Progress in trial registration in Latin America and the Caribbean, 2007–2013

ABSTRACT This descriptive study identifies trends in clinical trial registration in the World Health Organization International Clinical Trial Registry Platform (ICTRP) for Latin America and the Caribbean (LAC), from 2007–2013, and provides adjusted estimates for registration rates by population and publications (2007–2011). Trends and data are presented by subregion and language in interactive graphs, including annual registration rates by population (2007–2011) and publications (LILACS and MEDLINE) listed in SCIENTI Network (Science and Technology Indicators). Of the 11 945 clinical trials involving LAC countries, 8 282 were in South America, with Brazil leading at 4 070 (49%); 2 421 in North and Central America, with Mexico leading at 1 886 (78%); and 1 242 in the Caribbean, with Puerto Rico leading at 857 (69%). After adjusting by population and publication rates Chile, Panama, Argentina, and Peru led registration rates per 1 million inhabitants. Variations in the number of trials per year are quite substantial. Clinical trial registration increased in a steady yet inconsistent way. The implementation of the Policy on Research for Health has been followed by an increase in countries that require registration and have established clinical trial registries. However, there is room for improvement in adherence throughout LAC. Trial registration is offered gratis by Brazilian, Cuban, Peruvian, and United States registries, among others.

Research for health should inform sound investments in the health sector (1). Trial registration is used as a strategy for promoting transparency and trust in research and is promoted by PAHO and WHO policies.
In 2015, WHO reiterated its position on requiring clinical trial registration (2,3). Also, the BMJ called for research transparency for non-regulated interventions (e.g. diets, exercise programs, surgical procedures) to understand the harms and benefits of those interventions (4).
In 2009, PAHO/WHO Member States approved the "Policy on Research for Health" (the Policy; 5). The Policy called for better governance and transparency in research, and the strengthening of research systems that inform policies for health, prevention, and health care. This first-of-its-kind, Regional policy addressed nearly 30 years-worth of calls to improve research transparency using trial registration. It also called for a better use of the International Clinical Trial Registry Platform (ICTRP), the WHO meta-registry of clinical trials that organizes standardized information on 20 variables from approved data providers (6). Additionally, the "Strategy on Research for Health," adopted at the World Health Assembly in 2010, mandated clinical trial registration (7).
The Policy focuses on strengthening the building blocks of "research for health systems," and trial registration affects several of its key objectives. For example, trial registration contributes "to foster best practices and enhanced standards for research" and to promote transparency; it also improves access to information. It is essential to "maintaining public trust, confidence, and participation in research." Hence, it calls for Member States to "fulfill research registration and enable inventories and registries of research comparable and integrated with the ICTRP primary registries, and to adopt standard identifiers and data-set collections that contribute to international registration efforts and international ethics and publications standards" (1). Furthermore, clinical trial registration facilitates knowledge translation, the identification of competent human resources for research, and the development of partnerships.
Clinical trial registration had a boon during the 2004 Ministerial Forum on Health Research in Mexico leading to the subsequent development of the ICTRP. The initial assessment of registration in Latin America and the Caribbean (8) found a 17-fold, yet inconsistent, increase in clinical trial registration during first quinquennium of ICTRP (9).
Clinical trial registration has been promoted through different entry points ( Figure 1), such as national legislation requiring registration in publicly accessible databases, or through national clinical trial registries (10). In some countries, agencies request proof of trial registration to fund research. Some require registration to publish research reports, approve a project, or allow for ethical approval (9)(10)(11). The scientific community and the World Medical Association have been among the many stakeholders and ethics experts advocating for such transparency and governance measures (12,13), together with publishers and scientific journals that have made registration a requirement, even if compliance needs to be enhanced (9).
The implementation of the Policy is being assessed, as well as adherence to clinical trial registration since much has changed since ICTRP was set up. For example, the Ministerial Forum on Research for Health called for "…standards, regulations, and best practices for fair, accountable, and transparent research processes; guaranteeing quality and safety of patient care; registration and results reporting of clinical trials; open and equitable accessing to research data, tools, and information" (14).
Trial registration has remained a key strategy for transparency and dissemination of findings (15) and for informing initiatives, such as the Global Research for Health Observatory requested by the World Health Assembly in 2013 (16). Other stakeholders complement these efforts by improving transparency and public trust in research for health. For example, +AllTrials (17,18), the REWARD Initiative (19)(20)(21)(22)(23)(24), Equator Network (25), the Lancet, the BMJ, other leading jour-nals (26,27) and The Cochrane Collaboration all advocate for research registration.
The present study assessed registration trends in ICTRP, pulling together data from 15 primary registries and 3 other data providers (28). Primary registries meet WHO standards capturing data on the 20 items of the required minimum data set; other data providers meet some of these standards (10) and may not have a national or regional remit, or the support of the government, nor be open to registration by all.
The study objectives were: to describe the trends among clinical trials registered in ICTRP and being carried out in Latin America and the Caribbean (LAC) in 2007-2013; to provide adjusted registration rates by population and publication for 2007-2011; to deliver helpful recommendations for improving research governance; and to support transparency in research, increasing its value and reducing waste.

