In this case, I visualized the transition from initial diagnosis to final diagnosis by multidisciplinary discussion (MDD) in the diagnostic process for complex respiratory diseases. In a medical setting where diagnostic accuracy and the optimization of patient care are required, this shows how discussion among multiple specialists contributes to confirming the diagnosis.
Dr.DataScience illustrated the “flow” of the provided diagnostic data using an intuitive, information-dense chord diagram, clearly presenting the reality of the diagnostic process and the value of MDD.
Background and objective
In respiratory medicine, the client wished to grasp how the initial “facility diagnosis” (a diagnosis by a single department or physician) for particular respiratory diseases changes through the subsequent “multidisciplinary discussion (MDD) diagnosis” (a joint diagnosis by multiple specialists).
The aim was to identify bottlenecks and points for improvement in the diagnostic process and, by improving the final diagnostic accuracy, to achieve more appropriate and prompt treatment intervention for patients. Through this analysis, what was required was to clarify the degree of agreement between facility diagnosis and MDD diagnosis, which categories of respiratory disease are particularly prone to having their diagnosis changed in the transition from facility diagnosis to MDD diagnosis, and how multidisciplinary discussion (MDD) resolves the ambiguity and uncertainty of the initial diagnosis.
Data and variables
The data used in the analysis of this case were the diagnostic records of respiratory-disease patients over a certain period.
- Facility diagnosis: the diagnostic category a patient first received within the medical institution (e.g., particular respiratory disease A, respiratory disease B, undetermined diagnosis, etc.). This corresponds to one set of nodes in the chord diagram.
- MDD diagnosis: the final diagnostic category confirmed after the facility diagnosis through multidisciplinary discussion by multiple specialists such as respiratory physicians, radiologists, and pathologists. This corresponds to the other set of nodes in the chord diagram.
Analytical methods
In this case, I adopted a chord diagram as the main analytical method to visually capture the transition patterns from initial diagnosis to multidisciplinary discussion (MDD) diagnosis.
- Chord diagram
- It expresses the relationships and “flow” among diagnostic categories in a circular graph.
- Each category placed around the circumference is shown as a “node,” and the connections between categories are shown by the thickness of the “ribbons (chords).”
(Figure: Example of a chord diagram) - Purpose of the chord diagram
- Visualizing the diagnostic transition: I designed it so that one can grasp at a glance which MDD-diagnosis category each facility-diagnosis category transitioned to.
- Clarifying patterns of agreement and disagreement: When the diagnosis agreed within a category, I expressed it so that the ribbon cycles within the node (or connects in the same color); when the diagnosis was changed, the ribbon connects to a different node. The thickness of the ribbon represents the number of cases or the frequency corresponding to that pattern.
- Suggesting the role of MDD: By visually presenting the quantitative relationships of cases that were unclear at the initial diagnosis in particular, or cases whose diagnosis was clarified or changed by MDD, I suggested the influence that the intervention of multidisciplinary discussion has on the diagnostic process.
Overview of the main results and clinical considerations
As a result of the analysis using the chord diagram, important patterns in the diagnostic process for respiratory diseases were revealed. It was visually confirmed that in many cases the initial facility diagnosis and the MDD diagnosis agreed, suggesting that the standard diagnostic process is functioning. However, patterns of “diagnostic divergence,” in which cases transitioned from a particular facility-diagnosis category to several different diagnostic categories through MDD, were also clearly shown.
For example, important flows were visualized in which something initially diagnosed as a single disease was reclassified through MDD as a different or more detailed disease, or in which cases whose diagnosis was unclear at the initial stage were confirmed, through the intervention of MDD, into a particular disease category (e.g., a rare lung disease or a complex respiratory disease). This strongly suggests that multidisciplinary discussion (MDD) plays an essential role in resolving the uncertainty of the initial diagnosis and in reaching a more accurate and definitive diagnosis, especially in complex cases or cases where differential diagnosis is difficult.
These results provide the medical setting with concrete implications for identifying which initial-diagnosis paths are prone to diagnostic divergence, which disease groups MDD is especially effective for, and the challenges in the initial diagnosis.
Dr.DataScience’s contribution
In this case, Dr.DataScience contributed, through advanced visualization techniques and data-analysis expertise, to an important challenge in a medical institution’s diagnostic process. From a large volume of diagnostic data, I expressed the complex “flow” and “relationships” from facility diagnosis to MDD diagnosis using a chord diagram—an intuitive, information-dense graph.
This made it possible for specialists to grasp at a glance the degree of diagnostic agreement, the patterns of diagnostic divergence, and the degree to which MDD contributes to confirming the diagnosis. In particular, by visually showing the “quantitative relationships” and “qualitative changes” of the diagnostic flow that are hard to capture in figures or tables, the client was able to effectively identify bottlenecks and opportunities for improvement in the diagnostic process.
By not merely analyzing data but “making visible” the results in a form useful for decision-making in the medical setting, Dr.DataScience provided valuable insights contributing to improved diagnostic accuracy for respiratory diseases and the optimization of patient care.