Bayesian multivariate growth curve latent class models for mixed outcomes.

In many clinical studies, the disease of interest is multifaceted, and multiple outcomes are needed to adequately capture information about the characteristics of the disease or its severity. In the analysis of such diseases, it is often difficult to determine what constitutes improvement because of the multivariate nature of the outcome. Furthermore, when the disease of interest has an unknown etiology and/or is primarily a symptom-defined syndrome, there is potential for the disease population to have distinct subgroups. Identification of population subgroups is of interest as it may assist clinicians in providing appropriate treatment or in developing accurate prognoses. We propose multivariate growth curve latent class models that group subjects on the basis of multiple symptoms measured repeatedly over time. These groups or latent classes are defined by distinctive longitudinal profiles of a latent variable, which is used to summarize the multivariate outcomes at each point. The mean growth curve for the latent variable in each class defines the features of the class. We develop this model for any combination of continuous, binary, ordinal, or count outcomes within a Bayesian hierarchical framework. We use simulation studies to validate the estimation procedures. We apply our model to data from a randomized clinical trial evaluating the efficacy of Bacillus Calmette-Guerin in treating symptoms of interstitial cystitis where we are able to identify a class of subjects for whom treatment is effective.

Effectiveness of long-term acute care hospitalization in elderly patients with chronic critical illness.

BACKGROUND: For patients recovering from severe acute illness, admission to a long-term acute care hospital (LTAC) is an increasingly common alternative to continued management in an intensive care unit (ICU). OBJECTIVE: To examine the effectiveness of LTAC transfer in patients with chronic critical illness. RESEARCH DESIGN: Retrospective cohort study in United States hospitals from 2002 to 2006. SUBJECTS: Medicare beneficiaries with chronic critical illness, defined as mechanical ventilation and at least 14 days of intensive care. MEASURES: Survival, costs, and hospital readmissions. We used multivariate analyses and instrumental variables to account for differences in patient characteristics, the timing of LTAC transfer, and selection bias. RESULTS: A total of 234,799 patients met our definition of chronic critical illness. Of these, 48,416 (20.6%) were transferred to an LTAC. In the instrumental variable analysis, patients transferred to an LTAC experienced similar survival compared with patients who remained in an ICU [adjusted hazard ratio=0.99; 95% confidence interval (CI), 0.96 to 1.01; P=0.27). Total hospital-related costs in the 180 days after admission were lower among patients transferred to LTACs (adjusted cost difference=-$13,422; 95% CI, -26,662 to -223, P=0.046). This difference was attributable to a reduction in skilled nursing facility admissions (adjusted admission rate difference=-0.591; 95% CI, -0.728 to -0.454; P<0.001). Total Medicare payments were higher (adjusted cost difference=$15,592; 95% CI, 6343 to 24,842; P=0.001). CONCLUSIONS: Patients with chronic critical illness transferred to LTACs experience similar survival compared with patients who remain in ICUs, incur fewer health care costs driven by a reduction in postacute care utilization, however, invoke higher overall Medicare payments.

Acute alcohol consumption, alcohol outlets, and gun suicide.

A case-control study of 149 intentionally self-inflicted gun injury cases (including completed gun suicides) and 302 population-based controls was conducted from 2003 to 2006 in a major US city. Two focal independent variables, acute alcohol consumption and alcohol outlet availability, were measured. Conditional logistic regression was adjusted for confounding variables. Gun suicide risk to individuals in areas of high alcohol outlet availability was less than the gun suicide risk they incurred from acute alcohol consumption, especially to excess. This corroborates prior work but also uncovers new information about the relationships between acute alcohol consumption, alcohol outlets, and gun suicide. Study limitations and implications are discussed.

Impact of bacterial meningitis-associated conditions on pediatric inpatient resource utilization.

