analyzed death certificate data in NYC ulation using direct standardization (see the Supplementary during 1999–2015 in which pneumonia was listed as a cause Methods for details) . of death. We calculated annualized age-adjusted pneumonia and pneumonia-related death rates for each year and by month. We METHODS assumed a Poisson process in the generation of death counts and related standard errors. Rate ratios (RRs) and 95% confi- Data Source dence intervals (CIs) were calculated by sex, age categories, We analyzed death certificate data reported to the NYC race/ethnicity groups, neighborhood-level poverty, and bor- Department of Health and Mental Hygiene (DOHMH) for ough . NYC residents during 1999–2015. Underlying cause of death We assessed neighborhood-level poverty using underlying (the primary condition reported as most responsible for an cause of death data and residential Census tract information individual’s death) and multiple causes of death (all conditions available at the time of death. Neighborhood-level poverty was on the death certificate that were reported to have contributed defined as the percentage of residents with an income below the to an individual’s death) data were included in this analysis. federal poverty level within a defined Census tract . Poverty DOHMH determined that this investigation does not involve data were obtained from the 2000 Census and the American human subjects, and the Centers for Disease Control and Community Surveys from 2005–2014 [12, 13]. Neighborhood- Prevention determined this investigation to be public health level poverty for each Census tract was aggregated and cate- nonresearch. gorized into 4 groups: low (<10% of residents living below the Definitions federal poverty level), medium (10 to <20%), high (20 to <30%), We defined a pneumonia death as a death in which an ICD- and very high (≥30%) . Data before 2000 were excluded 10 code for pneumonia was listed as the underlying cause of from the neighborhood-level poverty analysis due to the lack of death on the death certificate. A pneumonia-related death was consistently constructed Census tract poverty data. one in which an ICD-10 code for pneumonia was listed among We examined accompanying information on death certif- any of the conditions on the death certificate that might have icates, such as place of death, type of institution where death contributed to the individual’s death. ICD-10 codes for pneu- occurred, autopsy performed, tobacco use contributing to monia were identified by searching the ICD-10 diagnostic code death, pregnancy status at time of death, birthplace, premature index for any diagnoses that included the word pneumonia . death (death in a person aged <65 years), and education level. Table 1 lists 41 pneumonia ICD-10 codes identified plus 2 add- Some death certificates had missing information on place of itional codes for Legionnaires’ disease and pneumonitis due to death (0.3%), type of institution where death occurred (0.3%), food and vomit . autopsy performed (5.4%), tobacco use contributing to death Pneumonia and pneumonia-related deaths were subcate- (70.7%), pregnancy status at time of death (52.3% among fe- gorized as having a specified etiology if an etiologic agent was male decedents), birthplace (1.8%), and education level (3.7%). identified on the death certificate (influenza viral pneumonia, For all accompanying information, we calculated the proportion noninfluenza viral pneumonia, bacterial pneumonia, or other of pneumonia deaths of specified and unspecified etiology and specified pneumonia) or as having an unspecified etiology if an used the chi-square test to assess the difference in proportion etiologic agent was not identified. . For pneumonia-related deaths, we identified the 10 lead- Data Analysis ing underlying causes of death listed on the death certificate. The unit of analysis was a person. We calculated the count and We performed 2 multivariable analyses, using a negative bi- percentage of pneumonia and pneumonia-related deaths using nomial regression, to assess the direct effects of demographic underlying cause of death and multiple cause of death data, factors on pneumonia death rate, while adjusting for potential respectively. confounders. Model 1 adjusted for borough and included data Annualized pneumonia and