Research Article

Development and Validation of a Simple Model to Predict the Risk of Nonmelanoma Skin Cancer on Screening Total Body Skin Examination

Table 1

Potential predictors of skin cancer on screening TBSE.

nn (%) with skin cancerUnadjusted odds ratio (95% CI)Adjusted odds ratio (95% CI)

Age
 <3525802 (0.1)1 (Reference)1 (Reference)
 35–4922218 (0.4)3.96 (0.97–16.24)3.84 (0.96–15.43)
 50–64195326 (1.3)14.17 (3.87–51.90)13.10 (3.64–47.13)
 ≥65174769 (3.9)42.68 (12.07–150.95)33.96 (9.72–118.67)

Gender
 Male368855 (1.5)1.44 (0.98–2.11)1.28 (0.87–1.89)
 Female481350 (1.0)1 (Reference)1 (Reference)

Previous NMSC
 Yes24819 (7.7)8.02 (4.82–13.35)3.01 (1.78–5.10)
 No825386 (1.0)1 (Reference)1 (Reference)

SPT
 I2489 (3.6)7.08 (0.89–56.39)2.57 (0.45–14.76)
 II255647 (1.8)3.52 (0.48–25.67)1.17 (0.23–6.07)
 III471943 (0.9)1.73 (0.24–12.62)0.66 (0.13–3.41)
 IV7895 (0.6)1.20 (0.14–10.32)0.54 (0.09–3.31)
 V-VI1891 (0.5)1 (Reference)1 (Reference)

TBSE-total body skin examination, CI-confidence interval; NMSC-nonmelanoma skin cancer; SPT-skin phototype. Exclusion criteria: history of melanoma, symptomatic (e.g., lesion of concern as chief complaint), missing values for above predictors. c-statistic = 0.831 (95% CI: 0.798–0.865) for multivariable model. c-statistic = 0.829 (95% CI: 0.794–0.863) for above multivariable model omitting gender.