APPENDIX D

Information on Cancer Registries, by State



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 141
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care APPENDIX D Information on Cancer Registries, by State

OCR for page 141
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care TABLE D-1 State Cancer Registries—Indicators of Data Quality   General Information Data Quality Indicators Year Casefindingg Began Using: Statea Year of Initial Operation Year Population-Based Datab Available Death Certificate Only(%)c Estimate of Completeness (%)d Included in U.S. Combined Rates?e Certified in 1999?f M.D. Offices Ambulatory Surgical Centers Arizona 1981 1995 2.2 86.0   1992 1992 California 1946 1988 1.2 100.4 1988 1988 Colorado 1968 1988 1.8 102.2 1995 1988 Connecticut 1935 1935 1.5 108.9 No No Delaware 1972 1972 5.7 92.9   1998 No Florida 1981 1981 NA 99.7   No 1995 Hawaii 1960 1960 0.4 112.8 1960 1978 Idaho 1969 1970 1.7 101.1 1980 1988 Illinois 1985 1986 5.4 93.0 No 1994 Indiana 1987 1987 NA NA     No No Iowa 1973 1973 1.2 101.9 No 1988 Kentucky 1991 1991 3.2 90.6 No 1995 Louisiana 1974 1988 1.6 94.4 1995 1988 Maine 1983 1983 NA 89.5     1995 1995 Maryland 1982 1982 NA NA     1996 1996 Massachusetts 1980 1982 NA 91.0     No 1982 Michigan 1985 1985 1.2 99.4   No No Minnesota 1988 1988 1.0 98.8 No No

OCR for page 141
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Montana 1979 1979 NA 83.5     No No Nebraska 1987 1987 0.3 92.0   1990 No Nevada 1979 1979 2.0 NA     No 1990 New Hampshire 1986 1987 NA 89.4     1986 1987 New Jersey 1978 1979 2.7 99.0 1978 1988 New Mexico 1966 1973 2.0 94.9 1973 1973 New York 1940 1976 4.8 93.3     No No North Carolina 1987 1990 NA 89.0     1990 1995 Pennsylvania 1982 1985 NA 97.2     No No Rhode Island 1986 1986 2.1 101.3 1986 1986 Tennessee 1986 1989 NA NA     No No Texas 1949 1992 6.5 92.9     No No Utah 1966 1966 0.2 98.7 1973 1994 Virginia 1970 1990 NA 81.5     1998 1998 Washington 1991 1992 3.5 103.4   1992 1992 West Virginia 1993 1993 2.1 92.5   No 1993 Wisconsin 1976 1978 3.7 103.8 1992 1992 Wyoming 1962 1962 0.4 91.0   1962 1997 a Thirty-six state registries responded to NAACCR's Call for Data. The nonparticipating registries include: Alaska, Alabama, Arkansas, Georgia, Kansas, Mississippi, Missouri, North Dakota, Ohio, Oklahoma, Oregon, South Carolina, South Dakota, and Vermont. b A population-based registry is defined as one that “includes information about all cases of a specific disease in a geographically defined area that relates to a specific population.” [From: Wallace RB (ed). Maxcy, Rosenau, Last-Public Health and Preventive Medicine, 14th ed. Stamford, CT: Appleton and Lange, 1998.]

OCR for page 141
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care c Standards for Cancer Registries, Vol. III, p. 41. Based on the experience of the SEER Program, 1% –1.5% death-certificate-only (DCO) cases are expected and acceptable. Values between 0% and 1% or 1.5% and 3% require analysis and explanation. If 0% are DCO, death clearance has not been performed. A high percentage of DCO cases may be the result of underreporting from other sources, incomplete investigation (or follow back) of the DCO cases due to limited resources, or both. In addition, when a population-based registry first begins death clearance, the percentage of DCO cases tends to be higher because some DCO cases were diagnosed prior to the operation of the registry and therefore are not linked to the registry database. dVol. I: Incidence, p. I-7. The adjusted NAACCR estimate of completeness was calculated using the following equation: Adjusted % Completeness = (Observed Cs–Ds) × 100% Expected Cs where observed Cs_= number of cancer cases for all sites in the registry, Ds = number of duplicate records (calculated using the NAACCR estimate of duplicates, based on the registry's results from completing the Protocol for Assessing Duplicate Cases), and Expected Cs = estimated number of cancer for all sites if completeness is 100%. [See Vol. I: Incidence, pp. 7–8 for a detailed derivation of this equation and its variables.] For registries that did not complete the Protocol for Assessing Duplicate Cases, the NAACCR adjusted estimate for completeness is omitted from the registry description. e Indicates states which meet all the following criteria for inclusion in the U.S. combined rates (Vol. I: Incidence, pp. 6–8): Data for all 5 years, 1991–1995, were submitted. The registry completed the Protocol for Assessing Duplicate Cases, developed by the Data Evaluation and Publication Committee, and submitted the results. If a registry had an estimate of duplicates that exceeded one per 1,000 records or 0.1% given its required sample size, the registry was not eligible for inclusion in the combined rates.

