People with the same diagnosis can have different symptoms and circumstances. Pattern diagnosis and constitutional classification in Korean medicine offer ways to bring those differences into treatment. But assigning people to different categories does not, by itself, show that the categories help select a better treatment. In this project, I separated evidence that a classification can be made from evidence that it improves treatment choice.
Start with the classification itself
Pattern diagnosis brings symptoms and examination findings together to classify a patient’s condition. Constitutional classification uses a different conceptual and diagnostic approach. I did not treat the two as a single test; the starting point was to establish how each divides people into groups.
I developed a staged framework for examining the claims carried by the word “personalized.” Whether different assessors classify the same patient similarly is one question. Whether changing treatment according to that classification helps patients is another. Evidence answering the first question should not silently become an answer to the second.
Predicting improvement is not yet choosing a treatment
An observation that one patient category improves more after treatment does not establish that a particular treatment is better for that category. The patients might have had milder symptoms initially, or other characteristics related to their course of illness. Predicting a patient’s course and choosing between treatments call for different comparisons and study designs.
I used that distinction to organize the literature and connect each claim to the evidence needed to support it. The manuscript follows the steps from classification to a treatment decision, rather than accumulating category names or theoretical descriptions. It is a literature-based evaluation framework, not a new patient experiment establishing treatment efficacy.
The preprint and the continuing manuscript
After making a preprint public, I continued the literature review and manuscript revisions. The public preprint and the later manuscript being evaluated are different versions, so I do not present material from the later manuscript as a finding already available in the preprint.
The progress record dated 3 September 2026 states that the preprint is public and evaluation is continuing. Developing an evaluation framework does not establish its clinical validity. Once the evaluated manuscript has a live version of record and an active DOI, I plan to explain the detailed evidence and process within that public scope.