Latent Class Models for Longitudinal Outcomes Heading
My statistical methodological work has been in the area of identification of subgroups using latent class models for longitudinal data. These methods are useful when multiple longitudinal outcomes of possibly varying types are needed to adequately characterize the effect of a treatment and it is of interest to identify the subset of patients for whom the treatment is effective.
Cognition, Depression, & Function in Older Adults
Mediational analysis delineating mechanism by which Behavior Activation (BA) þ low vision rehabilitation visits (LVR) prevent depression. BADS= Behavioral Activation for Depression Scale Social Impairment subscale; MOS =Medical Outcomes Study-6; PHQ-9=Patient Health Questionnaire-9.
A major focus of my collaborative work has been in the general area of geriatric psychiatry. Working with Dr. Barry Rovner and Dr. Robin Casten, I have served as the study statistician on projects evaluating Behavioral Activation for preventing depression in older adults with low vision, for increasing adherence to recommended vision care among diabetics, and for preventing cognitive decline in African Americans with mild cognitive impairment. In addition, I have collaborated with Drs. Rovner and Casten in exploring the relationship between activity loss and cognitive decline as well as the interplay of cognitive and noncognitive characteristics in contributing to variability in daily activity and cognitive decline. This work draws on my expertise in latent variable and structural equation modeling.
Translational Research in Prostate Cancer
Elevated nuclear Stat5a/b expression predicts early lethal disease recurrence of prostate cancer after radical prostatectomy.
As director of the Sidney Kimmel Cancer Center Biostatistics Shared Resource, I lead statistical support of basic, translational, and clinical research in cancer. A particular focus of my cancer-related work has been in the area of prostate cancer. I have collaborated with basic scientists and clinicians and am co-director of the Biostatistics core for the proposed Philadelphia Prostate SPORE led by Dr. Karen Knudsen.
Measurement of Function and Predictors of Change in Patients with Spinal Cord Injury
I have collaborated with Dr. Ralph Marino on questions surrounding measurement of upper-limb function and observational studies of change in function over time. The question of how to measure change in function in patients with spinal cord injury is still unresolved. Dr. Marino has developed a task-based test called the capabilities of upper extremity test (CUE-T) to attempt to quantify upper-limb function in these patients. We have demonstrated reliability and validity of this instrument and are currently assessing its sensitivity to change. We have also worked together to analyze data from large databases of spinal cord injury patients to evaluate the natural course of progression under current treatment methods and identify potential predictors of recovery.
Increasing Adherence to Eye Care in Patients with Diabetes and Glaucoma
This is a collaboration with Drs. Julia Haller, Ann Murchison, and Lisa Hark at Wills Eye Hospital and Dr. Laura Pizzi at Jefferson. We have studied factors associated with adherence to recommended eye care and tested interventions for improving adherence in patients with diabetes. Recent work has focused on improving access and follow-up to care for patients with Glaucoma.