How to write an ICS/IUGA conference abstract

In 2009 we held workshops at both meetings on good reporting of research findings. Speakers included chairs and members of ICS and IUGA scientific committees, specialty journal editors and expert methodologists. With the April 1 deadline for submissions to the joint 2010 ICS/IUGA meeting in Toronto looming, this article aims to condense the lectures and discussions from those workshops into practical advice for the novice researcher facing the daunting task of summarising their data for the first time.

It is an important principle that abstracts presented at ICS or IUGA should not have been previously presented internationally, nor should the abstracts have been indexed in a published journal. Every year there is some overlap of presentations between the two meetings, as well as with other urological or gynaecological meetings, which wastes the precious resource of podium time. It is a more strictly enforced rule that studies that have been published in full, even as an e-publication, should not be presented. Every year some abstracts slip through, despite prior publication [2, 3]. The reverse situation occurs when researchers choose to submit an abstract before conducting any analysis, or based on analysis of interim results. Each abstract should contain original data, not merely a description of methodology. Reports of interim analyses may be justified on safety grounds, but should ideally be led by an independent data monitoring committee. Particularly for randomised trials, conducting interim efficacy analyses before completion of recruitment jeopardises the power calculation of the study and compromises the equipoise of the researchers [4]. Even for observational studies, waiting for full recruitment and complete follow-up is usually associated with a more powerful message.

Every year, very large numbers of abstracts are submitted in the final hours before abstract submission closes. The writing of a conference abstract is usually the first chance for multiple authors to synthesise and reflect on complex findings. Sufficient time should therefore be allotted for multiple revisions, with a chance for all listed authors to approve the final draft. Working right up to the deadline may lead to gaps in the analysis and unnecessary approximations. Although we have ourselves each been guilty of last minute submissions, we would recommend leaving as much time as possible for abstract preparation.

The classic scientific format of Introduction, Methods, Results, and Discussion (IMRaD) is adapted for abstracts at both meetings, with extra stress placed on stating a clear objective or hypothesis for the study. Within the IMRaD structure more detailed guidance is available for the items that ought to be included in the full report of a trial. The EQUATOR Network collects and promotes such evidence-based guidance for research reporting. The best known of these recommendations is the CONSORT guideline [5]. Implementation of CONSORT has been shown to effectively improve reporting of randomised controlled trials [6], and it has been formally adopted for reporting of randomised trials at the ICS/IUGA meeting [7]. A core set of items suitable for inclusion in brief abstracts is also available [8] and should be considered as a minimum requirement. Most abstracts submitted to ICS and IUGA are observational cohort, case control or cross-sectional studies. The STROBE statement gives specific guidance for reporting research from each of these designs [9]. Other guidance relevant to the continence field includes PRISMA for systematic reviews and meta-analyses [10] and STARD for diagnostic accuracy studies [11]. The EQUATOR Network has collected many other useful examples [12]. Together these guidelines provide a framework for reporting of clinical research of most designs. Although, with the exception of the CONSORT for abstracts, these guidelines are intended for whole manuscripts, ICS and IUGA abstracts are relatively long. This should allow authors to fulfil most requirements. Failing to report these key items prevents full evaluation of the work, and therefore limits the impact of the study.

Both societies provide detailed abstract submission rules [7]. It is strongly advised to read these in detail before starting the process of writing. Very few abstracts are rejected outright, unless the authors failed to follow the abstract submission rules. Abstracts must be in the correct font, at the correct size, must not be too short or too long and must follow the recommended structure. Importantly, and seemingly obviously, abstracts for the ICS/IUGA meeting should be related to either continence or urogynaecology.

The scientific committee insists that abstracts are strictly anonymous, which enhances the perception of fairness, although it has an uncertain effect on the reliability of the scoring system [13]. Every year abstracts are rejected because authors disclose their names or their institution within the text of the abstract. Describing the country or region and the type of healthcare organisation should provide sufficient information about the setting. Although it is acceptable for abstracts to include self-citations, the cited paper should not be referred to in the first person possessive (e.g., ‘our previous work showed…’) in order to maintain anonymity.

Drafting a clearly worded, readable abstract is difficult even for native English speakers. While spelling errors and solecisms will not rule out acceptance of high quality science, there is evidence that reviewers are biased in favour of well-written abstracts. Non-native English speakers should therefore enlist native English speakers or better still, professional medical writers to review their drafts.

Abstracts are judged on three criteria: novelty, scientific merit and clinical impact, which are allotted equal importance. Authors should bear these categories in mind when drafting each section of the abstract. The process of scoring abstracts is very intense for members of the scientific committee. Each referee will be allocated approximately 300 abstracts to evaluate, with each abstract scored by multiple referees to increase reliability. This amounts to hundreds of thousands of words to be read over a 2-week period. Since most committee members also have full-time jobs as clinicians or scientists, the time available to score each abstract is extremely limited.

The title is the first, and perhaps main chance to grab the attention of the tired overworked referees. Formal guidance, such as STROBE, recommends that titles should indicate the study design (e.g., ‘A prospective cohort study of snake oil for stress incontinence’). Popular alternatives include phrasing the title as a question in order to highlight the novelty and aims (e.g., ‘Does snake oil cure stress incontinence?’); giving a conclusive statement summarising the main finding (e.g., ‘Snake oil is ineffective for stress incontinence’); or some combination of these approaches.

After the title, readers and reviewers tend to focus next on the aims and conclusions. It is important that the title fits with the aim and that the conclusion directly addresses the aim. The reviewers are unlikely to be familiar with the entire world literature on your topic, so the background is a chance to briefly explain the clinical or scientific uncertainties addressed and the rationale for the design. It is helpful to be specific about the main hypothesis.

When there are multiple hypotheses and multiple outcome measures it may be tempting to break the study into two or more abstracts, so called salami slicing [14]. This is explicitly forbidden in the abstract rules. The objective should be to present the single strongest abstract, with the best chance of being rewarded with a podium presentation or discussed poster. If one study was conducted, with one sample, then there should be just one abstract reported.

The scientific merit of an abstract is in part determined by the study design and in part by the clarity of the results and methods sections. The worst reported randomised trial will still score higher than the best case report. The formal reporting guidance (CONSORT, STROBE, STARD) appropriate to each study design will help decide what information is necessary in the limited space available. A well-designed table or figure will often be the clearest way to present the actual data. It is important that a table should neither contain too many data, making it unreadable, nor should it repeat the text.

The conclusion is the final chance to demonstrate the clinical impact of the work. Authors should therefore give explicit examples of the clinical or translational relevance. Many abstracts overstate the importance of their findings, making claims that would not withstand full peer review [2]. It is an egregious error to claim that an intervention tested is ‘safe and effective’ if it is based on one centre, one surgeon or only a small case series.

Table 1 summarises the types of information that might be included in each section of an ICS/IUGA abstract for a clinical trial, adapted from the CONSORT, and STROBE statements.