Results: The rate of death from cervical cancer was significantly higher among the .... standardized rates. Mantel's test was used for trend analysis.6. Results.
ORIGINAL RESEARCH * NOUVEAUTES EN RECHERCHE
Rate of death from cervical cancer among native Indian women in British Columbia Pierre R. Band, MD, FRCPC; Richard P. Gallagher, MSc; William J. Threlfall, MSc; Thomas G. Hislop, MD; Michele Deschamps, RN, PhD; John Smith, MPA Objective: To compare the rates of death from cervical cancer among native Indian women and non-native women in British Columbia from 1953 to 1984. Design: Retrospective analysis of data retrieved from the British Columbia Division of Vital Statistics. Main outcome measures: Age-standardized death rate and relative rate. Results: The rate of death from cervical cancer was significantly higher among the native women than among the non-native women throughout the study period. No deaths from cervical cancer were recorded in women under 20 years of age. Among those 20 to 64 the relative rate increased from 3.83 in 1953-62 to 6.53 in 1973-84; among those 65 or more it decreased slightly. For the entire study period the relative rate for women 20 to 64 years old was 5.95 and for those 65 or older 2.98. Conclusion: The rate of death from cervical cancer among native women in British Columbia is unacceptably high, probably because the provincial screening program does not reach as many native women as it does non-native women.
Objectif: Comparer les taux de mortalite causee par le cancer du col chez les femmes indiennes autochtones et les femmes non autochtones de la Colombie-Britannique, de 1953 a 1984. Conception: Analyse retrospective de donnees obtenues de la Division des statistiques demographiques de la Colombie-Britannique. Principales mesures des resultats : Taux de mortalite normalise selon l'age et taux relatif. Resultats: Le taux de mortalite causee par le cancer du col etait beaucoup plus eleve chez les femmes autochtones que chez les femmes non autochtones durant toute la periode d'etude. On n'a enregistre aucun deces cause par le cancer du col chez les femmes de moins de 20 ans. Chez les femmes de 20 a 64 ans, le taux relatif est passe de 3,83 en 1953-1962 a 6,53 en 1973-1984; il a diminue legerement chez celles de 65 ans ou plus. Au cours de toute la periode d'etude, le taux relatif s'est etabli a 5,95 chez les femmes de 20 a 64 ans et a 2,98 chez celles de 65 ans ou plus. Conclusion: Le taux de mortalite causee par le cancer du col chez les femmes autochtones de la Colombie-Britannique est inacceptablement eleve, probablement parce que le programme provincial de depistage n'atteint pas autant de femmes autochtones que de femmes non autochtones. A
centralized provincial cytology screening program has been in effect in British Columbia for over 30 years.' A study of data from 1964
to 1973 showed that the rate of death from cervical cancer among registered native Indian in British Columbia was four times that among white women.2
From the Division of Epidemiology, Biometry and Occupational Oncology, British Columbia Cancer Agency, Vancouver, BC
Reprint requests to: Dr. Pierre R. Band, Head, Division of Epidemiology, Biometry and Occupational Oncology, British Columbia Cancer Agency, 600 W 10th Ave., Vancouver, BC VSZ 4E6 1802
CAN MED ASSOC J 1992; 147 (12)
A recent analysis of the geographic distribution of cancer mortality rates in the province revealed a significantly elevated rate of death from cervical cancer in areas with a high native Indian population.3 This study was undertaken to determine whether the disparity in rates of death from cervical cancer has persisted and to compare the rates between registered native Indian women and non-native women in British Columbia from 1953 to 1984.
Methods Data on all deaths in British Columbia from 1953 to 1984 were obtained from the British Columbia Division of Vital Statistics. After 1984 native Indian status no longer had to be included on death registration forms, and death rates could not be calculated. Native Indians were distinguished from other ethnic groups by a special code on death registration forms. "Registered Indian" referred to all native Indians registered with an Indian band as recognized under the federal Indian Act. This category comprises most native Indians in British Columbia. The population figures for native Indians in 1951, 1956, 1961, 1966, 1971, 1976, 1981 and 1986 were obtained from the Department of Indian and Northern Affairs and the British Columbia Division of Vital Statistics. Corresponding figures for the non-native population, white people accounting for
over 90%, were obtained from Statistics Canada. Through linear interpolation population numbers for the study period were calculated. Causes of death had been coded according to the sixth, seventh, eighth and ninth revisions of the International Classification of Diseases by nosologists at the British Columbia Division of Vital Statistics. For our purposes we reconciled the codes according to the seventh revision.4 Rates of death from cervical cancer were calculated by 5-year age groups; age was standardized to the 1971 Canadian female population. The significance of differences in rates between the native and non-native women was assessed according to the Mantel-Haenszel method.5 Relative rates were calculated as the native to non-native ratio of agestandardized rates. Mantel's test was used for trend
Results Throughout the study period the rate of death from cervical cancer among native women significantly exceeded the rate among non-native women, the relative rate varying from 2.25 to 6.53 (Table 1). No deaths from this disease occurred in either population before the age of 20 years. Among women aged 20 to 64 the relative rate increased from 3.83 in 1953-62 to 6.53 in 1973-84. Among those aged 65 years or more the relative rate de-
Table 1: Rates of death from cervical cancer among registered native Indian women and non-native women in British Columbia from 1953 to 1984*
Non-native Time period; age, yr
1953-62 20-64 2 65 All
1963-72 20-64 > 65 All 1973-84 20-64 > 65 All 1953-84 20-64 2 65 All
375 228 603
5.53§ 2.25 4.24§
328 259 587
1066 685 1751
19 7 26
23 4 27
24 5 29
*Rates are standardized to the 1971 Canadian female population and are per 100 000 population. tNative to non-native ratio of age-standardized death rates. tp < 0.001. §P < 0.0001. lp < 0.05. DECEMBER 15, 1992
CAN MED ASSOC J 1992; 147 (12)
creased slightly over time. However, Mantel's test of trends showed no significant change in the relative rate over the three time periods for these age groups.
