Linking ClinicalTrials.gov and PubMed to Track Results of Interventional Human Clinical Trials

Objective

In an effort to understand how results of human clinical trials are made public, we analyze a large set of clinical trials registered at ClinicalTrials.gov, the world’s largest clinical trial registry.

Materials and Methods

We considered two trial result artifacts: (1) existence of a trial result journal article that is formally linked to a registered trial or (2) the deposition of a trial’s basic summary results within the registry.

Results

The study sample consisted of 8907 completed, interventional, phase 2-or-higher clinical trials that were completed in 2006-2009. The majority of trials (72.2%) had no structured trial-article link present. A total of 2367 trials (26.6%) deposited basic summary results within the registry. Of those , 969 trials (10.9%) were classified as trials with extended results and 1398 trials (15.7%) were classified as trials with only required basic results. The majority of the trials (54.8%) had no evidence of results, based on either linked result articles or basic summary results (silent trials), while a minimal number (9.2%) report results through both registry deposition and publication.

Discussion

Our study analyzes the body of linked knowledge around clinical trials (which we refer to as the “trialome”). Our results show that most trials do not report results and, for those that do, there is minimal overlap in the types of reporting. We identify several mechanisms by which the linkages between trials and their published results can be increased.

Conclusion

Our study shows that even when combining publications and registry results, and despite availability of several information channels, trial sponsors do not sufficiently meet the mandate to inform the public either via a linked result publication or basic results submission.     

The State of Infectious Diseases Clinical Trials: A Systematic Review of ClinicalTrials.gov

Background

There is a paucity of clinical trials informing specific questions faced by infectious diseases (ID) specialists. The ClinicalTrials.gov registry offers an opportunity to evaluate the ID clinical trials portfolio.

Methods

We examined 40,970 interventional trials registered with ClinicalTrials.gov from 2007–2010, focusing on study conditions and interventions to identify ID-related trials. Relevance to ID was manually confirmed for each programmatically identified trial, yielding 3570 ID trials and 37,400 non-ID trials for analysis.

Results

The number of ID trials was similar to the number of trials identified as belonging to cardiovascular medicine (n = 3437) or mental health (n = 3695) specialties. Slightly over half of ID trials were treatment-oriented trials (53%, vs. 77% for non-ID trials) followed by prevention (38%, vs. 8% in non-ID trials). ID trials tended to be larger than those of other specialties, with a median enrollment of 125 subjects (interquartile range [IQR], 45–400) vs. 60 (IQR, 30–160) for non-ID trials. Most ID studies are randomized (73%) but nonblinded (56%). Industry was the funding source in 51% of ID trials vs. 10% that were primarily NIH-funded. HIV-AIDS trials constitute the largest subset of ID trials (n = 815 [23%]), followed by influenza vaccine (n = 375 [11%]), and hepatitis C (n = 339 [9%]) trials. Relative to U.S. and global mortality rates, HIV-AIDS and hepatitis C virus trials are over-represented, whereas lower respiratory tract infection trials are under-represented in this large sample of ID clinical trials.

Conclusions

This work is the first to characterize ID clinical trials registered in ClinicalTrials.gov, providing a framework to discuss prioritization, methodology, and policy.     

Timing and Completeness of Trial Results Posted at ClinicalTrials.gov and Published in Journals

Background

The US Food and Drug Administration Amendments Act requires results from clinical trials of Food and Drug Administration–approved drugs to be posted at ClinicalTrials.gov within 1 y after trial completion. We compared the timing and completeness of results of drug trials posted at ClinicalTrials.gov and published in journals.

Methods and Findings

We searched ClinicalTrials.gov on March 27, 2012, for randomized controlled trials of drugs with posted results. For a random sample of these trials, we searched PubMed for corresponding publications. Data were extracted independently from ClinicalTrials.gov and from the published articles for trials with results both posted and published. We assessed the time to first public posting or publishing of results and compared the completeness of results posted at ClinicalTrials.gov versus published in journal articles. Completeness was defined as the reporting of all key elements, according to three experts, for the flow of participants, efficacy results, adverse events, and serious adverse events (e.g., for adverse events, reporting of the number of adverse events per arm, without restriction to statistically significant differences between arms for all randomized patients or for those who received at least one treatment dose).From the 600 trials with results posted at ClinicalTrials.gov, we randomly sampled 50% (n = 297) had no corresponding published article. For trials with both posted and published results (n = 202), the median time between primary completion date and first results publicly posted was 19 mo (first quartile = 14, third quartile = 30 mo), and the median time between primary completion date and journal publication was 21 mo (first quartile = 14, third quartile = 28 mo). Reporting was significantly more complete at ClinicalTrials.gov than in the published article for the flow of participants (64% versus 48% of trials, p<0.001), efficacy results (79% versus 69%, p = 0.02), adverse events (73% versus 45%, p<0.001), and serious adverse events (99% versus 63%, p<0.001).The main study limitation was that we considered only the publication describing the results for the primary outcomes.

