National Health Statistics Reports, Number 167, January 25, 2022

National Health Statistics Reports

Number 167  January 25, 2022

U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES

Centers for Disease Control and Prevention

National Center for Health Statistics

NCHS reports can be downloaded from: https://www.cdc.gov/nchs/products/index.htm.

National Hospital Care Survey Demonstration Projects:

Examination of Inpatient Hospitalization and Risk of

Mortality Among Patients Diagnosed With Pneumonia

Zachary J. Peters, M.P.H., Jill J. Ashman, Ph.D., Alexander Schwartzman, and Carol J. DeFrances, Ph.D.

Abstract

Objective—This report demonstrates the ability of data from the National Hospital

Care Survey (NHCS) linked to the National Death Index (NDI) to provide information

on inpatient hospitalizations and in-hospital and post-acute mortality among patients

hospitalized for a specific condition, in this case pneumonia.

Methods—The 2016 NHCS consists of information on all hospitalizations

provided by participating hospitals. NHCS collects personally identifiable information,

allowing for linkage across hospital settings and to outside data sources, including

NDI; 97 hospitals in the 2016 NHCS contributed qualifying patient records that were

eligible to be linked to NDI. Exploratory analyses were conducted to assess mortality

outcomes, describe cause of death, and model the risk of in-hospital and post-acute

mortality among hospitalized patients diagnosed with pneumonia. NHCS data are

unweighted and are not nationally representative.

Results—Nearly 77,000 adult patients were hospitalized for pneumonia in

the 2016 NHCS and were eligible for linkage to NDI. Overall, 35% of patients

hospitalized with pneumonia died either in the hospital or within 1 year of discharge.

Pneumonia patients who were admitted to the intensive care unit (ICU) and eventually

discharged had a 41% increased risk of death within 30 days of discharge than those

not admitted to ICU, after adjusting for sex, age, and length of stay (adjusted risk

ratio is 1.41, 95% confidence interval is 1.35–1.48).

Conclusion—Although the findings are not nationally representative, this report

demonstrates the potential of NHCS data linked to NDI to identify specific patient

cohorts, assess mortality outcomes, and estimate the risk of in-hospital and post-acute

mortality.

Keywords: inpatients • pneumonia • mortality risk • National Death Index

Introduction

Disease-specific mortality has

decreased among hospitalized patients

over the last 2 decades (1,2), but

information on posthospitalization, or

post-acute, mortality among patients is

limited as most national hospitalization

data sets do not capture subsequent

hospitalizations, emergency department

visits, or mortality outcomes.

To address these limitations, the

National Center for Health Statistics

(NCHS) launched the National Hospital

Care Survey (NHCS), which is designed

to collect inpatient and ambulatory

encounter data from a large sample of

hospitals in the United States. Through

NCHS’ data linkage program and

funding from the Patient-Centered

Outcomes Research Trust Fund, NCHS

has expanded the analytic use of NHCS

by linking it to the National Death Index

(NDI), which provides information such

as date of death and underlying and

multiple causes of death among patients

who died posthospital discharge.

NHCS-participating hospitals provided

data on all hospitalized patients

and patients who made emergency

department (ED) or outpatient

department (OPD) visits for all of 2016.

NHCS data contain the discharge status

for hospitalized patients, allowing for

the assessment of in-hospital mortality.

However, with NHCS data linked to

NDI, patients seeking care for a given

condition in EDs or inpatient settings

can be followed prospectively to assess

subsequent mortality status (and for

deceased patients, the timing and

underlying and multiple causes of death).

Few other data sets have the capability

to assess subsequent post-acute mortality

after a hospitalization. Linking health

care utilization data with mortality data

provides a platform to assess patients’

risk for death, thereby providing

Page 2 National Health Statistics Reports  Number 167  January 25, 2022

information to inform opportunities to

intervene during care delivery, improve

patient monitoring postdischarge, and

consider policies to improve mortality

outcomes.

This report demonstrates the ability

of NHCS-NDI linked data to perform

the following analyses: a) identify

hospitalized patients according to a

specific condition or disease, b) assess

mortality outcomes among those patients

out to 1 year postdischarge, c) evaluate

causes of death among deceased patients,

and d) model risk of death associated

with various patient and hospitalization

characteristics.

Specifically, this report assesses

in-hospital and post-acute mortality for

adult patients diagnosed with pneumonia.