participation in the provincial cervical cytology screening program. Barriers to participation in cervical cancer screening programs have been well documented10 and include cognitive, emotional and socioeconomic factors. We have undertaken a demonstration projDiscussion ect in selected native communities in British ColumBritish Columbia is recognized for the excel- bia to determine whether the provincial screening lence of its cervical cancer screening program. program serves native women at a level comparable The program comprises a central cytology labora- to that in the non-native population. tory for processing, screening and interpreting smears as well as a follow-up and recall system and This study was supported in part by grant 129(89-2) from a population-based cytology registry. In 1973 a the British Columbia Health Research Foundation. provincial colposcopy program was established to complement the cytology screening program.1"7 References Since the inception of the screening program the GH, Boyes DA, Benedet JL et al: Organization and provincial rate of death from cervical cancer has 1. Anderson results of the cervical cytology screening programme in decreased by 75%.' Despite this overall success, British Columbia, 1955-1985. BMJ 1988; 296: 975-978 the rates among native women have remained un- 2. Gallagher RP, Elwood JM: Cancer mortality among Chinese, Japanese and Indians in British Columbia, 1964-73. Natl acceptably high. Cancer Inst Monogr 1979; 53: 89-94 Epidemiologic studies have identified a number 3. Band PR, Spinelli JJ, Gallagher RP et al: Atlas of Cancer of risk factors for cervical cancer: early age at first Mortality in British Columbia, 1956-1983 (cat 85-565), Statistics Canada, Ottawa, 1989 sexual intercourse, early age at pregnancy, multiple World Health Organization: International Classification of pregnancies, multiple sexual partners, infection with 4. Diseases, 7th rev, WHO, Geneva, 1957 smoking.8 virus and cigarette human papilloma 5. Mantel N, Haenszel W: Statistical aspects of the analysis of Whether differences in the observed death rates are, data from retrospective studies of disease. J Natl Cancer Inst 1959; 22: 719-748 in part, due to differences in the distribution of risk Mantel N: Chi-square test with one degree of freedom: factors between the two study populations remains 6. extensions of the Mantel-Haenszel procedure. J Am Stat conjectural. In a study involving 437 patients in Assoc 1963; 58: 690-700 British Columbia with newly diagnosed invasive 7. Benedet JL, Anderson GH: Cervical intraepithelial neoplasia in British Columbia: a comprehensive program for detection, cervical cancer from 1985 to 1988, 39% had never and treatment. Gynecol Oncol 1981; 12: S280-S291 had a cytologic examination and 10% had had an 8. diagnosis, Miller AB, Anderson G, Brisson J et al: Report of a national examination more than 5 years before the diagnosis.9 workshop on screening for cancer of the cervix. Can Med Furthermore, 10% of the women were native InAssoc J 1991; 145: 1301-1325 dians; this is a disproportionate number, since native 9. Anderson GH, Benedet JL, LeRiche JC et al: Invasive cancer of the cervix in British Columbia: a review of the demography women account for only 2% of the province's popuand screening histories of 437 cases seen from 1985 to 1988. lation. Therefore, our data suggest that the signifiObstet Gynecol 1992; 80: 1-4 cantly higher death rate observed among the native 10. Peters RK, Bear MD, Thomas D: Barriers to screening for cancer of the cervix. Prev Med 1989; 18: 133-146 women in our study was likely the result of under-
Disease A disease also is farther on the road to being cured when it breaks forth from concealment and manifests its power. - Seneca (4? BC-AD 65) Moral Epistles to Lucilius, LVI (tr. by Richard M. Gummere)
CAN MED ASSOC J 1992; 147 (12)
For prescribing information see page 1876