A Delphi technique as described by Day and Bobeva (2005) was employed in this study.
It was a questionnaire survey technique with iterative feedback used to gather the
consensus ofGMexperts on the criteria for high-risk medication use. The study was carried
out from October 2005 to September 2006. In this study, high-risk medications, adapted
from the JCAHO definition (Rich, 2004) referred to ‘‘drugs that are involved in a high
percentage of medication errors or adverse outcomes in geriatric patients’’. The adverse
outcomes included potential adverse reactions, drug–disease interactions, and drug–drug
interactions. The criteria for the use of high-risk medications were defined as ‘‘explicit
practice statements of drug use with adverse outcomes that derive from the panel
consensus’’. The Delphi process comprised five steps: the plan for a Delphi study
(including the selection of an expert panel and a pilot study), three rounds of surveys, and
summary of a consensus.