BACKGROUND: Data on the two-year pattern of course of illness have been collected in the WHO study of the Determinants of Outcomes of Severe Mental Disorder (DOSMD). These data are reanalysed using recursive partitioning, a method not yet applied to psychiatric data to test the hypothesis that subjects from participating centres in developing countries had better outcomes than those in developed countries. METHOD: Subjects were those from the DOSMD study for whom two-year follow-up data were available (n = 1056). The classification and regression trees recursive partitioning technique was used to examine the predictor variables associated with the outcome variable two year pattern of course. RESULTS: Pattern of course was best predicted by centre, but two developed centres (Prague and Nottingham) grouped with the developing country centres excluding Cali, having better outcomes than in the remaining developed country centres and Cali. Type of onset (insidious v. non-insidious) was the next strongest predictor, but its effect differed across these two centre groupings. Effects for some groups were modified by other predictor variables, including age, child and/or adolescent problems, and gender. CONCLUSIONS: The predominant predictor effects on two-year pattern of course continued to be centre and type of onset, but complex interactions between these variables and other predictor variables are seen in specific centre groupings not strictly defined by 'developing' and 'developed'.
BACKGROUND: Data on the two-year pattern of course of illness have been collected in the WHO study of the Determinants of Outcomes of Severe Mental Disorder (DOSMD). These data are reanalysed using recursive partitioning, a method not yet applied to psychiatric data to test the hypothesis that subjects from participating centres in developing countries had better outcomes than those in developed countries. METHOD: Subjects were those from the DOSMD study for whom two-year follow-up data were available (n = 1056). The classification and regression trees recursive partitioning technique was used to examine the predictor variables associated with the outcome variable two year pattern of course. RESULTS: Pattern of course was best predicted by centre, but two developed centres (Prague and Nottingham) grouped with the developing country centres excluding Cali, having better outcomes than in the remaining developed country centres and Cali. Type of onset (insidious v. non-insidious) was the next strongest predictor, but its effect differed across these two centre groupings. Effects for some groups were modified by other predictor variables, including age, child and/or adolescent problems, and gender. CONCLUSIONS: The predominant predictor effects on two-year pattern of course continued to be centre and type of onset, but complex interactions between these variables and other predictor variables are seen in specific centre groupings not strictly defined by 'developing' and 'developed'.
Authors: Dean P McKenzie; Mark Creamer; Helen L Kelsall; Andrew B Forbes; Jillian F Ikin; Malcolm R Sim; Alexander C McFarlane Journal: Soc Psychiatry Psychiatr Epidemiol Date: 2009-09-10 Impact factor: 4.328
Authors: Dean P McKenzie; John W Toumbourou; Andrew B Forbes; Andrew J Mackinnon; Barbara J McMorris; Richard F Catalano; George C Patton Journal: J Affect Disord Date: 2011-06-13 Impact factor: 4.839
Authors: Brandon A Kohrt; Andrew Rasmussen; Bonnie N Kaiser; Emily E Haroz; Sujen M Maharjan; Byamah B Mutamba; Joop T V M de Jong; Devon E Hinton Journal: Int J Epidemiol Date: 2013-12-23 Impact factor: 7.196