Diagnostic discrepancies and their clinical impact in a neuropathology referral practice

Abstract

BACKGROUND During the course of their neuropathology practice, the authors received cases to review in consultation. In some cases, the patients came to the authors' hospital for therapy; in others, the primary pathologists requested a consultation. Because changes in diagnosis might significantly alter patient management, protocol entry, care costs, or the potential for physician liability, the authors determined the frequency and degrees of their disagreements with the original diagnoses submitted to them. METHODS The authors reviewed the first 500 brain or spinal cord biopsy cases that were submitted to their neuropathology consultation service for a second opinion in 1995. Disagreements were coded into 10 categories, but were grouped for this analysis as follows: serious (having immediate significance for therapy or intervention), less serious but potentially substantial (calling for a change in type or grade of glioma), minor (adding or deleting information), and those in which the authors made the first diagnosis themselves. RESULTS There was some degree of disagreement between the original and review diagnoses in 214 (42.8%) of the 500 cases. Disagreements were counted as serious in 44 cases (8.8%), less serious but substantial in 96 cases (19.2%), and minor in 50 cases (10.0%); the authors made the first diagnosis in 24 cases (4.8%). CONCLUSIONS Clinically important diagnostic errors that can affect immediate patient care decisions occur in a substantial number of brain and spinal cord biopsy cases. Thus, seeking expert neuropathology consultation is prudent and cost‐effective for pathologists who are less experienced with these types of cases. Cost savings in case management might result from confirmation of diagnosis before definitive therapy is administered to the patients. The rates of discrepancy between original diagnoses and second opinions in other subspecialties of pathology should be examined. [See editorial on pages 665‐7, this issue.] Cancer 1997; 79:796‐803. © 1997 American Cancer Society.

Verification and classification bias interactions in diagnostic test accuracy studies for fine‐needle aspiration biopsy

“…7 Major disagreement rates show wide variability, ranging from 0.3% to 10%, depending on many factors such as the tissue type, the type of referral (physician requested or routine transfer), and the institution. [8][9][10][11][12][13][14][15][16][17] We are not aware of any studies that have explored the interactions between these 2 types of bias in FNAB DTA studies. In addition and importantly, verification bias and classification bias may interact in unpredictable ways so that it is difficult to estimate the total bias based on the estimated bias due to each component.…”

Reducing False-Negative and False-Positive Diagnoses in Anatomic Pathology Consultation Material

“…nterinstitutional review of pathology material, either directed by the pathologist or by the patients or clinicians, has been previously studied [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18] and is often recommended or required by pathology organizations and medical institutions when patients change facilities for treatment. 7,[16][17] These studies have shown that the rate of disagreement depends not only on the type of tissue but also on the type of sampling.…”

Estratégias auxiliares para graduação dos tumores astrocíticos segundo os critérios histopatológicos estabelecidos pela OMS

“…Esse método foi escolhido por lidar elegante e eficientemente com a tarefa de encontrar modelos representados por árvores de decisão com complexidade limitada (3) . Ademais, o CART e outros métodos de inteligência artificial têm sido usados com sucesso na geração de modelos diagnósticos em medicina, notadamente através do uso de sistemas especialistas dedutivos e na indução de modelos de conhecimento a partir de uma coleção retrospectiva de casos (6,11,17) .…”

“…A despeito da aparente simplicidade na aplicação dos critérios histológicos de graduação e das notórias particularidades epidemiológicas e radiológicas observadas entre as diferentes gradações dos astrocitomas, a rotina diagnóstica histopatológica permanece laboriosa, sendo marcada pela baixa reprodutibilidade interobservador (3,7,12,19,20,22,24,25) . Mesmo assim, a análise histológica convencional permanece o melhor método diagnóstico e razoável indicador prognóstico para os tumores astrocíticos (3,11,19) .…”

“…Mesmo assim, a análise histológica convencional permanece o melhor método diagnóstico e razoável indicador prognóstico para os tumores astrocíticos (3,11,19) . Tal fato deve-se à relativa carência de correlações clínicas diretas entre as alterações genéticas implicadas na tumorigênese dos astrocitomas e o comportamento biológico desses tumores, especialmente na definição de conjecturas prognósticas (7,13,24) .…”