Data mining PCC annual reports

As submitted by the proposer: Investigation of drug-associated fatalities is an important component of forensic toxicology. A potential mitigating factor is drug interactions that increase the concentration of active drug, or active drug metabolite. Besides involvement in causing fatalities, this factor must also be considered in the forensic toxicology interpretation during case investigation. While we know many of the drugs that cause drug interactions; information on their actual co-occurrence in these cases is useful to direct investigations. Conversely, detection of co-administered drugs in the laboratory is time consuming and often limited to select agents; information on drugs not currently considered select agents but revealed for their drug interaction potential may be equally useful. The American Association of Poison Control Centers publishes an annual report on findings from the National Poison Control Data System. Table 21 in these reports provides a listing of fatal non-pharmaceutical and pharmaceutical exposures. Provided with each case is: the annual report ID, age/gender, substance(s), substance rank, cause rank, chronicity, route, reason, relative contribution to fatality, and analytical findings when available. Summarization of the data is currently limited, particularly for co-occurrence of drugs. Tables 5 in the annual reports list the number of substances involved (i.e., from1 to >9) in exposures and fatalities, but not specific co-exposures. We propose to mine the databases (collected Tables 21) from 2000 to 2013 and collate the co-occurrence of pharmaceuticals (and alcohols) in the listed fatalities. The primary outcomes will be the frequency of specific drug association with fatalities over time (currently provided for drug classes but not specific drugs), and the frequency of co-occurrences over time and over the total time period. A secondary outcome will be tabulation of the concentrations of the primary drug involved in cases. Preliminary mining for the annual occurrence of selected analgesics (e.g., methadone, oxycodone, hydrocodone, buprenorphine, etc), and the co-occurrence of benzodiazepines, provided familiarity with the database and the time needed to mine such data. The preliminary data also provide a framework for the structure of our plan to analyze the data and test for statistical differences. The results of these findings will be disseminated through presentations at an annual meeting of the Society of Forensic Toxicologists and through publication(s) in a peer-reviewed journal. In this manner we propose to add to the knowledge base concerning the basic science of drug interaction potentials. The PI and his colleagues are aptly suited to carry out these studies. ca/ncf