I am an epidemiologist and junior faculty member in the Division of Pharmacoepidemiology and Pharmacoeconomics at Brigham & Women's Hospital, and Instructor in Medicine at Harvard Medical School. My long term goal is to establish a career as an independent investigator in psychiatric pharmacoepidemiology. As new mental health products move from the highly-controlled pre-marketing setting into clinical practice, comparative safety and effectiveness information is required to inform optimal prescribing and policy decisions. Phase 3 trials usually do not provide sufficient evidence to inform these decisions due to their relatively narrow focus on efficacy defined as statistical superiority over placebo. Observational studies using electronic health- care data play a critical role in generating comparative information in routine care, but are typically conducted long after a drug has entered the market, and therefore can fail to identify serious safety issues in the early marketing experience. Prospective monitoring of medical products starting at the time of market entry offers a promising complementary approach to detect concerns as early as possible. Development of methods for use in a prospective monitoring framework for mental health applications is largely uncharted territory. The goal of the proposed research is to develop improved methods for understanding the comparative safety and effectiveness of new psychiatric medications, and is aligned with the priorities of Strategic Objective 3 in the NIMH Strategic Plan. In addition to pursuing training in mental health through coursework, hands-on experience, and close mentoring by a psychiatrist, I will pursue the following research aims:
Aim 1 : To implement our available monitoring system prototype, and study its applicability to the special characteristics of prescription drugs used in patients with psychiatric illness, using empirical examples.
Aim 2 : To customize the monitoring system to the methodological challenges that are specific to comparative research in mental health, in particular the study of drug-drug interactions and the use of disease risk scores and instrumental variables to mitigate confounding by measured and unmeasured factors.
Aim 3 : To explore the feasibility of using electronic medical records and natural language processing analysis to better capture mental health outcomes which are particularly difficult to identify using administrative data codes and to improve confounding control.
Aim 4 : To implement the new customized system and prospectively monitor the safety of selected mental health treatments newly marketed during the funding period as data accumulate in ear-real-time. Better detection of safety concerns early in the life cycle of mental health drugs is of great public health importance as it may expedite drug approval and make prescribers more comfortable using new drugs by providing a mechanism for ongoing post-marketing safety monitoring. It will also improve the evidence base for mental health providers, and help to contain use of medications with less favorable risk-benefit relations.
There is a critical need for comparative safety and effectiveness information for medical products soon after market entry to ensure safety alerts are raised as early as possible, and to inform prescribing and policy decisions. We will develop, adapt and rigorously validate different approaches for near-real-time monitoring and methods to address confounding and outcome misclassification that accommodate methodological issues arising from the specific needs and circumstances of people with mental illness. As such, this work will lay the foundation for establishing a successful prospective monitoring system optimally calibrated for use with mental health products and in mental health populations.
Showing the most recent 10 out of 46 publications
