|
KMID : 0603720130190030186
|
|
Journal of Korean Society of Medical Informatics
2013 Volume.19 No. 3 p.186 ~ p.195
|
|
|
Stratified Sampling Design Based on Data Mining
|
|
Kim Yeonkook J.
Oh Yoon-Hwan
Park Sung-Hoon
Cho Sung-Zoon
Park Ha-Young
|
|
|
Abstract
|
|
|
|
Objectives: To explore classifi cation rules based on data mining methodologies which are to be used in defi ning strata in stratifi ed sampling of healthcare providers with improved sampling effi ciency. Methods: We performed k-means clustering to group providers with similar characteristics, then, constructed decision trees on cluster labels to generate stratifi cation rules. We assessed the variance explained by the stratifi cation proposed in this study and by conventional stratifi cation to evaluate the performance of the sampling design. We constructed a study database from health insurance claims data and providers’ profi le data made available to this study by the Health Insurance Review and Assessment Service of South Korea, and popula-tion data from Statistics Korea. From our database, we used the data for single specialty clinics or hospitals in two specialties, general surgery and ophthalmology, for the year 2011 in this study. Results: Data mining resulted in fi ve strata in general surgery with two stratifi cation variables, the number of inpatients per specialist and population density of provider location, and fi ve strata in ophthalmology with two stratifi cation variables, the number of inpatients per specialist and number of beds. Th e percentages of variance in annual changes in the productivity of specialists explained by the stratifi cation in general sur-gery and ophthalmology were 22% and 8%, respectively, whereas conventional stratifi cation by the type of provider location and number of beds explained 2% and 0.2% of variance, respectively. Conclusions: Th is study demonstrated that data mining methods can be used in designing effi cient stratifi ed sampling with variables readily available to the insurer and government; it off ers an alternative to the existing stratifi cation method that is widely used in healthcare provider surveys in South Korea.
|
|
|
KEYWORD
|
|
|
Sampling Studies, Decision Trees, Data Mining
|
|
|
FullTexts / Linksout information
|
|
|
|
|
Listed journal information
|
|
  
|
|