75 Evaluation of strategies to prevent overdiagnosis of melanocytic skin lesion biopsies: a decision analysis

Abstract

Objectives The Melanocytic Pathology Study (MPath) reported variation in community pathologists’ interpretations of melanocytic skin lesions relative to reference diagnoses developed through consensus by a panel of three experts. Little is known about the impact of second (2nd) opinion strategies on false positives (FP- overdiagnosis relative to consensus diagnosis) or false negatives (FN-underdiagnosis relative to consensus diagnosis) in melanocytic lesion diagnosis, or on patient care costs incurred within one year of biopsy. Method Lesion severity was classified into five classes based on the nature of clinical follow-up care required. Relative to the reference diagnoses, community pathologists overcalled (FP) or undercalled (FN) as follows: Class I (FP: 7.8%), Class II (FN: 62.8%, FP: 12.5%), Class III (FN: 54.1%, FP: 5.5%), Class IV (FN: 48.1%, FP: 9.1%), Class V (FN: 27.9%). We assessed second opinion strategies on (1) concordance between community pathologists’ diagnoses and diagnoses rendered by the reference panel, and (2) patient care costs incurred during the first year following biopsy. Second opinion strategies assessed included: no 2nd opinion; 2nd opinion obtained for all lesions; 2nd opinion required for some lesions by institutional policyor based on pathologists’ preference. For each second opinion strategy, decision analysis was used to estimate the expected percent of concordant diagnoses, FN, and FP. Standardized care pathways were used to estimate care costs in the year following biopsy. Results Without a 2nd opinion, 83.2% of biopsies received a concordant diagnosis with 8.0% FP and 8.8% FN. Concordance increased under all 2nd opinion strategies and was highest (87.4%) with universally obtained 2nd opinions, resulting in 3.6% FP and 9.1% FN While the proportion of FN cases was fairly consistent across 2nd opinion strategies (range: 8.8% to 9.2%) the proportion FP cases ranged from 3.6% to 7.6%. Per 1 00 000 biopsies, the costs were estimated as $118.6 million with no 2nd opinions, and 127.6 million with 2nd opinions obtained for all lesions. Second opinion strategies based on institutional policy and/or pathologist preference reduced FP cases without appreciable change in FN cases, and led to lower costs in the year following diagnosis (approximately $117 million/100,000). Conclusions While 2nd opinion strategies did not appreciably alter the proportion of FN cases, they did result in fewer FP cases. If selectively implemented, 2nd opinion strategies have the potential to save resources and improve care in the year following biopsy. Such strategies could be mandated through regulatory channels.