Identifying Patterns of Association and Transition in the Use of Addictive Substances over Five Decades in the United States

ObjectivesTo conduct weighted itemset analysis to identify patterns of polysubstance use from National Survey on Drug Use and Health (NSDUH) data. MethodsWe computed weighted support for every combination of one or more substances, termed as a drugset, used by individuals in the nationally representative NSDUH data over 5 decades (1965-2014). A computational framework for efficient representation and search of patterns of association between drugsets and demographics of user groups over time was developed. A new method for mining rules of transition between pairs of substances used within a time-interval was given. ResultsWe identified the frequent drugsets from individual substance use data, and determined their representation among different demographic groups at different intervals. An interesting pattern of use of pain relievers and tranquilizers was detected for the age-group of 26-34 years. Transition rules for heroin use in the last decade (2004-2015) of data were mined. ConclusionsComputation of weighted supports over time for every possible drugset in the data and their association with specific user groups produced a framework for generation and testing new and interesting hypotheses. The framework is useful to explore different combinations of substances used among diverse demographic groups including those that have received less attention in this problem.