| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the City of Chicago Department of Public Health.
Correspondence to: Sandra D. Thomas, MD, MS, Epidemiology Program, Chicago Department of Public Health, 333 S State St, Room 2136, Chicago, IL 60604; e-mail: sandrathomas{at}worldnet.att.net
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
Design: Cross-sectional analysis of discharge data for 1996 and mortality time trend data for the period from 1990 to 1997.
Setting: The city of Chicago, IL, with Cook County, IL, and US data employed for comparisons.
Population studied: People who were hospitalized with a primary diagnosis of asthma and people whose underlying cause of death was asthma.
Interventions: None.
Measurements and results: The 1996 asthma hospitalization rate for Chicago was 42.8 per 10,000, more than twice as high as suburban Chicago or US rates. Medicaid patients were overrepresented. Length of stay was longer for older patients and Medicaid patients. Age-adjusted asthma mortality in Chicago was 4.7 times higher in non-Hispanic blacks than in non-Hispanic whites. The black/white asthma mortality ratio is 2.5:1 for the nation overall. Asthma mortality rates for Hispanics in Chicago were between those of non-Hispanic whites and blacks but have almost doubled during this decade.
Conclusions: The rising asthma mortality and high asthma hospitalization rates in Chicago constitute a significant public health problem. Comorbidities more common in urban environments, such as substance abuse, may play a unique role in determining the distribution of adverse outcomes within Chicago's population. Asthma hospitalizations and deaths may vary in their risk profiles, and this should be taken into account when developing research and intervention strategies.
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
Data Sources
Hospitalization data for Chicagoans and non-Chicago
residents of Illinois were obtained from the 1996 Research-Oriented
Data Set of the Illinois Health Care Cost Containment Council (IHCCCC)
in Springfield, IL. The IHCCCC collects and disseminates information on
virtually all acute care inpatient hospitalizations in Illinois. The
1996 data set contains all reported hospital discharges between January
1, 1996, and December 31, 1996. The information includes primary and
secondary diagnoses, residence, charges, payment sources, and lengths
of stay. There are no data available on race or ethnicity, and there
are no unique patient identifiers—observations are reported per
hospitalization event, not per patient.
Hospitalization data for the United States were obtained from published reports from the National Hospital Discharge Survey,4 which is administered by the National Center for Health Statistics in Hyattsville, MD. The National Hospital Discharge Survey is a multistage probability survey of hospital discharges from short-stay, acute-care hospitals in the United States.
Mortality data for Chicago residents were obtained from the electronic death files for 1990 through 1997, supplied by the Illinois Department of Public Health (IDPH) in Springfield, IL, as part of its vital records system. Data available include place of death, underlying cause of death, contributing causes of death, residence, age, race, and Hispanic ethnicity. Deaths occurring from January 1, 1990, to December 31, 1997, were included in the analysis.
Mortality data for the United States from 1990 to 1995 were obtained from published reports from the Centers for Disease Control and Prevention.5 The Centers for Disease Control and Prevention obtains mortality data from the underlying cause-of-death data set, which is part of the national vital records system.
Population data for Chicago were extracted from the 1990 US Census.6 Data for the Chicago Hispanic population were supplemented by 1998 projections supplied to the Chicago Department of Public Health by Claritas Data Services (Claritas Inc; Ithaca, NY).
Analysis
Asthma hospitalizations were defined as any hospital discharge
with a primary diagnosis code of 493 as given in the ninth edition of
the International Classification of Diseases-Clinical
Modification (ICD-9-CM).
Age is not supplied in the hospital data set, so age for hospitalizations was defined to be age at discharge, and was calculated by subtracting the date of birth from the discharge date.
To examine seasonal trends in asthma hospitalizations, the actual number of hospitalizations per month was compared with the expected number of hospitalizations per month (the number of hospitalizations that would occur in a given month if hospitalizations were equally distributed throughout the year). The daily hospitalization rate for the calendar year was calculated from the actual number of hospitalizations, and the expected number of hospitalizations per month was defined as the daily hospitalization rate times the number of days in the month.
