Antiretroviral drug expenditure, pricing and judicial demand: an analysis of federal procurement data in Brazil from 2004–2011

Overall drug procurement rose dramatically from a low of 124 DDDs/1000 persons-under-treatment/day in 2004 to 929 DDDs/1000 persons-under-treatment/day in 2005. While procurement from the private sector involved all three major classes: nucleoside reverse transcriptase inhibitors (NRTIs), non-nucleoside reverse transcriptase inhibitors (NNRTIs), and protease inhibitors (PIs) from the years 2004 until 2006, by 2007 the procurement of NNRTIs fell drastically and PIs became the predominant class of medications procured from the private sector by the federal government. This trend continued in subsequent years (with the exception of 2008, where a large purchase of tenofovir resulted in higher number of DDDs/1000 persons-under-treatment/day of NRTIs when compared to PIs). In 2008, the government began to procure newer classes of drugs such as integrase inhibitors (raltegravir), entry inhibitors (maraviroc) and fusion inhibitors (enfurvitide). In 2010, there was a significant reduction in overall drug procurement from the private sector (from 1534 DDDs/1000 persons-under-treatment/day to 238 DDDs/1000 persons-under-treatment/day), which was likely due to the fact that some intensively purchased or expensive medications such as lopinavir/ritonavir and darunavir were possibly procured in excess quantities during the previous year in order to cover the needs of the national treatment program for a two year span (data not shown). Additionally, the NRTI tenofovir (which represented the vast majority of NRTIs utilized in prior years) was not purchased from the private sector in either 2010 or 2011 due to domestic production in Brazil’s national public laboratories.

In 2004, the predominant PI procured was saquinavir. In 2005 and 2006, the most frequently procured PIs were atazanavir (362 DDDs/1000 persons-under-treatment/day) and nelfinavir (14 DDDs/1000 persons-under-treatment/day). In 2007, lopinavir/ritonavir became predominant PI (306 DDDs/1000 persons-under-treatment/day). This was also the case for the years 2009 and 2011. On alternating years (2008 and 2010), atazanavir was the most frequently procured PI.

The curve of federal expenditures follows that of drug procurement (Figure 1). Overall federal expenditures for private sector ARV procurement more than doubled from $114 million dollars in 2004 to $287 million dollars in 2005. Expenditures then came down slowly over the next three years (with a low of $176 million dollars in 2007). In 2009, federal expenditures for ARVs again doubled to reach a high of $375 million dollars before falling to $69 million dollars in 2010. In 2011, expenditures returned to 2006 levels ($262 million dollars).

In general, between 2007 to 2011, the price of medications per DDD were higher for drugs which were added to Brazil’s national treatment guidelines more recently when compared to older ARVs (see Table 1). For example, on average, medications which were introduced prior to 2007 (e.g. didanosine, saquinavir) had price per DDD ranging from $1.55 to $7.98. Medications procured more recently such as raltegravir and maraviroc had price per DDD ranges between $18.88 and $33.02.

Some medications experienced significant price reductions over time. For example, Brazil paid $80.06 per DDD for lopinavir 133 mg/ritonavir 33.3 mg capsules in 2004. By 2011, the government of Brazil purchased only the 200 mg/50 mg heat-stable formulation of lopinavir/ritonavir at a cost of $2.31 per DDD. While the exact rationale for this thirty-fold reduction in price per DDD cannot be derived from the SIASG database, our hypothesis is that Brazil’s involvement in an international movement demanding access to ARVs and the threat of a compulsory license for lopinavir/ritonavir allowed the Ministry of Health to negotiate significant price reductions for this medication with Abbott Laboratories [5].

Figure 2 shows the number of purchases per year of four medications (darunavir, etravirine, raltegravir and maraviroc) based on the justification of judicial action. As demonstrated, the overall number of judicial actions increased from 2007 to 2011, with a high in 2008 of 35 judicial actions resulting in ARV purchases.

A higher number of judicial actions for a drug made it more likely for that drug to be subsequently incorporated into the national treatment guidelines. For example, there were 20 judicial actions in 2007 for darunavir between January 7^th and December 31^st resulting in the purchase of 13,200 units of the drug. Darunavir was added to the treatment guidelines in October 2007. On December 26^th, shortly after incorporation, the government purchased 2.28 million units of darunavir. After a drug was added to the treatment guidelines, judicial demand falls dramatically. For example, there was only one darunavir purchase due to judicial action in the years following its incorporation. This trend continued for all other ARVs in the database.