Pneumonia patients were included in

this report as an example of a common

cause of hospitalization; in 2018,

pneumonia was the fourth most common

principal diagnosis among inpatients,

excluding births (3). Additionally, acute

lower respiratory disease (pneumonia

and influenza) routinely is one of

the top 10 leading causes of death in

the United States (4). Several cohort

studies have assessed various aspects

of mortality and mortality risk during

and after hospitalizations for pneumonia

in the United States (5–10), including

a recent study of a large, national,

electronic health record data set that

assessed overall mortality and risk

of mortality among adults in the year

following hospitalization for pneumonia

from 2012–2018 (11). All-cause mortality

within 1 year of admission was 18%

and increased with age and severity of

comorbidities; however, cause-specific

mortality was not assessed, and mortality

status was ascertained from the initial

and subsequent hospital admissions,

likely missing out-of-hospital deaths.

This study seeks to build on findings

from these studies by using a large

sample of hospitalizations linked to

official mortality data to assess overall

and cause-specific mortality, timing of

death, and risk factors for death after

pneumonia-related hospitalizations.

Although NHCS data are unweighted

and are not nationally representative, the

purpose of this report is to demonstrate

the analytical capabilities of NHCS and

NDI linked data set.

Methods

Data sources

2016 National Hospital Care Survey

The target universe for NHCS

is inpatient discharges and in-person

visits made to EDs and OPDs in

noninstitutional and nonfederal hospitals

in the 50 states and the District of

Columbia that have six or more staffed

inpatient beds. Data are extracted from

hospital billing or electronic health

record systems and then transmitted

electronically directly to NCHS or its

designated agent. The 2016 sample

consisted of 581 hospitals: 506 were

acute care hospitals and 75 were other

specialty hospitals, including children’s,

psychiatric, long-term acute care, and

rehabilitation hospitals. Participation in

NHCS is voluntary. Since the launch of

the survey in 2011, the participation rate

has remained relatively low and was 27%

in 2016, with 158 out of the initial 581

sample hospitals participating. Because

of the low participation rate, the data

are unweighted and are not nationally

representative.

Of the 158 participating hospitals,

93% were acute care, with a small

number of psychiatric, children’s,

long-term acute-care, and rehabilitation

hospitals making up the remaining

7% of the sample. This report focused

solely on data from inpatient stays in

the 2016 NHCS. Participating hospitals

were asked to provide data on all

hospitalizations during the 2016 calendar

year, representing an unweighted total

of 2,591,722 hospitalizations. More

information on NHCS methodology is

available elsewhere (12). NHCS collects

personally identifiable information (PII)

(for example, patient name and date of

birth), which allows for both following

patients across hospital settings and

linkage to external data sources.

2016 National Hospital Care Survey

data linked to the 2016–2017

National Death Index

Through its data linkage program,

NCHS has expanded the analytic use

of the data collected from NHCS by

linking it with mortality data from NDI, a

component of the National Vital Statistics

System (13). NDI is a centralized

database of death record information

compiled from state vital statistics

offices. In collaboration with the states,

NCHS established NDI as a resource

for epidemiological follow-up studies

and other types of health and medical

research that require determination of the

mortality status of study subjects. These

mortality data are provided by the states

under contract agreements with NCHS

that specify how these data may be used,

for what purposes, and at what cost.

Currently, NDI contains about 85 million

records from 1979 through 2018 from

the 50 states, the District of Columbia,

Puerto Rico, and the U.S. Virgin Islands.

The linkage between 2016 NHCS

patient records and 2016–2017 NDI

death certificate records was based on

both deterministic and probabilistic

methods. The probabilistic linkage

method performed weighting and link

adjudication as described in the Fellegi–

Sunter paradigm, the foundational

methodology used for record linkage

(14). More information on NHCS-NDI

linkage methodology and the process for

determining patient mortality status is

available elsewhere (15).

Analytic sample

This report includes patients aged 18

and over who were hospitalized with a

pneumonia diagnosis in the 2016 NHCS.

Overall, 158 hospitals participated in the

2016 NHCS, of which 145 submitted

inpatient hospitalization records.

Hospitalizations were excluded from the

analysis if age, sex, date of discharge,

diagnosis code, or discharge code were

missing (27%). Seventeen hospitals did

not submit any inpatient records that

met inclusion criteria and accounted for

26% of all excluded hospitalizations.

The remaining 128 hospitals contributed

qualifying inpatient records with complete

data in these fields, however, 121 of those

hospitals also submitted nonqualifying

records that did not meet inclusion criteria.

