Table 5 Outcome variables, results and limitations of ex-post studies

Abbott et al. [21]

Impact of market exclusivity and trade data protection on the number of generics, aggregate sales, and average price per daily dose for 29 essential medicines in Jordan.

Total cost of medicines increases from 81 million USD in 1999 to 125 million USD in 2004, a 53% increase in the total cost of medicines. Adjusting for increased sales volume and inflation, this represents an increase of 17% in the total cost of medicines.

Data used in this paper is for the year 1999 and 2004 and Jordan ratified new patent law in December 1999, registration of generic medicines may be artificially high in 1999 before the new patent law had become effective.

Comparing mean difference is a poor way to ascertain the rise in prices or costs to IPR provisions as equilibrium prices and quantity are influenced by many factors such as domestic and local economic factors, demographics, etc.

Alawi & Alabbadi [22]

Impact of data exclusivity on sales and cost saving in all pharmaceutical products in Jordan.

Following the expiration of data exclusivity, the medicine prices fall by about 56%.

Data is lacking both before and after the policy change. There is not even enough data for the trend analysis.

Trend analysis is not appropriate for causal analysis of data exclusivity. Outcome variables in this case are generally upward trending due to growth in population and disease prevalence and not due to policy change.

Borrell [23]

Impact of patent enforcement on pricing of HIV/AIDS’ ARV medicines in developing and least developed countries.

Medicine bundles containing at least one original medicine in a patent regime are on average priced 70% higher than medicine bundles containing only local copies marketed in no patent regimes.

Countries where there were patent laws in the pre 2000 era may not have same economic conditions and so treatment and control groups may also significantly differ in other dimensions in addition to patent policy.

Calculating price as sales divided by quantity is a very poor measure of actual prices or costs borne by the patients as HIV/AIDS medicines are publicly provided in developing countries.

Duggan, Garthwaite & Goyal [24]

Impact of patent laws on the average price, number of daily doses, and the number firms in India.

A small, negative, and statistically insignificant decrease (5.4%) in the quantity sold following a patent approval of a medicine

Regression analysis does not control for any economic or demographic variables that might significantly affect the outcome variables, prices, sales, quantity sold.

Jung & Kwon [25]

Impact of patent enforcement on the access to prescribed medicines and catastrophic medicine expenditure in developing countries.

A higher level of IPR protection is associated with higher probability that patients could not get access to their prescribed medicines.

Sample used in the analysis is from 2002 to 2003, when TRIPS implementation was not binding for the developing countries. Lack of access to medicine before 2005 cannot be attributed to IPR protection for most of the countries in the sample.

GP index is a poor measure of IPR protection as it does not consider the actual level of implementation of IPR laws.

Kyle & Qian [26]

Impact of patent and data exclusivity on launch speed, price level and quantity sold of medicines in 59 countries of varying degrees of development.

Products with expired patents sell in lower quantities and at lower prices than those that are on patent, but at higher prices and quantities relative to those that were never protected.

Difference-in-difference framework may not be a good framework to measure the effects of IPR on access to medicines as control and treatment groups of countries are very different.

Berndt & Cockburn [27]

Impact of patent laws on the difference in launch dates of new medicines in India, Germany and USA.

Almost one-quarter of the sample medicines were not available in India within the 10 years of their worldwide launch.

Launch date of medicines in a country is estimated by sales rather than using the official data of medicine approval.

No analysis is conducted to show that the difference in launch lag is due to IPR. The launch lags seem to be same before and after 2005 and so launch lag might be driven by other factors, which are not controlled in this study.

Shaffer & Brenner [28]

Impact of data protection on prices of all medicines provided through public sector health programs in Guatemala.

Medicine prices under data protection increase by 342 to 846% compared to equivalent generic medicines.

Only uses data of prices of public sector medicines, but prices could be higher or lower in private sector.

This paper uses trend analysis, but it does not test any structural break due to data protection and hence simply calculating changes in prices over time cannot be entirely attributed to the change in policy regime.