Publications (13)30.89 Total impact
-
Article: Mortality predictions on admission as a context for organizing care activities.
[show abstract] [hide abstract]
ABSTRACT: BACKGROUND: Favorable health outcomes are more likely to occur when the clinical team recognizes patients at risk and intervenes in consort. Prediction rules can identify high-risk subsets, but the availability of multiple rules for various conditions present implementation and assimilation challenges. METHODS: A prediction rule for 30-day mortality at the beginning of the hospitalization was derived in a retrospective cohort of adult inpatients from a community hospital in the Midwestern United States from 2008 to 2009, using clinical laboratory values, past medical history, and diagnoses present on admission. It was validated using 2010 data from the same and from a different hospital. The calculated mortality risk was then used to predict unplanned transfers to intensive care units, resuscitation attempts for cardiopulmonary arrests, a condition not present on admission (complications), intensive care unit utilization, palliative care status, in-hospital death, rehospitalizations within 30 days, and 180-day mortality. RESULTS: The predictions of 30-day mortality for the derivation and validation datasets had areas under the receiver operating characteristic curve of 0.88. The 30-day mortality risk was in turn a strong predictor for in-hospital death, palliative care status, 180-day mortality; a modest predictor for unplanned transfers and cardiopulmonary arrests; and a weaker predictor for the other events of interest. CONCLUSIONS: The probability of 30-day mortality provides health systems with an array of prognostic information that may provide a common reference point for organizing the clinical activities of the many health professionals involved in the care of the patient. Journal of Hospital Medicine 2012; © 2012 Society of Hospital Medicine.Journal of Hospital Medicine 12/2012; · 1.40 Impact Factor -
Article: Analyzing repeated measures semi-continuous data, with application to an alcohol dependence study.
[show abstract] [hide abstract]
ABSTRACT: Two-part random effects models (Olsen and Schafer,(1) Tooze et al.(2)) have been applied to repeated measures of semi-continuous data, characterized by a mixture of a substantial proportion of zero values and a skewed distribution of positive values. In the original formulation of this model, the natural logarithm of the positive values is assumed to follow a normal distribution with a constant variance parameter. In this article, we review and consider three extensions of this model, allowing the positive values to follow (a) a generalized gamma distribution, (b) a log-skew-normal distribution, and (c) a normal distribution after the Box-Cox transformation. We allow for the possibility of heteroscedasticity. Maximum likelihood estimation is shown to be conveniently implemented in SAS Proc NLMIXED. The performance of the methods is compared through applications to daily drinking records in a secondary data analysis from a randomized controlled trial of topiramate for alcohol dependence treatment. We find that all three models provide a significantly better fit than the log-normal model, and there exists strong evidence for heteroscedasticity. We also compare the three models by the likelihood ratio tests for non-nested hypotheses (Vuong(3)). The results suggest that the generalized gamma distribution provides the best fit, though no statistically significant differences are found in pairwise model comparisons.Statistical Methods in Medical Research 04/2012; · 2.44 Impact Factor -
Article: Duration of evidence-based medical therapy and the hazard for atherothrombotic events following percutaneous coronary intervention.
[show abstract] [hide abstract]
ABSTRACT: Patients who undergo coronary artery stent procedures are at risk for late atherothrombotic events, including stent thrombosis. The relationship between the duration during which evidence-based medical therapies are utilized after coronary artery stenting and the risk of late atherothrombotic events is not well characterized. In a retrospective cohort study linking a hospital-based percutaneous coronary intervention registry with a health maintenance organization claims dataset, we related the duration of medical therapy utilization during follow up to the hazard for death, myocardial infarction, unstable angina, transient ischemic attack or stroke following a coronary artery stent procedure. Multivariable Cox models were employed in which medical treatments were entered as time-varying covariates; data were stratified by stent type and time period. The median [interquartile range, IQR] duration of follow up was 832 [460, 1420] days. During this time, 86 ischemic events occurred in 84 of 386 patients at a median [IQR] of 260 [110, 658] days. The incidence of atherothrombotic events following coronary artery stenting was highest during the first post-procedure year and declined substantially thereafter. Multivariable predictors of incident ischemic events included multivessel coronary artery disease (HR 2.01 [95% CI 1.30-3.11], p=0.0018) and longer duration angiotensin converting enzyme (ACE) inhibitor/angiotensin receptor blocker (ARB), beta blocker or statin therapy (HR 0.52 [95% CI 0.28-0.99], p=0.045). The use of longer-term ACE inhibitor/ARB, beta blocker or statin therapy was associated with a significantly lower risk; these risk reductions were of greater magnitude than those associated with clopidogrel.International journal of cardiology 12/2011; 153(3):262-6. · 7.08 Impact Factor -
Article: Classical conditioning through auditory stimuli in Drosophila: methods and models.