OBJECTIVE: To define the epidemiology of systemic complications and focal infections associated with bacterial meningitis and quantify how the presence of such complications affects in-hospital healthcare resource utilization. METHODS: Retrospective cohort study using administrative data from 27 children's hospitals. Children <18 years of age diagnosed with bacterial meningitis from 2001 to 2006 were eligible. The primary exposure of interest was the presence of a bacterial meningitis-associated condition, classified as either systemic complications (eg, sepsis), associated focal infections (eg, pneumonia) or both. The primary outcomes were total in-hospital charges and length of stay (LOS). RESULTS: A total of 574 of 2319 (25%) of children had a systemic complication or an associated focal infection. Compared with children without complications, in-hospital charges were significantly higher in children with systemic complications (136% increase), associated focal infections (118% increase), and both conditions (351% increase). LOS was also significantly increased in those with systemic complications (by 72%), associated focal infections (by 78%), or both conditions (by 211%). The presence of systemic complications was more common in younger children while the presence of an associated focal infection was more common in older children. CONCLUSIONS: Children with bacterial meningitis often have additional morbidity due to systemic complications or associated focal infections indicated by increase use of acute in-hospital resource utilization. The apparent increase in in-hospital morbidity related to these conditions should be considered in future evaluations of vaccine efficacy, novel therapeutics, and hospital resource allocation.

Determinants of the optimal first-line therapy for follicular lymphoma: a decision analysis.

Combination immunochemotherapy is the most common approach for initial therapy of patients with advanced-stage follicular lymphoma, but no consensus exists as to the optimal selection or sequence of available regimens. We undertook this decision analysis to systematically evaluate the parameters affecting the choice of early therapy in patients with this disease. We designed a Markov model incorporating the three most commonly utilized regimens (RCVP, RCHOP, and RFlu) in combinations of first- and second-line therapies, with the endpoint of number of quality-adjusted life years (QALYs) until disease progression. Data sources included Phase II and Phase III trials and literature estimates of long-term toxicities and health state utilities. Meta-analytic methods were used to derive the values and ranges of regimen-related parameters. Based on our model, the strategy associated with the greatest number of expected quality-adjusted life years was treatment with RCHOP in first-line therapy followed by treatment with RFlu in second-line therapy (9.00 QALYs). Strategies containing RCVP either in first- or second-line therapy resulted in the lowest number of QALYs (range 6.24-7.71). Sensitivity analysis used to determine the relative contribution of each model parameter identified PFS after first-line therapy and not short-term QOL as the most important factor in prolonging overall quality-adjusted life years. Our results suggest that regimens associated with a longer PFS provide a greater number of total QALYs, despite their short-term toxicities. For patients without contraindications to any of these regimens, use of a more active regimen may maximize overall quality of life.

Costs of treating children with complicated pneumonia: a comparison of primary video-assisted thoracoscopic surgery and chest tube placement.

OBJECTIVES: To describe charges associated with primary video-assisted thoracoscopic surgery (VATS) and primary chest tube placement in a multicenter cohort of children with empyema and to determine whether pleural fluid drainage by primary VATS was associated with cost-savings compared with primary chest tube placement. STUDY DESIGN: Retrospective cohort study. SETTING AND PARTICIPANTS: Administrative database containing inpatient resource utilization data from 27 tertiary care children's hospitals. Patients between 12 months and 18 years of age diagnosed with complicated pneumonia were eligible if they were discharged between 2001 and 2005 and underwent early (within 2 days of index hospitalization) pleural fluid drainage. MAIN EXPOSURE: Method of pleural fluid drainage, categorized as VATS or chest tube placement. RESULTS: Pleural drainage in the 764 patients was performed by VATS (n = 50) or chest tube placement (n = 714). There were 521 (54%) males. Median hospital charges were $36,320 [interquartile range (IQR), $24,814-$62,269]. The median pharmacy and radiologic imaging charges were $5,884 (IQR, $3,142-$11,357) and $2,875 (IQR, $1,703-$4,950), respectively. Adjusting for propensity score matching, patients undergoing primary VATS did not have higher charges than patients undergoing primary chest tube placement. CONCLUSIONS: In this multicenter study, we found that the charges incurred in caring for children with empyema were substantial. However, primary VATS was not associated with higher total or pharmacy charges than primary chest tube placement, suggesting that the additional costs of performing VATS are offset by reductions in length of stay (LOS) and requirement for additional procedures.

Adaptive designs for randomized trials in public health.

In this article, we present a discussion of two general ways in which the traditional randomized trial can be modified or adapted in response to the data being collected. We use the term adaptive design to refer to a trial in which characteristics of the study itself, such as the proportion assigned to active intervention versus control, change during the trial in response to data being collected. The term adaptive sequence of trials refers to a decision-making process that fundamentally informs the conceptualization and conduct of each new trial with the results of previous trials. Our discussion below investigates the utility of these two types of adaptations for public health evaluations. Examples are provided to illustrate how adaptation can be used in practice. From these case studies, we discuss whether such evaluations can or should be analyzed as if they were formal randomized trials, and we discuss practical as well as ethical issues arising in the conduct of these new-generation trials.

Nested Markov compliance class model in the presence of time-varying noncompliance.

SUMMARY: We consider a Markov structure for partially unobserved time-varying compliance classes in the Imbens-Rubin (1997, The Annals of Statistics 25, 305-327) compliance model framework. The context is a longitudinal randomized intervention study where subjects are randomized once at baseline, outcomes and patient adherence are measured at multiple follow-ups, and patient adherence to their randomized treatment could vary over time. We propose a nested latent compliance class model where we use time-invariant subject-specific compliance principal strata to summarize longitudinal trends of subject-specific time-varying compliance patterns. The principal strata are formed using Markov models that relate current compliance behavior to compliance history. Treatment effects are estimated as intent-to-treat effects within the compliance principal strata.

Baseline patient characteristics and mortality associated with longitudinal intervention compliance.

Lin et al. (http://www.biostatsresearch.com/upennbiostat/papers/, 2006) proposed a nested Markov compliance class model in the Imbens and Rubin compliance class model framework to account for time-varying subject noncompliance in longitudinal randomized intervention studies. We use superclasses, or latent compliance class principal strata, to describe longitudinal compliance patterns, and time-varying compliance classes are assumed to depend on the history of compliance. In this paper, we search for good subject-level baseline predictors of these superclasses and also examine the relationship between these superclasses and all-cause mortality. Since the superclasses are completely latent in all subjects, we utilize multiple imputation techniques to draw inferences. We apply this approach to a randomized intervention study for elderly primary care patients with depression.

A mixed effects Markov model for repeated binary outcomes with non-ignorable dropout.

In many areas of research, repeated binary measures often represent a two-state stochastic process, where individuals can transition among two states. In a behavioural or physical disability setting, individuals can flow from susceptible or subthreshold state, to an infectious or symptomatic state, and back to a subthreshold state. Quite often the transition among the states happens in continuous time but is observed at discrete, irregularly spaced timepoints which may be unique to each individual. Methods for analyses of such data are typically based on the Markov assumption. Cook (Biometrics 1999; 55:915-920) introduced a conditional Markov model that accommodates the subject-to-subject variation in the model parameters with random effects. We extend this model by adding a non-ignorable dropout component to the model. Specification of the distribution of the random effects is made to guarantee a closed form expression of the marginal likelihood. This methodology is illustrated by applications to a data set from a parasitic field infection survey, a data set from a cocaine treatment study, and a data set from an aging study. Simulations suggest that the shared parameter model is robust with respect to at least one alternative non-ignorable model.

An assessment of non-randomized medical treatment of long-term schizophrenia relapse using bivariate binary-response transition models.

The analyses of observational longitudinal studies involving concurrent changes in treatment and medical conditions present difficulties because of the multitude of directions of potential relationships: past medication influences current symptoms; past symptoms influence current medication; and current medication is associated with current symptoms. In the context of a long-term study of non-randomized pharmacological treatment of schizophrenic relapse, we present an analysis of bivariate discrete-time transitional data with binary responses in an attempt to understand the transitional and concurrent relationships between schizophrenia relapse and medication use. A naive analysis does not show any association between previous medication and current relapse. However, we provide evidence suggesting that current treatment may impact current relapse for those who have previously taken medication, but not for those who haven't taken medication in the past. When univariate models are specified to assess these associations, the bivariate nature of the problem requires a choice of which response, relapse or medication, should be the dependent variable. In this case, the choice of relapse or medication as a dependent variable does matter. Hence, our results derive from models where both relapse and medication are treated as dependent variables. Specifically, we specify a bivariate log odds ratio for current relapse and current medication use and a separate univariate logit component for each of these outcomes. Each of these components contains transitional associations with previous relapse and medication. Such models represent extensions of univariate transitional association models (e.g. Diggle et al. (1994)) and correspond to bivariate transitional models (e.g. Zeger and Liang (1991)). We incorporate changes in transitional associations into the full-data parametric model for final inference, and investigate if these temporal changes are due to learning effects or the impact of drop-out. We also perform residual analyses and sensitivity analyses in the context of missing data patterns.