STUDY DESIGN

Overview study.

BACKGROUND

Abstracts of systematic reviews have presented 'spin' (i.e. overstated interpretation of study results) and inconsistency with the full text.

OBJECTIVES

1. Do abstracts of low back pain reviews contain spin? 2. Do these abstracts consistently represent the full text? 3. Is abstract spin associated with the type of conclusion?

METHODS

We searched the Physiotherapy Evidence Database (PEDro) on 10th January 2018. Data were extracted from systematic reviews of physiotherapy interventions for low back pain, published between 2015 and 2017. Spin was assessed using a 7-item checklist. We evaluated consistency by comparing information contained in the abstract and the full text using the 7-item checklist with Kappa coefficient analysis. We used logistic regression analysis to evaluate the association between spin in the abstract and type of conclusion. We evaluated methodological quality using the AMSTAR-2 (A MeaSurement Tool to Assess systematic Reviews).

RESULTS

We included 66 eligible systematic reviews, subdivided into Cochrane (n=9) and non-Cochrane (n=57) reviews. There was some form of spin in 80% of abstracts. Abstracts of non-Cochrane reviews were not consistent with the full text (fair to moderate agreement). Cochrane review abstracts had substantial to almost perfect agreement with the full text. Spin was not associated with the type of conclusion in all systematic reviews ( < 0.05). The methodological quality ranged from 'high' to 'critically low'.

CONCLUSIONS

The abstracts of systematic reviews evaluating physiotherapy interventions for low back pain need improvement. .