The analysis of stage-sequential process plays an important role in behavioral and biomedical applications, particularly in the study of developmental processes of substance use behavior. Using a stage-sequential process, researchers are able to explore sequential patterns of substance use behavior and find optimal opportunities for intervening in the onset process. The proposed research will develop an innovative statistical framework, the Latent Class-Profile Analysis (LCPA), which will be ideally suited to this type of research. LCPA will provide a set of principles for systematic identification of homogeneous subgroups of individuals by dividing identification processes into two levels: at the first level, LCPA identifies discrete classes of individuals who have similar responses to the items at each occasion (latent class); at the second level, LCPA examines individuals' class membership over the entire set of time points so as to classify the population into two or more subgroups. All individuals in a second-level subgroup are expected to have the similar sequential pattern of substance use behavior over time (latent class-profile). This approach allows us to make no further assumptions about the nature of the classes and their sequential patterns. For example, individuals classified into a specific class-profile can have both dimensional and configurational similarity in their substance abuse initiation and progression. In addition, no common precepts of behavioral patterns in substance use are required to identify stage-sequential processes of substance use. This type of information cannot be accomplished by the current available methods. We will establish the ability of LCPA to link potential risk factors to stage-sequential process of substance use behavior. This will allow us to investigate subgroup (e.g. gender/sex) differences in patterns of substance abuse initiation and progression along with subject-specific differences. An important goal of the proposed study is a better methodological understanding of how to investigate stage-sequential development in substance use behavior, and how to apply to real data examples in a responsible manner. This work will integrate recent advances in statistics with pressing needs for new tools in substance abuse research. With information gathered from this study, we will develop guidelines and conduct secondary analyses to provide practical recommendations in the context of drug abuse research. An open-source computer software will be developed to make LCPA accessible to a wide audience. This will provide a rich toolkit for investigating the substance abuse initiation and progression and its relationships to risk factors.
The analysis of stage-sequential process plays an important role in behavioral and biomedical applications, particularly in the study of developmental processes of substance use behavior. The proposed research will develop an innovative statistical framework to investigate stage-sequential process in substance abuse initiation and progression. The project includes dissemination of information to substance use scientists about how to apply the proposed methods for their research. ? ? ?
| Chung, Hwan; Anthony, James C; Schafer, Joseph L (2011) Latent class profile analysis: an application to stage-sequential process in early-onset drinking behaviours. J R Stat Soc Ser A Stat Soc 174:689-712 |