[show abstract] [hide abstract]
ABSTRACT: The role of sound in Drosophila melanogaster courtship, along with its perception via the antennae, is well established, as is the ability of this fly to learn in classical conditioning protocols. Here, we demonstrate that a neutral acoustic stimulus paired with a sucrose reward can be used to condition the proboscis-extension reflex, part of normal feeding behavior. This appetitive conditioning produces results comparable to those obtained with chemical stimuli in aversive conditioning protocols. We applied a logistic model with general estimating equations to predict the dynamics of learning, which successfully predicts the outcome of training and provides a quantitative estimate of the rate of learning. Use of acoustic stimuli with appetitive conditioning provides both an alternative to models most commonly used in studies of learning and memory in Drosophila and a means of testing hearing in both sexes, independently of courtship responsiveness.Journal of Experimental Biology 09/2011; 214(Pt 17):2864-70. · 3.00 Impact Factor -
Article: MM Algorithms for Minimizing Nonsmoothly Penalized Objective Functions
[show abstract] [hide abstract]
ABSTRACT: In this paper, we propose a general class of algorithms for optimizing an extensive variety of nonsmoothly penalized objective functions that satisfy certain regularity conditions. The proposed framework utilizes the majorization-minimization (MM) algorithm as its core optimization engine. The resulting algorithms rely on iterated soft-thresholding, implemented componentwise, allowing for fast, stable updating that avoids the need for any high-dimensional matrix inversion. We establish a local convergence theory for this class of algorithms under weaker assumptions than previously considered in the statistical literature. We also demonstrate the exceptional effectiveness of new acceleration methods, originally proposed for the EM algorithm, in this class of problems. Simulation results and a microarray data example are provided to demonstrate the algorithm's capabilities and versatility.01/2010; -
Article: Conditional GEE for recurrent event gap times.
[show abstract] [hide abstract]
ABSTRACT: This paper deals with the analysis of recurrent event data subject to censored observation. Using a suitable adaptation of generalized estimating equations for longitudinal data, we propose a straightforward methodology for estimating the parameters indexing the conditional means and variances of the process interevent (i.e. gap) times. The proposed methodology permits the use of both time-fixed and time-varying covariates, as well as transformations of the gap times, creating a flexible and useful class of methods for analyzing gap-time data. Censoring is dealt with by imposing a parametric assumption on the censored gap times, and extensive simulation results demonstrate the relative robustness of parameter estimates even when this parametric assumption is incorrect. A suitable large-sample theory is developed. Finally, we use our methods to analyze data from a randomized trial of asthma prevention in young children.Biostatistics 04/2009; 10(3):451-67. · 2.14 Impact Factor -
Article: Predictors and outcomes of emergency department visits within 30 days following percutaneous coronary intervention.
[show abstract] [hide abstract]
ABSTRACT: Our objective was to determine the frequency and predictive factors for cardiac-related emergency department (ED) encounters within 30 days after percutaneous coronary intervention (PCI). The data source was an electronic database of 2,731 patients who had PCI from 2002 to 2004. Almost all underwent stent placement. Risk factors for returning to the ED were identified from clinical, anatomic, and demographic candidate variables using multivariate logistic regression. Approximately 9% of the cohort (255 of 2,731 patients) returned to the ED for cardiac reasons within 30 days, peaking around 3 days. ED visits were more likely in those whose index PCI was emergent or urgent (odds ratio [OR] 2.0, 95% confidence interval [CI] 1.3 to 3.0), in women (OR 1.9, 95% CI 1.5 to 2.5), and in those who had previous encounters with the ED or hospital (OR 1.7, 95% CI 1.5 to 2.0). Patients receiving stents were somewhat less likely to return (OR 0.7, 95% CI 0.5 to 1.0). In conclusion, the clinical courses of the 255 returning patients were generally benign, but 12% had a subsequent myocardial infarction or repeat PCI within 30 days of the ED encounter.The American Journal of Cardiology 02/2007; 99(2):197-201. · 3.37 Impact Factor -
Article: Bayesian Inference for a Two-Part Hierarchical Model: An Application to Profiling Providers in Managed Health Care
Journal of the American Statistical Association 02/2006; 101(September):934-945. · 1.99 Impact Factor -
Article: Polymorphisms in cytoplasmic serine hydroxymethyltransferase and methylenetetrahydrofolate reductase affect the risk of cardiovascular disease in men.
[show abstract] [hide abstract]
ABSTRACT: Genetic variation in folate-regulating enzymes contributes to the risk of cardiovascular disease (CVD). The cytoplasmic serine hydroxymethyltransferase (cSHMT) enzyme is proposed to regulate a key metabolic intersection in folate metabolism. We hypothesized that a variant in cSHMT (cSHMT 1420C-->T) affects CVD risk, and that the effect depends on a linked step in the metabolic pathway catalyzed by methylenetetrahydrofolate reductase (MTHFR). A nested case-control study of incident CVD was conducted within the all-male Normative Aging Study cohort. Of the incident CVD cases, 507 had DNA samples; 2 controls/case were selected by risk set sampling (matched on age and birth year). A significant gene-gene interaction (P-values 0.0013, 0.0064) was found between MTHFR and cSHMT, and there was little or no change in the coefficients in covariate-adjusted models. The effect of MTHFR 677C-->T genotype on CVD risk varied by cSHMT 1420C-->T genotype. Among men with cSHMT 1420C-->T TT genotype, the odds ratios (OR) for CVD risk for MTHFR 677C-->T CT and TT genotypes compared with the MTHFR 677C-->T CC genotype were 3.6 (95% CI, 1.7-7.8) and 10.6 (95% CI, 2.5-46.0), respectively. Among men with the cSHMT 1420C-->T CC/CT genotype, the corresponding ORs were 1.0 (95% CI, 0.8-1.2) and 1.3 (95% CI, 0.9-1.8). Plasma total homocysteine concentrations were highest in the subgroup of men with both polymorphisms, MTHFR 677C-->T TT and cSHMT 1420C-->T TT, consistent with a higher risk of CVD in this subgroup. A more complete understanding of the molecular mechanism awaits identification of the functional effect of the polymorphism.Journal of Nutrition 08/2005; 135(8):1989-94. · 3.92 Impact Factor -
Article: Quantifying the physician contribution to managed care pharmacy expenses: a random effects approach.
[show abstract] [hide abstract]
ABSTRACT: Despite the availability of more sophisticated techniques, few alternatives to ordinary least squares (OLS) regression have been utilized to profile physician prescribing in managed care. It is not known to what extent the modest R values derived from OLS models reflect incomplete risk adjustment or widely varying physician prescribing patterns. To quantify the role of interphysician variability relative to overall variability in managed care pharmacy expenses, and to examine the extent to which different statistical approaches generate meaningful differences in profile results. Comparison of three basic statistical modeling approaches: OLS, fixed effects regression, and random effects (ie, hierarchical) regression models. Two managed care populations that differed more than 2-fold in per member pharmacy expenditures in 1999, one from the Midwestern United States, the other from three Western States. The intraclass correlation coefficient (ICC, the proportion of variability in expenses attributable to differences among physicians) and the range of projected expenses attributed to each physician's prescribing style. The ICCs were small for aggregated pharmacy expenditures, 0.04 or less in both populations. As determined by OLS, the most costly physician contributed 94,399 U.S. dollars in excess expenses to the organization whereas the most parsimonious saved 89,940 U.S. dollars. When derived from random effects models, the range in performance was 63% of that derived from OLS. In the populations studied, systematic prescribing differences among physicians were small relative to the overall variability in pharmacy expenses, suggesting other factors were more likely driving these costs. Random effects models generated smaller estimates of the individual physicians' contribution to costs, sometimes considerably, relative to those derived from OLS and fixed effects approaches.Medical Care 09/2002; 40(8):650-61. · 3.41 Impact Factor -
Article: Flexible hazard regression modeling for medical cost data.
