• Users Online: 193
  • Print this page
  • Email this page


 
 
Table of Contents
CASE REPORT
Year : 2020  |  Volume : 3  |  Issue : 1  |  Page : 50-53

Misdiagnosis features of ancient clinical records based on apriori algorithm


Department of Basic Theroy of TCM, Faculty of Basic Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai, China

Date of Submission09-Jan-2020
Date of Acceptance22-Jan-2020
Date of Web Publication27-Mar-2020

Correspondence Address:
Prof. Ling Yu
Faculty of Basic Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai
China
Login to access the Email id

Source of Support: None, Conflict of Interest: None


DOI: 10.4103/CMAC.CMAC_12_20

Rights and Permissions
  Abstract 


Objective: To analyze misdiagnosis features in clinical cases of “Classified Medical Cases of Famous Physicians” and “Supplement to Classified Case Records of Celebrated Physicians.” Materials and Methods: Two hundred and five ancient misdiagnosed cases were analyzed in aspects of locations (exterior-interior type, qi-blood type and Zang-Fu organs type) and patterns (heat-cold type and deficiency-excess type) by Apriori Algorithm Method. Results: The main types of misdiagnosis in those medical casesare as follows:: Zang-Fu location misjudgment, misjudging the interior as the exterior, misjudging deficiency pattern as excess pattern, and misjudging cold pattern as heat pattern. Among them, the most outstanding type is the misjudgment of deficiency–cold pattern as excess–heat pattern. Conclusions: (1) Accurate judgment of location and differentiation of deficiency and excess patterns are the key points in diagnosing the diseases correctly. The confusion of true deficiency–cold and pseudo-excess–heat pattern should be taken seriously. (2) Data mining on ancient clinical cases offers a new methodology for assisting clinical diagnosis of traditional Chinese medicine.

Keywords: Ancient clinical cases, apriori algorithm, classified medical cases of famous physicians, data mining, misdiagnosis features, supplement to classified case records of celebrated physicians


How to cite this article:
Yu L. Misdiagnosis features of ancient clinical records based on apriori algorithm. Chin Med Cult 2020;3:50-3

How to cite this URL:
Yu L. Misdiagnosis features of ancient clinical records based on apriori algorithm. Chin Med Cult [serial online] 2020 [cited 2020 Dec 4];3:50-3. Available from: https://www.cmaconweb.org/text.asp?2020/3/1/50/281474




  Introduction Top


Ancient medical cases are an important part of academic research in traditional Chinese medicine (TCM). They are not only a vivid portrayal of the clinical wisdom and skills of ancient famous doctors but also the comprehensive application of traditional Chinese medical theory and clinical practice. In addition to a large amount of skilled and careful thinking of successful cases, there are also various introspections on misdiagnosis and malpractice.[1] We can learn more from failure lessons than from successful experience. Hence, the analysis on misdiagnosed cases is an important link between theoretical and clinical studies and should not be ignored. There are two high-ranked medical cases that work in academy and practicability, being named “Classified Medical Cases of Famous Physicians” and “Supplement to Classified Case Records of Celebrated Physicians.”[2] In this article, misdiagnosis cases were taken as the research objective, and the misdiagnosis features were data-mined and analyzed in order to draw lessons from clinical cases and to provide more extensive materials, ideas, and methods for clinical application and subject research of TCM.


  Materials and Methods Top


Data sources

Misdiagnosis medical cases were selected from the “Classified Medical Cases of Famous Physicians”[3] edited by People's Medical Publishing House in 2005 and “Supplement to Classified Case Records of Celebrated Physicians”[4] edited by People's Medical Publishing House in 1997.

Inclusion criteria

The selected medical cases must (1) have complete symptoms and detailed process of pattern identification and treatment; (2) have misdiagnosing process and the amendatory results were clear, definite, and fitted to each other; and (3) be misdiagnosed and revised only once and the finally revised effect was assured.

Exclusion criteria

The medical records must be excluded if they were (1) unclear or have information missing in the pattern identification process; (2) diagnosed and treated repeatedly; (3) accurate diagnosis but improper medication or treatment; (4) predicted but actually did not occur; and (5) inexact or obscure and had no analyzing value.

Database construction

Two hundred and five misdiagnosed cases were selected and indexed according to the disorder type, true pattern name, and misjudged pattern name. In-depth index had been done in the aspects of viscera, qi and blood, exterior and interior, heat and cold, and deficiency and excess. The descriptive languages of the original medical cases were decomposed into a data unit which can be handled and analyzed by a computer, taking “China TCM Thesaurus” (the third edition)[5] compiled by the China Academy of TCM Information Research Institute as the indexing standard reference.

Data mining methods

The Apriori algorithm in IBM SPSS MODELER 14.1 software, (Chinese Version, USA) was used to mine and analyze the above data. The maximum number of the model parameters was 5, setting the minimum support threshold as 30% and the minimum confidence threshold as 60%. The network structure map was drawn by the same software.


