Exercise interventions are poorly reported

Exercise is effective for the prevention and management of acute and chronic health conditions. Exercise prescription is sensible when supported by high-quality evidence of effectiveness, and it is likely that the design of an exercise programme (eg, how long a person exercises each day, the duration of a programme, the level of intensity of the exercise) influences programme effectiveness. A meta-epidemiological review of 73 systematic reviews (1216 trials) of exercises reported that only 30% of trial reports provided information required to replicate the investigated exercise programme.1

Consequences of poor reporting

This means that researchers would, in most cases, be unable to replicate and validate trial outcomes for exercise programmes that have been reported to be effective. In addition, clinicians would be unable to accurately implement treatment based on the reported intervention.1 Furthermore, when trials of exercise are pooled in a meta-analysis, pooling of studies of unknown design may lead to incorrect conclusions regarding specific exercise effects.

A logical solution would be to report exercise programmes in a standardised and comprehensive manner.2

What is the CERT?

The Consensus on Exercise Reporting Template (CERT) is a 16-item checklist that prompts for exercise descriptions in seven sections: what (materials); who (provider); how (delivery); where (location); when, how much (dosage); tailoring (what, how) and how well (compliance/planned and actual).3 It was developed through consensus of an international team of 137 experts using Delphi consensus methods and it is accompanied by an explanation and elaboration statement that provides instruction on how to document each exercise element, including examples of good reporting from the literature.4 As the CONSORT Statement5 and template for intervention description and replication (TIDieR) checklist6 provide useful guidance on how to report some aspects of an exercise intervention, the CERT was designed so that overlapping items were aligned. The CERT extends these earlier recommendations in seeking more information about the type of exercise, dosage, intensity, frequency and supervision requirements. In addition, when an individualised rather than a ‘one-size fits all’ programme is implemented, information is requested on how exercises are tailored to the needs and abilities of individuals. The CERT is free and available at http://bjsm.bmj.com/content/50/23/1428.full (table 1 and supplementary appendix 1).

Unambiguous description of the elements in complex interventions is challenged by constraints on the length of journal publications. Solutions have included publishing research protocols separately or using online appendices. The CERT’s authors recommend that exercise programme details be reported with trial results as an online appendix, with the main body of the manuscript aligned with the recommended reporting for each study design, such as PRISMA,7 CONSORT,5 STROBE.8

How the CERT moves the field forward

Exercise interventions have many elements with the potential to vary, such as the equipment and facilities used, instructor expertise, types of motivation, compliance and adherence strategies, individualisation, exercise dosage parameters, estimates of capability and decision rules for programme progression, and the nature of associated adverse events. Prior to the CERT, there was no standardised and internationally endorsed consensus statement about which aspects of these components might be important to report and the nature of the information required for replication. The use of the CERT will move the field forward for researchers, peer reviewers, policy-makers and clinicians by facilitating an unambiguous description of an exercise programme.

Where to from here?

While there is a need for a standardised and internationally endorsed reporting template for exercise interventions, an ongoing challenge is integrating this into the reporting of multimodal interventions, especially those with a high degree of individualised care. The CERT is a step forward in accommodating this, as it provides a model for dissecting intervention elements and documenting each one using an agreed format. Other elements in multimodal care might be reported in a similar way.

One aspect of reporting that is not included in CERT is explicit statement of the aim of the exercise. This aspect is useful because different aims (pain relief, strength, endurance, etc) will influence the design and outcomes of the exercise intervention. Although it might be assumed that this would be included in the aims of the research, perhaps the inclusion of this criterion could be considered for addition in a future version of CERT.

The bottom line

Use of the CERT will facilitate comprehensive and accessible reporting of exercise programmes and accurate delivery of effective exercise programmes in clinical practice. It will enable validation of trial results through replication. It may also help us advance the science of exercise prescription. Over time, it may enable us to identify those elements in an exercise programme that are unimportant and those that are essential to intervention success.

Overall, less research waste and better translation into practice.