Biomedical subjects

Susan A Murphy

Publications and source records attributed to Susan A Murphy.

8 recordsLinked to original sources

PubMed2006

Methodological challenges in constructing effective treatment sequences for chronic psychiatric disorders.

Psychiatric disorders are often chronic conditions that require sequential decision making to achieve the best clinical outcomes. Sequential decisions are necessary to accommodate treatment response heterogeneity, a variable course of illness, and the often heavy burden associated with intensive or longer-term treatment. Yet, only a few studies in this field have been designed to address sequential decisions. Most of the experimental designs and data analytic methods that are best suited for improving sequential clinical decision making are often found in nonmedical fields such as engineering, computer science, and statistics. Promising designs and methods are surveyed with a focus on those areas most immediately useful for informing clinical decision making.

Chronic Disease↗

PubMed2006

Developing adaptive treatment strategies in substance abuse research.

For many individuals, substance abuse possesses characteristics of chronic disorders in that individuals experience repeated cycles of cessation and relapse; hence viewing drug dependence as a chronic, relapsing disorder is increasingly accepted. The development of a treatment for a chronic disorder requires consideration of the ordering of treatments, the timing of changes in treatment, and the use of measures of response, burden and adherence collected during treatment to make further treatment decisions. Adaptive treatment strategies provide a vehicle through which these issues can be addressed and thus provide a means toward improving and informing the clinical management of chronic substance abuse disorders. The sequential multiple assignment randomized trial (SMART) is particularly useful in developing adaptive treatment strategies. Simple analyses that can be used with the SMART design are described. Furthermore, the SMART design is compared with standard experimental designs.

Algorithms↗

PubMed2006

Assessing the total effect of time-varying predictors in prevention research.

Observational data are often used to address prevention questions such as, "If alcohol initiation could be delayed, would that in turn cause a delay in marijuana initiation?" This question is concerned with the total causal effect of the timing of alcohol initiation on the timing of marijuana initiation. Unfortunately, when observational data are used to address a question such as the above, alternative explanations for the observed relationship between the predictor, here timing of alcohol initiation, and the response abound. These alternative explanations are due to the presence of confounders. Adjusting for confounders when using observational data is a particularly challenging problem when the predictor and confounders are time-varying. When time-varying confounders are present, the standard method of adjusting for confounders may fail to reduce bias and indeed can increase bias. In this paper, an intuitive and accessible graphical approach is used to illustrate how the standard method of controlling for confounders may result in biased total causal effect estimates. The graphical approach also provides an intuitive justification for an alternate method proposed by James Robins [Robins, J. M. (1998). 1997 Proceedings of the American Statistical Association, section on Bayesian statistical science (pp. 1-10). Retrieved from http://www.biostat.harvard.edu/robins/research.html; Robins, J. M., Hernán, M., & Brumback, B. (2000). Epidemiology, 11(5), 550-560]. The above two methods are illustrated by addressing the motivating question. Implications for prevention researchers who wish to estimate total causal effects using longitudinal observational data are discussed.

Adolescent↗

PubMed2006

Examining clinical judgment in an adaptive intervention design: The fast track program.

Although clinical judgment is often used in assessment and treatment planning, rarely has research examined its reliability, validity, or impact in practice settings. This study tailored the frequency of home visits in a prevention program for aggressive- disruptive children (n = 410; 56% minority) on the basis of 2 kinds of clinical judgment: ratings of parental functioning using a standardized multi-item scale and global assessments of family need for services. Stronger reliability and better concurrent and predictive validity emerged for the 1st kind of clinical judgment than for the 2nd. Exploratory analyses suggested that using ratings of parental functioning to tailor treatment recommendations improved the impact of the intervention by the end of 3rd grade but using more global assessments of family need did not.

Aggression↗

PubMed2005

A strategy for optimizing and evaluating behavioral interventions.

BACKGROUND: Although the optimization of behavioral interventions offers the potential of both public health and research benefits, currently there is no widely agreed-upon principled procedure for accomplishing this. PURPOSE: This article suggests a multiphase optimization strategy (MOST) for achieving the dual goals of program optimization and program evaluation in the behavioral intervention field. METHODS: MOST consists of the following three phases: (a) screening, in which randomized experimentation closely guided by theory is used to assess an array of program and/or delivery components and select the components that merit further investigation; (b) refining, in which interactions among the identified set of components and their interrelationships with covariates are investigated in detail, again via randomized experiments, and optimal dosage levels and combinations of components are identified; and (c) confirming, in which the resulting optimized intervention is evaluated by means of a standard randomized intervention trial. To make the best use of available resources, MOST relies on design and analysis tools that help maximize efficiency, such as fractional factorials. RESULTS: A slightly modified version of an actual application of MOST to develop a smoking cessation intervention is used to develop and present the ideas. CONCLUSIONS: MOST has the potential to husband program development resources while increasing our understanding of the individual program and delivery components that make up interventions. Considerations, challenges, open questions, and other potential benefits are discussed.

Behavior Therapy↗

PubMed2005

A Generalization Error for Q-Learning.

Planning problems that involve learning a policy from a single training set of finite horizon trajectories arise in both social science and medical fields. We consider Q-learning with function approximation for this setting and derive an upper bound on the generalization error. This upper bound is in terms of quantities minimized by a Q-learning algorithm, the complexity of the approximation space and an approximation term due to the mismatch between Q-learning and the goal of learning a policy that maximizes the value function.

Journal Article↗

PubMed2004

A conceptual framework for adaptive preventive interventions.

Recently, adaptive interventions have emerged as a new perspective on prevention and treatment. Adaptive interventions resemble clinical practice in that different dosages of certain prevention or treatment components are assigned to different individuals, and/or within individuals across time, with dosage varying in response to the intervention needs of individuals. To determine intervention need and thus assign dosage, adaptive interventions use prespecified decision rules based on each participant's values on key characteristics, called tailoring variables. In this paper, we offer a conceptual framework for adaptive interventions, discuss principles underlying the design and evaluation of such interventions, and review some areas where additional research is needed.

Child↗

PubMed2002

Two-level proportional hazards models.

We extend the proportional hazards model to a two-level model with a random intercept term and random coefficients. The parameters in the multilevel model are estimated by a combination of EM and Newton-Raphson algorithms. Even for samples of 50 groups, this method produces estimators of the fixed effects coefficients that are approximately unbiased and normally distributed. Two different methods, observed information and profile likelihood information, will be used to estimate the standard errors. This work is motivated by the goal of understanding the determinants of contraceptive use among Nepalese women in the Chitwan Valley Family Study (Axinn, Barber, and Ghimire, 1997). We utilize a two-level hazard model to examine how education and access to education for children covary with the initiation of permanent contraceptive use.

Adolescent↗