Derivation and validation of a simple clinical risk-model in heart failure based on 6 minute walk test performance and NT-proBNP status--do we need specificity for sex and beta-blockers?