OCR for page 141
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care The registry ran the case records for 1991–1995 against the Call for Data metafile prepared for the EDITS software and made all corrections. The completeness of case ascertainment was completed using the formula above. Every registry included in the combined rates had an adjusted completeness estimate of at least 90%. The computed completeness estimate for all registries included in the combined incidence rates for the United States was about 99%. f See Table D-2 for certification criteria. Registries were certified in 1,000 based on cases reported as of 1996. g This date indicates the first year that cancer cases were reported from various sources. In addition to physicians' offices and ambulatory surgical centers, which are included in this summary table, cancer cases were also reported from the following sources: hospitals, death certificates, nonhospital pathology labs, radiation therapy sites, interstate data exchange, and nursing homes/hospices. Complete data for these sources can be found in the monograph. SOURCES: Chert VW, Wu XC, Andrews PA (eds.). 1999, Cancer in North America: 1991–1995. Volume One: Incidence. Sacramento, CA: NAACCR. Tucker TC, HL Howe and HK Weir. 1999. Certificationfor population-based registries. J Registry Manage Feb:24–27.

OCR for page 141
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care TABLE D-2 North American Association of Central Cancer Registries (NAACCR) Criteria and Standards for Certification of Cancer Registries Criteria Measure Rationale Gold Standard Silver Standard Completeness of case ascertainment 1. Compare actual incidence rate to expected incidence rate, using SEER Incidence to U.S. mortality ratio method* 2. Death clearance: Match all cancer deaths with registry records, and follow back on unmatched cancer deaths 3. Number of duplicate records Demonstrates the registry has identified a sufficient proportion of expected cases. Provides a more accurate count of cancer incidence by looking at unmatched cancer deaths. Duplicates should be consolidated to ensure that one case is not entered more than once from different institutions. 95% completeness Complete death clearance <1 duplicate per 1,000 90% completeness Complete death clearance <2 duplicates per 1,000 Completeness of information recorded 1. Sociodemographic Information (% missing) 2. Percentage of “death certificate only” (DCO) cases Includes: age at diagnosis, sex, race, county of residence at diagnosis DCO cases often lack information on key variables, which limits utility, so a minimal proportion of DCO cases is desired. <2% missing (3% for race) <3% <3% missing (5% for race) <5%

OCR for page 141
Enhancing DATA SYSTEMS to Improve the Quality of CANCER Care Accuracy of data EDITS metafile: an electronic editing procedure capable of identifying logical inconsistencies in case records Ensures that information is consistent and accurate enough to be useful, 99% cases passing EDITS 97% cases passing EDITS Timeliness All information and corrections must be entered within 23 months from close of diagnosis year. Timely submission of information Data submitted Data submitted NOTE: For certification at either the gold or silver standard, a registry must meet all criteria for that particular level of certification. * The incidence to mortality ratio method uses the ratio of SEER incidence (site, race, and sex-specific) to U.S. mortality, and applies that ratio to the site, race, and sex-specific mortality rates of the population served by the registry. The expected incidence rate is calculated by multiplying these rate ratios by the cancer-specific mortality rate for that population. The incidence to mortality rate ratio method provides a more accurate expected incidence rate because it allows for the possibility that different populations have lower incidence rates. In the past, expected incidence was calculated by applying incidence rates of one area with complete case ascertainment, to the area under evaluation; this method assumed that the cancer incidence rates were similar. SOURCE: Tucker TC, HL Howe and HK Weir. Certification for Population-BasedCancer Registries. J Registry Management. Feb 1999:24–27.