Development of a longitudinal model for characterizing adverse events of psychiatric drugs in routine clinical care
Lay Ahyoung Lim (1), Eun Lee (2), Kyungsoo Park (1)
(1) Department of Pharmacology, College of Medicine, Yonsei University, Seoul, Korea (2) Department of Psychiatry, College of Medicine, Yonsei University, Seoul, Korea
Objectives: In routine clinical care of psychiatric patients, the early treatment is important because adverse events (AEs) in this period often lead to noncompliance to a drug and lowering the therapeutic effect. This study aimed to develop a longitudinal model to describe early-phase AEs in Korean psychiatric patients in an effort to be used as a guide to improve medication compliance and drug efficacy.
Methods: Data were collected retrospectively from the medical records of outpatient clinic in Severance hospital, Seoul, Korea, involving 150 patients treated with anxiolytics or antidepressants. Data were censored on day 60 from their first visit. Using NONMEM 7, three different longitudinal models were developed within a mixed-effect model framework to describe the incidence, the time-to-event (TTE), and the count of AEs. The hazard function to describe censored data or dropout was chosen to be a constant for the incidence model and the Weibull function for TTE and count models. For the incidence model, a first order Markov element was also included. To evaluate the model performances, visual predictive check (VPC) was performed using 100 datasets simulated from each model using estimated parameters.
Results: The most frequently observed AE was drowsiness. About 30% of the patients reported AE more than once during the observed period. For the incidence model, a Markov element added in the baseline logit adequately described the data. Incorporating an exponential decay function as a time effect further improved the model, dropping OFV by134.28. VPC showed good the performance of the model. For TTE model, the estimated shape parameter of hazard function was 2.08, indicating the hazard probability increasing with time, and VPC showed the prediction somewhat overestimated. Of several models, the Weibull hazard model dropped the OFV most significantly in both TTE and count models.
Conclusions: Our preliminary results show that the incidence model described the data well whereas the TTE model needs to be further developed. To generalize our results, more work will be necessary, including assessing covariate influence on AEs with more patients. Including severity into the model will further improve the applicability of the model if such information is available.
References:
[1] Ette EI, Roy A, Nandy P. Population Pharmacokinetic/Pharmacodynamic Modeling of Ordered Categorical Longitudinal Data. In Ette EI and Williams PJ (Eds.), (2007) Pharmacometrics; John Wiley & Sons, Inc.: 655-688.
[2] Cox E H, Follet C, Beal SL, Fuseau E, Kenkare S, Sheiner LB. A Population Pharmacokinetic-Pharmacodynamic Analysis of Repeated Measures Time-to-Event Pharmacodynamic Responses: The Antiemetic Effect of Ondansetron; Journal of Pharmacokinetics and Biopharmaceutics (1999); 27(6): 625-644
[3] Plan EL, Karlsson KE, Karlsson MO. Approaches to Simultaneous Analysis of Frequency and Severity of Symptoms; Clinical Pharmacology & Therapeutics (2010); 88(2): 255-259