Additionally, only patients with

sufficient PII for linkage with NDI

were included. Linkage eligibility was

defined as having usable information for

two of the three data element groups:

Social Security number, name, and date

National Health Statistics Reports  Number 167  January 25, 2022 Page 3

of birth (year, month, day) (15); 97

hospitals contributed inpatient records

that met the definition of a qualifying

hospitalization among patients who were

linkage eligible. For linkage-eligible

patients with more than one qualifying

hospitalization for pneumonia in 2016,

analyses were based on their most

recent hospitalization. Two-thirds of

patients with a qualifying pneumonia

hospitalization (67%) were eligible to

be linked to NDI. Nearly all pneumonia

patients who were ineligible for linkage

were from hospitals that did not provide

any patient records with PII (98%).

Among linkage-eligible patients who

died in the hospital according to their

discharge code, 100% matched to an

NDI record indicating the patient died.

Although the linked NHCS-NDI data set

provides up to 2 years of mortality data,

the current study restricted mortality

assessment postdischarge to 1 year

so that all patients alive after a 2016

hospitalization had an equal follow-up

period in which death could occur.

Inpatient stays for pneumonia

Hospitalizations for pneumonia were

identified as having an International

Classification of Diseases, 10th Revision,

Clinical Modification (ICD–10–CM)

code between J12–J18 listed in any

diagnosis variable. For some NHCS

sampled hospitals it was not possible to

identify the primary diagnosis, and for

other hospitals there were no present

on admission indicators. Therefore, all

available diagnoses were used to identify

patients hospitalized with pneumonia

in this report. Among all pneumonia

hospitalizations with these indicators,

only 22% specified pneumonia as the

primary diagnosis, while 87% indicated

that the pneumonia diagnosis was present

on admission.

Mortality outcomes

Mortality was assessed for

all linkage-eligible patients with a

qualifying hospitalization in 2016.

Mortality outcomes include: in-hospital

death; post-acute deaths occurring

0–30, 31–60, 61–90, and 91–365 days

posthospital discharge; and alive at 366

days postdischarge. In-hospital death

was assigned to patients with an NHCS

discharge code signifying death. Among

patients who were alive at discharge,

post-acute mortality outcomes were

assigned when patients linked to NDI

and NDI death date minus the NHCS

discharge date was less than or equal to

365 days. Patients were treated as alive

after 1 year if their NHCS discharge

status did not indicate in-hospital death

and they did not link to a death record

in the NDI data within 365 days of their

hospital discharge.

Cause of death

Cause of death for in-hospital and

post-acute mortality was assessed using

the 113 Underlying Cause-of-Death

Recodes Adapted for Use by NCHS’

Division of Vital Statistics in NDI

(16). Cause-of-death categories were

included based on their prevalence

among deceased patients in this cohort

(for example, malignant neoplasms,

cardiovascular disease, and chronic lower

respiratory disease) or based on their

relevance to this analysis (for example,

pneumonia). A death with any mention

of the following codes in either the

contributing or underlying cause of death

was categorized as follows:

● Cardiovascular disease recodes

053–069, 071–075

● Chronic lower respiratory disease

recodes 082–086

● Malignant neoplasm recodes

019–043

● Pneumonia recodes 076, 078

All cause-of-death codes not listed

above were grouped into an “other”

category, including other respiratory

diseases (5%; includes influenza,

pneumonitis due to solids and liquids,

and pneumoconiosis), cerebrovascular

disease (4%), septicemia (3%), renal

failure (3%), and diabetes mellitus (3%).

All other specific cause-of-death codes

accounted for 2% or fewer of deaths in

this cohort.

Statistical analysis

Mortality outcomes were assessed

among linkage-eligible patients

hospitalized for pneumonia, overall and

by age cohort, including patients aged

18–44, 45–64, and 65 and over. Cause

of death among deceased patients was

assessed within each timing-of-death

category.

The risk of in-hospital and post-acute

mortality was then modeled using

hospitalization and patient characteristics.

Log-binomial regression was used to

estimate the following associations:

● Risk of in-hospital death by intensive

care unit (ICU) status (admitted to

ICU during hospitalization compared

with not admitted to ICU) among all

patients hospitalized with pneumonia

● Risk of death within 1 year by ICU

status among pneumonia patients

who were discharged from the

hospital.

Lastly, multivariate Poisson

regression with robust error variance was

used, as described by Zou (17) and Knol,

et al. (18), to model the adjusted risk of

death within 30 days among pneumonia

patients who were discharged from the

hospital. This model included independent

variables of age, sex, length of inpatient

stay, and ICU status. Age was treated as

continuous, with patients over age 97 set

equal to 97 years (n = 553, 1%). Length

of stay was treated as continuous with

hospital stays over 60 days set equal to

60 days (n = 480, 1%). Risk ratios (RRs),

standard errors, and 95% confidence

intervals (CIs) are shown.

