File
Port Nascimento D.1, Costa L.O.P.1,2, Gonzalez G.Z.1, Maher C.2, Moseley A.2
1Universidade Cidade de São Paulo, Masters and Doctoral Programs in Physical Therapy, São Paulo, Brazil, 2The George Institute for Global Health, Musculoskeletal Division, Sydney, Australia
Background: Abstracts of randomized controlled trials must be clear, concise and should reflect exactly what was performed in the study. Studies from many health-related areas have assessed the quality and accuracy of abstracts and found discrepancies such as omission of information, inconsistency of results when compared to full text, poor quality of abstracts´ structure, selective reporting and over-stating results (ie 'spin') in the results and conclusion sections of the abstracts. There are no studies that have investigated accuracy and reporting quality of abstracts in randomized controlled trials in physiotherapy.
Purpose: This study aims to analyse the reporting quality of abstracts of randomized controlled trials in physiotherapy interventions for low back pain, as well as to check the accuracy of results and conclusion of abstracts when compared to the full text.
Methods: A random sample of 100 randomized controlled trials in physiotherapy for low back pain, published between 2010 and 2015, and indexed in PEDro database was analysed. Information about reporting quality and accuracy of the abstracts and the full text were extracted. To analyse the reporting quality of abstracts, the items of the CONSORT recommendations for abstracts were classified as adequately reported, inadequately reported, not reported or not applicable. To analyse the accuracy of abstracts, items that indicate the presence of selective reporting and spin of information in the results and conclusion sections of the studies were evaluated. Two independent assessors extracted all data. Data was analysed descriptively.
Results: The percentage of agreement of the assessors ranged from moderate to almost perfect. The reporting quality of abstracts of randomized controlled trials was low. Most abstracts have several problems in adequately reporting information, such as primary outcome results (3% adequately reported), blinding (3%), participants allocation (4%), adverse events (4%), number of analysed participants (10%), results interpretation (33%), primary outcome (34%) and number of participants (37%). Lack of accuracy was highlighted by the omission of the primary outcome in 64% of the abstracts and failing to mention adverse events in 95% of the abstracts. In 70% of abstracts there was selective reporting and spin of the results in conclusions. Furthermore, the lack of consistency between abstract and text ranged from 21% to 70% for CONSORT recommendations and from 24% to 32% for the accuracy items.
Conclusion(s): Quality reporting and accuracy of abstracts in physiotherapy for low back pain is far from optimal.
Implications: The results of this study show that journal editors, reviewers and authors must pay closer attention to the way trial abstracts are written.
Funding acknowledgements: This study is supported by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP). Process Number: 2015/16953-2.
Topic: Research methodology & knowledge translation
Ethics approval: This study did not require Ethics approval.
All authors, affiliations and abstracts have been published as submitted.
