The aim of this study was to offer an efficient and unsupervised strategy for medical data exploration to find relationships between clinical tests and major as well as concomitant syndromes of a specific disease. A large data set consisting of a group of 100 patients suffering diabetes mellitus type 2 disease characterized by more than 30 clinical parameters was explored using self-organizing maps (SOM) and classified by the use of non-hierarchical K-means algorithm implemented in the SOM. An attempt was made to correlate the classification results with the parameters of the metabolic syndrome. The classification result revealed 4 specific patterns of patients, each presented by respective discriminating parameters: (i) patients with well-controlled state of the disease, (ii) patients with disturbed kidney function, (iii) patients who do not keep the required medical treatment regime, and (iv) patients who neglect the disease. All patterns included respective number of patients with different clinical status and attitude to the health problem. Patients with a well-controlled state of the disease, although being chronic for a long period of time were characterized by very low density proteins and triglycerides levels as well as by the lowest levels for hemoglobin A1c, total protein and creatine phosphokinase. Patients with disturbed kidney function were characterized by the highest averages of the uric acid, albumin, total protein and creatinine, and relatively low values for glucose. Patients who do not keep the required regime possessed the highest level of glucose along with maximal values for cholesterol and trombocytes. Patients who neglect the disease were characterized by high values of cholesterol and relatively high level of glucose, although the duration of the disease was the shortest as compared to the other groups of patients.
[1]
Plamen Nikolov,et al.
Economic Costs of Diabetes in the U.S. in 2002
,
2003,
Diabetes care.
[2]
W. Dietz,et al.
Prevalence of the metabolic syndrome among US adults: findings from the third National Health and Nutrition Examination Survey.
,
2002,
JAMA.
[3]
V. Simeonov,et al.
Multivariate statistical interpretation of laboratory clinical data
,
2007
.
[4]
M. V. Velzen,et al.
Self-organizing maps
,
2007
.
[5]
S. D. Jong,et al.
Handbook of Chemometrics and Qualimetrics
,
1998
.
[6]
M. Fujishima,et al.
Prevalence of type 2 diabetes mellitus and impaired glucose tolerance in a rural area of Japan. The Funagata diabetes study.
,
2000,
Journal of diabetes and its complications.
[7]
S. Wild,et al.
Global prevalence of diabetes: estimates for the year 2000 and projections for 2030.
,
2004,
Diabetes care.
[8]
Vasil Simeonov,et al.
Chemometrics as an option to assess clinical data from diabetes mellitus type 2 patients
,
2009
.
[9]
J. Shaw,et al.
Global and societal implications of the diabetes epidemic
,
2001,
Nature.
[10]
J. Tuomilehto,et al.
High Prevalence of NIDDM and Impaired Glucose Tolerance in Indian, Creole, and Chinese Mauritians
,
1990,
Diabetes.