ILLINOIS COUNTY CANCER STATISTICS REVIEW

INCIDENCE, 1998 – 2002

OVERVIEW

This report presents cancer incidence occurring in Illinois’ 102 counties for 1998 through 2002. The first three sections of the report include tables designed to facilitate county comparisons. Data for all races and whites are presented for all counties. Cancer incidence data for blacks are included for the 15 counties that have sufficiently large black populations to allow meaningful statistics for the race group. County-specific tables with expanded cancer sites are included in Section IV. County-level cancer incidence data by Hispanic ethnicity for five counties are included in Section V. Finally, Section VI contains projected county cancer incidence estimates for all sites combined and each of the five major cancer control sites. Details for each section follow.

Section I. Tables containing five-year aggregate incidence counts, average annual age-adjusted rates, and lower and upper 95 percent confidence intervals are presented for all sites combined, colon and rectum, and lung and bronchus by sex category and for female breast, cervix and prostate. These sites were chosen based on two considerations: the need to facilitate evaluation of various cancer prevention and control programs and sufficient numbers to allow meaningful presentation of rates.

Section II. Age-specific cancer incidence data are displayed with five-year aggregate counts, average annual age-adjusted rates, and lower and upper 95 percent confidence intervals for three age groups: less than 50 years, between 50 and 64 years, and 65 or more years. Similarly, the age-specific data are presented for all sites combined, colon and rectum, and lung and bronchus by sex category as well as for female breast, cervix and prostate.

Section III. Stage at diagnosis of cancer for counties is expressed as percentage localized, regional, distant and unstaged for cancers of the colon and rectum (both sexes) and for invasive cervix and prostate. Female breast cancer incidence data are displayed with in situ stage in addition to localized, regional, distant and unstaged stage categories.

Section IV. In separate tables for each of the state’s counties, cancer incidence data are presented including all sites combined, the sites in Sections I-III, as well as additional cancer sites included in the county-level public data set for 1998 – 2002. Data for female breast cancer diagnosed in the in situ stage that correlates highly with mammography screening usage also are reported in these tables. These tables contain five-year aggregate incidence counts, average annual age-adjusted rates, and lower and upper 95 percent confidence intervals for each cancer site by sex category for all races. For the 15 counties with large black populations, tables in the same format are presented for whites and blacks.

Section V. Hispanic and non-Hispanic cancer incidence data are featured in this section for Cook, DuPage, Kane, Lake and Will counties. These counties have sufficient Hispanic populations for meaningful cancer incidence statistics by Hispanic ethnicity. Data are formatted in the same manner as that presented in Sections I-IV.

Section VI. Projected cancer incidence for the five-year time periods 1999 – 2003, 2000 – 2004, 2001 – 2005 and 2002 – 2006 and the average annual projected counts for each respective time period are presented. Projections were calculated for all sites combined, colon and rectum, and lung and bronchus for all races, both sexes, and for female breast, cervix and prostate among all races.

TECHNICAL NOTES

Data Sources

Cancer Incidence

Cancer incidence data are from the Illinois Department of Public Health, Illinois State Cancer Registry (ISCR), the only source of population-based cancer incidence data for the state. Newly diagnosed cancer cases among Illinois residents are reported to ISCR by the health care facilities where the cancer is diagnosed and treated. Central cancer registries and facilities in other states also report data to ISCR on Illinois residents diagnosed and treated for cancer in their states. Most out-of-state cases come from Florida, Indiana, Iowa, Kentucky, Minnesota (Mayo Clinic), Missouri (state registry and St. Louis Barnes/Jewish Hospital) and Wisconsin. In addition, data exchange agreements are in effect with Arkansas, California, Michigan, Mississippi, North Carolina, Washington and Wyoming. For data used in this publication, almost 6 percent of ISCR cases are reported from out-of-state agencies and organizations.¹ A death certificate clearance process involving active follow back of cancer deaths in an effort to identify missed cases has served as an additional means of case identification since August 1993.

The preparation and release of this report is dependent on the completion of annual reporting by Illinois facilities. Although case reporting is mandated within six months of diagnosis, it has been the ISCR policy to keep database files open for late reporting of cases and to allow for the two-to four-year lag in case identification of Illinois residents from other state central cancer registries. For this report, the database files reflect the status of ISCR as November 2004.

