Who will benefit from antidepressants in the acute treatment of bipolar depression? A reanalysis of the STEP-BD study by Sachs et al. 2007, using Q-learning.

BACKGROUND: There is substantial uncertainty regarding the efficacy of antidepressants in the treatment of bipolar disorders. METHODS: Traditional randomized controlled trials and statistical methods are not designed to discover if, when, and to whom an intervention should be applied; thus, other methodological approaches are needed that allow for the practice of personalized, evidence-based medicine with patients with bipolar depression. RESULTS: Dynamic treatment regimes operationalize clinical decision-making as a sequence of decision rules, one per stage of clinical intervention, that map patient information to a recommended treatment. Using data from the acute depression randomized care (RAD) pathway of the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) study, we estimate an optimal dynamic treatment regime via Q-learning. CONCLUSIONS: The estimated optimal treatment regime presents some evidence that patients in the RAD pathway of STEP-BD who experienced a (hypo)manic episode before the depressive episode may do better to forgo adding an antidepressant to a mandatory mood stabilizer.

Typology of patients with fibromyalgia: cluster analysis of duloxetine study patients.

BACKGROUND: To identify distinct groups of patients with fibromyalgia (FM) with respect to multiple outcome measures. METHODS: Data from 631 duloxetine-treated women in 4 randomized, placebo-controlled trials were included in a cluster analysis based on outcomes after up to 12 weeks of treatment. Corresponding classification rules were constructed using a classification tree method. Probabilities for transitioning from baseline to Week 12 category were estimated for placebo and duloxetine patients (Ntotal = 1188) using logistic regression. RESULTS: Five clusters were identified, from "worst" (high pain levels and severe mental/physical impairment) to "best" (low pain levels and nearly normal mental/physical function). For patients with moderate overall severity, mental and physical symptoms were less correlated, resulting in 2 distinct clusters based on these 2 symptom domains. Three key variables with threshold values were identified for classification of patients: Brief Pain Inventory (BPI) pain interference overall scores of <3.29 and <7.14, respectively, a Fibromyalgia Impact Questionnaire (FIQ) interference with work score of <2, and an FIQ depression score of ≥5. Patient characteristics and frequencies per baseline category were similar between treatments; >80% of patients were in the 3 worst categories. Duloxetine patients were significantly more likely to improve after 12 weeks than placebo patients. A sustained effect was seen with continued duloxetine treatment. CONCLUSIONS: FM patients are heterogeneous and can be classified into distinct subgroups by simple descriptive rules derived from only 3 variables, which may guide individual patient management. Duloxetine showed higher improvement rates than placebo and had a sustained effect beyond 12 weeks.

Understanding heterogeneity in response to antidiabetes treatment: a post hoc analysis using SIDES, a subgroup identification algorithm.

BACKGROUND: Treatment response in patients with type 2 diabetes mellitus (T2DM) varies because of different genotypic and phenotypic characteristics. Results of clinical trials are based largely on aggregated estimates of treatment effect rather than individualized outcomes. This research assessed heterogeneity and differential treatment response using the subgroup identification based on differential effect search (SIDES) algorithm with data from a large multinational study. METHODS: This was a retrospective analysis of the DURABLE trial, a randomized, open-label, two-arm, parallel study. The trial enrolled 2091 insulin-naïve T2DM patients aged 30 to 80 years. Patients received twice-daily insulin lispro mix 75/25 (LM75/25) or once-daily insulin glargine (insulin glargine). The SIDES methodology was used to find subgroups where the treatment effect was substantially different from what was observed in the full population of the clinical trial. A subgroup identification tool implementing the SIDES algorithm was used to examine data for 1092 patients (584 LM75/25 and 508 insulin glargine), achieving at-goal 12- or 24-week glycated hemoglobin A1c (A1C) (≤7.0%). RESULTS: The overall at-goal population treatment difference (A1C reduction) was not clinically meaningful, but a clinically meaningful difference was observed (A1C reduction 2.31% ± 0.06% LM75/25 versus 2.01% ± 0.07% insulin glargine; p = .001) in patients with a baseline fasting insulin level >11.4 µIU/ml and age ≤56 years. CONCLUSION: The observation that younger patients with higher levels of fasting insulin may benefit from a regimen that includes short-acting insulin targeting postprandial glycemia helps explain the heterogeneity in response and warrants further investigation.

Comparison of several imputation methods for missing baseline data in propensity scores analysis of binary outcome.

We performed a simulation study comparing the statistical properties of the estimated log odds ratio from propensity scores analyses of a binary response variable, in which missing baseline data had been imputed using a simple imputation scheme (Treatment Mean Imputation), compared with three ways of performing multiple imputation (MI) and with a Complete Case analysis. MI that included treatment (treated/untreated) and outcome (for our analyses, outcome was adverse event [yes/no]) in the imputer's model had the best statistical properties of the imputation schemes we studied. MI is feasible to use in situations where one has just a few outcomes to analyze. We also found that Treatment Mean Imputation performed quite well and is a reasonable alternative to MI in situations where it is not feasible to use MI. Treatment Mean Imputation performed better than MI methods that did not include both the treatment and outcome in the imputer's model.

Defining "good" and "poor" outcomes in patients with schizophrenia or schizoaffective disorder: a multidimensional data-driven approach.

The study's goal was to characterize the typology of patient outcomes based on social and occupational functioning and psychiatric symptoms following antipsychotic drug treatment, and to explore predictors of group membership representing the best/worst outcomes. A hierarchical cluster analysis was used to define groups of patients (n=1449) based on endpoint values for psychiatric symptoms, social functioning, and useful work measured up to 30 weeks of treatment. Stepwise logistic regression was used to construct predictive models of cluster membership for baseline predictors, and with 2/4/8 weeks of treatment. Five distinct clusters of patients were identified at endpoint (Clusters A-E). Patients in Cluster A (25.6%, best outcome) had minimal psychiatric symptoms and mild functional impairment, while patients in Cluster D (14.3%) and E (14.8%) (worst outcome) had moderate-to-severe symptoms and severe functional impairment. Occupational functioning, disorganized thinking, and positive symptoms were sufficient to describe the clusters. Membership in the best/worst clusters was predicted by baseline scores for functioning and symptom severity, and by early changes in symptoms with treatment. Psychiatric symptoms and functioning provided complementary information to describe treatment outcomes. Early symptom response significantly improved the prediction of outcome, suggesting that early monitoring of treatment response may be useful in clinical practice.

Relationships among neurocognition, symptoms and functioning in patients with schizophrenia: a path-analytic approach for associations at baseline and following 24 weeks of antipsychotic drug therapy.

BACKGROUND: Neurocognitive impairment and psychiatric symptoms have been associated with deficits in psychosocial and occupational functioning in patients with schizophrenia. This post-hoc analysis evaluates the relationships among cognition, psychopathology, and psychosocial functioning in patients with schizophrenia at baseline and following sustained treatment with antipsychotic drugs. METHODS: Data were obtained from a clinical trial assessing the cognitive effects of selected antipsychotic drugs in patients with schizophrenia. Patients were randomly assigned to 24 weeks of treatment with olanzapine (n = 159), risperidone (n = 158), or haloperidol (n = 97). Psychosocial functioning was assessed with the Heinrichs-Carpenter Quality of Life Scale [QLS], cognition with a standard battery of neurocognitive tests; and psychiatric symptoms with the Positive and Negative Syndrome Scale [PANSS]. A path-analytic approach was used to evaluate the effects of changes in cognitive functioning on subdomains of quality of life, and to determine whether such effects were direct or mediated via changes in psychiatric symptoms. RESULTS: At baseline, processing speed affected functioning mainly indirectly via negative symptoms. Positive symptoms also affected functioning at baseline although independent of cognition. At 24 weeks, changes in processing speed affected changes in functioning both directly and indirectly via PANSS negative subscale scores. Positive symptoms no longer contributed to the path-analytic models. Although a consistent relationship was observed between processing speed and the 3 functional domains, variation existed as to whether the paths were direct and/or indirect. Working memory and verbal memory did not significantly contribute to any of the path-analytic models studied. CONCLUSION: Processing speed demonstrated direct and indirect effects via negative symptoms on three domains of functioning as measured by the QLS at baseline and following 24 weeks of antipsychotic treatment.

Identifying patterns in treatment response profiles in acute bipolar mania: a cluster analysis approach.

UNLABELLED: Patients with acute mania respond differentially to treatment and, in many cases, fail to obtain or sustain symptom remission. The objective of this exploratory analysis was to characterize response in bipolar disorder by identifying groups of patients with similar manic symptom response profiles. METHODS: Patients (n = 222) were selected from a randomized, double-blind study of treatment with olanzapine or divalproex in bipolar I disorder, manic or mixed episode, with or without psychotic features. Hierarchical clustering based on Ward's distance was used to identify groups of patients based on Young-Mania Rating Scale (YMRS) total scores at each of 5 assessments over 7 weeks. Logistic regression was used to identify baseline predictors for clusters of interest. RESULTS: Four distinct clusters of patients were identified: Cluster 1 (n = 64): patients did not maintain a response (YMRS total scores < or = 12); Cluster 2 (n = 92): patients responded rapidly (within less than a week) and response was maintained; Cluster 3 (n = 36): patients responded rapidly but relapsed soon afterwards (YMRS > or = 15); Cluster 4 (n = 30): patients responded slowly (> or = 2 weeks) and response was maintained. Predictive models using baseline variables found YMRS Item 10 (Appearance), and psychosis to be significant predictors for Clusters 1 and 4 vs. Clusters 2 and 3, but none of the baseline characteristics allowed discriminating between Clusters 1 vs. 4. Experiencing a mixed episode at baseline predicted membership in Clusters 2 and 3 vs. Clusters 1 and 4. Treatment with divalproex, larger number of previous manic episodes, lack of disruptive-aggressive behavior, and more prominent depressive symptoms at baseline were predictors for Cluster 3 vs. 2. CONCLUSION: Distinct treatment response profiles can be predicted by clinical features at baseline. The presence of these features as potential risk factors for relapse in patients who have responded to treatment should be considered prior to discharge. TRIAL REGISTRATION: The clinical trial cited in this report has not been registered because it was conducted and completed prior to the inception of clinical trial registries.

Initial symptoms of manic relapse in manic or mixed-manic bipolar disorder: post hoc analysis of patients treated with olanzapine or lithium.

UNLABELLED: A post hoc analysis of Young Mania Rating Scale (YMRS) item scores was conducted to identify symptoms that may predict impending relapse using prospectively collected data from a double-blind, randomized relapse prevention study of patients treated with olanzapine (N=200, 5-20 mg/d) versus lithium (N=201, 300-1800 mg/d). METHODS: Relapses (YMRS > or = 15, or hospitalization) included in this analysis occurred 3-52 weeks after randomization. Repeated measures logistic regression of increases (> or = 1) in YMRS item scores prior to the visit that preceded relapse was used to estimate the odds of relapse. RESULTS: A total of 31 patients relapsed during the first 3-16 weeks of the study (olanzapine, n=12; lithium, n=19). YMRS items that increased most frequently within a 2-week period preceding relapse were (olanzapine vs. lithium, respectively): increased motor activity/energy (58.3%, 21.1%), irritability (33.3%, 31.6%), decreased need for sleep (25.0%, 10.5%), increased speech (25.0%, 10.5%), and elevated mood (25.0%, 15.8%). YMRS items with significant odds ratios (OR) that predicted relapse in patients treated with olanzapine or lithium, respectively, were: increased motor activity/energy (OR, 35.7; OR, 7.8), irritability (OR, 9.5; OR, 7.8), elevated mood (OR, 8.1; OR, 4.2), and increased sexual interest (OR, 13.7; OR, 7.7). CONCLUSIONS: Early recognition of symptom exacerbation in bipolar mania, particularly increased motor activity-energy may permit clinical interventions to help avert relapse.