Aim: The aim of the present study was to detect factors associated with duration of untreated illness (DUI) in bipolar disorder (BD). Method: A total of 1575 patients were selected for the purposes of the study. Correlation analyses were performed to analyse the relation between DUI and quantitative variables. The length of DUI was compared between groups defined by qualitative variables through one-way analyses of variance or Kruskal-Wallis's tests according to the distribution of the variable. Linear multivariable regressions were used to find the most parsimonious set of variables independently associated with DUI: to this aim, qualitative variables were inserted with the numeric code of their classes by assuming a proportional effect moving from one class to another. Results: An inverse significant correlation between length of DUI and time between visits in euthymic patients was observed (r = −.52, P <.001). DUI resulted to be longer in patients with: at least one lifetime marriage/partnership (P =.009), a first psychiatric diagnosis of major depressive disorder or substance abuse (P <.001), a depressive polarity of first episode (P <.001), no lifetime psychotic symptoms (P <.001), BD type 2 (P <.001), more lifetime depressive/hypomanic episodes (P <.001), less lifetime manic episodes (P <.001), presence of suicide attempts (P =.004), depressive episodes (P <.001), hypomanic episodes (P =.004), hospitalizations (P =.011) in the last year. Conclusions: Different factors resulted to increase the length of DUI in a nationwide sample of bipolar patients. In addition, the DUI was found to show a negative long-term effect in terms of more suicidal behaviour, more probability of hospitalization and depressive/hypomanic episodes.