  Results Top


Classification of misdiagnosed medical cases

As for the analyzed misdiagnosis cases, all the cases were indexed from the aspects of both location and pattern, while location mistakes and pattern mistakes were approximately equal to each other, that is to say, there was no significance or clinical value merely according to frequency analysis [Table 1]. The first two misdiagnosis types were deficiency–excess misjudgment and visceral misjudgment, following by cold–heat misjudgment. Exterior–interior and qi–blood misjudgments were relatively scarce. The results illustrated that deficiency–excess nature is the most important distractor in diagnosing process due to its complicated changes and various false appearances. Furthermore, multiple-dimensioned relations and transmissions among the viscera increase the chance of misjudging results. We looked back on the original literature according to the indication from the above data and gained the following clues: qi–blood misjudgment always occurred in the gynecological cases and exterior–interior misjudgment always occurred in the exogenous febrile diseases.
Table 1: Classification of misdiagnosis cases

Click here to view


As for the misdiagnosing subtypes, four pairs of contrast are as follows [Figure 1]: (1) there are a small number of qi and blood misdiagnosing subtypes and no obvious difference between the two subtypes. In TCM clinic, Qi and blood often become disordered and be regulated together. As a result, misjudgment of qi and blood location appears relatively less likely. Even if it does exist, to some extent, it will not influence the therapeutic effect very seriously. (2) The cases of misjudging interior as exterior were much more than the cases of misjudging exterior as interior. If some doctors could not evaluate patients' constitutions or disorder complexity properly, many endogenous febrile diseases were always misdiagnosed as exogenous pathogens invasion. (3) The cases of misjudging cold as heat were much more than the cases of misjudging heat as cold. Based on the original literaturewriting and combined with the data analyzing results, a mixture of cold-heat type and true–false type was the main reasonleading the cold–heat misjudgment, among which false-heat(true-cold) patterns were much more than false- cold(true-heat) patterns. (4) The cases of misjudging deficiency as excess were much more than the cases of misjudging excess as deficiency. One reason is that chronic or prolonged or complex disorders were very easy to be selected when the literature was complied. Another reason may be that the situations of excess converted from deficiency and true deficiency with false excess were very common, with greater complexity, leading to a great amount of misjudging deficiency as excess.
Figure 1: Frequency of misdiagnosis subtype

Click here to view


Association rules analysis among the misdiagnosis types

According to [Table 2] and [Figure 2], location misjudgment and deficiency–excess misjudgment always occurred at the same time, due to the fact that the two mistake types are connected, influenced, and infiltrated to each other. Tracing indications from the original literature, we found that a high correlation between location misjudgment and deficiency–excess misjudgment existed in respiratory disorders (lung disorder often attributed to excess pattern, whereas kidney disorder often attributed to deficiency pattern). Furthermore, according to [Figure 2], it is clear that cold–heat misjudgment was relatively independent, without direct relationship with location misjudgment or deficiency–excess misjudgment [Table 2] and [Figure 2].
Table 2: Association rules data among the misdiagnosis types

Click here to view
Figure 2: Network map among the misdiagnosis types

Click here to view


Association rules analysis among the misdiagnosis subtypes

According to [Table 3] and [Figure 3], the top four closely connected pairs were visceral misjudgment versus misjudging deficiency as excess, misjudging deficiency as excess versus misjudging cold as heat, visceral misjudgment versus misjudging cold as heat, and misjudging deficiency as excess versus misjudging interior as exterior. Moreover, [Figure 3] explicitly points out that the tight linkage among the visceral location, deficiency–excess pattern, and cold–heat nature was the core trouble-causing misdiagnosis in ancient clinical cases. Hence, it is a significant guarantee that identifies the location and patterns clearly in diagnosing process. Combined with the original literature, we can conclude that paying attention to “excess converted from deficiency” and “true cold with false heat” is the key issue to enhance the diagnosing accuracy [Table 3] and [Figure 3].
Table 3: Association rules among the misdiagnosis subtypes

Click here to view
Figure 3: Network map among the misdiagnosis subtypes

Click here to view



  Discussion Top


Introspection on visceral location misdiagnosis

Visceral identification is the most important and commonly used diagnosing method. How to locate a certain Zang system precisely? It is always the diagnosing problem and the common misdiagnosing type along with the reason that TCM's five-Zang system is complicated integrally and correlated systematically. Five-Zang pathological changes vary widely and are obscure for the reasons of local disorder leading to overall disorder, one Zang disorder leading to another disorder, transmission/mixtures among visceral disorders, and indistinct or obscure boundary on symptoms. In TCM, visceral names are not the anatomical structure but the collection of some groups of function, having the feature of dynamic change. In TCM diagnosing process, visceral location points not to the morphological structure but to the general dysfunctional or imbalanced judgment. It is the primary factor which leads to the difference between TCM and Western medicine. Besides the complexity of TCM itself, inexperienced physicians themselves were in addition another reason. In comparison, identification of exterior–interior and qi–blood is simpler and their misdiagnosis occurrence is comparatively infrequent. Thus, it can be seen that we should pay more attention to visceral location in TCM diagnosing process. If more than one Zang system get diseased simultaneously, we should draw a clear distinction between the primary and the secondary. At the same time, deficiency–excess identification should be combined with it because the two types are always interfered mutually.