MATERIALS AND METHODS
This was a descriptive analysis of clinical trials from LAC countries registered in ICTRP, following the protocol developed by one of the authors (PRF) as part of a research internship with PAHO/WHO.

Inclusion and exclusion criteria
The data was comprised of clinical trials registered in the ICTRP and being conducted in LAC (29). The advanced search function of ICTRP was used to identify these trials. The searches were run on 11-22 August 2014. The study used the WHO definition of clinical trial: "any research study that prospectively assigns human participants or groups of humans to one or more health-related interventions to evaluate the effects on health outcomes (30)." Countries not listed in ICTRP would have no trials registered. The countries and overseas territories excluded were not listed in the ICTRP because they had no registered trials. 2  gorithms were excluded using the identification number and date of registration.

Data extraction and analysis
Registered trials were identified by year, whether conducted in a single country or as part of a multi-country, multicentric study. The data were stratified by geographic distribution within LAC, and by language, as follows:  Population-and publication-adjusted rates are available from the interactive graphs in the Supplementary Materials, enabling readers to explore trends by country, language, and geographic area (Box 1).

Geographic area
North/Central America. Mexico, Guatemala, and Panama led in this area with 1 886 (77.9%), 203 (8.4%), and 152 (6.3%) registered trials, respectively. Honduras with 28 (1.2%) and Nicaragua with 6 (0.2%) had the least (Table 1) Caribbean. Puerto Rico, Cuba, and the Dominican Republic led trial registration with 857 (69.0%), 218 (17.6%), and 91 (7.3%), respectively. Conversely, Suriname, Trinidad and Tobago, and Saint Kitts and Nevis had the lowest registration during these 7 years: 1 registration for Suriname and 0 from the other two. This most likely reflects poor compliance; research institutions in these countries conduct clinical trials on a regular basis (Table 1).
Puerto Rico led registration peaking in 2011 (151 trials), but decreased since then; only 74 trials were registered in 2013, a 49% reduction.
Brazil developed a national registry and doubled its registrations during the study period.

BOX 1. Internet links for viewing the clinical trials registration trends and yearly rates, adjusted by population or publications, in Latin America and the Caribbean
Other languages. This group had the lowest registration: 76 trials were registered for 11 countries during these 7 years: Jamaica, Haiti, and Belize led with 24 (31.6%), 13 (17.1%), and 12 (15.8%) trials, respectively. No trials were registered for Saint Kitts and Nevis nor for  (Table 1).

Yearly registration rates, by publication and population
Clinical trial registration rates per 100 publications were estimated using data in LILACS and MEDLINE per year, according to the Network on Science and Technology Indicators for 24 countries in LAC for 2007-2011 (31).

Adjusted rates for registration, by population and year
Clinical trial registration rates per 1 million inhabitants were analyzed. Chile led registration rates, with a range of 6.1-8.1 trials / 1 million inhabitants; then Panama, with 3.6-7.4; Argentina, 4.0-6.4; and Peru, 3.3-4.3.
As mentioned, Costa Rica's registration plunged following a moratorium on clinical trials. In 2007, Costa Rica was leading registration in LAC with 7.7 trials / 1 million inhabitants, but fell to 1.7 trials by 2011 (Table 2).

