The approach of Norm Balance in predicting pharmacists' intention to collaborate with physicians to improve medication therapy.

Objective: Norm Balance is an approach under the Theory of Planned Behavior (TPB) where subjective norm is weighted by the relative importance of others and self-identity is weighted by the relative importance of self. The relative importance was measured previously by a trade-off measure. In this study, we developed separate measures for the relative importance. The study objectives were to: 1) assess the construct validity of the separate measures; 2) examine the approach of Norm Balance in predicting pharmacists' intention to collaborate with physicians to improve medication therapy; and 3) establish a modified TPB. Methods: We selected a random sample of 750 Iowa pharmacists and conducted two surveys. The first survey was to examine intention prediction, and the second survey was to examine behavior prediction by measuring behavior among respondents to the first survey. The relative importance was measured by both the trade-off measure and the separate measures. Exploratory factor analyses were performed for the relative importance of others (separate measures) and subjective norm, and for the relative importance of self (separate measures) and self-identity. Regressions for intention prediction were conducted for TPB with self-identity and the approach of Norm Balance. The same regressions were also conducted for three subgroups according to the median of the relative importance of others (trade-off measure). Moreover, another regression was conducted for behavior prediction. Results: 239 practicing pharmacists responded to the first survey, and 188 responded to the second survey. The separate measures had cross factor loadings, whereas the trade-off measure had low correlations with other constructs. Both regressions for intention prediction explained 75% of the variance, with self-efficacy and attitude being strong predictors. Self-identity was not a predictor in the TPB with self-identify, but self-identity weighted by the relative importance of self was a significant predictor in the approach of Norm Balance. Regression coefficients of subjective norm and self-identify varied across subgroups. The regression for behavior prediction explained 30% of the variance, with intention and self-efficacy being two predictors. Conclusion: The trade-off measure was better than separate measures. The approach of Norm Balance appears to be a better model than the TPB with self-identity to examine pharmacist-physician collaboration.