Identification of some individual and social factors on relapse intervals in bipolar disorder:Bayesian analysis of log-skew-normal accelerated failure time model

Bipolar disorder is severe, chronic, pleomorphic and often recurrent, which can lead to severe abnormalities in lif's function. The purpose of this study was to determine some factors related to the time interval between relapse in bipolar disorder.
In a retrospective study, 606 patients with bipolar disorder at Avicenna hospital in Mashhad were collected. In this study, 329 patients with at least one relapse of bipolar disorder were included. Survival time was defined as elapsed time from discharge to readmission due to relapse of disease. These patients wereadmitted from the beginning of 2008 to the end of 2009 due to their illness and were include in the study and followed until the end of 2013. The log-skew-normal accelerated failure time model was fitted to identify the factors related to the time of relapse. Estimation of parameters was obtained based on the Bayesian approach using Markov chain Monte Carlo algorithm by Open BUGS.
The estimate of mean and median time between discharge and readmission were 25.6 and 15 months, respectively. Age, family history, drug use, and gender had a significant association with the time of relapse, such that, male, younger patients, drug users, and positive family history of disease experienced a recurrence of disease earlier. With the mentioned variables in the model, stress and education were not associated with the recurrence of disease.
Given that the time interval from discharge to recurrence of type 1 bipolar disorder is low in men, young people, family history of positive and drug users, it is recommended to identify strategies to prevent relapse or delay in these groups.