Rapid Penetration of COX2 Inhibitors in

1
Rapid Penetration of COX2 Inhibitors in
Non-Steroidal Antiinflammatory Drug Market an
Implication to Hospital Cost Containment Policy
Supon Limwattananon, MPHM, PhD Chulaporn
Limwattananon, MPharm, MSc, PhD Supasit
Pannarunothai, MD, PhD Faculty of
Pharmaceutical Sciences, Khon Kaen University
Center for Health Equity Monitoring, Naresuan
University - Thailand
2
Abstracts
Problem Statement Since their 1999 launch
claiming few gastrointestinal side effects,
cyclo-oxygenase-2 (COX2) inhibitors have become a
market competitor with conventional non-steroidal
antiinflammatory drugs (NSAIDs). Objectives To
examine variations in NSAID expenditures as
related to the use of COX2 inhibitors and to
assess patterns of drug channeling in the
hospital drug market. Design Retrospective,
secondary analysis of electronic drug use
databases. Setting and Population Eighteen
provincial hospitals in Thailand. A total of
1,558,633 prescriptions for oral NSAID solid
dosage forms given to ambulatory patients from
2000 to 2002. Outcome Measures Total and per
patient, quarterly and annual expenditures in
Thai Baht of expenditures and prescriptions
for COX2 inhibitors number of daily doses of
drugs prescribed odds ratios (OR) for receiving
COX2 inhibitors. Results In total, NSAID
expenditures increased by 47.6 from 23.2 million
Baht in 2000 to 34.3 million Baht in 2001, then
increased slightly after implementation of
Universal Health Care Coverage (UC) scheme to
38.0 million Baht in 2002. Part of this rapid
growth was driven by pricing. Expenditures for
COX2 inhibitors increased dramatically from 6.5
in the first quarter of 2000 to 52.1 in the last
quarter of 2002, and the prescription share
gradually increased from 0.5 to 4.6 during the
same period. Use of COX2 inhibitors rather than
other NSAIDs would raise the annual expenditures
per person as much as 1,101.2 (Plt0.001), using a
generalized linear model. The utilization
components did not show a consistent pattern
across health insurance schemes of drug
recipients. Based on logistic regression
analysis, patients covered by the Civil Servant
Medical Benefit Scheme (CSMBS) were 9.7-13.2 and
15.6-23.1 times (Plt0.001) more likely to receive
COX2 inhibitors in a year than Social Security
Scheme (SSS) and UC beneficiaries, respectively.
An analysis of the intensity-of-use component for
the NSAID recipients who visited the hospitals in
three consecutive years revealed a relatively
longer period of COX2 inhibitor use for CSMBS
patients as compared to the patients in SSS and
UC schemes. Conclusions Growth in NSAID
expenditures was largely driven by rapid
penetration of the expensive COX2 inhibitors. The
prime target for these patent-protected,
single-source products was fee-for-service
schemes like CSMBS. To contain hospital drug
costs with a generic substitution for COX2
inhibitors is unfeasible due to the market
exclusivity of the drugs. Therapeutic
substitution with the multi-source NSAID seems to
be a viable alternative in curbing NSAID
expenditure growth.
3

• Objectives
• 1. To examine variations in hospital NSAID
  expenditures as related to
• the use of COX2 inhibitors
• 2. To assess patterns of drug channeling for COX2
  inhibitors
• Design and Analysis
• Retrospective, secondary analysis of electronic
  databases
• Settings 18 MOPH-provincial hospitals in 4
  regions of Thailand
• Sample 1,558,633 prescriptions for oral NSAID
  solid forms
• rendered to ambulatory patients during fiscal
  years 2000 - 2002
• Statistical analysis
• For drug expenditures a generalized linear
  model (GLM)
• For propensity of drug use logistic regression
  analysis
• Control for the underlying differences in drug
  use patterns due to
• age groups and gender
• years of drug use (and interaction with health
  insurance schemes)
• hospital settings (proxy for variations in
  prescribing practice styles)

