The utility of self-report and biological measures of alcohol consumption in alcoholism treatment outcome studies