Higher number of items associated with significantly lower response rates in COS Delphi surveys

Abstract

ObjectivesThe Delphi method is commonly used to achieve consensus in core outcome set (COS) development. It is important to try to maximize response rates to Delphi studies and minimize attrition rates and potential for bias. The factors that impact response rates in a Delphi study used for COS development are unknown. The objective of this study was to explore the impact of design characteristics on response rates in Delphi surveys within COS development. MethodsPublished and ongoing studies that included Delphi to develop a COS were eligible. Second round voting response rates were analyzed, and multilevel linear regression was conducted to investigate whether design characteristics were associated with the response rate. ResultsThirty-one studies were included. Two characteristics were significantly associated with a lower response rate: larger panels and studies with more items included. ConclusionCOS developers should pay attention to methods when designing a COS development study; in particular, the size of the panels and the size of the list of outcomes. We identified other potential design characteristics that might influence response rates but were unable to explore them in this analysis. These should be reported in future reports to allow for further investigation.