Journal of Rehabilitation Medicine 51-9 | Page 5

629 RESULTS try), (ii) the characteristics of participants (sample size, study population, sex and age), (iii) diagnostic criteria of rotator cuff tendinopathy (physical examination/imaging and inclusion cri- teria), (iv) prevalence/incidence, (v) investigated risk/associated factors, and (vi) reported significant results published in odds ratios (ORs) with their corresponding 95% CI. Narrative synthesis was undertaken to categorize the iden- tified risk/associated factors into 6 domains (demographics, systematic diseases, anthropometrics, mechanical factors, sports activities and psychological factors). When data from at least 3 studies could be combined, a meta-analysis was performed to obtain a pooled estimate of the ORs with their corresponding 95% CI. A random-effects model was used, and dichotomous data were weighted by the Mantel-Haenszel method. The I 2 index was computed to represent the between-study hetero- geneity, and the thresholds used were “low” (I 2  = 25%–49%), “moderate” (I 2  = 50%–74%), and “high” (I 2 ≥ 75%) (32). A forest plot was made to show the ORs and 95% CI of each identified risk factor. The level of significance was set at 0.05 and Review Manager 5.3 (Cochrane Collaboration, Oxford, UK) was used for data analyses. Identification of studies The electronic search identified 433 studies. After ex- cluding 100 duplicates, 333 titles and abstracts were screened, and 303 irrelevant articles were excluded. The remaining 30 articles were obtained for full text, and another 16 articles were excluded based on the inclusion and exclusion criteria. Two additional articles were iden- tified through hand searching of the reference lists of the included studies. A total of 16 articles were included in this systematic review (34–49), of which 10 were eligible for meta-analysis (Fig. 1) (34, 36–38, 41–44, 46, 49). Methodological quality and risk of bias Identification The methodological quality scores of articles are reported in Table II. On the basis of the modified DB Level of evidence Based on the methodological quality ratings, the identified risk factors were classified into 5 levels of evidence, based on the method used by van Tulder et al. (33) and Hart et al. (31): (1) Strong evidence: pooled results derived from 3 or more studies with a minimum of 2 high-quality studies that were statistically homogenous (I 2  > 0.05); may be associated with a statistically significant or non-significant pooled result; (2) Moderate evi- dence: pooled results derived from multiple studies that were statistically heterogeneous, including at least one high-quality study; or from multiple moderate/low-quality studies, which were statistically homogenous; (3) Limited evidence: pooled results derived from 1 high-quality study or multiple moderate/ low-quality studies, which were statistically heterogeneous; (4) Very limited evidence: pooled results derived from one moderate/low-quality study; and (5) no evidence: pooled results that are insignificant and derived from multiple statistically heterogeneous studies (regardless of quality). Records identified through database searching (n=433) PUBMED (n=86) CINAHL (n=98) Scopus (n=249) Duplicate articles removed (n=100) Records screened on title and abstract (n=333) Irrelevant articles excluded (n=303) Additional records identified through other sourses (n=2) Risk factors for rotator cuff tendinopathy Full-text articles assessed for eligibility (n=30) Full-text articles excluded (n=16) - Participants not meeting criteria (n=14) - Review articles (n=2) Articles included in the review (n=16) Fig. 1. Preferred Reporting Items for Systematic Reviews and Meta- analyses (PRISMA) flow diagram of the search strategy. Table II. Methodological quality assessment using the modified Downs and Black Checklist a Downs and Black Checklist Items Included b Study 1 2 3 5 6 7 10 11 12 16 18 20 21 22 25 Total Applegate et al. (34) Bansal et al. (35) Bodin et al. (36) Bugajska et al. (37) Frost et al. (38) Frost et al. (39) Grzywacz et al. (40) Miranda et al. (41) Northover et al. (42) Rechardt et al. (43) Roquelaure et al. (44) Silverstein et al. (45) Silverstein et al. (46) Stenlund et al. (47) Sutinen et al. (48) Svendsen et al. (49) 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 1 0 0 0 0 1 0 1 0 0 0 1 0 0 0 1 0 0 1 1 1 1 1 1 0 0 0 0 1 0 1 1 1 0 1 1 1 0 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 0 1 1 1 1 0 0 0 1 1 1 0 1 1 1 0 1 1 0 1 1 1 1 1 1 1 12 12 15 13 10 13 13 13 8 14 15 12 12 11 11 14 Only criteria relevant to the included studies will be used here; therefore, several criteria will be excluded, yielding a checklist of 15 items with a maximum total of 15. A ‘’1’’ indicates a ‘’yes’’ score, and a “0” indicates a ‘’no’’ score. b Clear aim; 2=outcomes described; 3=subjects described; 5=confounders clearly described; 6=main findings clearly described; 7=estimates of random variability; 10=probability values reported; 11=subjects asked represent population; 12=included subjects represent population; 16=planned data analysis;18=appropriate statistics; 20=accurate outcome measures; 21=included subjects recruited from same population; 22=included subjects recruited from same period of time; 25=adjustment for confounding in analyses. a J Rehabil Med 51, 2019