Please activate JavaScript!
Please install Adobe Flash Player, click here for download

ISCB2014_abstract_book

54 ISCB 2014 Vienna, Austria • Abstracts - Oral PresentationsTuesday, 26th August 2014 • 11:00-12:30 Monday25thAugustTuesday26thAugustThursday28thAugustAuthorIndexPostersWednesday27thAugustSunday24thAugust The effect measure for this model is based on the mean number of events in the study period for a given treatment regime. This number is obtained by integration over hazards of hospitalization in discretized risk-periods. It allows a flexible modelling with multiple hazard-ratio parameters com- pared to the proportional hazard assumption in the standard marginal structural Cox model. Weights based on continuously updated vaccination probabilities adjust for the time-depending confounding. The weights are estimated using poisson regression models. We illustrate our approach in a multistate setup with 5 different vaccina- tion states allowing causal comparison, on a day to day basis, between a variety of vaccination schedules for children aged 15-24 months.   C28.5 Robustness and efficiency in instrumental variable models with covariates V Didelez1 , S Vansteelandt2 1 Maths, University of Bristol, Bristol, United Kingdom, 2 Ghent University, Ghent, Belgium   Instrumental variables provide an approach for consistent inference on causal effects even in the presence of unmeasured confounding. Such methods have for instance been used in the context of Mendelian ran- domisation, as well as in pharmaco-epidemiological contexts. In these and other applications, it is common that covariates are available, even if deemed insufficient to adjust for all confounding. As IVs allow inference when there is unobserved confounding, it appears that often the analyst assumes that even observed confounders / covariates do not need to or should not be taken into account. However, this is not generally the case. With view to the role of covariates, we here contrast two-stage least squares estimators, generalized methods of moment estimators and variants thereof with methods more common in biostatistics using G-estimation in so-called structural mean and distri- bution models. When using covariates, there are structural aspects to be considered, e.g. whether the covariates are prior to or potentially affected by the instruments. But in addition, one has to worry even more about efficiency versus model misspecification when modelling covariates. We discuss this for the IV procedures mentioned above, especially for linear and loglinear instrumental variable models. Our results motivate adaptive procedures that guarantee efficiency im- provements through covariate adjustment, without the need for covariate selection strategies. Besides theoretical findings, simulation results will be shown to provide numerical insight.   C29 Clinical trial designs C29.1 Patient-oriented randomization - a new clinical design C Schulz1 , J Timm1 1 University of Bremen, Bremen, Germany   The “gold standard” for clinical studies is a controlled, randomized and double-blinded trial usually comparing specific treatments. However, this procedure is far away from the physician´s daily routine. From an ethical point of view, the physician should be involved in the decision concerning treatment, taking risks and healing opportunities of each patient into ac- count. Therefore, the question arises: is there a way to combine random- ization and patient-oriented decisions in a clinical trial design? The answer is yes if strategies instead of specific treatments are compared. The idea is to randomize the strategies and let the physician decide be- tween treatments within these strategies. An example is the clinical trial NeSSy [1] with a randomized design com- paring efficacy and safety of the strategies to use either conventional or newer antipsychiotic drugs in patients suffering from schizophrenia. The new idea of randomization is generalized to the case of two different strategies with an arbitrary number of treatments within each strategy. Preliminary results will be displayed showing the behavior of this inno- vative design with respect to balance between strategies and between single treatments within each strategy. Main results cover the influence of treatment number in each strategy, number of patients and centres and the decision rules of the physicians. Results have been generated by theo- retical consideration as well as simulation studies. Furthermore, results re- garding the study design of the NeSSy study will be presented. Reference: [1] The Neuroleptic Strategy Study - NeSSy, funded under the program of clinical studies of the BMBF   C29.2 A novel modified standard-gamble task to measure patients’preferences for biomarker-led care J Harrington1,2 , M Morgan2 , A Cronin3,4 , R Hilton3 , S Sacks4 , M Hernandez-Fuentes1,4,5 , I Rebollo-Mesa4 1 NIHR Biomedical Research Centre, King’s College London, London, United Kingdom, 2 D.Health and Social Care Research, King’s College London, London, United Kingdom, 3 Guy’s and St Thomas’ NHS Foundation Trust, London, United Kingdom, 4 MRC Centre for Transplantation, King’s College London, London, United Kingdom, 5 GAMBIT Consortium, www.GAMBITstudy.co.uk, London, United Kingdom   Background: The success of personalized medicine is directly dependent upon the development of accurate clinical prediction models. However, predictive markers are rarely 100% accurate. Varying levels of accuracy are associated with different levels of risk in the subsequent clinical decision. Incorporating patients’ preferences into the development and validation of clinical prediction models is vital for a seamless translation into clinical practice. Aims: 1) To develop a novel modified Standard-Gamble task (MSG) to measure patients’ preferences for biomarker-led care (BLC), as a function of the marker’s sensitivity and specificity; 2) To use the MSG to obtain the Minimum Tolerated Specificity (MTSp) and Minimum Tolerated Sensitivity (MTSe) in a study of biomarkers of toler- ance in kidney transplant recipients (KTR); and 3) To study the association between MTSp, MTSe and patients’symptom burden. Methods: Participants: 87 KTR with stable function, and 13 operationally tolerant. Measures: 1) MSG task, 2) symptom burden questionnaire and 3) qualitative interview. Results: Preliminary results (N=57) show that KTR require a highly spe- cific test to accept BLC (Median=0.91). Sensitivity is of less concern (Median=0.67). No significant relationship was found between attitude towards risk and symptom burden, although immunosupression related symptoms tend to be higher among high-risk takers. Conclusion: Our findings show that the MSG can be used to measure patients’ treatment choice as a function of a (bio)marker’s specificity and sensitivity, and the risks associated with it. Further research is necessary to understand the factors associated with patients’choice, and incorporating them in the process of (bio)marker validation.

Pages Overview