Hindsight Bias
References and Abstracts

1. Unique Identifier 83274591
Author Muhm JR, Miller WE, Fontana RS, Sanderson DR, Uhlenhopp MA
Title Lung cancer detected during a screening program using four-month chest radiographs
Source Radiology Sept. 1983 148(3) p609-615
Abstract

Ninety-two lung cancers were detected in the Mayo Lung Project in patients undergoing chest radiography every four months for screening. Fifty patients had a peripheral nodule, 16 had a perihilar nodule, 20 had hilar or mediastinal enlargement, and six had pneumonitis. The peripheral cancers grew slowly. Ninety per cent were visible in retrospect for months or even years. Despite this, 70% of the peripheral cancers were classified as postsurgical American Joint Committee (AJC) Stage 1. The central cancers grew rapidly, usually presenting as hilar or mediastinal enlargements after normal findings on the previous radiograph obtained four months earlier. Most were classified as AJC Stage 3 tumors

2. Unique Identifier 20410967
Author Berlin L
Institution Department of Radiology, Rush North Shore Medical Center, 9600 Gross Point Rd., Skokie, IL 60076, USA
Title Hindsight bias
Source AJR Am J Roentgenol Sept. 2000 175(3) p597-601

3. Unique Identifier 95125245
Author Berbaum KS
Title Difficulty of judging retrospectively whether a diagnosis has been "missed"
Source Radiology Feb. 1995 194(2) p582-583

4. Unique Identifier 81215231
Author Arkes HR, Wortmann RL, Saville PD, Harkness AR
Title Hindsight bias among physicians weighing the likelihood of diagnoses
Source J Appl Psychol Apr. 1981 66(2) p252-254

5. Unique Identifier 89039082
Author Dawson NV, Arkes HR, Siciliano C, Blinkhorn R, Lakshmanan M, Petrelli M
Institution Department of Medicine, Case Western Reserve University, Cleveland Metropolitan General Hospital 44109
Title Hindsight bias: an impediment to accurate probability estimation in clinicopathologic conferences
Source Med Decis Making Oct. 1988 8(4) p259-264
Abstract

Although clinicopathologic conferences (CPCs) have been valued for teaching differential diagnosis, their instructional value may be compromised by hindsight bias. This bias occurs when those who know the actual diagnosis overestimate the likelihood that they would have been able to predict the correct diagnosis had they been asked to do so beforehand. Evidence for the presence of the hindsight bias was sought among 160 physicians and trainees attending four CPCs. Before the correct diagnosis was announced, half of the conference audience estimated the probability that each of five possible diagnoses was correct (foresight subjects). After the correct diagnosis was announced the remaining (hindsight) subjects estimated the probability they would have assigned to each of the five possible diagnoses had they been making the initial differential diagnosis. Only 30% of the foresight subjects ranked the correct diagnosis as first, versus 50% of the hindsight subjects (p less than 0.02). Although less experienced physicians consistently demonstrated the hindsight bias, more experienced physicians succumbed only on easier cases

6. Unique Identifier 98183480
Author Franken EA, Jr.
Title Individual error in radiology
Source Acad Radiol Mar. 1998 5(3) p147
Unique Identifier 97265188

7. Author Austin RM
Institution Department of Pathology and Laboratory Medicine, Roper Hospital and Medical University of South Carolina, Charleston 29401, USA
Title Results of blinded rescreening of Papanicolaou smears versus biased retrospective review
Source Arch Pathol Lab Med Mar. 1997 121(3) p311-314
Abstract

Review of Papanicolaou smear cases that are the focus of litigation typically takes place in a biased setting with foreknowledge of an adverse patient outcome (outcome bias) or litigation and with more time allotted for slide review than is available in normal screening situations. Factors that normally mitigate against overly aggressive slide interpretation, such as concern about false-positive diagnoses and possible unnecessary surgical procedures and expense, are absent. This results in a tendency toward overly aggressive interpretation of questionable or uncertain cytologic abnormalities. These factors can be minimized by a variety of blinded slide review formats with the goal of simulating normal, on-the-job, prospective screening as in actual practice. Despite some limitations, blinded rescreening can provide valuable insight into the relative degree of difficulty involved in interpreting specific slides. The difficulty of a case in question is arguably the second most important factor, after assessment of overall laboratory performance, in determining whether a reasonable standard of practice has been followed

8. Unique Identifier 92397696
Author Bosch MM, Rietveld-Scheffers PE, Boon ME
Institution Leiden Cytology and Pathology Laboratory, The Netherlands
Title Characteristics of false-negative smears tested in the normal screening situation
Source Acta Cytol Sept. 1992 36(5) p711-716
Abstract

It is common practice to rescreen false-negative (FN) smears. However, it is inevitable that this is done with some foreknowledge; at least it is known that the test smears contain one or more special cases. Therefore, we decided to test smears in the normal screening situation, when cytotechnologists are completely unaware of being tested. This experiment was done with five FNs and five true-positive (TP) smears. In a third experiment the FNs were tested with the cytotechnologists aware of their presence. Finally, 10 qualities of FNs and TPs were analyzed. In the normal screening situation, only in 1 of 25 tests was the FN recognized as malignant, while all the TPs were detected at the first testing. However, when the cytotechnologists were aware of being tested, the FN was detected in seven of eight tests. The FNs differed from the TPs in 5 of the 10 analyzed characteristics. FNs contained few (if any) large neoplastic epithelial fragments. Detached malignant cells were nearly lacking. The cancer cells had small nuclei that presented little anisokaryosis. The FNs possessed generally hypochromatic nuclei. These features explain why the malignant cells were almost never identified in the normal screening situation. We conclude that FNs and TPs differ and that it seems virtually impossible to avoid all false-negative diagnoses

Copyright © 2009 Electronic Handbook of Legal Medicine