pneumonia-related death rates from 1999–2015; model 2 adjusted for neighborhood-level pov- were derived 2 ways. First, we divided the annual age-specific erty and included data from 2000–2015. We used the negative 2 • OFID • Cordoba et al Downloaded from https://academic.oup.com/ofid/article-abstract/5/2/ofy020/4810679 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Table 1. Number and Frequency of Pneumonia and Pneumonia-Related Deaths—New York City, 1999–2015 ICD-10 Pneumonia Pneumonia Pneumonia- Pneumonia- Pneumonia Disease Category Code Deaths Death, % Related Deaths Related Death, % Total N/A 41 400 100.0 78 351 100.0 Specified N/A 7236 17.48 12 343 15.75 Influenza viral pneumonia N/A 92 0.22 82 0.10 Influenza with pneumonia, influenza virus J10.0 31 0.07 23 0.03 identified Influenza with pneumonia, virus not identified J11.0 61 0.15 59 0.08 Noninfluenza viral pneumonia N/A 98 0.24 163 0.21 Adenoviral pneumonia J12.0 6 0.01 8 0.01 Respiratory syncytial virus pneumonia J12.1 7 0.02 11 0.01 Parainfluenza virus pneumonia J12.2 3 0.01 6 0.01 Human metapneumovirus pneumonia J12.3 1 0.00 2 0.00 Other viral pneumonia J12.8 4 0.01 2 0.00 Viral pneumonia, unspecified J12.9 66 0.16 109 0.14 Varicella pneumonia B01.2 2 0.00 3 0.00 Measles complicated by pneumonia B05.2 0 0.00 0 0.00 Cytomegaloviral pneumonitis B25.0 7 0.02 18 0.02 Congenital pneumonia due to viral agent P23.0 2 0.00 4 0.01 Bacterial pneumonia N/A 6293 15.20 11 623 14.83 Pneumonia due to Streptococcus pneumoniae J13 108 0.26 156 0.20 Pneumonia due to Haemophilus influenzae J14 5 0.01 6 0.01 Pneumonia due to Klebsiella pneumoniae J15.0 426 1.03 728 0.93 Pneumonia due to Pseudomonas J15.1 218 0.53 379 0.48 Pneumonia due to Staphylococcus J15.2 348 0.84 615 0.78 Pneumonia due to Streptococcus, group B J15.3 0 0.00 0 0.00 Pneumonia due to other Streptococci J15.4 104 0.25 209 0.27 Pneumonia due to Escherichia coli J15.5 15 0.04 25 0.03 Pneumonia due to other aerobic Gram- J15.6 47 0.11 81 0.10 negative bacteria Pneumonia due to Mycoplasma pneumoniae J15.7 6 0.01 8 0.01 Other bacterial pneumonia J15.8 74 0.18 139 0.18 Bacterial pneumonia, unspecified J15.9 4873 11.77 9207 11.75 Chlamydial pneumonia J16.0 0 0.00 0 0.00 Legionnaires’ disease A48.1 63 0.15 63 0.08 Congenital pneumonia due to Chlamydia P23.1 0 0.00 0 0.00 Congenital pneumonia due to Staphylococcus P23.2 0 0.00 0 0.00 Congenital pneumonia due to Streptococcus, P23.3 0 0.00 0 0.00 group B Congenital pneumonia due to Escherichia coli P23.4 0 0.00 0 0.00 Congenital pneumonia due to Pseudomonas P23.5 1 0.00 1 0.00 Congenital pneumonia due to other bacterial P23.6 5 0.01 6 0.01 agents Other specified pneumonia N/A 753 1.82 475 0.61 Pneumonia due to other specified infectious J16.8 0 0.00 4 0.01 organisms HIV disease with Pneumocystis carinii pneumonia B20.6 751 1.81 471 0.60 Congenital pneumonia due to other organisms P23.8 2 0.00 0 0.00 Unspecified N/A 34 164 82.52 66 008 84.25 Bronchopneumonia, unspecified J18.0 1788 4.32 5744 7.33 Lobar pneumonia, unspecified J18.1 10 942 26.43 15 818 20.19 Hypostatic pneumonia, unspecified J18.2 2 0.00 78 0.10 Other pneumonia, organism unspecified J18.8 20 0.05 99 0.13 Pneumonia, unspecified J18.9 20 542 49.62 41 291 52.70 Abscess of lung with pneumonia J85.1 18 0.04 81 0.10 Congenital pneumonia, unspecified P23.9 32 0.08 57 0.07 Pneumonitis due to food and vomit J69.0 820 1.98 2840 3.62 Pneumonia deaths: the primary condition reported as most responsible for an individual’s death; pneumonia-related deaths: all conditions reported on the death certificate that contributed to an individual’s death. Abbreviations: ICD-10, International Classification of Diseases, Tenth Revision; N/A, not applicable. Deaths From Pneumonia—New York City, 1999–2015 • OFID • 3 Downloaded from https://academic.oup.com/ofid/article-abstract/5/2/ofy020/4810679 by Ed 'DeepDyve' Gillespie user on 16 March 2018 binomial model because it provided an improved t fi compared from that of non-Hispanic whites (RR, 1.0; 95% CI, 1.0–1.1). with the Poisson regression as well as addressing the issue of Neighborhoods with a very high poverty level had an annual- overdispersion in our data. RRs and 95% CIs were calculated. ized age-adjusted pneumonia death rate 1.6 (95% CI, 1.6–1.6) All analyses were done using SAS 9.4 software. A P value <.05 times