Conclusions

Our results highlight the need to search ClinicalTrials.gov for both unpublished and published trials. Trial results, especially serious adverse events, are more completely reported at ClinicalTrials.gov than in the published article. Please see later in the article for the Editors'' Summary     

Differences in Investigator-Initiated Trials between Japan and Other Countries: Analyses of Clinical Trials Sponsored by Academia and Government in the ClinicalTrials.gov Registry and in the Three Japanese Registries

Background

Following the amendment of the Pharmaceutical Affairs Law in Japan in 2003 researchers were permitted to begin investigator-initiated trials (IITs). In subsequent years, however, the number of IITs remained low. In other countries in Asia as well as in Europe, North America, and South Africa, the number of IITs has increased over the past decade. The differences in the characteristics of IITs between Japan and other countries are unknown. Some studies have analyzed the characteristics of all clinical trials according to registry databases, but there has been less research focusing on IITs.

Aims

The purpose of this study is to analyze the characteristics of IITs in the ClinicalTrials.gov registry and in the three Japanese registries, to identify differences in IITs between Japan and other countries.

Methods

Using Thomson Reuters Pharma™, trials sponsored by academia and government as IITs in 2010 and registered in ClinicalTrials.gov were identified. IITs from 2004 to 2012 in Japan were identified in the three Japanese registries: the University Hospital Medical Information Network Clinical Trials Registry, the Japan Pharmaceutical Information Center Clinical Trials Information, and the Japan Medical Association Center for Clinical Trials, Clinical Trials Registry. Characterization was made of the trial purposes, phases, participants, masking, arms, design, controls, and other data.

Results

New and revised IITs registered in ClinicalTrials.gov during 2010 averaged about 40% of all sponsor-identified trials. IITs were nearly all early-phase studies with small numbers of participants. A total of 56 Japanese IITs were found over a period of 8 years, and these were also almost nearly all early-phase studies with small numbers of participants.

Conclusion

There appear to be no great differences between Japan and other countries in terms of characteristics of IITs. These results should prompt a new review of the IIT environment in Japan.     

Molecular (RAPD) analysis on Italian taxa of theOphrys bertolonii aggregate (Orchidaceae)

The RAPD technique has been used for analyzing the genetic material of seven Italian populations of theOphrys bertolonii aggregate (Orchidaceae). For comparison the related speciesO. sphegodes andO. fuciflora have also been analyzed. A high genetic variability has been observed within the same populations, and on the whole onlyO. bertolonii can be safely differentiated from the other taxa of the aggregate, which are undistinguishable with this method.     

Thermodynamical model for insertion and aggregate binding of caffeine to the homopolymer poly(riboadenylate) and model choice by data analysis

This paper describes the model used to estimate the parameters of caffeine-poly(riboadenylate) (poly(A)) interactions from corresponding 1H-NMR measurements. The model of insertion and aggregate binding describes the non-cooperative insertion of a molecule C into an interspace between two monomers of a homopolymer in competition with aggregate binding. It contains two binding constants, K1 for insertion and K2 for the interaction of monomeric A units of the polymer with C molecules in bound aggregates, and two cooperativity parameters, Kcc for stacking of C molecules within aggregates and tau which is thought to be due to conformational adaptation of the polymer to those bound aggregates which cover more than one A unit. In contrast to other models, the size of a binding site (within the aggregates) is less than one monomeric unit, with n denoting the maximum number of C molecules per A unit in bound aggregates. The model is developed for general n by means of the method of sequence-generating functions. For n = 2 and n = 3, the correctness of the model treatment was checked by the matrix method. The model is applicable to the binding of aggregates to homopolymers, which are flexible enough to fit their structure to the aggregates.     

Yeast Protein database (YPD): a database for the complete proteome of Saccharomyces cerevisiae.

The Yeast Protein Database (YPD) is a database for the proteins of the budding yeast,Saccharomyces cerevisiae. YPD is the first annotated database for the complete proteome of any organism. Now that the complete genome sequence of yeast is available, YPD contains entries for each of the characterized proteins and for each of the uncharacterized proteins predicted from the sequence. Contained in YPD are the calculated properties of each protein such as molecular weight and isoelectric point, experimentally determined properties such as subcellular localization and post-translational modifications, and extensive annotations from the yeast literature. YPD contains 25 000 lines of textual annotation that describe the known functions, mutant phenotypes, interactions, and other properties for the approximately 6000 proteins in the yeast proteome. The information in YPD is updated daily, and it is available on the World Wide Web at http://www.proteome.com/YPDhome.html .     

The plant phenological online database (PPODB): an online database for long-term phenological data

We present an online database that provides unrestricted and free access to over 16 million plant phenological observations from over 8,000 stations in Central Europe between the years 1880 and 2009. Unique features are (1) a flexible and unrestricted access to a full-fledged database, allowing for a wide range of individual queries and data retrieval, (2) historical data for Germany before 1951 ranging back to 1880, and (3) more than 480 curated long-term time series covering more than 100 years for individual phenological phases and plants combined over Natural Regions in Germany. Time series for single stations or Natural Regions can be accessed through a user-friendly graphical geo-referenced interface. The joint databases made available with the plant phenological database PPODB render accessible an important data source for further analyses of long-term changes in phenology. The database can be accessed via www.ppodb.de.     

