A blind re-analysis of the Iowa family study of obsessive-compulsive disorder

Abstract We present results from a re-analysis of the Iowa family study of obsessive-compulsive disorder (OCD) that previously concluded the disorder was not familial. These conclusions were based on Diagnostic Interview Schedule results of first-degree relatives (FDRs) and not a best estimate diagnosis (BED). For the re-analysis we reviewed raw data on OCD and control probands and their FDRs. Relatives had been assessed through structured interviews, validated questionnaires, family history, and medical records in some cases. BEDs were assigned through a blind consensus procedure employing DSM-IV criteria. The data were analyzed using logistic regression with generalized estimating equations to account for within family correlations. BEDs were assigned to 32 OCD probands, 31 control probands, and 352 FDRs, including 249 FDRs who were interviewed directly and 103 FDRs who were unavailable or deceased. Lifetime prevalence of definite/probable OCD was significantly higher in the FDRs of OCD probands than controls (10.7% vs. 3.8%, OR=3.04, p =0.026). FDRs of OCD probands had significantly higher rates of depressive illness than relatives of controls. Depression of any type in relatives was predicted by the proband's depression history. We conclude that OCD is familial. The re-analysis highlights the importance of the BED procedure in family studies.