Sensitivity and Specificity for Skin Cancer Diagnosis in Primary Care Providers: a Systematic Literature Review and Meta-analysis of Educational Interventions and Diagnostic Algorithms.

Abstract

In areas without convenient access to dermatology care, primary care providers (PCPs) serve as an important patient resource for early skin cancer detection. To determine the most effective strategy for skin cancer detection training in PCPs, we conducted a systematic review of educational interventions and performed a meta-analysis on sensitivity and specificity outcomes in PCPs. To summarize data on skin cancer sensitivity and specificity outcomes for PCP-targeted training programs and diagnostic algorithms. Our PCP cohort included practicing physicians, trainee physicians, and advanced practice practitioners. A literature search was performed in MEDLINE, Embase, Web of Science, and the Cochrane Library for relevant English-language articles published worldwide from 2000 onward. Results were screened for eligibility, and overlapping datasets were reconciled. Data extracted included the educational intervention, diagnostic algorithm, and outcomes of interest (sensitivity and specificity). Outcomes were pooled across interventions that taught the same diagnostic algorithm. A bivariate model was fit to compare different interventions/algorithms. This review followed the PRISMA guidelines. In total, 21 articles were included in this review, encompassing over 58,610 assessments of skin lesions by about 1529 participants worldwide. Training programs that implemented the triage-amalgamated dermoscopic algorithm (TADA) demonstrated high pooled sensitivity (91.7%) and high pooled specificity (81.4%) among PCPs. Overall, this systematic review and meta-analysis showed that dermoscopy training in PCPs was generally associated with gains in skin cancer sensitivity without loss of specificity. Clinically, this correlates with fewer skin cancers overlooked by PCPs and fewer excisions of benign lesions.