Original Article Analysis of bipolar maintenance treatment with lithium versus olanzapine utilizing Multi-state Outcome Analysis of Treatments (MOAT) Tohen M, Mintz J, Bowden CL. Analysis of bipolar maintenance treatment with lithium versus olanzapine utilizing Multi-state Outcome Analysis of Treatments (MOAT) Bipolar Disord 2016: 18: 282–287. © 2016 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd. Objectives: Survival analysis has superseded most other analytic techniques for maintenance treatment studies over recent decades, despite providing results based solely on a single time-point predefined event. The aim of the present study was to develop the Multi-state Outcome Analysis of Treatments (MOAT), to provide more pragmatic information for clinicians and investigators in guiding maintenance treatment decisions. The present study was one of two published studies on the development of MOAT procedures, involving a one-year comparison of olanzapine versus lithium in recently manic patients. Methods: MOAT partitions total survival time into clinically distinct periods that are operationally defined by cut points on established rating scales. For bipolar disorders, the clinical states are remission, subsyndromal and syndromal mania, mixed states, and subsyndromal and syndromal depression. Results: MOAT re-analyses of the clinical trial revealed clinically important findings not identified when utilizing Kaplan–Meier survival analyses. Compared to patients treated with lithium, patients taking olanzapine experienced significantly more time in subsyndromal depression. Patients taking lithium spent significantly more time in mixed states than did patients taking olanzapine. Conclusions: MOAT provided detailed information on treatment outcomes that was not provided by Kaplan–Meier survival analysis. Its capability to identify and aggregate time in different clinical states of bipolar disorder may aid in identifying drug effects that are important in selecting and conducting maintenance treatment. Mauricio Tohen a , Jim Mintz b and Charles L Bowden b a Department of Psychiatry and Behavioral Sciences, University of New Mexico Health Sciences Center, Albuquerque, NM, b Division of Mood and Anxiety Disorders, Department of Psychiatry, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA doi: 10.1111/bdi.12383 Key words: bipolar disorder – outcome analysis – treatment Received 21 August 2015, revised 30 December 2015, revised and accepted for publication 8 February 2016 Corresponding author: Mauricio Tohen, MD, DrPH, MBA Department of Psychiatry and Behavioral Sciences University of New Mexico Health Sciences Center 2400 Tucker Avenue Albuquerque, NM 87131-0001 USA Fax: 505 272-2580 E-mail: mtohen@gmail.com A limitation of traditional survival analysis of time until a pre-specified clinical event is not consider- ing what happens between time zero and the speci- fied outcome. We developed the Multi-state Outcome Analysis of Treatments (MOAT) to cap- ture better the actual course of maintenance treat- ments in bipolar disorder. In the first of two developmental projects for MOAT, we analyzed combined data from the two registration studies of lamotrigine, lithium, and placebo in the mainte- nance treatment of bipolar disorder (1). We con- firmed longer time in study and more time in remission for both active drugs. Time spend in manic states, particularly at the subsyndromal level, was reduced with lithium, but not lamotrig- ine. In all three of the treatment conditions, 282 Bipolar Disorders 2016: 18: 282–287 © 2016 John Wiley & Sons A/S Published by John Wiley & Sons Ltd. BIPOLAR DISORDERS