The IHCCCC data set contains listings for up to eight secondary diagnoses per hospitalization. These diagnoses are listed by ICD-9-CM codes. A method of summarizing these codes for reporting was developed by the authors. The ICD-9-CM codes for the second listed diagnosis were examined, and categories of three-digit ICD-9-CM codes were developed from the most frequently occurring ICD-9-CM codes (codes that were in the top 50% of the frequency distribution). This selective approach was used because of the high number of individual diagnostic codes. These groupings were based on physiologic and clinical similarities of the diagnoses. The frequency of these categories among all secondary diagnosis listings was then calculated. A hospitalization was considered positive for a given secondary diagnosis category when a qualifying diagnostic code occurred in any of the secondary diagnosis fields.
Asthma deaths were defined as any deaths with the underlying cause of death attributed to ICD-9-CM code 493. The underlying cause of death for all death certificates is determined by the IDPH, according to standard nosologic criteria.
The IDPH electronic death files contain listings for up to four contributing causes of death. These diagnoses are listed by ICD-9-CM code. A method of summarizing these codes for reporting was also developed by the authors. The ICD-9-CM codes for all contributing causes of death were examined, and categories of three-digit ICD-9-CM codes were developed. These groupings were based on physiologic and clinical similarities of the diagnoses. Identical categorizations were used for ICD-9-CM codes that occurred in both the hospitalization and mortality datasets. The frequency of these categories among all contributing causes of death was then calculated. A death was considered positive for a given secondary diagnosis category when a qualifying diagnostic code occurred in any of the contributing cause of death fields.
Targonski et al2
indicated that asthma mortality rates in
Chicago were not constant over time, so we were interested in examining
time trends in the data. Published national data were available for
1990 to 1995, and were grouped in 3-year intervals (1990 to 1992 and
1993 to 1995).5
Using intervals that exactly corresponded
to those of Mannino et al5
would have significantly
reduced the numerators for the Chicago-specific calculations and made
the rates more unstable, so Chicago mortality rates were evaluated over
4-year periods (1990 to 1993 and 1994 to 1997). Chicago rates were
compared with the closest corresponding national time period: US data
for 1990 to 1992 for the 1990-to-1993 Chicago data, and US data for
1993 to 1995 for the 1994-to-1997 Chicago data. The data in Mannino et
al5
indicate that the total, age-specific, and
race-specific national death rates have changed relatively slowly in
the 1990s (ie, 0.3 to 5% per year); therefore,
reaggregation of the national data to correspond with the time periods
used to analyze the Chicago data would have affected the reported
average annual rates by 
5%. We considered this level of accuracy
to be sufficient for the purposes of this overview, as the construction
of confidence intervals and statistical testing are not being
performed. Chicago data were aggregated into four age groups that
corresponded with those used by Mannino et al5
to report
national asthma mortality data.
Chicago has a sizable Hispanic population, many of whom also self-identify racially as either white or black. The printed listing of US Census data referenced for this analysis6 did not contain simultaneous listing of both race and ethnicity by age group. The non-Hispanic white and non-Hispanic black populations for age and race mortality calculations were determined by calculating the estimated number of Hispanics in each age group for blacks and whites, and subtracting this correction factor from the 1990 Census totals for each age group for black and white race. Evidence suggests that Chicago's Hispanic population has undergone sizable growth since the 1990 US Census, so an intercensus estimate of the Hispanic population was used for determining Hispanic rates during the time period of 1994 to 1997.