'The surface management system' (SuMS) database: a surface-based database to aid cortical surface reconstruction, visualization and analysis.

Surface reconstructions of the cerebral cortex are increasingly widely used in the analysis and visualization of cortical structure, function and connectivity. From a neuroinformatics perspective, dealing with surface-related data poses a number of challenges. These include the multiplicity of configurations in which surfaces are routinely viewed (e.g. inflated maps, spheres and flat maps), plus the diversity of experimental data that can be represented on any given surface. To address these challenges, we have developed a surface management system (SuMS) that allows automated storage and retrieval of complex surface-related datasets. SuMS provides a systematic framework for the classification, storage and retrieval of many types of surface-related data and associated volume data. Within this classification framework, it serves as a version-control system capable of handling large numbers of surface and volume datasets. With built-in database management system support, SuMS provides rapid search and retrieval capabilities across all the datasets, while also incorporating multiple security levels to regulate access. SuMS is implemented in Java and can be accessed via a Web interface (WebSuMS) or using downloaded client software. Thus, SuMS is well positioned to act as a multiplatform, multi-user 'surface request broker' for the neuroscience community.     

The non-redundant Bacillus subtilis (NRSub) database: update 1998.

The non-redundant Bacillus subtilis database (NRSub) has been developed in the context of the sequencing project devoted to this bacterium. As this project has reached completion, the whole genome is now available as a single contig. Thanks to the ACNUC database management system and its associated retrieval system Query_win, each functional region of the genome can be accessed individually. Extra annotations have been added such as accession numbers for the genes, locations on the genetic map, codon adaptation index values, as well as cross-references with other collections. NRSub is distributed through anonymous FTP as a text file in EMBL format and as an ACNUC database. It is also possible to access NRSub through two dedicated World Wide Web servers located in France (http://acnuc. univ-lyon1.fr/nrsub/nrsub.html ) and in Japan (http://ddbjs4h.genes. nig.ac.jp/ ).     

The Yeast Protein Database (YPD): a curated proteome database for Saccharomyces cerevisiae.

The Yeast Protein Database (YPD) is a curated database for the proteome of Saccharomyces cerevisiae . It consists of approximately 6000 Yeast Protein Reports, one for each of the known or predicted yeast proteins. Each Yeast Protein Report is a one-page presentation of protein properties, annotation lines that summarize findings from the literature, and references. In the past year, the number of annotation lines has grown from 25 000 to approximately 35 000, and the number of articles curated has grown from approximately 3500 to >5000. Recently, new data types have been included in YPD: protein-protein interactions, genetic interactions, and regulators of gene expression. Finally, a new layer of information, the YPD Protein Minireviews, has recently been introduced. The Yeast Protein Database can be found on the Web at http://www.proteome.com/YPDhome. html     

Selection and subsequent analysis of sib pair data for QTL detection.

Haseman and Elston (1972) developed a robust regression method for the detection of linkage between a marker and a quantitative trait locus (QTL) using sib pair data. The principle underlying this method is that the difference in phenotypes between pairs of sibs becomes larger as they share a decreasing number of alleles at a particular QTL identical by descent (IBD) from their parents. In this case, phenotypically very different sibs will also on average share a proportion of alleles IBD at any marker linked to the QTL that is lower than the expected value of 0.5. Thus, the deviation of the proportion of marker alleles IBD from the expected value in pairs of sibs selected to be phenotypically different (i.e. discordant) can provide a test for the presence of a QTL. A simple regression method for QTL detection in sib pairs selected for high phenotypic differences is presented here. The power of the analytical method was found to be greater than the power obtained using the standard analysis when samples of sib pairs with high phenotypic differences were used. However, the use of discordant sib pairs was found to be less powerful for QTL detection than alternative selective genotyping schemes based on the phenotypic values of the sibs except with intense selection, when its advantage was only marginal. The most effective selection scheme overall was the use of sib pairs from entire families selected on the basis of high within-family variance for the trait in question. There is little effect of selection on QTL position estimates, which are in good agreement with the simulated values. However, QTL variance estimates are biased to a greater or lesser degree, depending on the selection method.     

Marfan Database (second edition): software and database for the analysis of mutations in the human FBN1 gene.

Fibrillin is the major component of extracellular microfibrils. Mutations in the fibrillin gene on chromosome 15 (FBN1) were described at first in the heritable connective tissue disorder, Marfan syndrome (MFS). More recently, FBN1 has also been shown to harbor mutations related to a spectrum of conditions phenotypically related to MFS. These mutations are private, essentially missense, generally non-recurrent and widely distributed throughout the gene. To date no clear genotype/phenotype relationship has been observed excepted for the localization of neonatal mutations in a cluster between exons 24 and 32. The second version of the computerized Marfan database contains 89 entries. The software has been modified to accomodate new functions and routines.