that of neighborhoods with a low poverty level. was considered statistically significant. By borough, the highest (38.3; 95% CI, 36.9–39.7; per 100 000 population) and lowest (23.0; 95% CI, 22.5–23.4; per 100 000 RESULTS population) annualized age-adjusted pneumonia death rates occurred in Staten Island and Queens, respectively (Table 2). During 1999–2015, there were 41 400 pneumonia deaths in e Th annualized age-adjusted pneumonia death rate was highest NYC, corresponding to a mean of 2435 pneumonia deaths per in Staten Island at 49.8 (95% CI, 47.6–52.1) per 100 000 popu- year or an annualized age-adjusted pneumonia death rate of lation during 1999–2007, but was later equal to or surpassed 29.7 (95% CI, 29.4–30.0) per 100 000 population. (Table 2) The by the annualized age-adjusted pneumonia death rate of 32.5 annualized age-adjusted pneumonia death rate declined from (95% CI, 31.4–33.6) per 100 000 population in the Bronx dur- 32.3 (95% CI, 31.6–33.0) per 100 000 population during 1999– ing 2008–2015. In contrast, Queens had the lowest annualized 2001 to 24.7 (95% CI, 24.1–25.3) per 100 000 population during age-adjusted pneumonia death rate during both time periods, 2013–2015; Figure 1A shows the trend over the study period. at 23.1 (95% CI, 22.5–23.8) in 1999–2007 and 22.8 (95% CI, This downward trend in annualized age-adjusted death rate was 22.2–23.5) in 2008–2015 (Figure 2). also observed for pneumonia-related deaths, from 67.6 (95% CI, 66.6–68.7) per 100 000 population during 1999–2001 to 50.5 Pneumonia Death Rate by Age Group (95% CI, 49.6–51.3) per 100 000 population during 2013–2015; Annualized pneumonia death rates increased with each age Figure 1B shows the trend over the study period. Pneumonia group above the 5–24-year age group, and most deaths were not deaths (Supplementary Figure 1) and pneumonia-related deaths premature (85.6%). However, persons aged <5 years (RR, 5.3; (data not shown) varied by season, with the highest pneumonia 95% CI, 4.3–6.6) had a higher annualized pneumonia death rate death rates occurring during winter months. compared with persons aged 5–24 years. Etiology of Pneumonia Deaths Multivariable Regression Estimates Pneumonia deaths of specified etiologies accounted for 17.5% Among persons of all age groups, pneumonia death RRs were (n = 7236) of all pneumonia deaths. The remaining 82.5% attenuated after adjusting for sex, race/ethnicity, and borough (n = 34 164) of pneumonia deaths were due to unspecified eti- for model 1 (which adjusted for borough), and were attenuated ologies. Of all deaths with pneumonia as the underlying cause, for those aged ≤64 years after adjusting for sex, race/ethnicity, 15.2% were attributed to bacterial infection, which was the most and neighborhood-level poverty for model 2 (which adjusted common etiologic agent listed among all specified etiologies. Of for neighborhood-level poverty). Additionally, in both multi- the 6293 bacterial pneumonia deaths, 63 (1.0%) were attributed variable models, pneumonia death RRs for non-Hispanic blacks to Legionnaires’ disease (Table 1). The annualized age-adjusted and males were larger compared with unadjusted RRs. In model pneumonia death rate for pneumonias of unspecified etiology 1, both Staten Island and the Bronx had the highest RR (RR, declined from 27.3 (95% CI, 26.6–28.0) per 100 000 popula- 1.2; 95% CI, 1.1–1.4; and RR, 1.3; 95% CI, 1.1–1.4; respectively), tion during 1999–2001 to 20.3 (95% CI, 19.8–20.9) per 100 000 whereas Queens had the lowest RR (RR, 0.8; 95% CI, 0.7–0.9), population during 2013–2015. Conversely, the annualized relative to Manhattan, after controlling for sex, race/ethnicity, age-adjusted pneumonia death rate for pneumonias of speci- and age (Table 3). In model 2, we observed an increasing gra- fied etiology remained statistically (weakly) unchanged from 5.0 dient in RR by neighborhood-level poverty, even after adjusting (95% CI, 4.7–5.3) per 100 000 population during 1999–2001 to for demographic characteristics (Supplementary Table 1). 