Since we did not start with a priori hypotheses, we do not report statistical tests of significance for the comparisons reported below.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
In a pattern similar to national data,4
hospitalizations
for Chicago residents showed a complex relationship with age, as shown
in Table 1
. The median age at discharge for Chicago residents was 27 years. Ten
percent of asthma hospitalizations for Chicago residents occurred in
persons < 21 months old, and 10% occurred in persons aged
66 years. Female patients had a higher hospitalization rate
than male patients, which is also consistent with national
data.4
|
|
The mean length of stay for Chicago asthma hospitalizations was 3.5
days, which is similar to the 1995 national mean of 3.7 days. Mean
length of stay for persons of age 65 years was nearly twice that of
the 18- to 34-year-old age group (5.5 vs 2.6 days). Length of stay was
also examined by payer status for persons aged < 65 years who were
covered by either Medicaid or commercial insurance (Table 1)
. Length of
stay was longer for Medicaid recipients than patients with commercial
insurance.
Comorbid conditions were present in the majority of asthma hospitalizations (68%). This was true for all ages, although the prevalence of any comorbidity and the mean number of comorbidities per hospitalization both increased with age. The three most frequently occurring secondary diagnosis categories are presented by age in Table 2 . Infectious disease is a common comorbidity in younger patients, while disorders of the cardiovascular system are the predominant source of comorbidity in older patients. Substance abuse is a prominent contributor to comorbidity in early and middle adulthood.
|
65 were 10 times more likely to die during hospitalization
(19 deaths; 1.4%) than patients < 35 years of age (five deaths;
0.11%).
Mortality
More than 950 deaths among Chicago residents were attributed to
asthma between 1990 and 1997, for an average of nearly 120 deaths per
year. Mortality rates for Chicago and the United States are presented
by selected demographic categories in Table 3 . Chicago mortality rates exceed those of the nation in every age
category and during both time intervals. Chicago's asthma mortality
rates have declined during the 1990s for younger children (< 15
years), but have continued to rise for adolescents and adults. While
white asthma mortality has risen at a faster rate than black asthma
mortality in the 1990s, the age-adjusted black/white mortality ratio is
still far higher in Chicago than in the United States overall: 4.7 for
Chicago from 1994 to 1997, compared with 2.5 for the United States from
1993 to 1995.
|
34 years.
|
35 years, women have higher mortality rates than
men. Data on Hispanic ethnicity have been available on Illinois death certificates since 1989. National mortality data were not reported for Hispanic ethnicity in Mannino et al.5 The age-adjusted (to the 1970 US population) average annual asthma mortality for Chicago Hispanics was 14.3 per million population from 1990 to 1993 (23 total deaths), and 25.7 per million from 1994 to 1997 (49 total deaths). This latter rate is somewhat higher than the non-Hispanic white rate (17.5), but much lower than the non-Hispanic black rate (82.5).
Comorbid conditions were not noted for the majority of asthma deaths in
younger people, but their prevalence increased with age. Less than 20%
of the deaths in people < 25 years of age had any comorbidities,
while 90% of the deaths in people 85 years of age had documented
comorbidities. Besides having a higher rate of any comorbidity, older
people were more likely to have complex patterns of comorbidity
(eg, contributing causes in both parts I and II of the death
certificate). The rank of order of comorbidities by age is presented in
Table 5
. Substance abuse was one of the most frequent contributing causes of
death in people < 65 years. Concurrent obstructive pulmonary disease
appears more frequently as a comorbidity in asthma deaths than in
hospitalizations. Although the presence of concurrent obstructive
pulmonary disease increases with age, the majority of asthma deaths,
even in the oldest age group, did not have an associated diagnosis of
other obstructive lung conditions.
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
Although Chicago has a higher asthma hospitalization rate than the nation overall, we were unable to detect evidence for unique environmental or demographic risk factors in Chicago that would account for the higher rates (although we were limited in the number of factors that we could investigate). The age distribution of Chicago hospitalization rates are concordant with national data. The seasonality distribution is also concordant with national data, and does not suggest any unique environmental exposures.
