Reach Us +441935548010


Using Outliers Detection in Policy Analysis: A Pilot Case Study of the Detection and Analysis of Average Healthcare Expense in China

Objective: To evaluate four outlier detection methods for choosing a relatively simple and accurate for predicting the tendency of average healthcare expense in China. Method: Dixon’s test, Hampel’s test, Grubbs’ test and T test were used to detect outliers from the average per capita health care costs in China from 1990 to 2013. Results and Conclusion: Our results showed Dixon’s and Hampel’s test methods to be more convenient to perform than T test and Grubbs’ method but they had poor sensitivity. There were many factors affecting medical expenses per capital trend in China, such as the aging population and the financial crisis, and these factors and events could be related to the observed trend and outlier. This showed that the use of simple outlier detection could contribute to policy analysis and research.

Author(s): Zhang F, Yang Y, Li S, Xiang R

Abstract | Full-Text | PDF

Share this  Facebook  Twitter  LinkedIn  Google+
30+ Million Readerbase
Recommended Conferences
Flyer image
Abstracted/Indexed in
  • Index Copernicus
  • Google Scholar
  • China National Knowledge Infrastructure (CNKI)
  • Cosmos IF
  • Directory of Research Journal Indexing (DRJI)
  • WorldCat
  • Publons
  • Geneva Foundation for Medical Education and Research