SAS version 9.4 (SAS Institute,

Cary, N.C.) was used for all data

analyses. These analyses are descriptive

only and are intended to represent

examples of analyses that can be

performed with these data. Any statistical

comparisons were included solely to

illustrate the capabilities of the data set,

not to produce official, representative

estimates for the condition examined.

Statements about the relative magnitude

of numbers or percentages in this report

do not imply statistical significance.

Results

There were 130,657 qualifying

pneumonia hospitalizations at an NHCS

hospital in 2016 among 114,090 patients

(data not shown). Of those, 89,448

qualifying hospitalizations were among

patients who were eligible to be linked to

NDI, resulting in a final cohort of 76,857

patients with a qualifying hospitalization

who were linkage-eligible (Table 1).

More than one-third of linkage-eligible

patients with pneumonia (35%) died

Page 4 National Health Statistics Reports  Number 167  January 25, 2022

during their hospitalization or within 1

year of their hospital discharge (Table 1).

Mortality outcomes

Figure 1 depicts mortality outcomes

among patients with a linkage-eligible

hospitalization for pneumonia.

● Among patients of all ages, 9% died

in the hospital, 9% died 0–30 days

after discharge, 16% died 31–365

days after discharge, and 65% were

alive at 366 days postdischarge.

● The percentage of patients who were

alive 1 year after discharge decreased

with age, with 86% of patients aged

18–44, 73% of patients aged 45–64,

and 57% of patients aged 65 and

over alive at 366 days postdischarge.

● The percentage of patients who

died during a hospitalization for

pneumonia increased with age

(5% of patients aged 18–44, 8% of

patients 45–64, and 11% of patients

65 and over).

Cause of death

Figure 2 presents the cause of death

for pneumonia patients who died in the

hospital or within 1 year of discharge.

● Pneumonia as a cause of death was

more prevalent among patients who

died in the hospital (11%) compared

with patients who died 0–30 days

(3%), 31–60 days (3%), 61–90 days

(3%), and 91–365 days (2%)

postdischarge.

● Cardiovascular disease as a cause

of death increased from 16% of

pneumonia patients who died in

the hospital to 24% of pneumonia

patients who died 90–365 days

postdischarge.

● Similarly, chronic lower respiratory

disease as a cause of death increased

from 7% of pneumonia patients

who died in the hospital to 11% of

pneumonia patients who died 90–365

days postdischarge.

● For all deaths combined, malignant

neoplasm was the most common

underlying cause of death for

pneumonia patients who died in

the hospital or within 1 year of

discharge (26%). This was true for

each timing-of-death category except

91–365 days, where malignant

neoplasms and cardiovascular

disease both accounted for 24% of

deaths occurring at that time.

Risk of death

Table 2 shows the risk of in-hospital

death by ICU status among all patients

hospitalized for pneumonia.

● Pneumonia patients who were

admitted to ICU had nearly four

times the risk of dying in the

hospital than patients who were not

admitted to ICU (RR: 3.77, 95%

CI: 3.57–3.98).

Table 3 shows the mortality risk

within 365 days after discharge by ICU

status among pneumonia patients who

were discharged alive.

Figure 1. Mortality outcomes among hospitalized pneumonia adult patients, by age

Postdischarge.

NOTES: This graph depicts mortality outcomes among 76,857 unique patients hospitalized for pneumonia. Pneumonia was defined by International Classification of Diseases, 10th Revision

diagnosis codes J12.0–J18.9, in any diagnosis field. Mortality outcome estimates within age categories may not add to 100% due to rounding.

SOURCE: National Center for Health Statistics, 2016 National Hospital Care Survey data linked to the 2016–2017 National Death Index.

Percent of patients

0 20 40 60 80 100

65 and over

45–64

18–44

All ages

Alive at 366 days

Died 91–365 days

Died 61–90 days

Died 31–60 days

Died 0–30 days

Died in hospital

National Health Statistics Reports  Number 167  January 25, 2022 Page 5

● Of pneumonia patients who were

discharged alive from the hospital,

those who were admitted to ICU

while hospitalized had a 30%

increased risk of death within 1

year than patients who were not

admitted to ICU (RR: 1.30, 95% CI:

1.27–1.33).

Table 4 depicts adjusted risk of

death within 30 days of discharge among

pneumonia patients who were discharged

from the hospital, modeled by sex, age,

length of stay, and ICU status.

