Example.

“Effectiveness of a strategy to improve adherence to tuberculosis treatment in a resource poor setting: a cluster randomized trial” [33].

Explanation.

The ability to identify a relevant report in an electronic database depends to a large extent on how it was indexed. Indexers may not classify a report as a RCT if the authors do not explicitly report this information [34]. To help ensure that a study is appropriately indexed and identified as a RCT, authors should state explicitly in the title that the participants were randomly assigned to their comparison groups.

Example.

“Correspondence to: Dr Sally Hopewell, UK Cochrane Centre, Summertown Pavilion, Middle Way, Oxford OX2 7LG, UK. Tel: +44 1865 516300; Fax: +44 1865 516311; Email: shopewell@cochrane.co.uk.”

Explanation.

Adequate contact details for the corresponding author are particularly important for RCTs reported in conference proceedings. These abstracts may be the only lasting source of information for many trials, as only half of RCTs reported in conference proceedings are subsequently published in full [1]. Adequate contact information would enable readers to contact trialists for additional information or clarifications regarding reported data. Adequate contact details should include the telephone number, postal, and email address of the principal investigator and, if available, the trial Web site.

Example.

“A cluster randomized controlled trial...” [33].

Explanation.

The design of the trial should be described, for example, parallel group, cluster randomized, crossover, factorial, superiority, equivalence or noninferiority, or some other combination of these designs. An important reason for identifying the design of the trial is to ensure appropriate indexing in electronic databases, thus ensuring greater ease of identification [34]. Alerting readers to the design of the trial also provides transparency as to the type of design used to conduct the trial and should reduce the likelihood of inadvertently misinterpreting data. For example, in a report of a cluster trial, readers might misinterpret a small sample size as the number of participants rather than the number of clusters, or vice versa [35].

Example.

“… conducted between June 2003 and January 2005, at 16 government district health centers in Senegal. Patients older than 15 years with newly diagnosed sputum smear-positive pulmonary TB were randomly assigned to the intervention or control group” [33].

Explanation.

Every RCT addresses an issue relevant to a particular population or group with the condition of interest. Trialists may further restrict this sample by using eligibility criteria and by performing the trial in a particular setting (for example primary, secondary, or tertiary care). Participant eligibility criteria may relate to demographics, clinical diagnosis, and comorbid conditions. A clear description of the trial participants and setting in which they were studied is needed so that readers may assess the external validity (generalisability) of the trial and determine its applicability to their own setting.

Example.

“Patients were randomized to receive either 100 mg hydrocortisone or matching placebo as follows: the first dose in the evening of the operative day, then 1 dose every 8 hours during the next 3 days. In addition, all patients received oral metoprolol (50–150 mg/d) titrated to heart rate” [36].

Explanation.

The essential features of the experimental and comparison interventions should be described. Authors should report details about the interventions, e.g., dose, route of administration, duration of administration, surgical procedure, or manufacturer of inserted device.

Example.

“To compare the effectiveness of an early switch to oral antibiotics with the standard 7 day course of intravenous antibiotics in severe community acquired pneumonia” [37].

Explanation.

The abstract should provide a clear statement of the specific objective or hypothesis addressed in the trial. If more than one objective is addressed, the main objective (i.e., based on the prespecified primary outcome) should be indicated and only key secondary objectives stated [26].

Example.

“Main outcome measure: all-cause mortality at 180 days” [38].

Explanation.

RCTs assess outcomes for which the interventions are being compared. Most trials have several outcomes, some of which are deemed more important than others. Such rankings are typically reported as primary and secondary outcomes. There is evidence of selective reporting with significant or favourable outcomes being more likely to be published than nonsignificant outcomes [39–41].

Authors should explicitly state the primary outcome for the trial and when it was assessed (e.g., the time frame over which it was measured). The primary outcome is the prespecified outcome considered of greatest importance and is usually the one used in the sample size calculation [22]. In some instances a publication may report an outcome different from the primary outcome. For example, conference abstracts are more likely to report interim analyses than are full publications [8,10], or to present different results for a single trial in a series of abstracts. If the abstract focuses on a secondary outcome of a trial, the abstract should identify both this outcome and the primary outcome of the trial.

Example.

“Randomization was computer-generated, with allocation concealment by opaque sequentially numbered sealed envelopes” [42].

Explanation.

It is important to conceal the allocation sequence from those assigning participants to the intervention groups. Allocation concealment prevents investigators from influencing which participants are assigned to a given intervention group (i.e., selection bias). Evidence shows that reports of trials reporting inadequate allocation concealment are associated with exaggerated treatment effects [43,44]. Research suggests that adequate allocation concealment is more important in preventing selection bias than are other components of the randomization process, such as the sequence generation (e.g., use of computer or random number table) [45].

