Evaluating pharmaceutical policies using cross-national comparisons and time series analysis

The aim of this thesis was to evaluate the effects of different pharmaceutical policies on the use of medicines (e.g. antibiotics). These policies were evaluated using diverse data sources from the public and private sector in countries in Africa, Latin America, and Western Europe. In addition, the studies contained in this thesis explore and strengthen methods for policy evaluation, with a focus on time series analysis.
The importance of cross-national comparisons is addressed together with the next steps to improve their reporting by adequate and detailed descriptions of data coverage and country characteristics such as marketing status of medicines and reimbursement policies to increase the validity and reliability of comparisons. Subsequently, the use of interrupted time series analysis (ITSA) and the application of this method in the evaluation of pharmaceutical policies is discussed.

As an example of a pharmaceutical policy evaluation, the intended and unintended effects of the over-the-counter antibiotic sales restriction in Mexico and Brazil were assessed. Firstly, the size of the effect of the policy was measured using ITSA. Next, changes in the seasonal variation of penicillins were estimated as a proxy of appropriateness of penicillin use. Subsequently, unintended effects of the policy were assessed by measuring the changes in the use of therapeutic groups that can be perceived as substitutes of antibiotics to relieve cold symptoms (NSAIDs, analgesics, and cough and cold medicines). It was found that the usage level of antibiotics decreased in both countries by about 1 defined daily dose per 1,000 inhabitants per day (DDD/TID), with a decreasing trend in Mexico and an increasing trend in Brazil. The seasonal variation in penicillins decreased in Mexico by 0.4 DDD/TID after the restriction mainly due to changes in seasonal variation of amoxicillin and ampicillin, whereas changes in seasonal variation were not found in Brazil. Lastly, in the two countries, NSAIDs-analgesics usage changes were related with antibiotic usage changes, while only in Mexico cough and cold medicines usage changes had a relation with the antibiotics usage changes. These results showed a substitution effect on the use of other medicines, especially NSAIDs and analgesics, after the reinforcement of OTC antibiotics sales restrictions which might have unintended clinical consequences. It was concluded that although after the regulations the level of antibiotics decreased in both countries, interventions to improve the appropriateness of antibiotics are required. Furthermore, this type of policies should be comprehensive and take into account the potential substitution effects on the use of other medicines.

Although ITSA has been one of the most used methodologies for the evaluation of pharmaceutical policies, further development and improvement of methodologies used in such studies is of great importance to guarantee evidence-based and effective policy-making. Applying other statistical tests in time series can improve the evaluation of policies by examining unintended effects and forecasting possible outcomes. The studies in this thesis show that the effects of interventions in the pharmaceutical sector need to be adequately quantified, and provide new approaches to do so which strengthens evidence-based policy making.