← vista completa

Quality of reporting and risk of bias of randomized clinical trials published in Spanish and Latin American journals

Calidad del reporte y riesgo de sesgo de los ensayos clínicos aleatorizados publicados en revistas de España y Latinoamérica

In the pyramid of the evidence quality to assess efficacy, effectiveness, and adverse reactions to interventions, randomized clinical/controlled trials (RCT) are conventionally at the top, right below systematic reviews of clinical trials and n-of-1 trials.[1] When RCTs are conducted properly, randomization of participants to treatments ensures that confounding by both known and unknown covariables will be controlled, thus reducing the risk of bias.[2] This is why we use RCTs to decide on the best interventions for our patients, which can be pharmacological, surgical, psychological, educational, informational or manual therapies, as well as changes in processes or the use of devices or medical devices.

Since the advent of the evidence-based medicine movement and the Cochrane Collaboration at the end of the eighties and beginning of the nineties, respectively, there has been a growing interest to improve the quality of the evidence from RCTs. Two main focuses have arisen: the methodological quality of the trials and the quality of the reporting of the findings. The former is concerned about the design and conduct of the RCT to reduce the risk of bias starting from the planning stage[3],[4] and for which a variety of scales, systems, and classifications have been developed to assess internal validity.[5] The latter concerns how the findings are published in the biomedical literature and how the methods and the results of an RCT are actually reported—imprecisions and gaps in the reporting will hinder the correct assessment of the internal validity of the trials[6], the consequence being that much of the medical literature is rendered useless and wasteful. As Chalmers and Glasziou bluntly stated: “Without accessible and usable reports, research cannot help patients and their clinicians.”[7]

Methodological quality of clinical trials has been defined as a set of parameters in the design and conduct of a study that reflects the validity of the outcome of interest[8] and primarily depends on the selection of the participants and sample size; on the random allocation of treatment; on blinding and masking; on patient follow-up with accounting for losses and drop-outs; and on the non-selective reporting of planned endpoints and outcomes.[9] Depending upon how each one of these phases of the clinical trials is planned, conducted, analyzed, and reported, the risk of bias may be higher or lesser. The most commonly used method to assess the risk of bias of RCTs is the Cochrane Collaboration tool.[5]

Acknowledging the need to improve the quality of reporting of the published clinical trials, in 1993 a meeting was held in Ottawa, Canada, that brought together editors of medical journals, trialists, epidemiologists, and methodologists. From this first effort, the first version of the Consolidated Standards of Reporting Trials (CONSORT) Statement came out, which was published in 1996. Subsequently, two other versions have been published, in 2001 and 2010. This last statement was published simultaneously in nine high-impact medical journals and is the version currently in use.[10] The CONSORT guideline sets minimum required standards for an adequate reporting of an RCT. It contains 25 items distributed in six topics: 1) title and abstract, 2) introduction, 3) methods, 4) results, 5) discussion, and 6) ancillary analyses.[11],[12]

After the publication of each CONSORT statement, the leading biomedical journals have gradually incorporated the recommendation to use CONSORT when reporting trial results in their author guidelines. How these recommendations and the integration of CONSORT adherence into the editorial workflow (endorsement and enforcement) have impacted the quality of publications has also been extensively studied. Two systematic reviews evaluated reporting completeness and concluded that journal endorsement and enforcement to CONSORT could lead to favorable results regarding reporting completeness, but that more research is needed.[13],[14] Likewise, another study with a research-on-research design concluded that while poor-quality reporting prevalence has gradually decreased, there is still much to be done, especially in lower impact factor journals.[15]

There are few studies on randomized clinical trials of Latin America. While the region accounts for a significant share of global disease burden, it does not have a corresponding number of published clinical trials in the five top biomedical journals.[16] One study looked at the prevalence of registration in the clinical trials databases of Latin American RCTs in 2010 and compared the methodological characteristics of the articles that reported registration before trial commencement versus those that did not.[17] The authors found that only 4% of the trials were prospectively registered. Another study compared methodological quality with funding source of trials conducted in Latin America but published elsewhere[18]—no statistically significant associations were found. Another descriptive study found that only 13% of a random sample of 101 RCTs published in 56 journals mentioned the CONSORT guideline.[19] More recently, a systematic hand search of RCTs published in dermatology journals in Spanish found a high risk of bias due to gaps in reporting and methodological deficiencies.[20]

The Iberoamerican Cochrane Network currently leads a large project to identify all Spanish-language journals that publish original clinical research in Spain and Latin America. In 2012, 1498 journals had been identified, of which only 3% had an impact factor; 4.1% were indexed in MEDLINE/PubMed; 3.7% were included in EMBASE.[21]Those journals are currently being hand-searched by Cochrane collaborators in order to identify all RCTs published in Spain and Latin America, by specialty. Additionally, an online database has been developed to centralize all hand searching activities and to facilitate submission of references to CENTRAL (Cochrane Collaboration Central Register of Controlled Trials) thus potentially feeding into the systematic reviews that Cochrane conducts on different health topics. This database is BADERI (Base de Datos de Ensayos y Revistas Iberoamericanas) and was officially launched in October 2015.[22] Until 2017, BADERI had included 6583 references to RCTs published in over 400 journals from Spain and Latin America on 46 medical specialties, covering a period from 1957 to 2017.[22] To date, manual hand searching has completed the following specialties: Obstetrics and Gynecology,[23]Dermatology,[20] Physiotherapy,[24] Ophthalmology,[25] Orthopedics and Traumatology,[26] and Dentistry.[27] Methodological quality of these clinical trials has been assessed only for Dermatology[20], Ophthalmology,[25] Orthopedics and Traumatology,[26] and in the clinical trials on reproductive techniques in obstetrics.[23]

We are currently beginning to undertake a broad research project that will assess both the methodological and the reporting quality of randomized controlled trials published in Spain and Latin America included in Cochrane’s BADERI database. The results of this research project will help journal editors, funders, and research communities of the region in their decision-making process on the planning, conduct, and publication of clinical trials.


DICYT Project 021901BN

Competing interests