The uncertainty involved in our conversion factor
The uncertainty involved in our conversion factor for past-year exposure to ever exposure is a limitation. Past year exposure is likely to be highly variable and susceptible to many different factors – rendering the conversion to ever exposure extremely difficult. A limited evaluation of the validity of our conversion factors comes from a study conducted in the Canadian province of Quebec that reports use in the past-year and the past 5-years . Conversion factors were generally lower (as expected) but of a similar magnitude to our lifetime conversion factors. For example, for the age group 16–24 the past-year to 5-year ratio was 1.7 compared to our conversion of 1.9. For those aged 25–54 the past-year to 5-year ratio was 1.9 compared to our conversion factor of 2.7. The past-year to 5-year ratio, particularly for older adults, does not capture individuals that no longer indoor tan, but that did in their younger years. The decrease in past-year year prevalence by age provides evidence that this Pertussis Toxin is common and that our conversion factor likely results in prevalence that is within a reasonable range of the true prevalence.
We have other limitations that must be acknowledged. The oldest age category (65+) had to be grouped because of the lack of reliable prevalence estimates from the NSS2 and an appropriate conversion factor for this age category. In addition, prevalence estimates from the NSS2 were based on self-reported behaviours, which could be susceptible to social desirability bias. For the meta-analysis component, we based the Canadian relevant exposure distribution on only one Canadian study, which may have led to misclassified inclusion or exclusion of some studies. Additionally, our meta-analyses for BCC and SCCincluded a small number of studies resulting in a large degree of uncertainty in the relative risk estimates. In addition, we relied on some older studies, however, a sensitivity analysis by publication date (before 2000 vs. in 2000 or after) found no difference in relative risk. Canada does not have a national surveillance system for NMSC and, consequently, many cases treated at clinics without confirmatory biopsies are not reported. This situation led to an underestimation of the number of NMSC cases attributable to indoor tanning. Moreover, because only the first NMSC occurring on each individual is reported and multiple NMSCs on the same individual may be caused by indoor tanning, our PAR estimates for NMSC would similarly be underestimated. Finally, Miettinen’s formula  is typically preferred when confounder adjusted relative risk estimates are used in the calculation of a PAR, but it was not possible to use this formula since it requires the availability of exposure distribution data among skin cancer cases.
While there is emerging evidence that moderate exposure to solar UVR may be protective against the risk of developing some non-cutaneous cancers , there is no evidence that a similar beneficial effect exists for indoor tanning. Indeed, it has previously been shown that the use of indoor tanning devices does not protect against solid non-cutaneous cancers and may even increase the risk of developing some hematologic malignancies . Given the lack of benefit of indoor tanning and the large burden of skin cancer attributable to it, efforts to reduce use are required or a total ban should be considered, as already been implemented in Australia and Brazil.
Authorship contribution statement
Conflicts of interest
Funding This research is supported by a Canadian Cancer Society Research Institute Partner Prevention Research Grant (#703106).
Introduction Colorectal cancer (CRC) is the third leading cause of cancer death in the United States and is expected to result in more than 50,000 deaths in 2018 . The CRC mortality rate has declined dramatically, by 48% from 1970 to 2011, with northeastern states showing greater decreases than southern states . The highest rates of CRC mortality have also shifted from northeastern states to the southern states [2,3]. A recent study further showed clusters (or hotspots) of elevated CRC mortality rates concentrated in areas of the Lower Mississippi Delta, west central Appalachia, and eastern Virginia and North Carolina (NC) . The hotspot of the 11 northeastern counties of NC are in geographic areas that have continually suffered from the highest CRC mortality in recent decades . To understand the driving forces of the excessive mortality rates and to inform tailored interventions for reducing mortality rates in these areas of NC, we sought to identify county-level determinants that contribute to the geographic variation of CRC mortality rates in NC.