Authors should clearly describe the method for assigning participants to interventions. Examples of approaches used to ensure adequate concealment include: centralised (e.g., allocation by a central office) or pharmacy-controlled randomization; sequentially numbered identical containers that are administered serially to participants; on-site computer system combined with allocations kept in a locked, unreadable computer file that investigators can access only after the characteristics of an enrolled participant are entered; and sequentially numbered, opaque sealed envelopes [46].

The method of allocation concealment is generally poorly reported in conference abstracts and in abstracts of journal articles [7,47–49]. For example, in a review of 494 abstracts presented at an oncology conference in 1992 and 2002, only nine (2%) abstracts reported the method of allocation concealment. This information was missing from the remaining 485 conference abstracts, with no improvements seen over the ten-year period [7].

Example.

“Children, parents, and the research assistants were blinded to group assignment” [50].

Explanation.

Blinding refers to the practice of keeping the trial participants, care providers, data collectors, and sometimes those analysing the data, unaware of which intervention is being administered to which participant, so that they will not be influenced by that knowledge. The term masking is sometimes used instead of blinding [51,52] and might be preferable when reporting studies involving eyes and vision. It is important that authors describe whether or not participants, those administering the intervention (usually health-care providers), and those assessing the outcome (the data collectors and analysts) were blinded to the group allocation. Authors should avoid using terms such as “single” or “double” blind as such terms are not well-understood [53].

Information on the method of blinding is poorly reported in conference and journal abstracts [7,8,47–49]. Such reporting is valuable as blinding may be important in protecting against bias [51]. Studies have shown that if investigators are aware of the treatment, their attitudes for or against an intervention can directly affect whether or not they include, or treat, participants in a trial [45]. Furthermore, there is evidence that participants who are aware of their assignment status are more likely to report symptoms, leading to biased results [51]. Perhaps most importantly, if outcome assessors are not blinded to the intervention they are more likely to report favourable outcomes for the intervention which they believe is better [54]. However, unlike allocation concealment, blinding of the participants, health-care providers, and outcome assessors may not always be appropriate or possible, such as in many surgical trials. In this case, authors should report if any form of blinding (such as blinding of data analysts) was used.

Example.

“Children (n = 633) aged 1–3 randomly allocated to receive fortified milk (n = 316) or control milk (n = 317)” [55].

Explanation.

The number of participants randomized to each intervention group is an essential element of the results of a trial. This number defines the sample size, and readers can use it to assess whether all randomized participants were included in the data analysis. Again, this may be particularly important for conference abstracts reporting interim analyses, if a trial is still open to participant accrual or follow-up [8,10]. Here authors should report the period of recruitment on which the data are based.

Example.

“An interim analysis was performed because of slow accrual” [56].

Explanation.

Authors should describe the status of the trial and whether it is still ongoing, closed to recruitment, or closed to follow-up. This information is particularly important for conference abstracts, which are more likely than full articles to report interim analyses [10].

If the trial has stopped earlier than planned it is important to say why. Possible reasons for early termination include: slow accrual rates, poor data quality, poor adherence, resource deficiencies, unacceptable harms or large benefits, or emerging information that makes the trial irrelevant, unnecessary, or unethical. If a trial stops early for apparent benefit, the estimates of treatment effect are more likely to be biased and prone to exaggeration [57,58].

Example.

“… 300 were included in the analysis of the primary outcome (100 in the acetaminophen group, 100 in the ibuprofen group, and 100 in the codeine group)” [50].

Explanation.

Authors should report the number of participants included in the analysis for each intervention group. These data permit an assessment of whether participants were analysed according to their original group assignment, which is important, because failure to include all participants in the analysis may bias the results of the trial [22].

Several studies have reported deficiencies in journal and conference abstracts in reporting the number of participants included in the analysis [6–8,13,48,59]. In a review of RCTs in acute brain injury reported in journal abstracts, only 43% reported the number of participants included in the analysis [48]. In another evaluation of trials reported in abstracts for an oncology conference, only 40% reported the number of participants analysed, and only 6% indicated intention to treat analysis [8].

Example.

“Treatment was successful for 682 (88%) of 778 patients recruited in the intervention group, and for 563 (76%) of 744 patients recruited in the control group (adjusted risk ratio [RR], 1.18; 95% confidence interval [CI], 1.03–1.34)” [33].

Explanation.