4.4 (95% CI, 4.2–4.7) per 100 000 population during 2013–2015. Accompanying Data on Death Certificates Pneumonia Death Rate by Sex, Race/Ethnicity, Neighborhood-Level Among persons who experienced a pneumonia death, 34.1% Poverty, and Borough of Residence were born outside of the United States. The majority of pneu- The annualized age-adjusted pneumonia death rate was greater monia deaths occurred during an inpatient hospital stay in males than females (RR, 1.5; 95% CI, 1.5–1.5) (Table 2). (86.7%). More than half of pneumonia deaths (53.8%) occurred Non-Hispanic blacks (RR, 1.2; 95% CI, 1.2–1.2) had signifi- among persons with a high school diploma/general educational cantly higher annualized age-adjusted pneumonia death rates development certificate or some college. An autopsy was per- compared with non-Hispanic whites. Non-Hispanic Asians/ formed in 4.4% of pneumonia deaths (Supplementary Table 2). Pacific Islanders (RR, 0.8; 95% CI, 0.8–0.8) had a significantly Pneumonia-Related Deaths lower annualized age-adjusted pneumonia death rate compared A total of 78 351 pneumonia-related deaths were reported dur- with non-Hispanic whites. The annualized age-adjusted pneu- ing 1999–2015, with 15.8% of these deaths due to pneumonias monia death rate for Hispanics was not significantly different 4 • OFID • Cordoba et al Downloaded from https://academic.oup.com/ofid/article-abstract/5/2/ofy020/4810679 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Deaths From Pneumonia—New York City, 1999–2015 • OFID • 5 Downloaded from https://academic.oup.com/ofid/article-abstract/5/2/ofy020/4810679 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Table 2. Annualized Age-Adjusted Pneumonia Death Rates in Bivariate Analysis—New York City, 1999–2015 Overall Specified Etiology Unspecified Etiology a a a Category No. % Death Rate Rate Ratio 95% CI No. % Death Rate Rate Ratio 95% CI No. % Death Rate Rate Ratio 95% CI Total 41 400 100.0 29.7 - - 7236 17.5 5.2 - - 34 164 82.5 24.5 - - Sex Male 19 041 46.0 37.4 1.5 1.5–1.5 3525 48.7 6.6 1.5 1.5–1.6 15 516 45.4 30.8 1.5 1.5–1.5 Female 22 359 54.0 24.9 Ref Ref 3711 51.3 4.3 Ref Ref 18 648 54.6 20.7 Ref Ref Age category, y <5 201 0.5 2.2 5.3 4.3–6.6 77 1.1 0.8 5.5 3.9–7.8 124 0.4 1.4 5.2 4.0–6.8 5–24 149 0.4 0.4 Ref Ref 55 0.8 0.2 Ref Ref 94 0.3 0.3 Ref Ref 25–44 993 2.4 2.2 5.4 4.6–6.4 452 6.2 1.0 6.7 5.1–8.9 541 1.6 1.2 4.7 3.8–5.8 45–64 4615 11.1 14.2 34.4 29.2–40.5 1234 17.1 3.8 24.9 19.0–32.6 3381 9.9 10.4 39.9 32.5–49.0 65–74 5343 12.9 59.2 143.4 121.9–168.7 1023 14.1 11.3 74.4 56.7–97.6 4320 12.6 47.8 183.8 149.8–225.4 75–84 11 513 27.8 206.1 499.3 424.8–586.9 1820 25.2 32.6 213.8 163.5–279.6 9693 28.4 173.5 666.4 543.9–816.5 ≥85 18 586 44.9 803.9 1947.8 1657.8–2288.6 2575 35.6 111.4 731.1 559.7–954.9 16 011 46.9 692.6 2659.8 2171.7–3257.6 Race/ethnicity Non-Hispanic 2456 5.9 22.6 0.8 0.8–0.8 342 4.7 3.0 0.7 0.6–0.8 2114 6.2 19.6 0.8 0.8–0.9 Asian/ PI Non-Hispanic black 9579 23.1 33.8 1.2 1.2–1.2 2194 30.3 7.4 1.8 1.7–1.9 7385 21.6 26.3 1.1 1.1–1.1 Hispanic 6866 16.6 29.2 1.0 1.0–1.1 1392 19.2 5.5 1.3 1.2–1.4 5474 16.0 23.7 1.0 1.0–1.0 Non-Hispanic white 21 702 52.4 28.3 Ref Ref 3131 43.3 4.2 Ref Ref 18 571 54.4 24.1 Ref Ref Other/unknown 797 1.9 - - - 177 2.4 - - - 620 1.8 - - - Neighborhood-level poverty Low: <10% 11 333 29.3 24.1 Ref Ref 1587 23.4 3.4 Ref Ref 9746 30.6 20.7 Ref Ref Medium: 10–<20% 10 846 28.1 27.8 1.2 1.1–1.2 1802 26.5 4.6 1.3 1.3–1.4 9044 28.4 23.1 1.1 1.1–1.1 High: 20–<30% 7877 20.4 32.3 1.3 1.3–1.4 1472 21.7 5.9 1.7 1.6–1.9 6405 20.1 26.3 1.3 1.2–1.3 Very high: ≥30% 8560 22.2 38.7 1.6 1.6–1.6 1933 28.5 8.5 2.5 2.3–2.6 6627 20.8 30.2 1.5 1.4–1.5 Borough of residence during time of death Bronx 7220 17.4 35.5 1.2 1.1–1.2 1825 25.2 8.9 1.9 1.7–2.0 5393 15.8 26.6 1.0 1.0–1.1 Brooklyn 12 912 31.2 31.4 1.0 1.0–1.1 2299 31.8 5.6 1.2 1.1–1.2 10 613 31.1 25.8 1.0 1.0–1.0 Manhattan 8850 21.4 30.2 Ref Ref 1385 19.1 4.8 Ref Ref 7465 21.9 25.5 Ref Ref Queens 9399 22.7 23.0 0.8 0.7–0.8 1319 18.2 3.2 0.7 0.6–0.7 8080 23.7 19.7 0.8 0.8–0.8 Staten Island 3019 7.3 38.3 1.3 1.2–1.3 408 5.6 5.1 1.1 1.0–1.2 2611 7.6 33.2 1.3 1.2–1.4 Abbreviations: CI, confidence interval; PI, Pacific Islander; Ref, reference. Rate expressed as annualized age-adjusted and age-specific death rate per 100 000 population of the corresponding population per year. Age-specific pneumonia death rates. For neighborhood-level poverty, the number of pneumonia deaths was calculated for only those deaths with available Census tract data during 2000–2015. Overall age-adjusted death rates AB Age-adjusted death 40.0 80.0 rates due to unspeciﬁed etiology