● Male patients had a 21% higher risk

of death within 30 days than female

patients (adjusted RR [aRR]: 1.21,

95% CI: 1.16–1.26), after adjusting

for age, length of stay, and ICU

status.

● For every 1-year increase in age, risk

of death within 30 days of discharge

among pneumonia patients increased

by 4% after adjusting for sex, length

of stay, and ICU status (aRR: 1.04,

95% CI: 1.04–1.04).

● For every 1-day increase in length of

stay, risk of death within 30 days of

discharge among pneumonia patients

increased by 2% after adjusting for

sex, age, and ICU status (aRR: 1.02,

95% CI: 1.02–1.03).

● Patients who were admitted to ICU

while hospitalized for pneumonia

had a 41% higher risk of death

within 30 days than patients who

were not admitted to ICU (aRR:

1.41, 95% CI: 1.35–1.48), after

adjusting for sex, age, and length of

stay.

Summary

This report adds to previous efforts

(19–22) that describe and demonstrate

the capability of NHCS data linked to

NDI to assess relationships between

hospital care experiences and subsequent

mortality. Although descriptive in nature,

these analyses showcase the richness of

these linked data in describing in-hospital

and post-acute mortality for specific

patient populations, as well as assessing

mortality risk based on patient and

hospitalization characteristics.

In this report, the results showed that

over one-third of patients hospitalized

for pneumonia in the 2016 NHCS died in

the hospital or within 1 year of discharge,

which is somewhat higher than other

cohort studies from this timeframe that

measured all-cause mortality among

patients hospitalized for pneumonia

1 year postdischarge (5,11). However,

these studies were also not nationally

representative and had stricter inclusion

criteria for their cohort, such as requiring

a primary pneumonia diagnosis or

radiological evidence of pneumonia.

Additionally, one study did not use linked

mortality data (11) and the other assessed

pneumonia hospitalizations from only

one city in the United States (5), all of

which likely contributed to differences in

mortality rates.

Results in this report also showed

that only 5% of deaths among patients

Figure 2. Causes of death among deceased pneumonia adult patients, by timing of death

Postdischarge.

NOTES: There were 26,962 hospitalized pneumonia patients who died in the hospital or within 365 days of discharge. A pneumonia hospitalization was defined by International Classification

of Diseases, 10th Revision diagnosis codes J12.0–J18.9, in any diagnosis field. Cause of death was determined using underlying cause-of-death recodes (UCOD_113), available from:

https://www.cdc.gov/nchs/data/datalinkage/underlying_and_multiple_cause_of_death_codes.pdf. Cause-of-death estimates within timing-of-death categories may not add to 100% due

to rounding.

SOURCE: National Center for Health Statistics, 2016 National Hospital Care Survey data linked to the 2016–2017 National Death Index.

Percent of deaths

Timing of death

Other

Pneumonia

Chronic lower respiratory

disease

Cardiovascular disease

Malignant neoplasms

31–60 days

61–90 days

91–365 days

All deaths combined

In hospital

0–30 days

0 10 20 30 40 50

Page 6 National Health Statistics Reports  Number 167  January 25, 2022

hospitalized for pneumonia listed

pneumonia as an underlying or

contributing cause of death. Studies

have also found pneumonia to be an

infrequent contributing cause of death

for pneumonia patients who died in

the hospital (23) or postdischarge (24).

Lastly, these linked data allow for the

assessment of risk of death using both

patient and hospitalization characteristics.

For example, pneumonia patients who

were admitted to ICU were four times

more likely to die in the hospital than

patients not admitted to ICU; even among

patients who were discharged alive,

admission to ICU predicted risk of death

within 1 year. This aligns with previous

studies that have shown increased risk

of mortality among pneumonia patients

admitted to ICU (9,25) or as length of

ICU stay increased (8).

Findings should be considered

within the context of this study’s

limitations. First, the 2016 NHCS data

are not nationally representative and are

therefore not generalizable beyond the

responding sample of hospitals. Second,

models of mortality risk included a small

number of predictor variables; other

contributing factors that were not or

could not be assessed in this study may

further explain these results. Additionally,

not all NHCS sampled hospitals provided

data on primary or first-listed diagnoses

or whether diagnoses were present on

admission. Consequently, this study

could neither assess hospitalizations

with pneumonia as a primary diagnosis

nor distinguish between hospitalizations

for community-acquired compared with

hospital-acquired pneumonia. Lastly,

this study lacked the ability to examine

findings by patient race and ethnicity,

as the 2016 NHCS did not collect this

information.

