Studies meeting the eligibility criteria were assessed for methodological quality using a 7-item checklist based on the STROBE guidelines (Pengel et al 2003): use of a representative sample, having a defined sample, use of blinding, having a follow-up rate greater than 85%, appropriate choice of outcome measures, reporting outcome data at follow-up, and control for confounding via statistical adjustment. Screening for eligible studies, methodological quality assessment, and data extraction were conducted independently by two assessors with disagreement resolved by discussion. Data extracted from each study included:

descriptive data on gender, sample size, age, and source of participants (ie, patients and clinicians); verbal, nonverbal and/or interaction style factors; and the association estimates (eg, correlation value) between communication factors

Selleck BGJ398 and selleck chemicals llc satisfaction with care. Correlations between communication factors and satisfaction that were reported as Pearson’s r, Spearman’s rho or Pointbiserial correlation were grouped as verbal, nonverbal and interaction style factors. Meta-analysis was carried out for homogeneous constructs. Pooled analyses were performed using random-effects for trials presenting an I2 of 50% or more (Higgins et al 2003). Correlation values were reported on a common –1 to 1 point scale with 95% CIs. Analytic softwarea was used to conduct all analyses. Correlations were considered poor for values before < 0.21, fair for values ≥ 0.21 but < 0.41, moderate for values ≥ 0.41 but < 0.61, substantial for values ≥ 0.61 but < 0.81, and high for values ≥ 0.81 (Landis and Koch 1977). Individual communication factors that could not be pooled were presented separately. Factors used by clinicians were categorised by two assessors using the Verona medical interview classification, which is based on clinician interview performance considering its main functions and the corresponding patient/ clinician-centred interview techniques (Del Piccolo et al 2002). Disagreements were resolved by discussion. This categorisation allowed data synthesis,

given that different studies employed different systems to code communication factors (Zimmermann et al 2011, Zimmermann et al 2007). The Verona medical interview classification (Del Piccolo et al 2002) categorises clinician responses during the interaction as: information gathering (ie, closed and open questions used by clinicians), patient facilitating (ie, clinicians using facilitators, transitions, and conversation), patient involving (ie, clinicians asking for information and checking for clarification), patient supporting (ie, responses of clinicians supporting, agreeing, or reassuring), and patient education (ie, clinicians informing about the condition or psychosocial issues). The database searches yielded a total of 3414 titles, of which 27 studies in 28 publications were included in the review (Figure 1).