349) (Table 1). A two-factor solution emerged with 69.02% of the variance explained. The data were suitable for PCA as the Kaiser–Meyer–Oklin value was 0.90, exceeding the recommended value of 0.6, and Bartlett’s test of sphericity was statistically significant (P<0.001). The first eight items loaded more

strongly on the first component, corresponding to the process of shared decision-making and patient involvement, and the last two items loaded more strongly on the second component, corresponding to the process of making the final medical decision. Cronbach’s α reliability estimate was high for the 10 items at 0.91. Cronbach’s α was 0.92 for the first eight items and 0.72 for the last two items. Given that the concordance items loaded on two correlated factors, analyses were performed for summed scores Pim inhibitor of the 10 items (referred to as ‘concordance’) as well as summed scores of the first eight items (referred to as ‘shared decision-making process’) and summed scores of the last two items (referred to as

‘medical decision’). Spearman correlations were used to investigate relationships between concordance (as well as shared decision-making and medical decision) and continuous variables. Mann–Whitney and Kruskal–Wallis tests were used to investigate relationships AZD1152-HQPA nmr between concordance (as well as shared decision-making and medical decision) and categorical variables. Nonparametric tests were selected as concordance, shared decision-making and medical decision scores were skewed. Six linear regressions investigated the relationship between each independent variable (concordance, shared decision-making and medical

decision) and the dependent variables (CD4 cell count at baseline and CD4 cell count at 6–12 months post-study) controlling for treatment status GNA12 (on treatment/stopped treatment), baseline CD4 cell count (for CD4 cell count at 6–12 months post-study as dependent variable) and any demographic variable related to concordance, shared decision-making or medical decision and CD4 cell count at P<0.25. Treatment status was included in regression analyses looking at concordance or shared decision-making because it was associated with these variables and CD4 cell count (at baseline and at follow-up) in univariate analyses at P<0.25. Ethnicity was included in regression analyses looking at medical decision because it was associated with this variable in univariate analyses and CD4 cell count (at baseline and at follow-up) at P<0.25. White patients scored lower on medical decision and reported higher CD4 cell counts than non-White patients. None of the other demographic variables was associated with medical decision and CD4 cell count at P<0.25.