Age-adjusted death 35.0 70.0 rates due to speciﬁed etiology 30.0 60.0 25.0 50.0 20.0 40.0 15.0 30.0 10.0 20.0 5.0 10.0 0.0 0.0 Year Year Figure 1. Annual age-adjusted pneumonia and pneumonia-related death rates using (A) underlying cause of death and (B) multiple causes of death data—New York City, 1999–2015. Rate expressed as annual age-adjusted death rate per 100 000 population of the corresponding population per year. of specified etiology. During this period, the 5 most common not routinely presented in mortality reports for NYC . We underlying causes of death among pneumonia-related deaths found that the death rate from pneumonia declined in NYC were (in descending order): chronic ischemic heart disease, during 1999–2015 and that an average of 2400 New Yorkers unspecified dementia, other chronic obstructive pulmonary were reported to have died from pneumonia annually. We disease, malignant neoplasm of bronchus and lung, and stroke, also observed differences in pneumonia death rates by various not specified as hemorrhage or infarction (Supplementary sociodemographic subgroups. Most pneumonia deaths had no Table 3). etiology specified on the death certificate. Health disparities by age, sex, and race/ethnicity have been DISCUSSION previously documented in NYC, and we observed similar dis- We conducted an exploratory analysis of pneumonia deaths parities for pneumonia deaths in our analysis [15–17]. For in NYC using death certificate data, given that pneumonia is example, older adults and males experienced higher pneumo- a leading cause of death in NYC. The data presented here are nia death rates in NYC compared with younger persons and 80.0 Bronx 70.0 Brooklyn Manhattan 60.0 Queens 50.0 Staten Island 40.0 30.0 20.0 10.0 0.0 Year Figure 2. Annual age-adjusted pneumonia death rates by borough of residence—New York City, 1999–2015. Rate expressed as annual age-adjusted death rate per 100 000 population of the corresponding population per year. 6 • OFID • Cordoba et al Downloaded from https://academic.oup.com/ofid/article-abstract/5/2/ofy020/4810679 by Ed 'DeepDyve' Gillespie user on 16 March 2018 Age-adjusted pneumonia death rate per 100 000 population Age-adjusted pneumonia death rate per 100 000 population Age-adjusted pneumonia-related death rate per 100 000 population Table 3. Multivariable Model 1 of Annualized Age-Adjusted Pneumonia Death Rate Ratios, Including Borough—New York City, 1999–2015 Category Parameter Estimate Parameter Standard Error P Rate Ratio 95% CI Intercept –12.5479 0.1152 <.0001 NA NA Sex Male 0.4474 0.0448 <.0001 1.6 1.4–1.7 Female Ref Ref Ref Ref Ref Age category, y <5 1.1360 0.1388 <.0001 3.1 2.4–4.1 5–24 Ref Ref Ref Ref Ref 25–44 1.4363 0.1119 <.0001 4.2 3.4–5.2 45–64 3.2466 0.1068 <.0001 25.7 20.8–31.7 65–74 4.8126 0.1074 <.0001 123.1 99.7–151.9 75–84 6.0887 0.1068 <.0001 440.8 357.6–543.5 ≥85 7.4773 0.1072 <.0001 1767.5 1432.7–2180.7 Race/ethnicity Non-Hispanic Asian/PI –0.3448 0.0669 <.0001 0.7 0.6–0.8 Non-Hispanic black 0.4996 0.0621 <.0001 1.6 1.5–1.9 Hispanic 0.0633 0.0608 .2977 1.1 0.9–1.2 Non-Hispanic white Ref Ref Ref Ref Ref Borough Bronx 0.2263 0.0705 .0013 1.3 1.1–1.4 Brooklyn 0.0381 0.0675 .5720 1.0 0.9–1.2 Manhattan Ref Ref Ref Ref Ref Queens –0.2444 0.0689 .0004 0.8 0.7–0.9 Staten Island 0.1991 0.0763 .0091 1.2 1.1–1.4 Abbreviations: CI, confidence interval; NA, not applicable; PI, Pacific Islander; Ref, reference. females, respectively, even aer ad ft justing for confounding fac- double and triple the rates of Manhattan and Queens, respect- tors, and this is consistent with national data . Similarly, we ively. Aer 2007, t ft he annual age-adjusted pneumonia death rate found that non-Hispanic blacks had a higher pneumonia death in Staten Island began to approach the rates of other boroughs rate compared with non-Hispanic whites (also consistent with and was no longer ranked the highest from 2008–2015. The national data) in both the unadjusted and adjusted models . reason for this trend in Staten Island is unknown. One possi- We also found that neighborhoods characterized by higher bility is that there are systematic cause of death reporting issues levels of poverty had higher pneumonia death rates relative to within medical institutions in different boroughs. Given that neighborhoods characterized by lower levels of