A geospatial analysis of persons opting out of an exception from informed consent out-of-hospital clinical trial

Abstract

For trials involving exception from informed consent, some IRBs require that community members be allowed to "opt out" prior to enrollment. We tested for geospatial clustering of opt-out requests and the associated census tract characteristics in one study region. This was a retrospective study at a single site of a multicenter exception from informed consent resuscitation trial. We collected and geocoded mailing addresses for persons requesting opt-out bracelets over 16 months, then tested for geospatial clustering using geographic information systems (GIS) analysis. Characteristics for tracts with and without bracelet clustering were compared using univariate tests, multivariable regression, and classification and regression tree (CART) analysis. We received 395 phone calls requesting 718 bracelets, of which 673 were analyzable. Of 397 census tracts in the region, 208 (52%) had at least one request and 38 (10%) demonstrated clustering. In multivariable models, an increasing proportion of family households (OR .90, 95%CI .85-.93), veterans (OR .91, 95%CI .81-1.02), and renters (OR .96, 95%CI .92-.99) were associated with lower odds of requesting an opt-out bracelet, while census tracts with higher income had higher odds of opting-out (OR 1.07, 95%CI 1.02-1.11). Using CART, the proportion of family households and graduate education identified the majority of opt-out requests by census tracts (cross-validation sensitivity 92%, specificity 56%). Opt-out requests for an exception from informed consent trial at one study site were geographically clustered and associated with certain population demographics. These findings may help identify key target groups for community consultation in future trials.