Public ARV Procurement Data: Using Information to Obtain Lower Prices

1
Public ARV Procurement DataUsing Information to
Obtain Lower Prices

• Brenda Waning
• March 6, 2008
• Open Society Institute Seminar
• Kiev, Ukraine

Funded by United Kingdom Department of
International Development, Medicine Transparency
Alliance (MeTA) Project
2
Background to Procurement Practices

• Countries typically struggle with medicine
  procurement
• Unreliable suppliers, unknown medicine quality
• Insufficient funds and financial systems
• Unsure of fair, market price
• (information asymmetry)
• Result is enormous variation in
• price and across countries
• Global Fund (GF) Offers Level Playing Field
• Sufficient financing
• Public posting of prices paid by all recipients
• Limited pre-qualified ARVs and suppliers
• Expect little price variation in this environment

3
Reality and Remedy

• Extreme price variation exists for identical ARV
  products across countries
• 2 Publicly available databases can provide price
  information historically impossible to obtain
• Global Fund Purchase Price Report
• WHO Global Price Reporting Mechanism

4
Public ARV Procurement Databases

• Global Fund Purchase Price Report
• ARV procurement transactions reported by GF
  principle recipients
• WHO Global Price Reporting Mechanism
• Includes GF data PLUS ARV procurements reported
  by international procurement agencies (IDA,
  UNICEF), WHO country offices, John Snow, MSH,
  etc.
• Both available on the web in Excel format

5
Using Existing Data to Assess ARV Prices

• Merged GF Global Price Reporting Mechanism
  (GPRM) Data Sets
• Country
• Date of order
• Generic name
• Strength
• Dosage form
• Total units
• Price per unit
• Type of package
• Manufacturer
• Country of manufacturer
• INCO term

• Expanded data set created at Boston University
• Price per unit calculated
• Year of procurement
• WHO Region of country
• World Bank Income of country
• ARV drug class
• Fixed-Dose Combination
• Generic/Brand
• PEPFAR country
• Clinton Foundation eligible
• Differential Price eligible
• WHO 1st/2nd line regimens 2003 2006 guidelines
• Daily dose

Final Data Set 9,731 procurements
July 02 Oct 07 103 countries
636 million
Extensive Data Cleaning
6
High price outliers screens to assess
efficiency, policy, governance
stavudine 15mg Guyana high outlier prices
verified
volume
price per tablet
lamivudine 150mg 76 similar high price
outliers removed
nevirapine 200mg Russia 20x times more expensive
than median price
volume
volume
price per tablet
price per tablet
7
Extreme Price VariabilityHighest Price/Lowest
Price
8
Percentile Definition Comparisons

• 25th percentile
• 25 of prices paid for that particular ARV were
  equal to or less than the price you paid
• 75th percentile
• 75 of prices paid for that particular ARV were
  equal to or less than the price you paid
• If we compare each procurement made by a country
  with the global price distribution for that exact
  same product, the desire is to be in the lt25th
  percentile (few countries paid less than you)

9
Country Benchmarking
of Country Procurements Across Quartiles of Global Price Distribution of Country Procurements Across Quartiles of Global Price Distribution of Country Procurements Across Quartiles of Global Price Distribution of Country Procurements Across Quartiles of Global Price Distribution
Country () lt25th percentile 25th-50th percentile 51st -75th percentile gt75th percentile
Armenia (19) 5.3 10.5 15.8 68.4
Belarus (14) 7.1 42.9 14.3 35.7
Estonia (11) 0 0 0 100
Georgia (15) 20 20 20 40
Kazakhstan (7) 28.6 28.6 0 42.9
Russian Fed. (32) 0 6.3 0 93.8
Tajikistan (9) 22.2 22.2 22.2 33.3
Ukraine (47) 17 19.1 34 29.8
Uzbekistan (8) 0 0 50 50
Prices for July 2006-June 2007
10
Amount of Money Spent in Excess of Global Median
Price (7/06-6/07)
Country ( procurements) Amount Spent Above Median
Armenia (19) 18,252
Belarus (14) 148,216
Estonia (11) 1,075,273
Georgia (15) 33,483
Kazakhstan (7) 74,208
Russian Federation (32) 12,399,841
Tajikistan (9) 3,773
Ukraine (47) 3,615,215
Uzbekistan 79,804
11
Generic Procurements
Country Total Procurements Generic ()
Armenia 19 4 (21)
Belarus 14 7 (50)
Estonia 11 0 (0)
Georgia 15 9 (60)
Kazakhstan 7 5 (71)
Russian Federation 32 0 (0)
Tajikistan 9 9 (100)
Ukraine 47 22 (47)
Uzbekistan 8 7 (88)
12
stavudine 40mgGlobal Median Price 0.05/tablet
Estonia Russian Federation Ukraine Uzbekistan
3.03 3.17 0.69 0.04 0.05 2.89
July 2006 - June 2007 Brand Name ARV
13
efavirenz 600mgGlobal Median Price 0.65/tablet
Estonia Kazakhstan Russian Feder-ation Tajikistan Uzbekistan
2.07 0.21 0.65 0.98 0.67 0.92
July 2006 - June 2007 Brand Name ARV
14
zidovudine 300mg/lamivudine 150mgGlobal Median
Price 0.19/tablet
Estonia Kazakh-stan Russia Ukraine Tajiki-stan Uzbeki-stan
4.32 0.38 1.33 1.71 0.16 0.18 0.21 0.21 0.23
July 2006 - June 2007 Brand Name ARV
TJ 151 pp/year KZ 274 pp/yr Russia 958 pp/yr
15
lopinavir 133mg/ritonavir 33mgGlobal Median
Price 0.33/tablet
Kazakhstan Russian Federation Ukraine
1.95 1.87 1.44 0.52 1.12
July 2006 - June 2007 Brand Name ARV
16
Comparison of WHO 2003, 2006, and 2008 ART
Guidelines