There were only 6 judicial actions for raltegravir in 2009. However, in the year prior to its incorporation, there were 23 judicial actions resulting in 9,360 units purchased. On November 18^th, 2008 there was a purchase of 720,000 units of raltegravir, likely in anticipation of its official incorporation into the consensus guidelines by January 2009.

There were 5 judicial actions in 2010 for etravirine resulting 3,240 units purchased. Etravirine was incorporated into the treatment guidelines on October 2010. On August 12, 2010, the government procured 403,200 units of etravirine.

From 2009 to 2011, there were over 20 judicial actions each year resulting in purchases of maraviroc. In 2011, there were 29 judicial actions for maraviroc resulting in the purchase of 24,600 units of maraviroc. This medication was not included until the most recent supplement to the treatment guidelines (July 2012).

Our results show that the pricing of drugs has not remained stable. Medicines which were introduced at very high prices such as and lopinavir/ritonavir achieved significant price reductions over time. In general, ARVs first introduced in Brazil after 2007 cost significantly more per DDD than ARVs introduced prior to 2007.

Our results suggest that the government of Brazil should be prepared for a trend towards the use of newer, more expensive medications. Although these newer medications are reserved for cases of treatment failure or for salvage regimens, the fact that the AIDS population in Brazil is one of the oldest treatment cohorts among developing countries suggest that newer medicines will be needed. Our results indicate a trend towards growth in the procurement of fusion inhibitors, integrase inhibitors and newer PIs which have no generic competition. As such, in order for the program to remain sustainable, efforts should be aggressively pursued to reduce prices through price negotiation, exercising TRIPS flexibilities and local production [22].

Our results indicate that the number of judicial actions is related to timing of incorporation into Brazil’s national treatment guidelines. With the exception of etravirine (which was incorporated after published data showed improved outcomes for treatment-experienced patients), the government of Brazil usually timed its purchases of large quantities of newer ARVs following the results of numerous court cases, often greater than 20 a year, in favor of plaintiffs [23]. Shortly after these large purchases by the Ministry of Health, the medication was incorporated into the following years’ treatment guidelines.

These results suggest one of three possibilities: 1) the relationship between number of judicial actions and timing of incorporation into the national treatment guideline is coincidental (unlikely), 2) judicial demand is a reflection of established treatment preferences by prescribers (which prompts review by the expert committees who meet annually to draft treatment guidelines), or 3) judicial action may exert a previously undescribed degree of influence on a process which is presumed to be objective and evidence-based.

Our work supports previous literature which has described the impact of litigation for access to medicines [24‐27]. While many have suggested that drug companies may be using patient advocacy groups to expand market share through litigation, our study is the first which directly examines the impact of judicial cases on national drug procurement [28]. Overall, our results indicate that judicial demand has been highly successful in granting access to newer ARVs.

One of the strengths of our study was that we were able to monitor trends in expenditures for private sector purchases of ARVs by the federal government. These medications often account for the bulk of expenditures for AIDS treatment because they are patented, and have little or no available competitors in the national or international market. Additionally, this type of drug procurement may be a proxy for measuring external dependency in the medicines market as all recorded drugs were purchased either directly from foreign suppliers, or from domestic private pharmaceutical companies operating under licensing agreements with foreign companies. Another strength is that our data is extracted from only one ministry located in the global South, and does not include figures from international sources such as the Global Fund, PEPFAR, or Clinton Health Access Initiative [29].

Limitations of our study include the fact that we could only describe one measure of drug utilization (procurement), and did not have access to other measures such as number of prescriptions or level of consumption. We could not describe relative use of drug classes in the population directly because we did not have access to treatment data. Additionally, because the SIASG database does not include public domestic production, we could not estimate overall ARV procurement as a proxy of drug utilization for the entire country. However, we did validate our data by comparing our expenditures with what was available from the literature and the national AIDS program.

Our work opens the possibilities for many future studies. For example, the SIASG database could be examined to compare prices paid with patent status of medications in Brazil [30]. Additionally, if this data could be combined with drug procurement measures from public national manufacturers and drug prescribing and dispensing (a proxy to consumption), a correlation would be shown between increased drug utilization and improved clinical outcomes such as decreased AIDS morbidity/mortality and reduced transmission. Another study could further explore the causes for lopinavir/ritonavir’s thirty-fold reduction in price from 2004 to 2007.