NCHS expects the number hospitals

participating in NHCS to increase in

future years, as it moves toward collecting

patient data through electronic health

records. As participation in NHCS

grows and these data become more

representative, data linked to NDI will

provide more comprehensive assessment

of the relationships between hospital care

utilization and mortality among various

patient cohorts.

About the authors

Zachary J. Peters, Jill J. Ashman,

Alexander Schwartzman, and Carol J.

DeFrances are with the National Center

for Health Statistics, Division of Health

Care Statistics.

References

1. Hines AL, Heslin KC, Jiang HJ,

Coffey R. Trends in observed adult

inpatient mortality for high-volume

conditions, 2002–2012. Healthcare

Cost and Utilization Project Statistical

Brief, no 194. 2015.

2. Hall MJ, Levant S, DeFrances C.

Trends in inpatient hospital deaths:

National Hospital Discharge Survey,

2000–2010. NCHS Data Brief, no

118. Hyattsville, MD: National Center

for Health Statistics. 2013.

3. Agency for Healthcare Research

and Quality. Healthcare cost and

utilization project fast stats—Most

common diagnoses for inpatient stays.

2021. Available from: https://www.

hcup-us.ahrq.gov/faststats/NationalD

iagnosesServlet?year1=2018&charac

teristic1=0&included1=0&year2=&c

haracteristic2=0&included2=1&expa

nsionInfoState=hide&dataTablesState

=hide&definitionsState=hide&export

State=hide.

4. Xu JQ, Murphy SL, Kochanek KD,

Arias E. Mortality in the United

States, 2018. NCHS Data Brief, no

355. Hyattsville, MD: National Center

for Health Statistics. 2020.

5. Ramirez JA, Wiemken TL, Peyrani P,

Arnold FW, Kelley R, Mattingly WA, et

al. Adults hospitalized with pneumonia

in the United States: Incidence,

epidemiology, and mortality. Clin Infect

Dis 65(11):1806–12. 2017.

6. Ruhnke GW, Coca-Perraillon M,

Kitch BT, Cutler DM. Marked

reduction in 30-day mortality among

elderly patients with community-

acquired pneumonia. Am J Med

124(2):171–8.e1. 2011.

7. Falcone M, Tiseo G, Russo A, Giordo

L, Manzini E, Bertazzoni G, et al.

Hospitalization for pneumonia is

associated with decreased 1-year

survival in patients with Type 2

Diabetes: Results from a prospective

cohort study. Medicine 95(5):e2531.

2016.

8. Moitra VK, Guerra C, Linde-Zwirble

WT, Wunsch H. Relationship between

ICU length of stay and long-term

mortality for elderly ICU survivors.

Crit Care Med 44:655–62. 2016.

9. Restrepo MI, Mortensen EM,

Rello J, Brody J, Anzueto A. Late

admission to the ICU in patients with

community-acquired pneumonia is

associated with higher mortality.

Chest 137(3):552–7. 2010.

10. Peyrani P, Arnold FW, Bordon J,

Furmanek S, Luna CM, Cavallazzi R,

Ramirez J. Incidence and mortality of

adults hospitalized with community-

acquired pneumonia according to

clinical course. Chest 157:34–41.

2020.

11. Averin A, Shaff M, Weycker D,

Lonshteyn A, Sato R, Pelton SI.

Mortality and readmission in the

year following hospitalization for

pneumonia among US adults. Respir

Med 185:106476. 2021.

12. National Center for Health Statistics.

National Hospital Care Survey.

Available from: https://www.cdc.gov/

nchs/nhcs/about_nhcs.htm.

13. National Center for Health Statistics.

National Death Index. Available from:

https://www.cdc.gov/nchs/ndi/index.

htm.

14. Fellegi IP, Sunter AB. A theory for

record linkage. J Am Stat Assoc

64(328):1183–210. 1969.

15. National Center for Health

Statistics. Office of Analysis and

Epidemiology. The linkage of the

2016 National Hospital Care Survey

to the 2016/2017 National Death

Index: Methodology overview and

analytic considerations, January 2019.

Hyattsville, Maryland. Available

from: https://www.cdc.gov/nchs/data/

datalinkage/NHCS16_NDI16_17_

Methodology_Analytic_Consider.pdf.

16. National Center for Health Statistics.

Underlying and multiple cause

of death codes. 2018. Available

from: https://www.cdc.gov/nchs/

data/datalinkage/underlying_and_

multiple_cause_of_death_codes.pdf.

