Nine-Year Trends in Cancer Incidence and Mortality
To view PDF files you will need Adobe Acrobat Reader or for screen reader accessibility Adobe Acrobat Access (free downloads from Adobe Web site.)
The Minnesota Department of Health and the Minesota Cancer Surveillance System (MCSS) are often asked, Are the cancer rates changing in Minnesota? Since the MCSS started operation on January 1, 1988, long-term trends cannot yet be meaningfully assessed. It is interesting, however, to look at nine-year trends (1988-1996) as a partial answer to this question. Determination of changes in incidence at the state level must account for the year-to-year variability, and the fact that most underlying changes in the natural history of a cancer (e.g., changes in exposure or risk factors such as smoking) manifest themselves as changes in incidence slowly. On the other hand, data from the MCSS are and will be directly comparable over the years; and since the existence of incidence trends may be interesting to many people, the nine-year trends in cancer incidence (and mortality) are provided in this Section.
The incidence of several cancers appears to have changed in the nine-year period. Non-Hodgkin's lymphoma among males and females, for example, appears to have increased throughout the period 1988-1996. In order to assess whether any of these changes are real (i.e., not due to chance variation), the statistic referred to as the "estimated annual percent change" (EAPC) is used. The EAPC is a straight line (regression) that best approximates all of the data points. For example, an EAPC of 2.1 percent means that on the average, over the nine-year period, the cancer incidence increased 2.1 percent per year. After the EAPC is calculated, the next question that must be addressed is whether the EAPC is statistically different than zero. That is, is there evidence that the apparent change is real or is it consistent with the normal variability of the cancer incidence and no change in incidence has occurred? Appendix D contains the methodological details of these calculations.
Table 4.1 (and 4.2 for mortality data) contains the EAPCs for the cancer incidence data given in Figures 4.1-4.35 PDF (174kb\18 pages). A dagger () indicates that the EAPC is statistically different from zero. In Minnesota, colon cancer incidence has decreased significantly an average of 2.66 percent per year among males and 1.37 percent per year among females. Similarly, rectal cancer incidence has declined among males and females, although the decline among females is not statistically significant. Decreases in colorectal mortality rates are also apparent. Colon cancer mortality has declined significantly (3.1 percent per year among males and 2.7 percent per year among females) while rectal cancer rates have declined nonsignificantly among both males and females. Similar downward trends in colorectal cancer incidence and mortality have been reported nationwide. These downward trends in cancer incidence and mortality are probably due to the early detection of colorectal cancers through screening.
Cancers strongly related to smoking have generally decreased among males and increased among females. These trends are thought to reflect the changing smoking patterns of males and females over time. Among Minnesota males, the incidence and mortality of cancers of the lung and bronchus have decreased an average of about one percent per year, and laryngeal cancer incidence has declined 3.27 percent per year. Oral cavity cancers have also decreased among males (a significant 2.65 percent per year for incidence and a nonsignificant 2.44 percent per year for mortality). Curiously, the incidence of cancer of the esophagus has increased in males since 1988 an average of 3.27 percent per year. Nationally, esophageal cancer incidence rates for males have increased since 1988, with the suggestion of a plateau starting in 1994. Among Minnesota females, incidence and mortality of cancers of the lung and bronchus have increased significantly about three percent per year, and cancers of the larynx have increased nonsignificantly an average of one to two percent per year. Incidence and mortality rates for cancer of the oral cavity and cancer of the esophagus among females show little significant change over the nine-year period.
Trends in breast, cervical and ovarian cancers are notable. Breast cancer incidence in females increased an average of 0.2 percent per year. Although this finding is not significant, it is comparable to national estimates of approximately -0.2 percent per year. Between 1950 and 1991, there was no evidence that breast cancer mortality in females changed. This was true nationally, as well as in Minnesota. The lack of improvement in mortality experience was puzzling since it was hoped and believed that screening would produce a more favorable outcome. In 1992, the mortality decreased for the first time. The mortality rate has continued to decline in Minnesota, yielding a statistically significant decline in breast cancer mortality averaging 2.0 percent per year over the nine-year period (Table 4.2, Figure 4.16 PDF (174kb\18 pages)). Historically, the incidence of cervical cancer has declined since the 1950s, when the "Pap test" was first introduced as a method of screening for this disease. More recent incidence rates for cervical cancer have continued to decline, albeit not as dramatically as in previous decades. Ovarian cancer incidence increased an average of 2.6 percent per year. A change in coding was instituted in 1992, so that ovarian tumors of "borderline malignancy" began to be counted as malignancies, where they had previously not been included in cancer incidence rates. The observed increase is in all likelihood an artifact of this coding change. Mortality rates for ovarian cancer were stable over the nine-year period.
Since 1988, the incidence of melanomas of the skin among males increased significantly an average of 4.5 percent per year. Statistically significant increases are seen in males under age 65 years (2.7 percent per year) and in males 65 years of age and older (7.8 percent per year). In contrast, incidence among females of both age groups showed little change during the same time period. Mortality in males from melanoma increased, but not significantly, while mortality from melanoma in Minnesota females fell by about 5 percent per year.
The incidence of non-Hodgkin's lymphoma has increased significantly an average of nearly 2.0 percent per year in males and 1.5 percent per year in females. The increasing incidence of non-Hodgkin's lymphoma is reflected in increasing mortality and is also seen nationally.
A dramatic change in prostate cancer is shown in Table 4.1. Between 1988 and 1992, prostate cancer incidence increased by nearly 68 percent, going from 113.4 to 190.3 cancers per 100,000 per year (Figure 4.22 PDF (174kb\18pages)). The rate has declined dramatically by 25 percent from the 1992 high to 142.0 per 100,000 in 1996. Most authorities attribute these dramatic changes, which are also seen nationally, to the practices associated with using the prostatic specific antigen (PSA) test for prostatic cancer screening. It is believed that the decreases are due to two factors: widespread use of PSA in the 1980s and 1990s found many cancers that would have been found later; and the use of PSA may be declining due to the controversy associated with its efficacy for preventing death from prostate cancer.
Figures 4.1-4.35 PDF (174kb\18pages) contain the age-standardized (1970 U.S. Standard) cancer incidence and mortality for each year 1988-1996 which are summarized in Tables 4.1 and 4.2. Where appropriate, the rates are portrayed by sex also. Sex-specific, age-standardized rates are directly comparable across the nine-year period. There is a surprising amount of year-to-year variability in the cancer incidence rates. The rates of the more common cancers such as lung, colon, and breast were more stable. This observation highlights an important fact of cancer statistics. With the exception of the most common cancers, there is a substantial year to year variability for state rates, often exceeding 20-25 percent. Because of this variability, it is difficult to identify and even more difficult to interpret trends in cancer incidence or mortality over short time periods, particularly for relatively rare cancers.
Figure 4.36 PDF (174kb\18 pages) contains the age-standardized cancer incidence and mortality for children less than 15 years of age for each of the years 1988-1996. During this period, overall, both male (EAPC = 2.4 percent) and female (EAPC = 4.5 percent) childhood cancer mortality declined. Although the declines were not statistically significant, they were indicative of one of the greatest success stories in the battle against cancer - the dramatic improvement of cancer treatment for children.
Figure 4.36 PDF (174kb\18 pages): Trends in Pediatric (0 - 14 years) Cancer Incidence 1988 - 1996, Minnesota Cancer Surveillance System: Estimated Annual Percent Change Over Nine-Year Period
For questions about this page, please contact email@example.com