poverty, even there are fewer medical institutions located in Staten Island aer ft adjusting for individual sociodemographic characteristics. relative to the other 4 boroughs, any systematic reporting dif- Additional investigation of risk factors at the neighborhood ferences concentrated within Staten Island medical institutions level could potentially reveal some that are modifiable. For ex- would be magnified compared with the other 4 boroughs . ample, air pollution levels in NYC vary by neighborhood; they To our knowledge, there are no data to confirm this hypothesis, are higher in higher-poverty than lower-poverty neighbor- and further investigation is needed. hoods and have been linked to higher hospitalization rates for Similar to prior investigations, we found that the majority respiratory disease and higher risk of pneumonia death [19–21]. of pneumonia deaths were of unspecified etiology, and autop- u Th s, strategies that target neighborhood-level risk factors for sies were performed only rarely, even though >80% of deaths pneumonia deaths such as air pollution might have an indirect occurred while inpatient [1, 5]. Because pneumonia can be impact in NYC. caused by numerous pathogens, identifying the etiologic Variation in pneumonia death rates were also observed by agent can help clinicians tailor patient-specific treatment . NYC borough, with Staten Island having the highest and Queens Additional investigation is needed into the frequency with which having the lowest annualized age-adjusted pneumonia death appropriate microbiologic diagnostics are performed among rates, independent of individual sociodemographic factors. It is patients with pneumonia (such as testing for Legionella and re- notable that Staten Island had a disproportionately high annu- spiratory tract viral pathogens), and linking these data to the alized age-adjusted pneumonia death rate from 1999–2000 and causes of death listed on death certificates. Moreover, knowing 2003–2006 compared with the other 4 boroughs. In 2005, Staten the setting of pneumonia acquisition (community vs hospital) Island reported its highest annual age-adjusted pneumonia can assist public health authorities in developing targeted inter- death rate of 68.3 per 100 000 population, which was nearly ventions to reduce pneumonia deaths . Unfortunately, data Deaths From Pneumonia—New York City, 1999–2015 • OFID • 7 Downloaded from https://academic.oup.com/ofid/article-abstract/5/2/ofy020/4810679 by Ed 'DeepDyve' Gillespie user on 16 March 2018 from death certificates currently do not indicate the setting of potential interventions, and gaps or systemic patterns in pneu- pneumonia acquisition. Taken together, we suggest that future monia cause of death reporting. studies should investigate whether improvements made to iden- Supplementary Data tify etiologic agents and the settings of pneumonia acquisition Supplementary materials are available at Open Forum Infectious Diseases can help to reduce the risk of pneumonia deaths by helping to online. Consisting of data provided by the authors to benefit the reader, tailor patient-specific treatments. the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corre- er Th e are several limitations to our study. First, the data we sponding author. used were limited to data reported on the death certificate and did not involve medical chart review. Thus, we did not analyze Acknowledgments how cases were managed, such as timing of antibiotic admin- Disclaimer. The findings and conclusions in this article are those of the istration, types of antibiotics used, and advanced life support. authors and do not necessarily represent the official position of Department Second, because we did not review medical charts or autopsy of Health and Mental Hygiene or Centers for Disease Control and Prevention. reports, we do not know the accuracy of death certificate data Financial support. This work was supported by the Epi Scholars pro- for pneumonia. In NYC, physician training on completion of gram and the New York City Department of Health and Mental Hygiene. death certificates is not a requirement, and while DOHMH Potential conifl cts of interest. All authors: no reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of does oer a ff n electronic learning module, more comprehensive Potential