• WHO 2003 ART Guidelines
• 12 ARVs
• 10 regimens 4 first line, 6 second line
• WHO 2006 Revised ART Guidelines
• 23 ARVs
• 108 regimens 24 first line, 84 second line
• WHO 2008 2nd Line Prioritization
• 20 ARVs (1st and 2nd line)
• 42 2nd line regimens (10 Urgent, 22 High, 10
  Important)

17
WHO First line regimens generic price lt
brand (Median Prices July 2005-June 2006)
2003 WHO regimens Generic 170-431 Brand
500-619
Low range of cost of brand regimens is higher
than upper end of cost generic regimens
2006 WHO regimens Generic 170-1,234 Brand
431-1,393
Upper end of cost of generic regimens is 3 times
higher than 2003
2006 STG big potential impact, even for 1st line
18
WHO 2nd line regimens generic price gt brand
(Median Prices July 2005-June 2006)
2003 WHO regimens Generic 2,913 - 4,117 Brand
1,714 - 2,351
2006 WHO regimens Generic 948 - 4,245 Brand
865 - 2,577
Mostly due to price of protease inhibitors
19
WHO 1st and 2nd line RegimensMedian Prices Paid
July 2006-June 2007

• Prices changing dramatically
• Decisions about country adoption MUST involve
  price, among other things
• Median 1st line prices
• Generics range 102-730 pp/year
• Brand range 595-1,015 pp/year
• Median 2nd line prices range
• Generics range 540-6231 pp/year
• Brands range 880-2,044 pp/year

Generic PIs more competitive with brands, except
Lop/r
20
Multiple Regression to Examine Predictors of
Price

• Regression model included 7 variables
• volume (divided in tertiles)
• year of procurement
• INCO terms
• generic/brand
• CHAI-eligible
• Differential price-eligible
• Country World Bank income stage

21
Higher Volume does NOT mean Lower Price
High Volume Middle Volume Low Volume
Efavirenz 600mg Ref 3.75 6.85
Loprit 133.333.3mg Ref 12.18 23.87
Stalamnvp 40150200mg Ref 12.87 19.03
Zidovudine 300mg Ref 4.57 2.00
Only 4 of 24 dosage forms showed volume-price
association
22
Remaining 20 Dosage FormsNo Price-Volume
Association

• Ritonavir 100mg
• Stavudine 20mg
• Stavudine 30mg
• Stavudine 40mg
• Stalam 30150mg
• Stalam 40150mg
• Stalamnvp 30150200mg
• Tenofovir 300mg
• Zidovudine 100mg
• Zidlam 300150mg

• Abacavir 300mg
• Didanosine 100mg
• Didanosine 200mg
• Didanosine 400mg
• Efavirenz 50mg
• Efavirenz 200mg
• Indinavir 400mg
• Lamivudine 150mg
• Nelfinavir 250mg
• Nevirapine 200mg

23
CHAI Generic Prices Compared to Non-CHAI Generic
Prices

• 11 comparisons
• 1 (EFV 600mg) CHAI 22 less expensive Non-CHAI
• 4 ARVs CHAI 8-13 less expensive Non-CHAI
• 6 ARVs CHAI price was lt5 less than non-CHAI

24
CHAI Inferences CHAI Setting Market Prices for
All?
Oct 2003
Jan 2006
25
Differential Prices (DP)

• 19 ARVs available under DP scheme
• Most less expensive than brand prices when DP not
  available
• 7 of the 19 DP-ARVs compared to generic prices
• 5 of these 7 DP ARVs more expensive than generics

DP Offers little savings over generics for most
ARVs
26
Summary Conclusions

• Transparency of medicine prices using electronic
  databases can potentially lead to increased
  access to medicines
• Knowledge can be power
• BUT only if information is accurate,
  comprehensive (this data represents 40-50 of
  all GF procurements),
• High level political commitment needed from
  donors to mandate data be reported
• No Policy from World Bank or PEPFAR to report
  (although SCMS is voluntarily reporting)

27
Summary Conclusions (2)

• Databases have potential to inform and create
  accountability
• donors, countries, and programs
• Rapidly changing market place requires constant
  monitoring
• Need to understand how data can be most useful to
  implementers
• User-friendly interfaces
• Accessible in low resource settings
• Incorporate into procurement system

28

• Acknowledgements
• Funding United Kingdom Department for
  International Development
• Michael Borowitz (DFID)
• Research team Warren Kaplan, Matthew Fox, Alexis
  King, Danielle Lawrence, Richard Laing, Bert
  Leufkens, Rose Radin, Sapna Mahajan
• Thank you!
• Brenda Waning bwaning_at_bu.edu