Page 49 - THE PERCEPT STUDY Illness Perceptions in Physiotherapy Edwin de Raaij

In addition, the references of the included articles and recently published review articles were screened for additional publications.
Study selection
In the first round, 2 authors (EDR and HW) independently reviewed all titles and abstracts and excluded all studies that did not fulfill the inclusion criteria. If an abstract was noninformative but potentially relevant, the full-text article was read. In the second round, the full texts of all articles were read for fulfillment of all inclusion criteria and selected by 2 independent authors. Articles on psychometric properties or with qualitative designs were excluded. Any disagreement was resolved by discussion and consensus with a third author (RO).
The studies had to meet the following criteria for final inclusion:
1) study population of individuals with musculoskeletal pain, as defined by the European Musculoskeletal Conditions Surveillance and Information Network11, 2) measures of illness perceptions with questionnaires based on the CSM, 3) measures of pain and physical function with self-reported questionnaires and 4) study designs that included cohort studies, cross-sectional studies, and randomized controlled trials. To answer the research questions concerning associations of illness perceptions with pain intensity and physical function, the authors considered crosssectional studies or longitudinal studies most appropriate.
To answer the research questions concerning prognoses by illness perceptions of increased pain intensity and increased limitation in physical function, the researchers considered longitudinal studies most appropriate. They excluded qualitative studies.
Data extraction and quality assessment
Two authors developed and independently completed the data-extraction form, which included author, publication year, study design, number of participants, study setting, characteristics of the study population (eg, musculoskeletal disorder, type of illness perceptions), measurement instruments of pain and limitation in physical function, and outcome and statistical measures (correlations, odds ratios, and regression coefficients).
The risk of bias was assessed using the Quality in Prognosis Studies (QUIPS) tool20 by 2 authors independently. This tool has 31 items that are scored as “yes” (fulfilled), “/”(partial), “no” (not fulfilled), or “?” (unclear whether criterion is ful- filled). The 31 items cover 6 domains for potential bias: study participation, study attrition, prognostic factor measurement, outcome measurement, study confounding, and statistical analysis. These domains are labeled “high,” “moderate,” or “low” risk of bias, based on the individual item’s score within each domain, as described by Hayden et al22. A study has a low risk of bias if all 6 domains are rated as low risk
A systematic literature review
47
3