Utility of Non-rule-based Visual Matching as a Strategy to Allow Novices to Achieve Skin Lesion Diagnosis

Abstract

Non-analytical reasoning is thought to play a key role in dermatology diagnosis. Considering its potential importance, surprisingly little work has been done to research whether similar identification processes can be supported in non-experts. We describe here a prototype diagnostic support software, which we have used to examine the ability of medical students (at the beginning and end of a dermatology attachment) and lay volunteers, to diagnose 12 images of common skin lesions. Overall, the non-experts using the software had a diagnostic accuracy of 98% (923/936) compared with 33% for the control group (215/648) (Wilcoxon p < 0.0001). We have demonstrated, within the constraints of a simplified clinical model, that novices' diagnostic scores are significantly increased by the use of a structured image database coupled with matching of index and referent images. The novices achieve this high degree of accuracy without any use of explicit definitions of likeness or rule-based strategies.

Assessment of Smartphone Application for Teaching Intuitive Visual Diagnosis of Melanoma

“…This study reinforces the importance of visual pattern recognition in clinical diagnosis and supports the premise that intuitive diagnosis is superior to rule-based algorithms. Despite its small sample size, this study supports that Skinder is a more effective learning tool for accurate diagnosis of melanoma than traditional rule-based methods.…”

Dermatology undergraduate skin cancer training: a disconnect between recommendations, clinical exposure and competence

“…Diagnosing skin cancer is largely a perceptual skill, relying little on formal or explicit rules, but rather on prior exposure and feedback either in a training environment or in the clinic. Following Norman’s terminology, the skills involved are largely thought to be those of non-analytical pattern recognition (NAPR), and this core skill can be viewed as being able to attach semantics to images or percepts [ 2 - 10 ].…”

Mitigating diagnostic performance bias in a skin‐tone balanced dermatology curriculum

“…Arguing against this, however, is the fact that the mean number of patient cards completed prior to the first assessment equated to more than 10 cards per skin condition, and both the duration of study and number of images per skin condition were similar to the study by Aldridge et al that reported a large improvement in diagnostic performance with NAT. 24 Two alternative and overlapping explanations for our results are that CT may have built upon NAT so that, in effect, we were comparing more training to less training or that the benefits of NAT were delayed and might not have been apparent on the first formative assessment. 37 Anticipating these possibilities, we intentionally provided opposite skin tone NAT training to our two groups and compared the skin tone difference between groups on both assessments.…”

“…There is disagreement on the relative merits of these different learning experiences with some favouring the reduced cognitive load of non-analytic training and supporting this with experimental data suggesting shortterm superiority of training in non-analytic processing alone. 23,24 Others might argue that biomedical knowledge in an encapsulated form (the 'fault' component) is an essential part of experts' illness scripts and that incorporating biomedical explanations into training enhances longer-term diagnostic performance. 20,[25][26][27][28] This study started with a question: If an imbalance in training opportunities on LS and SoC can contribute to diagnostic performance bias, can we mitigate this bias by providing students with training on SoC cases?…”

“…Consequently, learning experiences in dermatology can be divided into those that provide training in non‐analytic processing alone versus those that combine non‐analytic and analytic processing (also referred to as rule‐based training 22 ). There is disagreement on the relative merits of these different learning experiences with some favouring the reduced cognitive load of non‐analytic training and supporting this with experimental data suggesting short‐term superiority of training in non‐analytic processing alone 23,24 . Others might argue that biomedical knowledge in an encapsulated form (the ‘fault’ component) is an essential part of experts' illness scripts and that incorporating biomedical explanations into training enhances longer‐term diagnostic performance 20,25–28 …”