Conflicts of Interest. Conflicts that the editors consider relevant to education on completion of death certificates would likely be the content of the manuscript have been disclosed. beneficial . Without any required standard training, it is possible that unorthodox practices of death certificate comple- References tion can develop within medical institutions, such as at teach- 1. Jain S, Self WH, Wunderink RG, et al; CDC EPIC Study Team. Community- acquired pneumonia requiring hospitalization among U.S. adults. N Engl J Med ing hospitals where there are clinicians with less experience 2015; 373:415–27. completing death certificates. For example, a study at an NYC 2. Garg S, Jain S, Dawood FS, et al. Pneumonia among adults hospitalized with lab- oratory-confirmed seasonal influenza virus infection United States, 2005–2008. hospital investigating deaths in which unspecified pneumonia BMC Infect Dis 2015; 15:369. (ICD-10 code J18.9) was originally listed as the underlying 3. New York City Department of Health and Mental Hygiene. New York City cause of death found that aer a ft mending the death certificate, summary of vital statistics issues 1990 to 2015. https://www1.nyc.gov/site/ doh/about/about-doh/publications.page?keyword=&topic=NYC Public >90% of these deaths had a change in the assigned underly- Health&title=Summary of Vital Statistics&type=&language=0. Accessed 1 ing cause of death . While this study is not generalizable December 2017. 4. Centers for Disease Control and Prevention. Classification of diseases, function- to all of NYC, it does suggest that hospital-level investigation ing, and disability. ICD-10 (mortality). International Classification of Diseases, can be helpful in examining entrenched systematic bias sur- Tenth Revision (ICD-10). ftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/ Publications/ICD10/allvalid2011%20(detailed%20titles%20headings).pdf. rounding death certificate completion. Another study of the Accessed 30 December 2015. quality of cause of death reporting in NYC showed that >60% 5. Jain S, Williams DJ, Arnold SR, et al; CDC EPIC Study Team. Community- acquired pneumonia requiring hospitalization among U.S. children. N Engl J Med of death certificates listing pneumonia as the underlying cause 2015; 372:835–45. of death were of limited quality . When completing death 6. Farnham A, Alleyne L, Cimini D, Balter S. Legionnaires’ disease incidence and certificates, physicians use their best judgment and the medical risk factors, New York, New York, USA, 2002-2011. Emerg Infect Dis 2014; 20:1795–802. information available at the time of death to work back from 7. The New York City Department of Health and Mental Hygiene. 2015 alert #38: the immediate cause of death to the underlying cause of death Legionnaires’ disease cluster in east Bronx. https://a816-health30ssl.nyc.gov/sites/ NYCHAN/Lists/AlertUpdateAdvisoryDocuments/HAN_Legionella%20East%20 . There are situations in which the medical information Bronx%200928.pdf. Accessed 30 December 2015. initially available may be limited and may result in a cause of 8. The New York City Department of Health and Mental Hygiene. 2017 alert #12: investigation of a cluster of Legionnaires’ disease in the Lenox Hill area of death being listed that may not accurately describe the chain of Manhattan. https://www1.nyc.gov/assets/doh/downloads/pdf/han/alert/legionel- events. Finally, although not a limitation, in conducting this in- la-cluster-lenox-hill.pdf. 9. Corrado RE, Lee D, Lucero DE, et al. Burden of adult community-acquired, vestigation we used an expanded set of ICD-10 codes to define health-care-associated, hospital-acquired, and ventilator-associated pneumonia: pneumonia in such a way that reflects local circumstances (such New York City, 2010 to 2014. Chest 2017; 152:930–42. as Legionnaires’ disease) [6–8]. Because we used an expanded 10. Klein RJ, Schoenborn CA. Age adjustment using the 2000 projected US popula- tion. Healthy People 2010 Stat Notes 2001; (20):1–10. https://www.cdc.gov/nchs/ definition for pneumonia deaths, caution should be taken when data/statnt/statnt20.pdf. Accessed 5 September 2016. comparing the rates presented here with those in other studies 11. Kleinbaum DG, Kupper LL, Muller KE, Nizam A. Applied Regression Analysis and Other Multivariable Methods. 