The concentration of asthma hospital discharges among Medicaid recipients agrees with findings by other researchers studying populations in California, Maryland, and New York City, NY.9 10 11 The robustness of this finding defies a unidimensional explanation. The combination of higher asthma prevalence among persons of lower socioeconomic status12 and the higher percentage of children in the Medicaid-eligible population compared with the general population may explain some of the observed findings. Wissow et al9 found, however, that even in children within the same racial groups, Medicaid recipients had a higher asthma discharge rate than children with other forms of insurance. Examination of all hospital discharges by Carr et al10 and selected common diagnoses by Wissow et al9 also found that the association of Medicaid eligibility status with hospitalization may be greater in asthma than in other illnesses. The discharge rates in this paper do not reflect the experience of individuals. It is possible that similar percentages of persons are being hospitalized in the Medicaid and non-Medicaid populations, but that the Medicaid patients are more likely to have multiple hospitalizations in a year.
We do not have hospitalization comorbidity data on other populations for comparison purposes, but the prominence of comorbidities in Chicago asthma hospitalizations underscores the need for quality primary care for asthma patients. The frequency of substance abuse as a comorbidity is disturbing. Further research is needed to determine the best methods for categorizing comorbidities, and which of the common comorbidities are most specific to asthma hospitalization risk.
In contrast to the hospitalizations, Chicago's asthma mortality shows marked differences from national patterns, as well as from prior Chicago data. Our data show stable or declining mortality in ages 0 to 14 years, while national data show rises in mortality for all age groups throughout the decade. Chicago's child asthma mortality is still above national rates, however, and the positive trends in this age group were more than offset by the steeper than average rises in adult mortality during this time. Earlier Chicago data from Targonski et al2 found a rise in black mortality while white mortality remained stable. In this decade, we have found a concordance in mortality trends between blacks and whites. In spite of this, Chicago still has the highest black-to-white asthma mortality ratios documented in the literature, to our knowledge.2 6 9 13 Poverty, racism, and/or lack of access to care may play a role in these intergroup differences, although it should be noted that Chicago's interracial differences in the percent of the population below the poverty line are very similar to national data, for both children and adults. Notably, any explanation of Chicago's high interracial mortality ratio for asthma must also take into account that interracial differences in mortality are much lower for other chronic diseases in Chicago, which are usually of the magnitude of 2:1.14 Furthermore, some causes of death with strong social determinants, such as infant mortality, have declined dramatically during the decade in Chicago.14
Even after adjusting for the expansion in Hispanic population in Chicago during the 1990s, Hispanic asthma mortality in Chicago appears to have increased at a higher rate than for blacks or whites. Although this might be an artifact of the instability of rates based on small numbers of events, or differences in ethnicity reporting on the death certificate, it warrants close monitoring. There are relatively few epidemiologic data available for Hispanic asthma hospitalizations and mortality.10 11 13 15 16 We hope the data reported here will stimulate additional investigations in this area.
Limitations
We do not have data available on hospitalizations for Chicago
residents that occurred in states other than Illinois, so the
hospitalization rates presented here may be slightly underestimated. We
only had 1 year of data available for analysis of hospitalizations, so
we are not able to evaluate any trends or the representativeness of our
findings. The lack of individual patient identifiers in the records
makes it difficult to correlate hospitalization rates with risks of
hospitalization for individuals in the population, as a patient can
have more than one hospitalization in a given year. This also makes it
more difficult to interpret intergroup differences in hospitalization
rates, as the distribution of hospitalizations per patient will be
skewed toward high utilizers, and group differences in outcomes may be
driven by a small minority of exceptional individuals rather than group
characteristics.
Race and ethnicity data are not available for hospitalizations in Illinois, and insurance coverage or other surrogate measures of income/employment are not available for deaths. The lack of common demographic and socioeconomic variables in the hospitalization and mortality data sets limits the comparisons that can be made across the findings.
