Details

Comparisons and Applications of Quantitative Signal Detections for Adverse Drug Reactions (ADRs): An Empirical Study Based On The Food And Drug Administration (FDA) Adverse Event Reporting System (AERS) And A Large Medical Claims Database

by CHEN, YAN

Abstract (Summary)
This dissertation study, consisting of three chapters, aims to compare the four commonly used data mining algorithms (DMAs) in terms of the sensitivity and the manner of early signal detection; to demonstrate the applications of the DMAs in adverse drug reactions (ADRs) signals detection; and to evaluate the risk of neuroleptic malignant syndrome (NMS) associated with the use of antipsychotic agents. In the first chapter, four DMAs, including the Reporting Odds Ratio (ROR), the Proportional Reporting Ratio (PRR), the Information Component (IC), and the Gamma Poisson Shrinker (GPS), are compared in terms of the sensitivity and the manner of ADRs signals early detection based upon the Adverse Event Reporting System (AERS) of the Food and Drug Administration (FDA). In the second chapter, two examples are studied to demonstrate the applications of the four DMAs in ADRs signal detection: the risk of NMS associated with the use of antipsychotics and the risk of liver injuries associated with the use of telithromycin (Ketek®). In the third chapter, a population-based nested case-control study is conducted to examine risk factors of NSM associated with the use of antipsychotic among patients with bipolar disorder. In summary, this study is the first one to compare the four commonly used DMAs in terms of the sensitivity and the manner of early ADRs signals detection. Our findings appear to be more robust and reliable than previous studies, since the use of confirmed DECs as reference standards eliminates possible confounding effects due to the inclusion of false-positive DECs. Although the sensitivities vary across the four DMAs, there is no statistically significant difference. The ROR seems to detect more number of DECs than the other algorithms. The sensitivity increases when the number of reports per DEC increases. When compares the index date of detection (IDD, defined as the time when the DEC was detected) with the index date of withdrawal (IDW, defined as the time when the FDA removed the drugs from the market), the ROR is able to detect the ADRs signals earlier than the other algorithms. As expected, the results from the ADRs signals detection through the AERS clearly and consistently shows a higher-than-expected number of reports of NMS in patients with the use of antipsychotics. It supports the belief that the use of antipsychotics is in an association with an increased risk of NMS. The four DMAs also highlight the hypothesis that the use of telithromycin may be associated with an increased risk of liver injuries. In the case-control study, the observation that an increased risk of NMS is associated with the use of an antipsychotic, particularly among patients with the use of a high D2 potent antipsychotic, not only supports the belief that the use of an antipsychotic plays a critical role in the occurrence of NMS, but further suggests the magnitude of the risk of NMS seems to be related to an antipsychotic…#8482; potency of D2 inhibition . Besides the risk associated with the use of antipsychotics, our study suggests that other risk factors, including the gender (being male), the pre-existing delirium, confusion, dehydration, and extrapyramdial signs, are associated with an increased risk of NMS as well.
Bibliographical Information:

Advisor:

School:University of Cincinnati

School Location:USA - Ohio

Source Type:Master's Thesis

Keywords:data mining algorithms adverse drug reactions event reporting system signal detection case control study antipsychotic bipolar disorder

ISBN:

Date of Publication:01/01/2008

© 2009 OpenThesis.org. All Rights Reserved.