For the primary outcome, authors should report trial results as a summary of the outcome in each group (e.g., the number of participants with or without the event, or the mean and standard deviation of measurements), together with the contrast between groups known as the effect size. For binary outcomes, the effect size could be the relative risk, relative risk reduction, odds ratio, or risk difference. For survival time data, the measurement could be the hazard ratio or difference in median survival time. For continuous data, the effect measure is usually the difference in means. Authors should present confidence intervals for the contrast between groups and as a measure of the precision (uncertainty) of the estimate of the effect [22]. For abstracts not reporting the “primary” outcome of the trial (e.g., abstracts focusing on safety data or economic impacts), the secondary nature of the outcomes should be indicated, and, where possible, sufficient details of the primary outcome should be included to allow other findings to be taken in the proper context.

Several studies have observed deficiencies in the reporting of statistical results in journal abstracts [57,60–62]. For example, Pocock and colleagues [57] found that journal abstracts of RCTs tended to overemphasize statistically significant outcomes compared to the full journal article, leading to problems in interpretation of the results. Poor reporting of results is also a problem for trials presented in conference abstracts [7,8,59]. A study of 494 reports of RCTs in oncology found that only 26% of conference abstracts reported the size of the effect and significance of the result [7].

Example.

“Adverse events were more common with topiramate vs placebo, respectively, including paresthesia (50.8% vs 10.6%), taste perversion (23.0% vs 4.8%), anorexia (19.7% vs 6.9%), and difficulty with concentration (14.8% vs 3.2%)” [63].

Explanation.

Most interventions have unintended and often undesirable effects as well as intended and beneficial effects. In order to make rational and balanced decisions, readers need information about the relative benefits and harms of an intervention. Authors should describe any important adverse (or unexpected) effects of an intervention in the abstract. If no important adverse events have occurred, the authors should state this explicitly [20].

Explicit reference to the reporting of harms in the title or abstract is also important for appropriate database indexing and information retrieval. Derry and colleagues [64] found that only 66 of 107 RCTs that reported data on adverse events in the full publication mentioned harms in the title or abstract; thus, harms could not have been identified for many of the articles in a search of titles and abstracts in an electronic bibliographic database.

Harms are also poorly reported in conference abstracts. A recent examination of over 800 ophthalmology conference abstracts reporting trials found that the majority (71%) did not report harms related to the treatment intervention, and harms were reported as a primary outcome measure in only 6% of abstracts [9].

Example.

“Multivitamin supplementation reduced the incidence of low birth weight and small-for-gestational-age births but had no significant effects on prematurity or fetal death” [65].

Explanation.

The conclusions of the trial, consistent with the results reported in the abstract, should be clearly stated along with their clinical application (avoiding over-generalisation). Authors should balance the benefits and harms in their conclusions. Where applicable, authors should also note whether additional studies are required before the results are used in clinical settings [26].

Example.

“Trial Registry: www.clinicaltrials.gov; Identifier: NCT00412009” [33].

Explanation.

Nonpublication of entire trials and selective reporting of outcomes within trials has been well-documented [39,41,66]. Covert redundant publication can also cause problems in systematic reviews when results from the same trial are inadvertently included more than once [67]. To minimize or avoid these problems there have been many calls for trial registration [68]. Due to more recent serious problems of withholding data [69] there has been a renewed effort to register RCTs. By registering a RCT, authors typically report a minimal set of information and obtain a unique trial registration number.

In September 2004 the International Committee of Medical Journal Editors (ICMJE) indicated a change in their policy for publishing RCTs, saying that they would consider trials for publication only if they had been registered before the enrolment of the first patient (as of 1 July 2005) [70]. This position has resulted in a dramatic increase in the number of trials being registered [71].

In an abstract reporting a trial, authors should provide details of the trial registration number and name of trial register. Registration information will be particularly important for abstracts reported in conference meetings, as not all of them are subsequently published [1]. Such trial registration provides readers with a way to obtain more information about the trial and its results. Registration information will also help to link abstracts with subsequent full publications (or multiple abstracts from the same trial) and thus reduce the risk of inadvertent double-counting in systematic reviews.

Example.

“Funded by The Breast Cancer Research Foundation” [72].

Explanation.

Authors should report the source of funding for the trial as this is important information for readers assessing a trial. A recent systematic review showed that studies funded by the pharmaceutical industry had four times (odds ratio 4.05; 95% confidence interval 2.98–5.51) the odds of having outcomes favouring the sponsor than studies funded by other sources [73]. Similarly, authors should report any other sources of support, such as in the preparation of the abstract, presentation, or manuscript [74].