Population Estimates

For 1998 through 2002, estimates of Illinois’ resident populations of all races, whites, blacks, Asian/other races, Hispanics and non-Hispanics were provided by NCI’s SEER program. The methodology for these estimates is available at the following Web site: <www.seer.cancer.gov/popdata/methods.pdf>. These population estimates are displayed in Appendix A.

Definitions

Cancer Incidence Sites

The International Classification of Diseases for Oncology version 2 (ICD-O-2, which was used for cancer cases diagnosed prior to 2001)² or version 3 (ICD-O-3, which was used for cancer cases diagnosed in and after 2001)³ codes and the major and minor cancer site groups of the SEER program were used to define cancer sites. The sites group definitions for major and minor sites are those established by the SEER program of the National Cancer Institute (NCI) and also are used by the North American Association of Central Cancer Registries (NAACCR). These standardized classification schemes allow direct comparisons of Illinois data with international, national and state publications.^4-8 In this year’s report, both Kaposi sarcoma and mesothelioma were classified as separate site groups. This change will have slight impact on cancer incidence rates for a few specific cancers, compared to using the previous site grouping method.

Counts and rates were calculated only for invasive cancers with the exception of carcinoma in situ occurring in the urinary bladder. Counts and rates for carcinoma in situ of the breast are displayed separately in tables but were not included in the calculation of counts or incidence rates for all sites combined.

Incidence Rates

Rates are expressed per 100,000 population and are age-adjusted by the direct method to the 2000 U.S. standard million population. The SEER*Stat^® software package, developed by Information Management Services Inc. for the NCI, was used to calculate average annual age-adjusted cancer incidence rates for 1998 – 2002. Rates are rounded to the nearest tenth and very small rates (e.g., 0.04) are shown as 0.0. Rates are presented with the lower and upper confidence intervals computed at the 95 percent level. The formulas for rate calculations are displayed in Appendix B.

Race Categories

The race-specific categories in this report are all races and whites for all 102 counties. Data for blacks are presented for 15 counties (Champaign, Cook, DuPage, Kane, Kankakee, Lake, Macon, Madison, Peoria, Rock Island, St. Clair, Sangamon, Vermilion, Will and Winnebago) with sufficient black population estimates and annual cancer incidence for blacks to allow meaningful statistics. Cases reported as “other” or “unknown” race are included in the “all races” category.

Hispanic Ethnicity

Hispanic ethnicity was determined by the NAACCR Hispanic identification algorithm (NHIA).⁹ NHIA is a generally reliable method to enhance the ethnic identification of the Latino population in the United State.¹⁰ Section V contains tables with cancer incidence occurring among Hispanic and non-Hispanics residing in five counties (Cook, DuPage, Kane, Lake and Will) with sufficient Hispanic populations as well as in all counties over 1998 – 2002.

Projections of Future Cancer Incidence in Illinois Counties

The age-sex-specific rates for the 19 standard five-year age groups for all sites combined and selected sites were applied to either aggregated population estimates or extrapolated estimates by individual year for the same age-sex groups to calculate expected number of new invasive cancer incidence counts in Illinois counties. Cancer incidence projections were made for 1999 – 2003, 2000 – 2004, 2001 – 2005 and 2002 – 2006. (Detailed methodology can be found on the IDPH Web site at <http://www.idph.state.il.us/about/epi/index.htm >.

Quality Control

Ongoing quality control procedures are integral components of ISCR operations that assure high quality cancer incidence data. In 1997, NAACCR developed a certification process that reviews registry data for completeness, accuracy and timeliness of reporting (starting with cases diagnosed in 1995). ISCR has submitted data each year to the NAACCR for registry certification. Based on the certification criteria shown in the following table, ISCR has been awarded gold certification for 1996 through 2002.

