Page 50 - THE PERCEPT STUDY Illness Perceptions in Physiotherapy Edwin de Raaij

Chapter 3
of bias. Any disagreement was resolved by discussion and consensus. It is recommended not to report a total score of the 31 items for overall study quality22.
Data synthesis and analysis
Descriptive statistics were used to summarize the medical condition, number of participants, age, sex, study design, and questionnaires used for illness percepions/pain intensity/physical function across all included studies. Extraction of results focused on obtaining unadjusted and adjusted correlations, regression co- efficients, relative risks, odds ratios, and 95% confidence intervals. To explore possible publication bias or outcome reporting bias, funnel plots were made by plotting all extracted data against the number of participants in each study.
To assess statistical heterogeneity, the I2 test was used. As proposed by Higgins et al24, a value higher than 50% was considered an indicator of substantial heterogeneity. As outcomes were considered too heterogeneous, the authors refrained from statistical pooling and summarized the evidence qualitatively, according to Hayden et al21 (Table 2).
Table 2: Levels of evidence
Level
Description
Strong Moderate
Limited Conflicting No evidence
Consistent findings (defined as greater than 75% of studies showing the same direction of effect) in multiple low-risk-of-bias studies
Consistent findings in multiple high-risk-of-bias studies and/or 1 study with low risk of bias
1 study available
Inconsistent findings across studies
No association between variables for association or prognosis Inconsistent findings across studies
No association between variables for association or prognosis
Results
Study selection
The literature search generated a total of 1418 references: PubMed, 411; PsycINFO, 381; Embase, 314; CI-NAHL, 253; and SPORTDiscus, 59. A total of 114 references were identified in the gray literature. After screening for duplicates (J.M.), 1045 were included for screening. Two authors (EDR and HW) independently screened all 1159 studies for eligibility, using the inclusion/exclusion criteria. A total of 26 studies met these criteria and were included in the review (Figure 1).
48