Developmental psychopathologists have documented a number of patterns of heterotypic continuity from one disorder to another across the transition to adolescence (a.k.a successive comorbidity;Angold et al., 1999). One purpose of documenting these patterns was to reduce treatment delays for children who might otherwise be disregarded as """"""""outgrowing"""""""" transient disorders. However, the implications of documented successive comorbidity patterns for treatment and prevention work remain unclear because etiological explanations of these patterns continue to be widely disputed. One common etiological theory contends that successive comorbidity is indicative of an underlying liability that manifests in different forms across development-resulting in probabilities of symptom endorsement that change systematically across time. A competing etiological theory contends that successive comorbidity is indicative of multiple disorder- types. Types manifest distinct disease progressions, or pathways, and symptom endorsement probabilities vary between types. Refining etiological explanations of successive comorbidity requires adjudicating between these two competing theories. Several obstacles have arisen. The particular models traditionally used to test the first theory and those used to test the second theory have contradictory assumptions. So, one theory is assumed false to test the other theory. The models decompose the correlations among symptoms in opposite ways (analytically equivalent under certain conditions), precluding us from identifying in what respects each theory is supported. This study does three things. First, it evaluates a promising new integrative methodology for simultaneously modeling changes in the structure of psychopathology over time, and population heterogeneity in symptom co-occurrence pathways. The performance of the proposed model will be studied in a Monte Carlo simulation, under conditions typically encountered in comorbidity research with DSM symptoms (e.g. symptom non-normality). Second, this study seeks to replicate several of the most often studied successive comorbidity patterns across the transition to adolescence, in a longitudinal, epidemiological dataset with detailed psychiatric interview measurement of psychopathology-using each traditional model. Since there has been little demonstration, outside of the methodological literature, that the same data can produce results consistent with either etiological theory, merely due to the analysis chosen, this demonstration can propel researchers to consider new, integrative methods of analysis. Third, the proposed integrative model will be applied to provide the first simultaneous test of both etiological theories of successive comorbidity across the transition to adolescence, to our knowledge.
