Clinical Inquiries

Does screening by primary care providers effectively detect melanoma and other skin cancers?

Author and Disclosure Information


Possibly. No trials have directly assessed detection of melanoma and other skin cancers by primary care providers.

Training a group comprised largely of primary care physicians to perform skin cancer screening was associated with a 41% increase in skin cancer diagnoses but no change in melanoma mortality.

Visual screening for melanoma by primary care physicians is 40% sensitive and 86% specific (compared with 49% and 98%, respectively, for dermatologists and plastic surgeons).

Melanomas found by visual screening are 38% more likely to be thin (≤ 0.75 mm) than melanomas discovered without screening, which correlates with improved outcomes.

Visual skin cancer screening overall is associated with false-positive rates as follows: 28 biopsies for each melanoma detected, 9 to 10 biopsies for each basal cell carcinoma, and 28 to 56 biopsies for squamous cell carcinoma. False-positive rates are higher for women—as much as double the rate for men—and younger patients—as much as 20-fold the rate for older patients (strength of recommendations for all foregoing statements: B, cohort studies).




No trials have directly assessed skin cancer morbidity associated with physician visual skin screening. A 2018 ecologic cohort study found no difference in melanoma mortality in a population undergoing a national screening program, although screening was associated with 41% more diagnoses of skin cancer.1 A 2012 cohort study found a reduction in melanoma mortality over 7 years associated with a population-based visual skin cancer screening program compared with similar populations that didn’t undergo specific screening.2 At 12-year follow-up, however, there was no longer a difference in mortality.

Primary care visual screening doesn’t decrease melanoma mortality

German researchers trained 1673 non-­dermatologists (64% of general practitioners, obstetrician-gynecologists, and urologists in that region of Germany) and 116 dermatologists (98% in the region) to recognize skin cancer through whole-body visual inspection.1 They recruited and screened 360,000 adults (19% of the population older than 20 years; 74% women) and followed age- and sex-adjusted melanoma mortality over the next 10 years. Non-dermatologists performed most screening exams (77%); 37% of screened positive patients were lost to follow-up.

Melanoma mortality ultimately didn’t change in the screened region, compared with populations in other European countries without national screening programs. Screening detected approximately half of melanoma cases (585/1169) in the region and was associated with 41% greater detection of skin cancers compared with other countries.

Researchers recorded age-adjusted increases in incidence per 100,000 of melanoma from 14.2 (95% confidence interval [CI], 13.3-15.1) to 18 (95% CI, 16.6-19.4), melanoma in situ from 5.8 (95% CI, 5.2-6.4) to 8.5 (95% CI, 7.5-9.5), squamous cell carcinoma from 11.2 (95% CI, 10.6-11.8) to 12.9 (95% CI, 12.0-13.8), and basal cell carcinoma from 60.5 (95% CI, 59.0-62.1) to 78.4 (95% CI, 75.9-80.8).

Visual screening by primary care providers vs screening by ­dermatologists

A cohort study of 16,383 Australian adults found that visual screening by primary care physicians detected melanoma over 3 years with a sensitivity of 40.2% (95% CIs not supplied) and specificity of 86.1% (95% CI, 85.6-86.6%; positive predictive value = 1.4%).3

A second cohort study, enrolling 7436 adults, that evaluated visual screening by dermatologists and plastic surgeons over 2 years found a sensitivity for melanoma of 49% (95% CI, 34.4-63.7%) and a specificity of 97.6% (95% CI, 97.2-97.9%) with a positive predictive value of 11.9% (95% CI, 7.8-17.2%).4

Visual screening more often detects thinner melanomas

A 3-year case-control study (3762 cases, 3824 controls) that examined the association between visual skin screening by a physician (type of physician not specified) and thickness of melanomas detected found that thin melanomas (≤ 0.75 mm) were more common among screened patients compared with unscreened patients (odds ratio [OR] = 1.38; 95% CI, 1.22-1.56) and thicker melanomas (≥ 0.75 mm) were less common (OR = 0.86; 95% CI, 0.75-0.98).5

Continue to: A systematic review...


Evidence-based answers from the Family Physicians Inquiries Network

Next Article: