To describe and evaluate policies implemented in Chile, Colombia, Venezuela and Mexico (1995–2009) to prohibit antibiotic OTC sales and explore limitations in available data.
To describe and evaluate policies implemented in Chile, Colombia, Venezuela and Mexico (1995–2009) to prohibit antibiotic OTC sales and explore limitations in available data.
We searched and analysed legislation, grey literature and peer-reviewed publications on regulatory interventions and implementation strategies to enforce prohibition of OTC antibiotic sales. We also assessed the impact using private sector retail sales data of antibiotics studying changes in level and consumption trends before and after the policy change using segmented time series analysis. Finally, we assessed the completeness and data quality through an established checklist to test the suitability of the data for analysis of the interventions.
Whereas Chile implemented a comprehensive package of interventions to accompany regulation changes, Colombia's reform was limited to the capital district and Venezuela's limited to only some antibiotics and without awareness campaigns. In Mexico, no enforcement was enacted. The data showed a differential effect of the intervention among the countries studied with a significant change in level of consumption in Chile (−5.56 DID) and in Colombia (−1.00DID). In Venezuela and Mexico, no significant change in level and slope was found. Changes in population coverage were identified as principal limitations of using sales data for evaluating the reform impact.
Retail sales data can be useful when assessing policy impact but should be supplemented by other data sources such as public sector sales and prescription data. Implementing regulatory enforcement has shown some impact, but a sustainable, concerted approach will be needed to address OTC sales in the future.
Décrire et évaluer les politiques mises en œuvre au Chili, en Colombie, au Venezuela et au Mexique (1995–2009) visant à interdire les ventes libres d'antibiotiques et explorer les limitations des données disponibles.
Nous avons recherché et analysé la législation, la littérature grise et des publications évaluées par les pairs sur les interventions réglementaires et les stratégies d'implémentation pour faire respecter l'interdiction des ventes libres d'antibiotiques. Nous avons également évalué l'impact à l'aide des données des détaillants d'antibiotiques du secteur privé en étudiant l’évolution des tendances des degrés et de consommation avant et après le changement de politique, en utilisant l'analyse de séries chronologiques. Enfin, nous avons évalué la qualité et l'exhaustivité des données par le biais d'une liste de contrôle mise en place pour tester l'adéquation des données pour l'analyse des interventions.
Alors que le Chili a mis en œuvre un ensemble complet d'interventions pour accompagner les changements de réglementation, la réforme Colombienne a été limitée au district de la capitale et au Venezuela, elle a été limitée à seulement quelques antibiotiques et sans campagnes de sensibilisation. Au Mexique, aucune application n'a été promulguée. Les données ont montré un effet différentiel de l'intervention entre les pays étudiés avec un changement important du niveau de consommation au Chili (−5,56 DJ) et en Colombie (−1,00 DJ). Au Venezuela et au Mexique aucun changement significatif dans le degré et la tendance n'a été trouvé. Des changements dans la couverture de la population ont été identifiés comme principales limitations de l'utilisation des données de ventes pour l’évaluation de l'impact de la réforme.
Les données de vente des détaillants peuvent être utiles pour évaluer l'impact des politiques, mais doivent être complétées par d'autres sources de données telles que celle des ventes du secteur public et les données des prescriptions. L'implémentation de la réglementation a montré un certain impact, mais une approche durable et concertée sera nécessaire pour répondre aux ventes sans prescription à l'avenir.
Describir y evaluar las políticas implementadas en Chile, Colombia, Venezuela y Méjico (1995–2009) para prohibir la venta de antibióticos sin receta médica y explorar las limitaciones en los datos disponibles.
Hemos buscado y analizado la legislación, la literatura gris y las publicaciones con revisión por pares sobre intervenciones de regulación y estrategias de implementación para reforzar la prohibición de venta de antibióticos sin receta médica. También evaluamos el impacto, utilizando los datos de venta de antibióticos al por menor en el sector privado, estudiando cambios en el nivel y en las tendencias del consumo antes y después del cambio de política mediante un análisis de regresión segmentada de series de tiempo. Finalmente, hemos evaluado la integridad y la calidad de los datos mediante una lista de control establecida para evaluar la idoneidad de los datos para el análisis de las intervenciones.
Mientras que en Chile se implementó un paquete completo de intervenciones para acompañar los cambios de regulación, la reforma de Colombia se limitó a la capital y en Venezuela estuvo limitada a solo algunos antibióticos y no se realizaron campañas de concienciación. En Méjico no se aprobó ninguna ley. Los datos mostraban un efecto diferencial de la intervención entre los países estudiados con un cambio significativo en el nivel de consumo en Chile (−5.56 DID) y en Colombia (−1.00DID). En Venezuela y en Méjico no se halló un cambio significativo en el nivel ni en la pendiente. Los cambios en la cobertura de la población se identificaron como la principal limitación de utilizar datos de ventas para evaluar el impacto de la reforma.
Los datos de ventas al por menor pueden ser útiles a la hora de evaluar el impacto de políticas sanitarias pero debería complementarse con otras fuentes de datos tales como el de las ventas del sector público o los datos de prescripción. El implementar medidas de control ha mostrado tener algún impacto, pero se requiere un enfoque sostenible y concertado para abordar de cara al futuro el tema de las ventas sin receta médica.
As antibiotic resistance increases worldwide, international organisations have proposed strategies to combat this global public health problem. Among them are prohibiting over-the-counter (OTC) antibiotic sales. Although OTC sales are prohibited by law in many Latin American countries, in practice these laws are often not enforced (PAHO 2004), and antibiotics are still requested and sold without prescription.
Achieving compliance with OTC laws requires enforcement measures either through incentives or sanctions. Sanctions can include penalties for non-compliant pharmacies, retention of the operating manager's license or closure of the business. To make legal surveillance more effective, pharmacies could be requested to retain antibiotic prescriptions and maintain sales records, as is the case for controlled medicines; regular inspections can be used to monitor regulation adherence (WHO 2001). Such interventions should be accompanied by strategies to promote the appropriate use of antibiotics such as training prescribers or dispensers and public awareness campaigns (Leung et al. 2011).
As several Latin American countries (e.g. Chile, Colombia and Venezuela) implemented measures to promote adequate use of antibiotics including prohibiting OTC sales, it is relevant to analyse the impact of regulatory interventions on consumption. There are single country studies in Chile (Bavestrello et al. 2002; Bavestrello & Cabello 2011) and in Venezuela (Rivas & Alonso 2011) using retail sales data, but to our knowledge, there is no cross-national comparison of policy changes and the feasibility of analysing their impact using retail sales data.
Our study therefore describes policies implemented in Chile, Colombia, Venezuela and Mexico since 1995 to enforce prohibition of OTC antibiotic sales, evaluate their impact on antibiotic consumption and explore limitations in available data.
We reviewed legislation related to antibiotic sales in the four study countries, in addition to other documents published between 1995 and 2009 describing sales regulations and related sanctions for infringement. We also searched government websites and other relevant sites including information related to interventions to promote adequate use of antibiotics such as public awareness campaigns; prescriber training; and involvement of pharmacists, pharmacy managers or staff and other professional associations (e.g. general practitioner associations). To complement the information gathered from the grey literature, we searched peer-reviewed publications using PubMed and Web of Science search engines. We used the free search function as well as MeSH terms such as antibacterial agents, over the counter drugs, country names, drug utilisation and drug legislation. Finally, we interviewed two to five key informants such as infectious disease specialists, representatives of general practitioners’ associations, administrators at the Ministry of health, and pharmaceutical policy analysts, and representatives from non-governmental organisations promoting the rational use of antibiotics per country except Mexico where not policy change was implemented in the study period. These interviews were carried out to inquire about specific aspects of regulatory changes that we did not find in the grey literature or peer review publications but which were necessary for meaningful comparison between the four countries.
In Latin America, medicines distribution is separated by private and public channels. The public sector channel supplies public health services running their own pharmacies and dispensing outlets in clinics and hospitals. To obtain an antibiotic in the public sector, a patient must present a prescription issued by physicians working in the institution to which the pharmacy or dispensing outlet belongs. Therefore, a policy change related to sales restriction is not expected to show any direct effect on antibiotic dispensing in the public sector. There could be an indirect effect as more patients might use public sector sources to obtain prescriptions if a prescription is obligatory in the private sector. The private (retail) sector channel is in charge of distributing pharmaceutical products to pharmacies, private clinics and hospitals outside the public health institutions. One would expect a sales restriction to affect the private sector as sales without prescription and clinically not justified would be prohibited and therefore result in lower total sales. Our main study hypothesis was that the consumption of antibiotics in the private sector would decrease following enforcement because there is some evidence that a significant proportion of sales of antibiotics occurs without a prescription; for instance, in Mexico 46% of sales of antibiotics occurred without a prescription (Wirtz et al. 2007). Another hypothesis was that through prohibition of OTC sales the consumption of some therapeutic groups would be affected more than others.
Medicines consumption can be measured using sales, prescription or dispensing data. To measure antibiotic consumption following enforcement, it would be ideal to have total sales data at the national level in both public and private sectors, with the possibility of stratification by therapeutic group, hospital versus ambulatory care, and sales with or without prescription. Unfortunately, such comprehensive data sets described for many European countries (Vander Stichele et al. 2004) are not available for the majority of other countries.
Access to national prescription or dispensing data on medicines consumption in the study countries is difficult to obtain because public and private sectors are very fragmented and electronic record keeping differs by country. However, sales data for the private sector are recorded for marketing study purposes at the national level, so we requested sales data on antibiotic consumption from IMS Health for Chile, Colombia and Venezuela (IMS Health 2011a). We added a request for data from Mexico as a comparator because no change in antibiotic policies had been introduced before August 2010. A study protocol submitted to IMS Health was approved. Sales data were provided free of charge for the purpose of academic research. All the information in the data set was retrieved at global level in the same manner for each country to ensure comparability. The set consisted of wholesaler data and in two countries (Colombia and Chile) of data collected at pharmacy level in addition to wholesaler data. As the data provided for the study was sales data from each study country, it was not possible to stratify between medicines sold with prescription and without prescription.
We defined antibiotics by the anatomic therapeutic classification (ATC) (class J01) (WHO CC, 2011) and converted quarterly reported kilogram sales into daily defined doses (DDD) per 1 000 inhabitants (DID) per day for each country. The DDD is a standard measure to compare consumption of medicines across setting assuming an average maintenance dose per day for a given medicine in adults (WHO Collaborating Centre for Drug Statistics Methodology Norwegian Institute of Public Health 2011). Antibiotics for which no DDDs were available as well as products for which no DDD could be calculated (e.g. dose of active substance) were excluded. For products containing a combination of active antimicrobial substances without a particular DDD assigned to the combination – such as amoxicillin in combination with dicloxacillin – the DDD for each active substance was taken into account. Combinations for which no DDD was available represented less than 10% of the total DDD consumed in each study country.
We studied changes in consumption by main antibiotic therapeutic groups (penicillins, sulphonamides (single and in combination with trimethoprim), macrolides-lincosamides, quinolones, first/second-generation cephalosporins and third-generation cephalosporins). The therapeutic groups were selected according to sales restrictions implemented: enforcement in Chile and Colombia affected all antibiotics; in Venezuela enforcement affected three therapeutic groups (macrolides, quinolones and third-generation cephalosporins). In Mexico, no enforcement was enacted. The impact of regulation on antibiotic consumption was estimated using interrupted time series analysis (Wagner et al. 2002) taking into account 12 quarterly volume sales values before and 12 after the intervention (for Chile only 11 before and 11 afterwards were taken into account as data were available from 1997 onwards). The method has been described elsewhere in more detail and found particularly suitable in the context of the impact evaluation of promoting appropriate use of antibiotics (MacDougall & Polk 2005). Briefly, the interrupted time series analysis permits to measure the changes in the level and the slope of the consumption over time comparing the data before and after the intervention (Wagner et al. 2002). We limited the timeframe for data analysis as we hypothesised that sales in periods before or after the intervention would potentially be confounded by factors other than the intervention.
We assessed data completeness and quality and the implications for feasibility and validity of analysis using a checklist developed by Vander Stichele et al. (2004). Developed for the purpose of drug use research and in particular for cross-national comparison in relation to the European Surveillance of Antibiotic Consumption (ESAC) project (Vander Stichele et al. 2004), the list recommends checking for problems related to (i) ambulatory/hospital care mix, (ii) drug coverage, and (iii) population coverage (Table 1). We used descriptions of the data sets for each country provided by IMS Health to check whether and to what extent each potential problem applies to our data set.
|Ambulatory/hospital care mix||Stratification by type of care||Difference in utilisation of antibiotics according to type of care; ambulatory care of particular relevance for the policy measure under study (prohibition of OTC sales)|
|Drug coverage|| |
|Under-detection of OTC sales or other products excluded from reimbursement results in errors|
|Population coverage|| |
Pharmaceutical market coverage
|Information on the percentage that the retail sales represents out of the total pharmaceutical market at country level as well as the number of inhabitants accessing medicines via private retail sector|
According to the literature review and key informant interviews, OTC antibiotic sales have been prohibited in all 4 study countries for more than 20 years. But, the law was only enforced in Chile from the last quarter of 1999. In Colombia, enforcement officially started in the second quarter of 2005 although a regulation restricting dispensing of antibiotics was already in force in the first quarter of 2005. In Venezuela, enforcement was announced in the first quarter of 2006 (Table 2). Mexico did not enforce any regulations within the study time frame.
|Country||Date of publication||Geographical spread||Nature of the legal enforcement||Therapeutic groups of antibiotics affected||Additional implementation strategies|
|Inspections of pharmacies||Retention of prescription||Information campaign||Involvement of pharmacies in the promotion of the enforcement|
|Colombia||May 2005||Capital District||+||−||All||+||−|
In Chile, the policy change was motivated by the concern of infectious diseases specialists about the increasing consumption of antibiotics in the 1990s and resulting rising levels of resistance; after several years of preparation, a steering committee took the lead in developing and implementing enforcement measures to prohibit OTC antibiotics sales (Bavestrello & Cabello 2011). In Colombia, antimicrobial resistance also triggered policy changes, which included the prohibition of dividing packages and selling of antibiotics in single pill form. A ban on antibiotics sales without a prescription came into effect in the first and second quarter of 2005, respectively (Secretaria Distrital de Salud de Bogotá 2005; Concejo de Bogota 2005). In Venezuela, official documents show that the enforcement of sales restrictions of antibiotics had been discussed as early as in 2002 (Gaceta Oficial de Venezuela 2002), but was officially introduced only in January 2006 (Gaceta Oficial de la Republica Bolivariana de Venezuela 2006).
While in Chile and Venezuela enforcement was applied nationwide, in Colombia, it affected only the capital district. In Chile and Colombia, all antibiotics were affected by the prohibition; in Venezuela, the regulation targeted only quinolones, macrolides-lincosamides, third-generation cephalosporins and antibiotics including rifampicin (Gaceta Oficial de la Republica Bolivariana de Venezuela 2006). Chile and Colombia (Lopez et al. 2009) also used public information campaigns to inform the population before and during the implementation. No such campaigns were carried out in Venezuela and Mexico.
Only in Chile was there evidence that pharmacies were actively involved in preparing the implementation. Antibiotics were removed from the list of medicines having sales incentives provided for pharmacy staff (Bavestrello & Cabello 2011).
Figure 1 shows antibiotic consumption for each study country between 1997 and 2009. In Chile antibiotic consumption decreased sharply from 12.3 DID before the intervention in 1999 to 8.5 DID just after the enforcement in 2000. The change in the level of consumption immediately after the enforcement was significant (−5.56; P < 0.05), but not the change in the trend of consumption after the intervention compared with before the intervention. From 2002 on, consumption data show a slow increase. In Colombia, antibiotic consumption decreased during the total study period (9.16 DID to 6.76 DID). Between the last quarter of 2004 and the first quarter of 2005, there was a significant change in level (−1.00; P = 0.001) but not in the trend. Meanwhile, in Venezuela antibiotic consumption, data show an increase from 9.99 DID to 15.38 DID during the entire study period, and there is no statistically significant change in level or trend of consumption before and after enforcement. For Mexico consumption, data indicate a steady decrease over time, from 11.6 in 1997 to 9.2 in 2009.
Figure 2 shows antibiotic consumption by therapeutic group and country of intervention. For Chile and Colombia, a decrease in the groups of penicillins consumed is visible. In contrast, consumption for penicillins increased sharply in Venezuela and was nearly constant in Mexico. The other therapeutic groups studied show very similar trends in all countries: whereas, there was decreasing consumption of sulphonamides, for macrolides and particularly quinolones overall consumption increased during the study period; cephalosporin consumption slightly increased or remained constant.
Problems due to ambulatory/hospital care mix were not an important issue with respect to study data because they captured only retail sales to private pharmacies. Excluded were all sales from wholesalers or manufacturers to hospitals except for Venezuela where 0.4% of the total sales data were hospital data (Table 3).
|1. Healthcare level in the private sector captured by the data source||Ambulatory care (AC)||Yes||Yes||Yes||Yes|
|Hospital care (HC)||No||No||Yesa||No|
|2. Drug coverage||Underreporting of OTC sales or products excluded from reimbursement||No||No||No||No|
|3. Data coverage of total pharmaceutical market volumea||Beginning of the study periodb(coverage in%)||84.0 (1997)||85.7(2002)||68.9(2003)|| |
|End of the study period (coverage in%)||84.0 (2001)||67.8(2007)||68.9c(2008)|| |
We did not find bias in our data related to drug coverage for two reasons: first, the data represent retail sales independent of medicines reimbursement; second, the authors assigned ATC/DDD status during data analysis rather than from external data sources. Assignments were double checked by the authors to avoid bias.
According to IMS Health information, pharmaceutical volume coverage varied among countries during the study period. It was highest in Chile, with 84% before and after the intervention. In Colombia, coverage was 85.7% in 2005 at the time of the intervention and decreased to 67.8% 2 years (2007) after the intervention took place.
In Venezuela, distribution estimates were updated only in 1996, 2001 and 2010, so it was assumed that no significant change in the percentage of volume had occurred between these dates. This was supported by the fact that the percentage between the estimation in 2001 and 2010 was very similar (68.8 versus 67.8%). In Mexico, pharmaceutical volume coverage was the lowest (46% in 2006 and 41% in 2010).
Evaluating the impact of prohibiting OTC sales of antibiotics is important for informing policy makers about the effectiveness of measures taken, including undesired effects of interventions. Few analyses have studied policy changes in prohibiting OTC antibiotic sales in low- and middle-income countries. To our knowledge, the present study is the first to analyse systematically sales restrictions in various countries to compare the effects of interventions.
With respect to our first two study objectives, we think that the differences of the policies and their implementation between countries might explain the variation in their impact on consumption. Chile's effort to prohibit OTC antibiotic sales was accompanied by a public awareness campaign on prudent use of antibiotics and the involvement of key stakeholders in the preparation of the intervention (Bavestrello & Cabello 2011). Whether the reduction of 30% antibiotic consumption after the prohibition in Chile means that the intervention was effective in prohibiting all OTC antibiotic sales is difficult to tell, as robust national population wide estimates are lacking but a study of a paediatric population prior to the policy change in 1995 in Chile show that more than 30% of antibiotics were sold without a medical prescription (Soto et al. 1999). In Colombia, the enforcement took effect in 2005 and was restricted to the Capital District of Bogota and not nationwide. Our data showing a small effect of the regulation are confirmed by a recent publication reporting that adherence to the sales restrictions was very low in 2008, 3 years after implementation: of all pharmacies visited, 80.3% did not request a prescription before selling antibiotics (Vacca et al. 2011). Our data from Venezuela indicate very limited involvement by various stakeholders in preparing the enforcement and an absence of involvement by pharmacy owner associations. It appears that enforcement in Venezuela was made through formal government publication but was not followed by strategies to ensure implementation such as pharmacy supervision, pharmacy closures or financial sanctions for non-compliance. Interestingly, the policy interventions in Chile and Colombia affected mainly penicillins, which are antibiotics commonly used by self-prescription, as other studies have pointed out (Wirtz et al. 2007). The use of some wide-spectrum antibiotic (quinolones) is increasing and does not seem to be affected by the intervention.
