00 to 0.25), fair relationship (from 0.25 to 0.50), moderate to good relationship (from 0.50 to 0.75), and good

to excellent relationship (above 0.75) ( Portney and Watkins 2000). We aimed to pool correlation coefficients when studies were homogenous. When pooling was not possible due to the heterogeneity of measures of communication factors and constructs of therapeutic alliance, communication factors were tabulated and descriptive analyses conducted. After removing duplicates, a total of 3063 titles was identified with the electronic searches. Of these, 69 were selected as potentially eligible VX-770 molecular weight on the basis of their title/abstract and were retrieved as full articles. Following examination of the full text, 12 papers were included (Figure 1). All included studies provided cross-sectional observational data collected after or during the medical encounter. One study (Thom 2001) also included a longitudinal analysis see more one month and six months after the first encounter but only data related to the first encounter were included in this review to allow comparison with other included studies. Another study conducted a cross-sectional analysis with all patients from a randomised clinical trial using

baseline measurements (Ommen et al 2008). Quality: A detailed description of the methodological quality of all included studies is presented in Table 1. Briefly, most of the studies stated explicitly that patients were selected as consecutive or random cases. Coders

were blinded in only one study ( Harrigan et al 1985). Eight of 12 studies reported details of assessment methods including reliability measures. Study characteristics: The study settings included general practices ( Carter et al 1982, Fiscella et al 2004, Harrigan et al 1985, Keating et al 2002, Tarrant et al 2003, Thom 2001), hospital outpatient clinics ( Perry 1975), and within tertiary hospital outpatients ( Berrios-Rivera et al 2006, Garcia-Gonzalez et al 2009, Keating et al 2004, Takayama and Yamazaki 2004) and inpatients ( Ommen et al 2008). Participants: Patients interacted with physicians whatever in six studies ( Carter et al 1982, Fiscella et al 2004, Harrigan et al 1985, Keating et al 2002, Tarrant et al 2003, Thom 2001), with specialist physicians in five studies ( Berrios-Rivera et al 2006, Garcia-Gonzalez et al 2009, Keating et al 2004, Ommen et al 2008, Takayama and Yamazaki 2004), and with physiotherapists in one study ( Perry 1975). Only four studies reported the health conditions of the patients, which included rheumatic diseases ( Berrios-Rivera et al 2006, Garcia-Gonzalez et al 2009, breast cancer ( Takayama and Yamazaki 2004), and severely injured patients ( Ommen et al 2008). Communication factors: Among the 12 included studies we identified 36 interaction styles in nine studies, 17 verbal factors in five studies, and 14 non-verbal factors in three studies.