Completeness (NAACCR Method)	Pass EDITS	DCO	Timeliness	Unresolved Duplicate	Missing Data Fields				Certification Status
					Sex	Age	County	Race
³ 90%	³ 97%	£ 5%	Within 23 months	£ 2/1000	£ 3%	£ 3%	£ 3%	£ 5%	SILVER
³ 95%	100%	£ 3%	Within 23 months	£ 1/1000	£ 2%	£ 2%	£ 2%	£ 3%	GOLD

Constantly updating registry data is a standard operation in ISCR. As of November 2004, ISCR quality control data for the diagnosis years considered in this report (1998 – 2002) are as follow:

Year	Completeness (NAACCR Method) (%) (As of 11-04)	Pass EDITS (%)	DCO (%)	Unresolved Duplicate (%)	Missing Data Fields
					Sex (%)	Age (%)	County (%)	Race (%)
1998-2002	100	100	2.22	0.02	0.0	0.0	0.0	1.32

Data Interpretation

Observed differences in cancer incidence among counties may be real, reflecting differences in risk factor modifications or consequences of screening and early detection programs within the county. However, county cancer incidence differences may instead be the result of other factors. Any conclusions should be made only after carefully considering the following factors that influence the average annual age-adjusted cancer incidence rates:

• Aggregate cancer case counts for 1998 – 2002 produce more stable age-adjusted cancer incidence rates than those calculated for an individual year. However, counties with smaller populations and smaller numbers of cancer incidence cases will still have less stable age-adjusted rates than larger counties or the entire state. Where the number of cases is less than or equal to 16, the relative standard error for the rate in these instances exceeds 25 percent. At this level, interpretation of the rate is limited by excessive uncertainty and these rates should be evaluated cautiously.

• The 95 percent confidence intervals are included with reported rates to help put the rate in perspective and to facilitate county comparisons. Observed differences may not be statistically significant. The range between the lower confidence interval and the upper confidence interval defines with 95 percent probability where the “true” rate for the county or the state actually falls. The comparison of two sets of confidence intervals is approximately equivalent to statistical significance tests for differences between two county rates and is more conservative than the standard significance test when the null hypothesis is true.¹¹

• Population estimates used for denominators may lack precision. These data are estimates based on demographic characteristics of the population rather than actual counts. Incidence rates produced using these population estimates would be expected to exhibit more error than those calculated using 1990 or 2000 census population counts.

REFERENCES

1. Lehnherr M, Havener LA. Assessment of interstate exchange of cancer data, 1986 – 1998. ISCR Quality Control Series 02:1. Springfield, Ill.: Illinois Department of Public Health, January 2002.

2. Percy C, Van Holten V, Muir C (eds). International Classification of Diseases for Oncology. 2nd edition. Geneva: World Health Organization, 1990.

3. Fritz A, Percy C, Jack A, Shanmugaratnam K, Sobin L, Parkin DM, Whelan S (eds). International Classification of Diseases for Oncology. 3rd edition. Geneva: World Health Organization, 2000.

4. Havener L and Hultstrom D (eds). Standards for Cancer Registries, Vol II, Data Standards and Data Dictionary. Version 11, Tenth Edition. Springfield, Ill..: North American Association of Central Cancer Registries, October 2004.

5. NAACCR (North American Association of Central Cancer Registries). Standards for Cancer Registries, Vol III, Standards for Completeness, Quality, Analysis and Management of Data. Springfield, Ill.: North American Association of Central Cancer Registries, October 2004.

6. Ellison JH, Wu XC, Howe HL, McLaughlin C, Lake A, Firth R, Roney D, Sullivan S, Cormier M, Leonfellner S, Kosary C (eds). Cancer in North America, 1998-2002. Volume One: Incidence. Springfield, Ill.: North American Association of Central Cancer Registries, April 2005.

7. Ellison JH, Wu XC, Howe HL, McLaughlin C, Lake A, Firth R, Roney D, Sullivan S, Cormier M, Leonfellner S, Kosary C (eds). Cancer in North America, 1998-2002. Volume Two: Mortality. Springfield, Ill.: North American Association of Central Cancer Registries, April 2005.

8. Ries LAG, Eisner MP, Kosary CL, Hankey BF, Miller BA, Clegg L, Mariotto A, Feuer EJ, Edwards BK (eds). SEER Cancer Statistics Review, 1975-2002, National Cancer Institute. Bethesda, MD, http://seer.cancer.gov/csr/1975_2002, 2005.

9. NAACCR Expert Panel in Hispanic Identification. Report of the Expert Panel on Hispanic Identification 2003. Springfield, Ill.: North American Association of Central Cancer Registries, October 2003.

10. Howe HL. Evaluation of NHIA Submissions for 1997 – 2001. Springfield, Ill.: North American Association of Central Cancer Registries, October 2004.

11. Schenke N, Gentleman JF. On judging the significance of differences by examining the overlap between confidence intervals. The American Statistician 2001;55:182-186.