1. Investigating the Impact of Medicare, Part D on the Diabetes Medications Using Enterprise Miner and Survival Analysis
Xiao Wang , PhD Candidate,
Department of Mathematics,
University of Louisville, KY
Copyright © 2010, SAS Institute Inc. All rights reserved.

2. Introduction to the study Methods in detail Conclusion
Outline
Introduction to the study
Methods in detail
Conclusion
Copyright © 2010, SAS Institute Inc. All rights reserved.

3. Introduction Aim of the paper: Data Source:
To investigate the influence of Medicare, Part D on diabetes medications since its introduction in 2006
Data Source:
Data sets about prescription drugs, demographic and physician visits information
All data sets are MEPS survey data for the year 2005 & 2006
Copyright © 2010, SAS Institute Inc. All rights reserved.

4. Methods Summary Statistics Clustering Analysis
Kernel Density Estimation
Association Analysis
Survival Analysis
SAS SQL and some data step procedures such as the transpose procedure
Copyright © 2010, SAS Institute Inc. All rights reserved.

5. Summary Statistics
Table1. Average Overall Payment and Medicare Payment in 2005 & 2006
Year
Variable
Mean N
2005*
SUM OF PAYMENTS
MEDICARE(IMPUTED)
501.66
20.58
1759
2006
558.11
129.02
1994
Copyright © 2010, SAS Institute Inc. All rights reserved.

6. Summary Statistics Figure1. Pie charts of payments in 2005 &2006
Copyright © 2010, SAS Institute Inc. All rights reserved.

7. Cluster Analysis Figure 2. Clusters of Drugs in 2005 Using Text Miner
Copyright © 2010, SAS Institute Inc. All rights reserved.

8. Cluster Analysis Figure 3. Clusters of Drugs in 2006 Using Text Miner
Copyright © 2010, SAS Institute Inc. All rights reserved.

9. Kernel Density Estimation
General Case:
Figure4.Kernel Density Estimation in 2005
Male: Female:
Copyright © 2010, SAS Institute Inc. All rights reserved.

10. Kernel Density Estimation
General Case:
Figure5.Kernel Density Estimation in 2006
Male: Female:
Copyright © 2010, SAS Institute Inc. All rights reserved.

11. Kernel Density Estimation
Medicare Case:
Figure6.Kernel Density Estimation for Medicare in 2005
Male: Female:
Copyright © 2010, SAS Institute Inc. All rights reserved.

12. Kernel Density Estimation
Medicare Case:
Figure7.Kernel Density Estimation for Medicare in 2006
Male: Female:
Copyright © 2010, SAS Institute Inc. All rights reserved.

13. Association Analysis Figure 8.Link Graphs for the drugs in 2005
Copyright © 2010, SAS Institute Inc. All rights reserved.

14. Association Analysis Figure9.Link Graphs for the drugs in 2006
Copyright © 2010, SAS Institute Inc. All rights reserved.

15. Survival Analysis Figure 10. Diabetes Medications in 2006
Copyright © 2010, SAS Institute Inc. All rights reserved.

16. Survival Analysis /*Define 0-1 indicators and status*/
If CHMED=‘ ’ THEN STATUS=1;
IF CHMED=‘GLYBURIDE’ THEN GLYBURIDE=0
AND STATUS=1; ELSE GLYBURIDE=1;…
/*Define the variable days*/
If DATE_1^=. THEN SDATE=DATE_1;
ELSE SDATE=‘01 JAN2006’D;
If DATE_2^=. THEN EDATE=DATE_2;
ELSE EDATE=31DEC2006’D;
FORMAT SDATE EDATE DATE9.;
DAYS=DATDIF(SDATE, EDATE, ‘ACT/ACT’);
Copyright © 2010, SAS Institute Inc. All rights reserved.

17. Survival Analysis Figure11. Survival Data for 2006
Copyright © 2010, SAS Institute Inc. All rights reserved.

18. Survival Analysis SAS Code for Survival Analysis:
PROC LIFETEST DATA=SASUSER.ST06 OUTSURV=SASUSER.GP06
ALPHA=0.05
METHOD=LIFE WIDTH=10;
STRATA CHMED;
TIME=DAYS*STATUS(0);
Copyright © 2010, SAS Institute Inc. All rights reserved.

19. Survival Analysis
Table 2. Summary of Censored/uncensored values for 2005
Stratum
CHMED
Total
Failed
Censored
Percent 1
GLYBURIDE 4
0.00 2
GLYBURIDE-METFORMIN
100.00 3
INSULIN 8 5
62.50
METFORMIN 90 10 80
88.89
PRECOSE
66.67 6
ROSIGLITAZONE 7
TOLAZAMIDE
112 94
83.93
Copyright © 2010, SAS Institute Inc. All rights reserved.

20. Survival Analysis
Table 3. Summary of Censored/uncensored values for 2006
Stratum
CHMED
Total
Failed
Censored
Percent 1
GLYBURIDE 2
100.00
INSULIN 7 3
METFORMIN
100 17 83
83.00 4
PIOGLITAZONE
0.00 5
PRANDIN 6
PRECOSE
60.00
STARLIX
66.67
120 23 97
80.83
Copyright © 2010, SAS Institute Inc. All rights reserved.

21. Survival Analysis
Figure12. Survival distribution function for the year 2005
Copyright © 2010, SAS Institute Inc. All rights reserved.

22. Survival Analysis
Figure13. Survival distribution function for the year 2006
Copyright © 2010, SAS Institute Inc. All rights reserved.

23. To Summarize: Medicare, Part D increases the expenditures
The drug plan reduces the beneficiaries’ choices of their drugs
The usages of metformin and insulin are stable in both years;
More drugs switches in 2006
Need more information for further study
Copyright © 2010, SAS Institute Inc. All rights reserved.

24. Thank you!
Xiao Wang Department of Mathematics University of Louisville
Copyright © 2010, SAS Institute Inc. All rights reserved.