4th ed. Pacific Grove: Duxbury Press; 2008. that use the conventional pneumonia and influenza definition 12. Toprani A, Hadler JL. Selecting and applying a standard area-based socioeconomic (ICD-10 codes J09–J18). The expanded pneumonia diagnoses status meausre for public health data: analysis for New York City. New York, NY: New York City Department of Health and Mental Hygiene: Epi Research Report; 2013. that we included only accounted for 2.0% of the total pneu- 13. Greene SK, Levin-Rector A, Hadler JL, Fine AD. Disparities in reportable com- monia deaths reported in our analysis. municable disease incidence by Census tract-level poverty, New York City, 2006– 2013. Am J Public Health 2015; 105: e27–34. In conclusion, pneumonia is an important cause of reported 14. Gambatese M, Kelley D, Kennedy J, et al. Summary of Vital Statistics 2010 The mortality in NYC. More research into the epidemiology of City of New York. Population and Mortality. NYCDOHMH Bureau of Vital pneumonia in NYC could reveal modifiable risk factors, Statistics. 2011. http://www.nyc.gov/vitalstats. Accessed 9 August 2017. 8 • OFID • Cordoba et al Downloaded from https://academic.oup.com/ofid/article-abstract/5/2/ofy020/4810679 by Ed 'DeepDyve' Gillespie user on 16 March 2018 15. Torian LV, Wiewel EW. Age: continuity of HIV-related medical care, New York 22. New York State. Department of health. Richmond County. https://www.health. City, 2005–2009: do patients who initiate care stay in care? AIDS Patient Care ny.gov/professionals/ems/counties/richmond.htm. Accessed 28 November 2017. STD 2011; 25:79–88. 23. Pilishvili T, Bennett NM. Pneumococcal disease prevention among adults: strate- 16. Izmirly PM, Wan I, Sahl S, et al. The incidence and prevalence of systemic lupus gies for the use of pneumococcal vaccines. Am J Prev Med 2015; 49:S383–90. erythematosus in New York County (Manhattan), New York: the Manhattan 24. Kalil AC, Metersky ML, Klompas M, et al. Management of adults with hospi- Lupus Surveillance Program. Arthritis Rheumatol 2017; 69:2006–17. tal-acquired and ventilator-associated pneumonia: 2016 clinical practice guide- 17. Fei K, Rodriguez-Lopez JS, Ramos M, et al. Racial and ethnic subgroup dispari- lines by the Infectious Diseases Society of America and the American Thoracic ties in hypertension prevalence, New York City health and nutrition examination Society. Clin Infect Dis 2016; 63:e61–111. survey, 2013–2014. Prev Chronic Dis 2017; 14:E33. 25. New York City Department of Health and Mental Hygiene. Improving cause of 18. American Lung Association. Trends in pneumonia and influenza morbidity and death reporting. http://www.nyc.gov/html/doh/media/video/icdr/index.html. mortality. 2015. http://www.lung.org/assets/documents/research/pi-trend-re- Accessed 15 June 2017. port.pdf. Accessed 30 November 2016. 26. Korin L, Das T, Madsen A, Soto A, Begier E. Test of an electronic program to 19. Laumbach RJ, Kipen HM. Respiratory health effects of air pollution: update on query clinicians about nonspecific causes reported for pneumonia deaths, New biomass smoke and traffic pollution. J Allergy Clin Immunol 2012; 129:3–11; York City, 2012. Prev Chonic Dis 2014; 11:140282. quiz 12–3. 27. Falci L, Argov EJL, Van Wye G, Plitt M, Soto A, Huynh M. Examination of 20. New York City Health. Air pollution and the health of New Yorkers: the impact of cause-of death data quality among NYC deaths due to cancer, pneumonia, or dia- fine particles and ozone. http://www1.nyc.gov/assets/doh/downloads/pdf/eode/ betes from 2010 to 2014. Am J Epidemiol 2018; 187:144–152. eode-air-quality-impact.pdf. Accessed 30 November 2016. 28. US Department of Health and Human Services. Physician’s Handbook on Medical 21. Zelikoff JT, Chen LC, Cohen MD, et al. Effects of inhaled ambient particulate mat- Certification of Death 2003 Revision. DHHS Publication No. (PHS) 2003-1108. ter on pulmonary antimicrobial immune defense. Inhal Toxicol 2003; 15:131–50. Hyattsville, MD: Department of Health and Human Services; 2007. Deaths From Pneumonia—New York City, 1999–2015 • OFID • 9 Downloaded from https://academic.oup.com/ofid/article-abstract/5/2/ofy020/4810679 by Ed 'DeepDyve' Gillespie user on 16 March 2018
Open Forum Infectious Diseases – Oxford University Press
Published: Feb 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera