-
[show abstract]
[hide abstract]
ABSTRACT: We explore a Bayesian approach to selection of variables that represent fixed and random effects in modeling of longitudinal binary outcomes with missing data caused by dropouts. We show via analytic results for a simple example that nonignorable missing data lead to biased parameter estimates. This bias results in selection of wrong effects asymptotically, which we can confirm via simulations for more complex settings. By jointly modeling the longitudinal binary data with the dropout process that possibly leads to nonignorable missing data, we are able to correct the bias in estimation and selection. Mixture priors with a point mass at zero are used to facilitate variable selection. We illustrate the proposed approach using a clinical trial for acute ischemic stroke.
Biometrical Journal 11/2012; · 1.25 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: A common problem in the longitudinal data analysis is the missing data problem. Two types of missing patterns are generally considered in statistical literature: monotone and non-monotone missing data. Nonmonotone missing data occur when study participants intermittently miss scheduled visits, while monotone missing data can be from discontinued participation, loss to follow-up, and mortality. Although many novel statistical approaches have been developed to handle missing data in recent years, few methods are available to provide inferences to handle both types of missing data simultaneously. In this article, a latent random effects model is proposed to analyze longitudinal outcomes with both monotone and non-monotone missingness in the context of missing not at random. Another significant contribution of this article is to propose a new computational algorithm for latent random effects models. To reduce the computational burden of high-dimensional integration problem in latent random effects models, we develop a new computational algorithm that uses a new adaptive quadrature approach in conjunction with the Taylor series approximation for the likelihood function to simplify the E-step computation in the expectation-maximization algorithm. Simulation study is performed and the data from the scleroderma lung study are used to demonstrate the effectiveness of this method.
Statistical Methods in Medical Research 05/2012; · 2.44 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: We propose a semiparametric joint model for bivariate longitudinal ordinal outcomes and competing risks failure time data. The association between the longitudinal and survival endpoints is captured by latent random effects. This approach generalizes previous joint analysis that considers only one response variable at the longitudinal endpoint. One unique feature of the proposed model is that we relax the commonly used normality assumption for random effects and leave the distribution completely unspecified. We use a modified version of the vertex exchange method in conjunction with an expectation-maximization algorithm to estimate the random effects distribution and model parameters. We show via simulations that robust parameter estimates are obtained from the proposed method under various scenarios. We illustrate the approach using cough severity and frequency data from a scleroderma lung study.
Statistics in Medicine 02/2012; 31(16):1707-21. · 1.88 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: Abstract BACKGROUND:Cough is a significant symptom in patients with scleroderma interstitial lung disease (SSc-ILD), affecting 73 % of the 158 patients enrolled in the Scleroderma Lung Study (SLS), a multicenter randomized trial of oral cyclophosphamide (CYC) vs. placebo (PLA) in patients with active ILD. METHODS:We examined the correlation of cough frequency, severity and phlegm production at baseline in 156 SLS participants with other baseline variables representing SSc-ILD disease activity and the cough response to one year of treatment with CYC vs. placebo (PLA). RESULTS:Patients with cough at baseline had significantly lower DLCO, dyspnea, the quality of life physical component summary, and the maximal fibrosis (MAXFIB) score on HRCT compared with non-coughers at baseline. Cough severity and frequency correlated with % FVC predicted. After 12 months of treatment, cough frequency decreased in the CYC group compared with PLA, and was significantly different from PLA at 18 months (6 months after discontinuation of CYC). However, the decreases in cough frequency did not correlate with the changes in FVC or DLCO observed in the CYC group. Treatment-related improvements in cough frequency, as well as in FVC, were no longer apparent 12 months after discontinuation of CYC. CONCLUSIONS:Cough is a common symptom in SSc-ILD and correlates with the extent of fibrosis. Cough frequency decreases significantly in response to treatment with CYC but returns to baseline 1 year after withdrawal of treatment. Cough may be a symptom of ongoing fibrosis and an independent variable in assessing therapeutic response to CYC.
Chest 12/2011; · 5.25 Impact Factor
-
Michael D Roth,
Chi-Hong Tseng,
Philip J Clements,
Daniel E Furst,
Donald P Tashkin,
Jonathan G Goldin,
Dinesh Khanna,
Eric C Kleerup, Ning Li,
David Elashoff,
Robert M Elashoff
[show abstract]
[hide abstract]
ABSTRACT: To identify baseline characteristics of patients with scleroderma-related interstitial lung disease (SSc-ILD) that could serve as predictors of the most favorable response to 12-month treatment with oral cyclophosphamide (CYC).
Regression analyses were retrospectively applied to the Scleroderma Lung Study data in order to identify baseline characteristics that correlated with the absolute change in forced vital capacity (FVC) (% predicted values) and the placebo-adjusted change in % predicted FVC over time (the CYC treatment effect).
Completion of the CYC arm of the Scleroderma Lung Study was associated with a placebo-adjusted improvement in the % predicted FVC of 2.11% at 12 months, which increased to 4.16% when patients were followed up for another 6 months (P=0.014). Multivariate regression analyses identified the maximal severity of reticular infiltrates (assessed as maximum fibrosis scores) on high-resolution computed tomography (HRCT) at baseline, the modified Rodnan skin thickness score (MRSS) at baseline, and the Mahler baseline dyspnea index as independent correlates of treatment response. When patients were stratified on the basis of whether 50% or more of any lung zone was involved by reticular infiltrates on HRCT and/or whether patients exhibited an MRSS of at least 23, a subgroup of patients emerged in whom there was an average CYC treatment effect of 9.81% at 18 months (P<0.001). Conversely, there was no treatment effect (a -0.58% difference) in patients with less severe HRCT findings and a lower MRSS at baseline.
A retrospective analysis of the Scleroderma Lung Study data identified the severity of reticular infiltrates on baseline HRCT and the baseline MRSS as patient features that might be predictive of responsiveness to CYC therapy.
Arthritis & Rheumatism 05/2011; 63(9):2797-808. · 7.87 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: The Scleroderma Lung Study (SLS) was a 1-year, randomized, controlled trial of oral cyclophosphamide for scleroderma-related pulmonary alveolitis. It concluded that oral cyclophosphamide slowed the decline in the forced vital capacity (% predicted) and had a beneficial effect on dyspnea, skin changes, and several quality of life measures of systemic sclerosis. We now report an in-depth assessment of the toxicity of cyclophosphamide during the year of therapy and the year after therapy was completed, during which time the investigators were still masked to the treatment assignment.
One-year, double-blind, randomized controlled trial of oral cyclophosphamide versus placebo with 1-year masked follow-up. Adverse events (AEs) were tabulated, described, and compared using descriptive statistics (eg, mean and median) and t, Wilcoxon rank sum, chi-squared, or Fisher's exact tests as appropriate.
During year 1, treatment-related overall AEs occurred more frequently in cyclophosphamide (CYC)-treated patients (overall AEs for CYC=154 events vs placebo=60 events; P=0.002), and especially for mild to moderate leukopenia (CYC=19 subjects vs placebo=0 subjects; P < .0001). For cancer, we followed patients beyond 2 years. There were no differences in the occurrence of cancer (CYC=4 subjects vs placebo=2 subjects), serious related AEs (CYC=8 events vs placebo=13 events), or deaths (CYC=6 subjects vs placebo=6 subjects).
Over 2 years, cyclophosphamide was associated with more AEs than placebo, including overall AEs and relative leukopenia. There were no differences in other AEs, including serious AEs, cancers, or deaths.
The American journal of medicine 05/2011; 124(5):459-67. · 4.47 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: To assess whether serum concentrations of surfactant protein D (SP-D) and Krebs von den Lungen-6 (KL-6), glycoproteins expressed by type II pneumocytes, correlate with the presence of "alveolitis" and measures of lung function in patients enrolled in the Scleroderma Lung Study (SLS).
Serum obtained at baseline screening of patients with systemic sclerosis (SSc, scleroderma) in the SLS was assayed. "Alveolitis" was defined by either bronchoalveolar lavage or thoracic high-resolution computed tomography (HRCT) by SLS criteria. SP-D and KL-6 levels were measured by ELISA in 66 SSc patients (44 with "alveolitis," 22 without "alveolitis") and in 10 healthy controls. These were compared to clinical measures of lung disease and "alveolitis" in the SLS patients.
SP-D levels were 300+/-214 ng/ml (mean+/-SD) in the SSc patients compared to 40+/-51 ng/ml in controls (p<0.0001). KL-6 levels were 1225+/-984 U/ml in the SSc patients and 333+/-294 U/ml in controls (p<0.0001). SSc patients with "alveolitis" had higher levels of both SP-D and KL-6 than those without "alveolitis." The level of SP-D was 353+/-219 ng/ml in patients with "alveolitis" and 161+/-143 ng/ml without "alveolitis" (p=0.0002). The level of KL-6 was 1458+/-1070 U/ml in patients with "alveolitis" and 640+/-487 U/ml without "alveolitis" (p=0.0001). Receiver operator characteristic curve analysis demonstrated high sensitivity and specificity of both SP-D and KL-6 for the determination of "alveolitis." KL-6 and SP-D were positively correlated with maximum fibrosis scores, but not with maximum ground-glass opacities, on HRCT.
Serum levels of SP-D and KL-6 appear to be indicative of "alveolitis" in SSc patients as defined by the SLS, and are significantly higher than in SSc patients without "alveolitis." Serum SP-D and KL-6 may serve as noninvasive serological means of assessing interstitial lung disease in patients with SSc.
The Journal of Rheumatology 05/2009; 36(4):773-80. · 3.69 Impact Factor
-
Donald P Tashkin,
Robert Elashoff,
Philip J Clements,
Michael D Roth,
Daniel E Furst,
Richard M Silver,
Jonathan Goldin,
Edgar Arriola,
Charlie Strange,
Marcy B Bolster, [......],
Kamal Mubarak,
M Kari Connolly,
Jeffrey Golden,
Mitchell Olman,
Barri Fessler,
Naomi Rothfield,
Mark Metersky,
Dinesh Khanna, Ning Li,
Gang Li
[show abstract]
[hide abstract]
ABSTRACT: The Scleroderma Lung Study enrolled 158 patients with scleroderma-related interstitial lung disease in a placebo-controlled trial of oral cyclophosphamide (CYC). Although treatment-related benefits in pulmonary function, skin scores, and patient-centered outcomes were demonstrated after 1 year of therapy, the duration of benefit beyond 1 year was unclear.
A second year of follow-up was performed to determine if these effects persisted after stopping treatment.
A detailed analysis of data obtained over the two years of the study was performed.
Using a longitudinal joint model, we analyzed FVC, total lung capacity, transitional dyspnea index, Rodnan skin scores, and the Health Assessment Questionnaire-Disability Index during the second year, after adjusting for baseline values, baseline fibrosis score, and nonignorable missing data. Evaluable subjects (72 CYC; 73 placebo) included 93 who completed all visits plus 52 who completed at least 6 months of therapy and returned at 24 month or had their 24-month data imputed. The beneficial effects of CYC on pulmonary function and health status continued to increase through 18 months, after which they dissipated, whereas skin improvements dissipated after 12 months. In contrast, the positive effect on dyspnea persisted through 24 months. Adverse events were uncommon.
One year of CYC improved lung function, skin scores, dyspnea, and health status/disability, effects which either persisted or increased further for several months after stopping therapy. However, except for a sustained impact on dyspnea, all of these effects waned and were no longer apparent at 24 months. Treatment strategies aimed at extending the positive therapeutic effects observed with CYC should be considered. Clinical trial registered with www.clinicaltrials.gov (NCT 000004563).
American Journal of Respiratory and Critical Care Medicine 12/2007; 176(10):1026-34. · 11.08 Impact Factor