17. Zou G. A modified Poisson regression

approach to prospective studies with

binary data. Am J Epidemiol 159(7):

702–6. 2004.

National Health Statistics Reports  Number 167  January 25, 2022 Page 7

18. Knol MJ, Le Cessie S, Algra A,

Vandenbroucke JP, Groenwold RHH.

Overestimation of risk ratios by odds

ratios in trials and cohort studies:

Alternatives to logistic regression.

CMAJ 184(8):895–9. 2012.

19. Spencer MR, Flagg LA, Jackson G,

DeFrances C, Hedegaard H. National

Hospital Care Survey Demonstration

Projects: Opioid-involved emergency

department visits, hospitalizations,

and deaths. National Health Statistics

Reports; no 141. Hyattsville, MD:

National Center for Health Statistics.

2020.

20. National Center for Health Statistics

Research Data Center. Linked data on

hospitalizations, mortality, and drugs:

Data from the National Hospital Care

Survey 2016, National Death Index

2016–2017, and the Drug-Involved

Mortality 2016–2017. 2020. Available

from: https://www.cdc.gov/nchs/data/

nhcs/Task-3-Doc-508.pdf.

21. Jackson G, Brown AM, DeFrances C.

Opioid-involved emergency department

visits in the National Hospital Care

Survey and the National Hospital

Ambulatory Medical Care Survey.

National Health Statistics Reports; no

149. Hyattsville, MD: National Center

for Health Statistics. 2020.

22. Ashman JJ, DeFrances CJ, Cairns C,

Schwartzman A. Respiratory illness

emergency department visits in the

National Hospital Care Survey and

the National Hospital Ambulatory

Medical Care Survey. National Health

Statistics Reports; no 151. Hyattsville,

MD: National Center for Health

Statistics. 2021.

23. Stefan MS, Jaber R, Lindenauer PK,

Garb JL, Fitzgerald J, Rothberg MB.

Death among patients hospitalized

with pneumonia: Implications for

hospital outcome measures. JAMA

Intern Med 175(5):851–3. 2015.

24. Bruns AHW, Oosterheert JJ,

Cucciolillo MC, El Moussaoui

R, Groenwold RHH, Prins JM,

Hoepelman AIM. Cause‐specific

long‐term mortality rates in patients

recovered from community‐acquired

pneumonia as compared with the

general Dutch population. Clin

Microbiol Infect 17(5):763–8. 2011.

25. Mortensen EM, Restrepo MI,

Anzueto A, Pugh JA. Antibiotic

therapy and 48-hour mortality for

patients with pneumonia. Am J Med

119(10):859–64. 2006.

Page 8 National Health Statistics Reports  Number 167  January 25, 2022

Table 1. Criteria for dening qualifying pneumonia hospitalizations and linkage-eligible patients

Characteristic N Percent

Hospitalizations, total and by compounding inclusion criteria

Total hospitalizations in 2016 National Hospital Care Survey . . . . . . . . . . . . . . . . . . . . . . . . . 2,591,722 100

With a pneumonia diagnosis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159,945 6

Among patients aged 18 and over . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144,317 6

With complete sex and discharge data (qualifying hospitalizations) . . . . . . . . . . . . . . . . . . 130,657 5

Among patients eligible to be linked to the National Death Index . . . . . . . . . . . . . . . . . . . . 89,448 4

Patients, total and by mortality status

Total linkage-eligible patients with a qualifying pneumonia hospitalization. . . . . . . . . . . . . . . 76,857 100

Who died in hospital or within 1 year of discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26,962 35

Who were alive 1 year after discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49,895 65

NOTES: Table 1 shows the progression of inclusion criteria applied to all 2016 National Hospital Care Survey hospitalizations. There were 89,448 qualifying pneumonia hospitalizations among 76,857

linkage-eligible patients. For patients who had more than one qualifying hospitalization, their most recent qualifying hospitalization was used for analysis.

SOURCE: National Center for Health Statistics, 2016 National Hospital Care Survey data linked to the 2016–2017 National Death Index.

Table 2. Risk of in-hospital death, by Intensive Care Unit status among all pneumonia patients

ICU admission

In-hospital death

Risk Risk ratio Standard error

95% condence

intervalYes No

Yes . . . . . . . . . . . . . . . . . . . . . . 5,649 31,315 0.15 3.77 0.10 3.57–3.98

No. . . . . . . . . . . . . . . . . . . . . . . 1,617 38,276 0.04 … … …

… Category not applicable.

