Cancer Incidence in Shoshone County, Idaho
Ron Roizen, Ph.D.
March 18, 2001
According to a quotation appearing in a Cascadia Times article published late last year, "...sparsely populated Shoshone County ranks number one in the whole state of Idaho in the prevalence of cancers." The quotation was attributed to Dr. John F. Rosen.
I contacted Dr. Rosen by email earlier this month and I asked him where he got his data.
He responded that the quotation was only "partially accurate" and that his actual assertion had been that "...this county, with 1.16% of the total population of the State of Idaho, does rank #1 (or very, very close to #1), collectively, in cancers of the lung, kidney, bladder, colon and larynx." His data, he continued, were drawn from Idaho's cancer registry for years 1993-1997 and 1998.
In the following very brief analysis I've examined Shoshone County's cancer data with an eye to comparing cancer incidence in this county with cancer incidence in the remainder of the state.
The result is a little less disconcerting than Dr. Rosen's claims suggested, though this county's marginally elevated cancer rate also rightfully merits, in my opinion, further examination and public discussion.
As it happens, the Cancer Data Registry of Idaho (CDRI, see http://www.idcancer.org) does not report cancer prevalence (i.e., the total number of cancer cases at a given time) but instead reports cancer incidence (the number and rate of new cancer cases) and cancer mortality (number and rate of deaths from cancer) per year for a single year or group of years.
According to the latest County Profile put out by the CDRI, Shoshone County does indeed appear to have a higher-than-average cancer incidence and mortality rates. The lion's share of that difference, however, is contributed by a higher-than-average lung cancer incidence rate in the county. As it happens, Shoshone County residents also report a higher-than-average proportion of current smokers in the population.
Looking a little more closely.
The county has a small population and therefore it is necessary to combine data for a number of years to reach stable incidence and mortality figures. Even then, the CDRI warns that incidence and mortality figures based on 10 or fewer cases should be regarded with caution because of year-to-year fluctuations.
The CDRI report I'm going to summarize reports cumulative figures for five years, 1994-1998, inclusive.
In Idaho as a whole, 24,023 new cancer cases were reported over the 1994-1998 period, for an average of 4,805 per year. Idaho's estimated mean population over this same period was 1,184,320. This translates into about four new cancers for every 1,000 population per year. Epidemiologists usually report incidence in terms of new cases per 100,000 population -- by that metric, Idaho's cancer incidence rate was 405.68 (that is, per 100,000 population).
Shoshone County, on the other hand, reported 469 cases over the same 1994-1998 period, for an average of 93.8 per year. The county's estimated population in this period was 13,946. And this in turn translates into an incidence rate of 6.73 per 1,000 population or (using the conventional metric) 672.58 per 100,000 population.
In terms of "crude rates" (unadjusted for age or sex), then, Shoshone County's reported cancer incidence rate appears to be about 66% greater than the rate found in Idaho as a whole.
But age is a key factor in cancer incidence -- cancer incidence increases with increasing age. As it happens, Shoshone County has a somewhat older population than the state as a whole -- for instance, whereas people 65 years and older comprise 11.3% of the state's population, in Shoshone County 17.1% of our population is 65 and older. In other words, Shoshone County has a 51% greater proportion of residents 65 and older than Idaho as a whole.
The impact of that feature of our population on the apparent cancer rate is substantial. When Shoshone County's population is standardized for age and sex, the incidence rate drops to 487.32, which is 28% lower than the crude rate of 672.58. Put in another way, when age and sex distributions in the population are taken into account, Shoshone County's cancer incidence rate drops from 66% greater to 20% greater than Idaho's as a whole.
Where, then, are the remaining "extra" cancer cases coming from in terms of bodily cancer sites?
Rosen mentioned five sites in his compilation of high cancer rates: lung, kidney, bladder, colon and larynx.
Here is what the most recent CDRI data show regarding these sites:
Cancer of the larynx was infrequent in terms of absolute numbers (4 cases over
5 years) in the county, and this site did not generate a statistically significant difference when the county's rate was compared with the rate found in the remaining population of the state.
Cancer of the kidney (and renal pelvis) reported 13 cases over 5 years and did not generate a statistically significant difference with the remainder of the state's population.
Cancer of the colon, though more frequently reported (37 cases over 5 years) than larynx or kidney cancers, also did not generate a statistically significant difference with the remainder of the state.
Bladder cancer, on the other hand, did generate a statistically significant difference at the .05 level -- Shoshone County's 29 cases over 5 years "should have been" 19 cases to match the remainder of the state's incidence rate. Removing those 10 excess cases would in turn have marginally reduced the county's overall cancer incidence from 469 to 459 cases -- a reduction of only a shade over 2%.
Lung cancer, however, was another matter entirely. There were 110 new lung (and bronchus) cancer cases reported over 5 years. This was the leading bodily cancer site and accounted for almost a quarter of all cancers reported in the county. The county's age- and sex-controlled lung cancer rate was 108.74 cases per 100,000 population -- more than double the rate found in the remainder of the state.
By my rough estimate, if the county's lung cancer rate were reduced to the rate found in the remainder of the state, then Shoshone County's excess total cancer mortality (in comparison with the rest of the state) would be reduced to about 8%. In other words, what began (see above) as a 66% difference when age and sex were uncontrolled, and became a 20% difference when age and sex standardization was imposed, would further become reduced to about an 8% difference if lung cancer incidence in the county were brought into conformity with the rate found in the remainder of the state.
Why the big difference in lung cancer incidence in Shoshone County? That question, unfortunately, cannot be answered by the CDRI data. As already noted above, however, smoking is more prevalent among Shoshone County residents than among Idahoans in general. According to data drawn from the Centers for Disease Control and Prevention, the rate of "current smokers" in the county is almost half-again as high (at 29.6%) as the rate found in the state as a whole (20.5%). Three in ten in Shoshone County smoke; in Idaho as a whole, two in ten smoke.
Still other factors of course -- socioeconomic level and diet -- would doubtless also help account for Shoshone County's marginally higher cancer incidence rate.
The cancer mortality rate is another measure of the cancer burden in a county or state, and I will review cancer mortality figures for Shoshone County and Idaho in a subsequent analysis. According to CDRI's figures, 23.8% of Shoshone County deaths from 1994-1998 were reported as cancer deaths; in Idaho as a whole, 22.9% of all deaths were reported as cancer deaths.
This less-than-one-percentage-point difference affords yet another perspective on the comparative cancer burdens in Shoshone County and Idaho as a whole.