With respect to the third objective of analysing the completeness and quality of data, we found a series of limitations affecting a precise analysis of antibiotic consumption at the national level and also for interpreting the results. First, there was a lack of detailed information on policy development and implementation for Colombia and Venezuela, which certainly limits our ability to explain the reasons for the differential effects of the interventions in the study countries. Complementing the present analysis with a qualitative study of interventions to comprehend potential causes of differential effects is warranted. Second, with the exception of a very small proportion of hospital sales in Venezuela, the study data represented ambulatory care sales data. We cannot entirely exclude the possibility that some private pharmacies might sell medicines to hospitals or vice versa, but we are confident that the overwhelming majority of sales are for ambulatory use as confirmed by utilisation patterns of the different therapeutic classes. Bias due to medicines benefit coverage as reported for some European countries where data represented a mix of reimbursed and non-reimbursed medicines (Vander Stichele et al. 2004) was not found to be an issue in our study data.
Finally, our biggest concern regarding the validity of our consumption data analysis is related to changes in population coverage as changes in the number of people obtaining their medicines through the retail sector would affect the estimated consumption level. In Colombia and Mexico, increasing numbers of individuals reported having health insurance, which might explain the apparent slow and steady decrease in antibiotic consumption over time as more and more individuals regularly use the public sector, which provides prepaid services offered by health insurance and rely less on private services which require out-of-pocket expenditure. According to World Bank data (2012) on household health insurance coverage, population coverage increased in Colombia from 66% in 2000 to 93% in 2010 and in Mexico from 53% in 2000 to 78% in 2010. In Venezuela, no decrease in antibiotic consumption was found. Instead, antibiotic consumption increased which might indicate that there are influencing factors in addition to the regulatory change. It may be speculated that more individuals use private sector sources as the public health sector in Venezuela has been found fragile and very vulnerable to changes in economic conditions (Bonvecchio et al. 2011).
As the precise number of individuals using the retail sector in a given time is unknown for the study countries, we used the total population as denominator. Another possible method would be to use the percentage of the population with health insurance, assuming that it is an approximation of the extent to which the retail sector is used to access medicines. However, in the Latin American context, some caution is warranted in using population insurance coverage as an approximation for estimating access: for instance, in Mexico, 30% of those with Social Security insurance use private services when having an acute health problem (one reason among others is long waiting times). Alternatively, many without health insurance use public rather than private sector facilities to obtain medicines, sometimes incurring a small fee (Olaiz et al. 2006). In addition, decisions on whether to use private or public sectors probably depend also on geographical location and health condition. To accurately determine medicines utilisation at population level one would need to either obtain total pharmaceutical market volume for the entire population – which at present (2012) is not commercially nor publically available for most Latin American countries – or to extrapolate consumption to 100% of the total pharmaceutical market (Wirtz et al. 2010).
One study assumption was that the prevalence of infections was stable during the study time. Our, literature searches revealed that the increasing consumption of antibiotics in Venezuela cannot be explained by an increase in the prevalence of infectious diseases.
The retail sales data used in this study is based on reported sales volume by wholesalers and – for some of the countries – by individual retail pharmacies. Systematic or random under – as well as overreporting of sales by wholesalers or pharmacies can result in errors, which has an effect on the estimates of consumption a retail level. A validation of the data is conducted annually for each country by IMS Health to verify the estimations and results are used to correct for errors (IMS Health 2011b). Currently, there are no other information sources to corroborate the retail sector sales estimates for concurrent validity.
We used times series, a standard methodology for analysing pharmaceutical policy impact for our analysis (Wagner et al. 2002). However, this method does not permit excluding confounders causing changes in the level or trend of antibiotic consumption.
We found differences not only in the process of implementing the prohibition of over-the-counter antibiotics sales in the four Latin American countries studied but also in their impact on antibiotic consumption. It seems that a more comprehensive set of interventions including the involvement of stakeholders, promoting changes in prescribing behaviour and a public education campaign can enhance the uptake of the policy measure. However, the results also show that interventions have to be sustained over time. This implies developing periodical public awareness campaigns and activities to engage pharmacies, as well as systems to monitor compliance. Even though using retail sales data have a number of limitations, these data can be used to inform policy making as long as those limitations are taken into account. They should be complemented with other data sets that allow a more complete quantification of the overall pharmaceutical consumption at the national level.