NOTES: ICU is intensive care unit. There were 76,857 linkage-eligible patients with a qualifying pneumonia hospitalization. Pneumonia was dened by International Classication of Diseases, 10th

Revision diagnosis codes between J12.0–J18.9, in any diagnosis eld.

SOURCE: National Center for Health Statistics, 2016 National Hospital Care Survey data linked to the 2016–2017 National Death Index.

Table 3. Risk of death within 365 days, by Intensive Care Unit status among discharged pneumonia patients

ICU admission

Post-acute death

Risk Risk ratio Standard error

95% condence

intervalYes No

Yes . . . . . . . . . . . . . . . . . . . . . . 10,150 21,165 0.32 1.30 0.02 1.27–1.33

No. . . . . . . . . . . . . . . . . . . . . . . 9,546 28,730 0.25 … … …

… Category not applicable.

NOTES: ICU is intensive care unit. There were 69,591 linkage-eligible patients with a qualifying pneumonia hospitalization who were discharged alive. Pneumonia was dened by International

Classication of Diseases, 10th Revision diagnosis codes between J12.0–J18.9, in any diagnosis eld.

SOURCE: National Center for Health Statistics, 2016 National Hospital Care Survey data linked to the 2016–2017 National Death Index.

Table 4. Risk of death within 30 days among discharged pneumonia patients

Predictor Adjusted risk ratio Standard error

95% condence

interval

Sex (male) . . . . . . . . . . . . . . . . . . . . . . 1.21 0.03 (1.16–1.26)

Age †‡ . . . . . . . . . . . . . . . . . . . . . . . . . 1.04 0.00 (1.04–1.04)

Length of stay†§ . . . . . . . . . . . . . . . . . . 1.02 0.00 (1.02–1.03)

ICU admission (yes) . . . . . . . . . . . . . . . 1.41 0.05 (1.35–1.48)

Intercept . . . . . . . . . . . . . . . . . . . . . . . . -5.34 0.07 …

† Continuous predictor variable.

‡ Patients over age 97 were set equal to 97 years (553 patients, 1%).

§ Hospitalization stays longer than 60 days were set equal to 60 days (480 patients, 1%).

… Category not applicable.

NOTES: ICU is intensive care unit. There were 69,591 linkage-eligible patients with a qualifying pneumonia hospitalization who

were discharged alive. Pneumonia was dened by International Classication of Diseases, 10th Revision diagnosis codes J12.0–

J18.9, in any diagnosis eld. Risk ratios for individual predictors are adjusted for all other variables in this table.

SOURCE: National Center for Health Statistics, 2016 National Hospital Care Survey data linked to the 2016–2017 National Death Index.

National Health Statistics Reports  Number 167  January 25, 2022

For more NCHS NHSRs, visit:

https://www.cdc.gov/nchs/products/nhsr.htm.

FIRST CLASS MAIL

POSTAGE & FEES PAID

CDC/NCHS

PERMIT NO. G-284

U.S. DEPARTMENT OF

HEALTH & HUMAN SERVICES

Centers for Disease Control and Prevention

National Center for Health Statistics

3311 Toledo Road, Room 4551, MS P08

Hyattsville, MD 20782–2064

OFFICIAL BUSINESS

PENALTY FOR PRIVATE USE, $300

Acknowledgments

The authors thank Christine Cox, Linda Hermer, and Shawn Linman for their assistance in the early phase of this project.

Suggested citation

Peters ZJ, Ashman JJ, Schwartzman A,

DeFrances CJ. National Hospital Care Survey

demonstration projects: Examination of

inpatient hospitalization and risk of mortality

among patients diagnosed with pneumonia.

National Health Statistics Reports; no 167.

Hyattsville, MD: National Center for Health

Statistics. 2022.

DOI: https://dx.doi.org/10.15620/cdc:112080.

All material appearing in this report is in

the public domain and may be reproduced

or copied without permission; citation as to

source, however, is appreciated.

National Center for Health Statistics

Brian C. Moyer, Ph.D., Director

Amy M. Branum, Ph.D., Associate Director for

Science

Division of Health Care Statistics

Carol J. DeFrances, Ph.D., Acting Director

Alexander Strashny, Ph.D., Associate Director

for Science

For e-mail updates on NCHS publication releases, subscribe online at: https://www.cdc.gov/nchs/email-updates.htm.

For questions or general information about NCHS: Tel: 1–800–CDC–INFO (1–800–232–4636) • TTY: 1–888–232–6348

Internet: https://www.cdc.gov/nchs • Online request form: https://www.cdc.gov/info • CS328331