Skin cancer rates in North Rhine-Westphalia, Germany before and after the introduction of the nationwide skin cancer screening program (2000–2015)

Abstract

Germany is the first nation that implemented a nationwide skin cancer screening program in 2008. The aim is to study the effect of the program on skin cancer rates and to estimate the number needed to screen for an unselected and a hypothetical high-risk population in Germany. We used population-based data on skin cancer incidence (2000–2014), mortality, hospitalization and sick leave (2000–2015) from North Rhine-Westphalia, Germany (18 million population). We calculated annual age-standardized rates per 100,000 person years and calculated the relative change of the rates (%) including 95% confidence intervals (95% CI). Between 2007 and 2014, the estimated annual percentage change (EAPC) of the age-standardized incidence rate of skin melanoma was 3.8% among men and women. These increases were accompanied by increases of the age-standardized mortality rates (EAPC men 3.2%, women 2.0%) and age-standardized sick leave rates (EAPC men 11.0%, women 6.1%). Hospitalization rates showed barely any change. All types of rates for nonmelanoma skin cancer showed marked increases. The number needed to screen for skin melanoma death would be 34,000 if the risk reduction due to screening would be 50%. In a hypothetical high-risk approach with 10% of the population at high risk, that is, a relative risk of melanoma death of 4.0, a skin melanoma mortality risk reduction of 50% among these people due to screening would result in a reduction of the skin melanoma mortality by 15% in the total population. However, this reduction would require a number needed to screen of 11,141. Seven years after the introduction of the skin cancer screening program, there is no discernible beneficial effect at population level. The estimated number needed to screen for skin melanoma in an unselected approach is high and a realistic high-risk approach is currently not feasible.

Appendix

The following formula is based on the assumption that g × 100% of the deaths in high risk people can be prevented by SCS with g coding the fraction of skin melanoma deaths among high risk people being prevented by screening. If N is the population size, π denotes the proportion of the screened high risk group, τ is the mortality risk of melanoma in the total population, RR the relative risk of death due to skin melanoma in the high risk group and D the total number of deaths due to skin melanoma, then

$${\text{D}} = {\text{N }}\pi {\text{ RR }}\tau + \left( {1 - \, \pi } \right){\text{N}}\tau$$

and yields

$$\tau = {{\text{D}} \mathord{\left/ {\vphantom {{\text{D}} {\left( {{\text{N }}\pi {\text{ RR}} + \left( {1 - \, \pi } \right){\text{N}}} \right)}}} \right. \kern-0pt} {\left( {{\text{N }}\pi {\text{ RR}} + \left( {1 - \, \pi } \right){\text{N}}} \right)}}.$$

(1)

The number needed to screen NNS in the high risk group is

$${\text{NNS}} = {\text{ N }}{\pi \mathord{\left/ {\vphantom {\pi {\left( {{\text{g N }}\pi {\text{ RR }}\tau } \right) = \, 1/\left( {{\text{g RR }}\tau } \right)}}} \right. \kern-0pt} {\left( {{\text{g N }}\pi {\text{ RR }}\tau } \right) = \, 1/\left( {{\text{g RR }}\tau } \right)}}$$

which transforms to

$${\text{RR = }}{ 1\mathord{\left/ {\vphantom { 1{\left( {{\text{g NNS }}\tau } \right)}}} \right. \kern-0pt} {\left( {{\text{g NNS }}\tau } \right)}}.$$

(2)

Inserting (1) into (2) yields

$${\text{RR = }}{{\left( {{\text{N }}\pi {\text{ RR + }}\left( { 1 { - }\pi } \right){\text{ N}}} \right)} \mathord{\left/ {\vphantom {{\left( {{\text{N }}\pi {\text{ RR + }}\left( { 1 { - }\pi } \right){\text{ N}}} \right)} {\left( {\text{g D NNS}} \right)}}} \right. \kern-0pt} {\left( {\text{g D NNS}} \right)}}.$$

(3)

Solving (3) for RR gives

$${\text{RR = }}{{\left( { 1 { - }\pi } \right){\text{ N}}} \mathord{\left/ {\vphantom {{\left( { 1 { - }\pi } \right){\text{ N}}} {\left( {{\text{g D NNS - }}\pi {\text{ N}}} \right)}}} \right. \kern-0pt} {\left( {{\text{g D NNS}}}-\pi {{\text{ N}}} \right)}}.$$

(4)