[show abstract] [hide abstract]
ABSTRACT: The modeling of lifetime (i.e. cumulative) medical cost data in the presence of censored follow-up is complicated by induced informative censoring, rendering standard survival analysis tools invalid. With few exceptions, recently proposed nonparametric estimators for such data do not extend easily to handle covariate information. We propose to model the hazard function for lifetime cost endpoints using an adaptation of the HARE methodology (Kooperberg, Stone, and Truong, Journal of the American Statistical Association, 1995, 90, 78-94). Linear splines and their tensor products are used to adaptively build a model that incorporates covariates and covariate-by-cost interactions without restrictive parametric assumptions. The informative censoring problem is handled using inverse probability of censoring weighted estimating equations. The proposed method is illustrated using simulation and also with data on the cost of dialysis for patients with end-stage renal disease.Biostatistics 04/2002; 3(1):101-18. · 2.14 Impact Factor -
Article: Majorization-Minimization algorithms for nonsmoothly penalized objective functions
[show abstract] [hide abstract]
ABSTRACT: The use of penalization, or regularization, has become common in high-dimensional statistical analysis, where an increasingly frequent goal is to simultaneously select important variables and estimate their effects. It has been shown by several authors that these goals can be achieved by minimizing some parameter-dependent “goodness-of-fit” function (e.g., a negative loglikelihood) subject to a penalization that promotes sparsity. Penalty functions that are singular at the origin have received substantial attention, arguably beginning with the Lasso penalty [62]. ¶ The current literature tends to focus on specific combinations of differentiable goodness-of-fit functions and penalty functions singular at the origin. One result of this combined specificity has been a proliferation in the number of computational algorithms designed to solve fairly narrow classes of optimization problems involving objective functions that are not everywhere continuously differentiable. In this paper, we propose a general class of algorithms for optimizing an extensive variety of nonsmoothly penalized objective functions that satisfy certain regularity conditions. The proposed framework utilizes the majorization-minimization (MM) algorithm as its core optimization engine. In the case of penalized regression models, the resulting algorithms employ iterated soft-thresholding, implemented componentwise, allowing for fast and stable updating that avoids the need for inverting high-dimensional matrices. We establish convergence theory under weaker assumptions than previously considered in the statistical literature. We also demonstrate the exceptional effectiveness of new acceleration methods, originally proposed for the EM algorithm, in this class of problems. Simulation results and a microarray data example are provided to demonstrate the algorithm’s capabilities and versatility. -
Article: A flexible two-part random effects model for correlated medical costs
[show abstract] [hide abstract]
ABSTRACT: In this paper, we propose a flexible “two-part” random effects model ( [35] and [40]) for correlated medical cost data. Typically, medical cost data are right-skewed, involve a substantial proportion of zero values, and may exhibit heteroscedasticity. In many cases, such data are also obtained in hierarchical form, e.g., on patients served by the same physician. The proposed model specification therefore consists of two generalized linear mixed models (GLMM), linked together by correlated random effects. Respectively, and conditionally on the random effects and covariates, we model the odds of cost being positive (Part I) using a GLMM with a logistic link and the mean cost (Part II) given that costs were actually incurred using a generalized gamma regression model with random effects and a scale parameter that is allowed to depend on covariates (cf., Manning et al., 2005). The class of generalized gamma distributions is very flexible and includes the lognormal, gamma, inverse gamma and Weibull distributions as special cases. We demonstrate how to carry out estimation using the Gaussian quadrature techniques conveniently implemented in SAS Proc NLMIXED. The proposed model is used to analyze pharmacy cost data on 56,245 adult patients clustered within 239 physicians in a mid-western U.S. managed care organization.Journal of Health Economics.
Top co-authors
Top Journals
- The American Journal of Cardiology (1)
- Journal of Nutrition (1)
- Medical Care (1)
- Biostatistics (1)
- Biostatistics (1)
Institutions
-
2009
-
Cornell University
- Department of Statistical Science
Ithaca, NY, USA
-
-
2002
-
Saint Joseph Hospital
Chicago, IL, USA -
University of Michigan
- Department of Biostatistics
Ann Arbor, MI, USA
-