4
Number of Prescriptions for NSAID
Rxs
4.6
4.0
2.6
0.5
COX2 inhibitors
Other NSAID-NED
Meloxicam
Other NSAID-ED
5
Expenditures by types of NSAID
Baht
COX2 inhibitors
33.9
52.1
46.5
6.5
Other NSAID-NED
Meloxicam
Other NSAID-ED
6
Factors Affecting NSAID Expenditures per Capita
(Competing Models)
Model with interaction terms
Main effect model Coefficienta P value
Coefficienta P value COX2 inhibitors
2.486 lt 0.001 2.488
lt 0.001 Age 36 49 years b 0.368
lt 0.001 0.370 lt 0.001 Age 50 years b
0.798 lt 0.001
0.805 lt 0.001 Male - 0.158 lt
0.001 - 0.158 lt 0.001 CSMBS c 0.864
lt 0.001 0.847 lt 0.001 LIC/UC c
- 0.001 0.954 - 0.053 lt
0.001 ROP c - 0.022 0.188 -
0.084 lt 0.001 Year 2001 d - 0.035
0.065 0.038 lt 0.001 Year 2002 d
0.186 lt 0.001 0.025
0.002 CSMBS x Year 2001 0.083
0.002 CSMBS x Year 2002 - 0.093 lt
0.001 LIC/UC x Year 2001 0.123 lt
0.001 LIC/UC x Year 2002 - 0.205 lt
0.001 ROP x Year2001 0.070
0.002 ROP x Year2002 - 0.249 lt 0.001 a
Based on generalized linear model (GLM) using log
link, gamma distribution , adjusted for hospital
indicators b Age of 18-35 years as the reference
category c SSS as the reference category d Year
2000 as the reference category
7
Effects on Difference in NSAID Expenditure
difference a 95
CI COX2 inhibitors 1,101.2
1,056.5 to 1,147.6 vs. other NSAID Age 36-49
years 44.5 42.3 to
46.7 vs. 18-35 years Age 50 years 122.0
118.5 to 125.6 vs. 18-35 years Male
-14.6 - 15.7 to -13.5 vs.
Female a difference due to an indicator
variable exp(Coefficient) - 1
8
Effects on Difference in NSAID Expenditure
(Trends for Each Scheme)
difference a LIC/UC
SSS ROP CSMBS Year 2001 vs. 9.2
-3.4 3.5 4.9 Year 2000 Year 2002 vs.
-1.9 20.4 -6.1
9.7 Year 2000 a difference due to an
indicator variable exp(Coefficient) - 1
Based on GLM with interaction of schemes and years
9
Effects on Difference in NSAID Expenditure
(Comparison between Schemes for Each Year)
difference a Year 2000
Year 2001 Year 2002 CSMBS vs.
SSS 137.2 157.7
116.1 ROP vs. SSS -2.2 4.9
-23.7 LIC/UC vs. SSS -0.1 13.0
-18.6 a difference due to an indicator
variable exp(Coefficient) - 1 Based on GLM
with interaction of schemes and years
10
Propensity to Receive COX2 Inhibitors (Competing
Models)
Model with interaction terms
Main effect model Coefficienta P value
Coefficienta P value Age 36 49 years b
0.619 lt 0.001 0.617 lt 0.001 Age
50 years b 1.267 lt
0.001 1.270 lt 0.001 Male - 0.302
lt 0.001 - 0.301 lt 0.001 CSMBS c
2.279 lt 0.001 2.434 lt 0.001 LIC/UC
c - 0.845 lt 0.001
- 0.585 lt 0.001 ROP c - 0.407 lt 0.001
0.178 lt 0.001 Year 2001 d 1.105
lt 0.001 1.200 lt 0.001 Year 2002 d
1.145 lt 0.001 1.512 lt
0.001 CSMBS x Year 2001 - 0.009
0.936 CSMBS x Year 2002 0.303
0.003 LIC/UC x Year 2001 0.367
0.009 LIC/UC x Year 2002 0.285
0.038 ROP x Year2001 0.461 lt 0.001 ROP
x Year2002 0.853 lt 0.001 a Based on
logistic regression analysis, adjusted for
hospital indicators b Age of 18-35 years as the
reference category c SSS as the reference
category d Year 2000 as the reference category
11
Odds of Receiving COX2 Inhibitors
CSMBS
Odds (in log scale)
SSS
ROP
LIC/UC
Odds exp(constantbAgebGenderbSchemebYearb
SchemexYearbHosp)
12
Odds Ratio of Receiving COX2 Inhibitors(Compariso
n between Schemes for Each Year)
Odds Ratio a Year 2000
Year 2001 Year 2002 CSMBS vs.
LIC/UC 22.74 15.62 23.14 CSMBS vs. SSS
9.77 9.68 13.22 ROP vs. LIC/UC 1.55
1.70 2.73 LIC/UC vs. SSS 0.43 0.62
0.57 a Based on logistic regression model with
interaction of schemes and years
13
Conclusion

• Growth in NSAID expenditures was largely driven
  by
• rapid penetration of the expensive COX2
  inhibitors.
• The prime target for the patent-protected,
  single-source drugs
• was patients covered by fee-for-service scheme
  like CSMBS.
• To contain hospital drug costs, a generic
  substitution for
• COX2 inhibitors is unfeasible due to market
  exclusivity nature.
• Therapeutic substitution with the multi-source
  NSAID is
• a viable alternative in curbing the
  expenditure growth.