PhD, University of Cincinnati, 2019, Medicine: Biostatistics (Environmental Health)

There is growing interest in ambulatory blood pressure monitoring (ABPM) in clinical studies, which creates intensive amounts of longitudinal data from individual patients. This allows close study of the clinical course over different monitoring occasions, which results in nested repeated measures (NRM) data. These real-life data from the clinical setting can be influenced by several factors including subject withdrawal from a study. Informative drop-out is a common challenge in longitudinal studies. Failure to adjust a model for such a drop-out mechanism can lead to biased parameter estimates. Joint models for longitudinal data to account for informative drop-out have improved analysis by enabling estimation of direct and indirect effects of treatment, thereby resulting in improved efficacy assessment. This research proposes a novel joint model incorporating NRM with informative dropout. We present a shared parameter modeling approach that links NRM with potentially informative drop-out by a set of random effects that summarize each subject's specific deviation in terms of their intercepts, slopes, and measurement occasions. The longitudinal and survival (i.e., drop-out) processes are conditionally independent given the random effects. We use a Gaussian linear mixed-effects model for the NRM. In order to fit nonlinearity of ABPM trajectories, orthonormal cubic splines were used in the Gaussian linear mixed-effects model. For the time-to-drop-out process, we incorporate Weibull proportional hazards. Parameter estimates are obtained by a fully Bayesian approach using Markov Chain Monte Carlo (MCMC) models. The performance of the proposed model is also evaluated via simulation studies. The joint model is compared to a standard na¨ive two-stage approach, which incorporates estimates of the fitted values from the longitudinal model as covariates in the survival model. Misspecification of covariates for the linear mixed-effects and time-to-dropout models (open full item for complete abstract)

Committee: Roman Jandarov Ph.D. (Committee Chair); Raouf Amin Ph.D. (Committee Member); Jane Khoury Ph.D. (Committee Member); Rhonda Szczesniak Ph.D. (Committee Member) Subjects: Biostatistics

MS, University of Cincinnati, 2012, Medicine: Clinical and Translational Research

Background: In normotensive subjects, an exaggerated blood pressure response to exercise is associated with the development of resting hypertension. We sought to 1) determine the prevalence of elevated blood pressures during exercise in post-operative coarctation patients with normal resting blood pressure and 2) investigate associations with exercise induced hypertension in this population. Method: 38 subjects status post end to end anastomosis repair and resting normotension were prospectively enrolled. All patients underwent anthropometric and blood pressure measurements, echocardiographic evaluation of function, arterial stiffness assessment by pulse wave velocity and a graded exercise test. An abnormal response was defined as a maximum systolic blood pressure greater than the 95th percentile of published normal values. Correlation analyses and stepwise regression analyses were performed. Results: Mean age was 12.7 years, 79% male. Mean resting systolic blood pressure was 111.3 mmHg, mean exercise systolic blood pressure was 178.1 mmHg. The prevalence of a systolic blood pressure greater than the 95th percentile was 16.7%. In multivariate analysis, exercise systolic blood pressure index was associated with body mass index, age, aortic valve annulus, shortening fraction and pulse wave velocity (R2 = 0.79, p = 0.0009). Estimates of ventricular filling and indexed left ventricular mass were elevated. Conclusions: There is a risk of elevated systolic blood pressure during exercise in normotensive patients after coarctation repair. Resting blood pressures are useful but not sufficient. Echocardiography demonstrated abnormalities suggestive of a chronic cardiac burden despite resting normotension. Regular imaging may be necessary to improve long term outcomes. New paradigms for the continued follow-up of these patients are necessary.

Committee: Erin Nicole Haynes PhD (Committee Chair); Phillip Khoury MS (Committee Member); Thomas Kimball MD (Committee Member) Subjects: Surgery