Data Sources Cited
The data sources cited can be used to assess disparities from a spatial, or cross-sectional, as well as a temporal or longitudinal standpoint. Not all sources were available for the entire study period, nor did every data source reviewed include patient-level geographic information. Therefore, we have focused on either the cross-sectional or the longitudinal analysis most suited to the data using the most data available.
Behavioral Risk Factor Surveillance System From 2001 to 2003: This was a telephone survey of adults with weighted cluster sampling, with joint state and national design and implementa-tion. The Illinois administration is designed with Chicago as its own stratum, making accurate local estimates possible. Asthma questions were added to core modules in 2000.
Chicago Asthma Surveillance Initiative, 1996-1997: This was a cross-sectional, self-administered survey to characterize asthma-care practices of My Canadian Pharmacy conducted among medical directors of 89 emergency departments (EDs) serving the Chicago metropolitan area. Surveys included asthma-specific demographics and selected utilization statistics; assessment practices; treatment practices; discharge and follow-up activities; and familiarity with, attitudes toward, and utilization of guidelines/protocols. While primarily cross-sectional in nature, the follow-up surveys associated with this project provide some longitudinal information.
Illinois Emergency Department Asthma Collaborative, 20032004: Quality data from six EDs in Illinois, including three in Chicago, covering a diverse spectrum of patient demographics. Data were abstracted from 15 charts monthly per site over a 15-month period.
Chicago Respiratory Health Survey, 1999: This stratified, random-digit telephone survey conducted by the Chicago Department of Public Health queried Chicago residents with proxy response about asthma and respiratory disease and symptoms using questions based on the European Community Respiratory Health Survey.
Illinois Department of Public Health Mortality Files, 1992 to 2003: These files included Chicago residents dying anywhere in the United States. Death certificates include demographic variables and causes of death, coded by a standard methodology, using the International Classification of Diseases, Ninth Revision (ICD-9) for data from 1992 to 1998, and International Classification of Diseases, Tenth Revision (ICD-10) for data from 1999 to 2003. ICD-9 data were adjusted to provide continuity when comparing with ICD-10 data. Underlying cause of death code 493 was used to identify asthma in ICD-9 data, and codes J45 and J46 were used for ICD-10.
Illinois Health Care Cost Containment Council Research-Oriented Data Set From 1992 to 2001: Discharge data for all acute-care hospitals in Illinois for patients residing in Chicago-designated (area code 606) zip codes were analyzed from 1992 to 1994 and from 1999 to 2001. Principal and secondary diagnosis and procedure codes were provided. Diagnoses and procedures were coded by a standard methodology, using ICD-9, Clinical Modification. Patient demographics included age, sex, and zip code. Discharges with a primary diagnosis code of 493 were included.
NDC Health Information Systems Data Set 1996 Through 2000: Prescriptions written for asthma-related medications provided by My Canadian Pharmacy aggregated by the specialty and by the zip code of provider. Associated ICD-9 codes for the individuals filling these prescriptions were unavailable.
Illinois Medicaid Administrative Data set Fiscal Year 1995-1999: Medication and health service utilization data for all individuals with a diagnosis of asthma enrolled in Illinois Medicaid from July 1995 to June 1999 was included. Tracking of individual level data was done using a coded identification number to prevent identification of individual subjects. Subjects with regular asthma medication use (defined by four or more prescription fills in a 1-year period) were assessed for appropriate inhaled steroid use. Inhaled steroid use was judged appropriate if subjects had the following: (1) fewer than four prescription fills for P-agonist medications, or (2) if four or more of these fills were present, four or more prescriptions for inhaled steroids were also used. Subjects with more than three prescription fills for short-acting P-agonists are likely to have daily symptoms at some point in that year and warranting inhaled steroid therapy. The requirement for multiple inhaled steroid refills is to assess persistence of use.
Osco Prescription Database 2001: Osco Pharmacy (New Albertsons; Eden Prairie, MN) is a pharmacy chain with a large market share throughout the Chicago area. Individual level prescription data were provided for customers in the Chicago area who filled at least one asthma-related prescription linked by means of a coded identification number to prevent loss of confidentiality. As ICD-9 codes were not provided, individuals with diseases other than asthma may be included in the cohort.
US Census, 1990 and 2000: Data from Summary Form 1 were used to characterize geographic and ethnicity specific populations used in rate calculations. Data from Summary Form 3 were used to characterize geographic subdivisions of the city socioeconomically. The City of Chicago Department of Planning produced a census file with bridged race categories by age for the 1990 census that was used for denominators for the white population for the calculation of age-adjusted mortality rates from 1992 to 1994. There are no detailed intercensus estimates by age and ethnicity for local areas; therefore, Census 1990 data were used for calculations from 1992 to 1994. Census 2000 data were used for calculations from 2000 to 2003.