We did not validate the diagnostic accuracy of asthma for either hospitalizations or deaths, so our numerators are subject to the usual misclassification errors. The extent of misclassification bias in asthma deaths has been investigated more thoroughly than hospitalizations. More research has been published on false-positive misclassification of asthma events (other diagnoses being recorded as asthma),17 18 than on false-negative misclassification (asthma events being recorded under other diagnoses) and the net effect of misclassification bias in this population is uncertain. Asthma is more difficult to diagnose in older adults and it would be expected that the extent of death certificate misclassification may increase with age, but we feel that is important not to completely ignore data on middle-aged and older adults. In data sets in which an asthma diagnosis was established in vivo, such as our hospitalization data and in other studies, middle-aged and older adults experienced appreciable mortality.19 20
The diagnostic coding scheme used to analyze comorbidities was developed by the authors, and it has not been tested on other data sets. The validity and reliability of secondary diagnoses relative to primary diagnoses in hospitalization and death records is not clear. Data from other health-care systems indicate that secondary diagnoses in hospitalizations may be undercoded,24 and that coding errors in hospitalization records increase as the total number of diagnoses increases.25
The lack of intercensus population data for non-Hispanic black and white populations decreases the accuracy of our rate calculations for these groups. Besides decreased accuracy in absolute rates, interpretation of intergroup differences may be altered to the extent that there were differential population changes between groups during the intercensus period. Data from Claritas, Inc, suggest that the non-Hispanic black population in Chicago has remained stable while the non-Hispanic white population has declined approximately 20% since the last census. If this is true, it would serve to decrease the disparity between black and white mortality rates, but we believe that this ratio would still remain much higher than the national ratio.
Prior research indicates that there may be substantial heterogeneity among Hispanic subgroups with regard to asthma outcomes,11 21 22 23 so the rates presented here may not accurately identify mortality for higher-risk groups within the Hispanic community.
|
Conclusion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
|
Acknowledgements |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. K. Paasche-Orlow, K. A. Riekert, A. Bilderback, A. Chanmugam, P. Hill, C. S. Rand, F. L. Brancati, and J. A. Krishnan Tailored Education May Reduce Health Literacy Disparities in Asthma Self-Management Am. J. Respir. Crit. Care Med., October 15, 2005; 172(8): 980 - 986. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Krishnan, K. A. Riekert, J. V. McCoy, D. Y. Stewart, S. Schmidt, A. Chanmugam, P. Hill, and C. S. Rand Corticosteroid Use after Hospital Discharge among High-risk Adults with Asthma Am. J. Respir. Crit. Care Med., December 15, 2004; 170(12): 1281 - 1285. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. G. Szilagyi, E. Shenkman, C. Brach, B. J. LaClair, N. Swigonski, A. Dick, L. P. Shone, V. A. Schaffer, J. F. Col, G. Eckert, et al. Children With Special Health Care Needs Enrolled in the State Children's Health Insurance Program (SCHIP): Patient Characteristics and Health Care Needs Pediatrics, December 1, 2003; 112(6): e508 - 520. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Gehlbach, J. P. Kress, J. Kahn, C. DeRuiter, A. Pohlman, and J. Hall Correlates of Prolonged Hospitalization in Inner-city ICU Patients Receiving Noninvasive and Invasive Positive Pressure Ventilation for Status Asthmaticus Chest, November 1, 2002; 122(5): 1709 - 1714. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. N. Grant, T. Li, C. S. Lyttle, and K. B. Weiss Characteristics of Asthma Care Provided by Hospitals in a Large Metropolitan Area: Results From the Chicago Asthma Surveillance Initiative Chest, October 1, 1999; 116(suppl_2): 162S - 167S. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. T. Naureckas, R. L. Wolf, M. J. Trubitt, K. B. Weiss, E. Hernandez-Thomas, S. Thomas, J. Fink, H. J. Zeitz, L. Coover, and J. S. Scharf The Chicago Asthma Consortium: A Community Coalition Targeting Reductions in Asthma Morbidity Chest, October 1, 1999; 116(suppl_2): 190S - 193S. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