DISCUSSION
Overall, LAC increased its absolute and relative registration during the study period. Clinical trial registration rates adjusted by publications also increased, but there is room for improvement since many eligible trials appear to be missing.
Costa Rica was an outlier, with high registration rates by population and a sharp decline following the moratorium of 2010 (32). In 2011, the country's legislative assembly approved the Ley Reguladora de Investigación Biomédica (Act Regulating Biomedical Research) to regulate biomedical research with humans, and created the National Health Research Council to guarantee research quality, the protection of human research subjects, and their rights. The Government of Costa Rica has to implement this act by developing further legislation (29, 32). The moratorium had a lasting impact and the country will need time and effort to regain its leadership in trial development and registration. It is unclear whether the dip in Costa Rica reflects a loss of clinical trial volume, a loss in registration, or both.
Clinical trials registration fluctuated in Argentina. In documenting barriers to trials registration, various issues were found: researchers were unfamiliar with registration platforms, other than Clinicaltrials.gov; few investigators knew the registration process; and responsibilities, especially for multicentric studies sponsored by the industry, were diluted. In a nutshell, researches did not feel compelled to check the registration status of their trials (33).
Brazil increased clinical trial registration during the 7-year study period and launched a WHO-accredited national primary registry; the country invested in its national health research governance and is decidedly supporting registration (34,35). However, it faces challenges in increasing adherence, such as improving the standards for data collection in fields subject to interpretation (e.g., free text fields), and streamlining the registration process to make it swift and competitive (36).
Additional national and regional WHO-accredited primary registries may be necessary in the Americas since registries serve specific governance and monitoring requirements. However, efficiencies may be found in regional registries, and perhaps these can be aligned with regional integration agreements, such as CARICOM, or the Central American Integration System. Some countries may have the volume of trials to justify investing in a sustainable national registry, while  Special report Rodríguez-Feria & Cuervo • Trial registration in Latin America and the Caribbean others may find efficiency and sustainability in supporting a regional registry within an integrated system. Two national registries have been accredited as Primary Registries for ICTRP (the Cuban Public Registry of Clinical Trials and Brazilian Clinical Trials Registry); Peru is expected to complete its accreditation soon, but there is still room for growth (28).
Africa offered a collaborative model solving many challenges for trials registration through the Pan African Clinical Trial Registry (PACTR) established in 2007. This registry brought efficiencies, with only marginal increments in cost, and increased coverage from a national to a continental registry. PACTR also offered innovative options for registration that address local needs with adaptations for poor Internet connections (e.g., manual, email, and postal mail registration). By 2010, PACTR had doubled registration, with South Africa leading the trend (37).
An analysis done in Egypt of ICTRP and PACTR showed that registration began in 1999 and peaked for ICTRP in 2012 with 686 trials, and for PACTR in 2013 with 56 (38). Egypt trailed South Africa in registration.
In LAC, the launch of national registries has been followed by increased registration. Similarly, when Choi and colleagues assessed registration in the Korean Clinical Research Information Service established in 2010, they found 1 323 trials registered in 2010-2014 and a growing trend (39). While in Australia, Lam and colleagues found that researchers preferred the Australian New Zealand Clinical Trials Registry in 2008 -2012, over clinicaltrials.gov (3 379 versus 1 764); many had duplicate registrations (40). Researchers in India found 1 826 trials registered in 2007-2011 using Clinical Trial Registry-India, with a growth trend of 3.7%. They concluded that the gradual increase was associated with having their own registry (41).
Globally, and following the launch of ICTRP, registration has increased sevenfold: from 3 294 trials in 2004 to 23 384 in 2013; this reflects improved adherence as trials have not grown at the same rate. Europe peaked in 2005; North America in 2005 and 2008; and Asia has seen a gradual, but steady increase since 2005 (42). In the Americas, registration is mandated by the Policy, and several countries have made it a legal requirement. Adjusting ICTRP registration by popula-tion, Chile, Panama, and Argentina are leading countries; however, absolute figures are dominated by Brazil and Mexico. Costa Rica, Nicaragua, Puerto Rico, and Peru had decreasing registration trends. In 2016, Peru launched its Primary registry. It would be interesting to assess how this influences registration rates. Monitoring and assessment, as conducted in Argentina, helps to identify and address adherence issues.
Although LAC countries have been improving their registration trends, the growth is inconsistent, dominated by some countries and leaving substantial room for improvement by others. Improvement requires working with countries, regulatory agencies, research sponsors, funding agencies, publishers, and ethics review committees. This means addressing limitations in the governance systems, promoting adherence, and providing incentives to increase the value of research and reduce research waste. Therefore, additional legal and technological efforts are needed, and we recommend replicating this assessment in 3 years. Monitoring and collecting data will provide actionable data to guide progress in Latin America and the Caribbean.
ICTRP allows the public, research observatories, sponsors, consumers, and researchers to link research proposals with outputs. It enables systematic reviewers to include unpublished or ongoing research, a practice that is becoming an increasing feature in Cochrane Reviews. Research institutions can organize and give visibility to their clinical trials, identify research partners, peers, and sponsors. This is being done, for example, in a cancer research systems mapping under way (43). Tutorials are becoming available to facilitate registration (44). Still, identifying duplicate clinical trials is a challenge and a science, requiring complex algorithms, something ICTRP advisors and developers had already foreseen. Duplicates are frequently identified comparing variables, such as countries of recruitment, inclusion and exclusion criteria, health condition(s), and intervention(s). However, sometimes these change over time or from one registry to another.
This study provides a benchmark for progress made by implementing the mandate of the PAHO Policy on Research for Health. The assessment can be replicated, and readers can further explore the data using the interactive graphs accessible from Box 1. It also offers actionable data, and even if it is impossible to get an accurate count of clinical trials in each country, registration is a good alternative and a tool for better governance.
Growth in registration, without substantial growth in publications, suggests better adherence to registration. This study informs progress on this aspect of the Policy and the WHO Strategy on Research for Health (4,11). This guidance and further analysis, such as the extent of characteristics of registered research reports on specific topics, enables a variety of various consumers to better determine the links between clinical trials and health research agendas. This helps to better understand if research sponsors and national agendas are aligned with national and regional priorities.