Introspection on deficiency–excess or cold–heat pattern misjudgment

If we refer to the clinical cases' text based on the data mining results, we can find that the type of deficiency–cold patterns being misjudged as excess–heat patterns was definitely the most common and being emphasized. The reasons might be as follows: first, “true deficiency–cold with false excess–heat” pattern or mixture of various patterns has a high frequency compared with other misdiagnosing patterns exactly. Second, those patterns always occur in prolonged disease or critical phase. Hence, they are prone to be emphasized by the writers of clinical cases. Third, those patterns cause confusion and lead the physician easily to draw some wrong conclusions. In any case, it teaches us a lesson that we should make diagnosis very carefully when we face with the excess–heat pattern which cannot be alleviated by cold therapies.

Introspection on data mining-related issue in ancient clinical case research

TCM clinical cases have a large amount of scholarly thinking and accompanied by the clinical experience of famous doctors. Based on the research of clinical cases, it is important to promote the TCM discipline advancement and elevate the clinical effect.[6],[7] With the development of information technology and data processing, massive data in TCM clinical cases need to be mined to attain new knowledge and regularities in practice.[8],[9] The ancient clinical literature has many characters such as massive information, various sentence changes, and indistinctive contents. Hence, the data mining on ancient clinical cases is on its early stage. We start the research from the pattern identification which is relatively easy to outline and try to explain the buried knowledge using statistics and a mathematical language. Despite many complexities and some controversy, we believe that this approach is a basis and direction for future-related research.

Financial support and sponsorship

  1. Budget Foundation of Shanghai University of TCM


  2. (A1-GY010130)

  3. Philosophy and Social Science Foundation of Shanghai (2019BTQ005)


Conflicts of interest

There are no conflicts of interest.



 
  References Top

1.
Zhu YY, Liu GS. Design and implementation of the training system for the thought on TCM symptom differentiation based on ancient medical records. World J Integr Tradit West Med 2016;11(8):1162-5.  Back to cited text no. 1
    
2.
Wang J, Huang H, Jiang HJ. Top ten Xin'an medicine books. China J Tradit Chin Med Pharm 2013;28:1008-15.  Back to cited text no. 2
    
3.
Jiang G. Classified Medical Cases of Famous Physicians. People's Medical Publishing House; 2005.  Back to cited text no. 3
    
4.
Wei ZX. Supplement to Classified Case Records of Celebrated Physicians. People's Medical Publishing House; 1997.  Back to cited text no. 4
    
5.
Wu LC. China TCM Thesaurus. 3rd ed. Beijing: TCM Ancient Books Publishing House; 2008.  Back to cited text no. 5
    
6.
Mi HY, Li CY, Li HR, Chang LP, Wei C. Analysis method of Chinese medical records. Chin J Exp Tradit Med Formul 2017;23(13):226-30.  Back to cited text no. 6
    
7.
Dong XY, Qi S, Jiang M, Han DY, Dong SH, Cui Y. Application of data mining in the research of traditional Chinese medicine. Glob Tradit Chin Med 2017;10(3):364-638.  Back to cited text no. 7
    
8.
Wang Y, Yu Z, Chen L, Chen Y, Liu Y, Hu X, et al. Supervised methods for symptom name recognition in free-text clinical records of traditional Chinese medicine: An empirical study. J Biomed Inform 2014;47:91-104.  Back to cited text no. 8
    
9.
Lucini FR, Fogliatto FS, Silveira DG, Lucini, Filipe R, Fogliatto, et al. Text mining approach to predict hospital admissions using early medical records from the emergency department. Int J Med Inform 2017;100:1-8.  Back to cited text no. 9
    


    Figures

  [Figure 1], [Figure 2], [Figure 3]
 
 
    Tables

  [Table 1], [Table 2], [Table 3]



 

Top
 
  Search
 
    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
    Access Statistics
    Email Alert *
    Add to My List *
* Registration required (free)  

 
  In this article
Abstract
Introduction
Materials and Me...
Results
Discussion
References
Article Figures
Article Tables

 Article Access Statistics
    Viewed642    
    Printed50    
    Emailed0    
    PDF Downloaded56    
    Comments [Add]    

Recommend this journal