Limitations
This analysis does not address variations in the quality of registration for the 20-item dataset within or between registries. ICTRP registrations can be influenced by retrospective registration (done after recruitment) or when a registry is launched, a law enforced, a requirement is set up (by publishers, for example), or when study start-dates bring inaccuracies. Clinical trials are a subset of all the studies published, and compliance can be complicated and demand many resources. Considering that the denominator is unknown, the figures are a limited indicator. Yet, trends provide actionable data and an idea of whether a plateau has been being reached.

Conclusions and recommendations
Significant progress was made with research registration in PAHO Member States, but these efforts must expand and scale up to achieve consistent adherence throughout the Region of the Americas. Trial registration brings transparency in research for health and has gained momentum and credibility in countries, such as Cuba and Brazil, where leadership in implementing WHO-endorsed primary registries has produced the first Spanish and Portuguese interfaces, respectively, while submitting data in English to the ICTRP. These two registries are open to trial registration from other countries and are now being complemented by the Peruvian registry. Furthermore, publishers are instrumental to increasing adherence when requiring proof of registration, and when displaying registration identification numbers, prominently in publications. Additionally, regulatory agencies can require that applications for medicines and technologies be backed by registered trials. Governments can promote trial registration when collaboration results in efficiencies, as done in Africa with the Pan African Clinical Trial Registry.
We recommend the development of regional registries especially for countries with low volumes of research, where efficiencies and sustainability are key. All countries should promote legislation and regulation that encourages research registration and transparency. For countries with high volumes of research and the need for additional specific information, it may make sense to have a national registry, even though the key components of registration may be satisfied by using the existing registries.
There are pros and cons to having a regional versus several national registries, especially in terms of logistics, efficiencies, and governance. High volumes of trials may justify having national registries, while efficiencies may favor regional registries when volumes are low. Typically, a registry requires at least two fulltime skilled professionals, a cost that makes sense when the volume justifies it. National registries can capture additional information of relevance to a country, becoming a tool of the regulatory and stewardship processes. Regional registries can facilitate standardization (27).
Countries with low trial volumes may benefit from using an international registry, thereby investing fewer resources in the technological tool, and focusing more on legislation and other means of improving adherence. Registration should be expanded to cover observational studies.
Considering the recommendations and local situation, stakeholders can advance strategies to promote transparency in research and adherence to international standards and mandates that foster best practices and standards for research, seek efficiencies and enhanced impact and appropriation of research, strengthen research governance and promote research agendas, and build on public trust and engagement in research.