V. Haroutunian, Ph.D.The Mount Sinai School of Medicine and Bronx VA MedicalCenter Department of Psychiatry Brain Bank
Traditionally brain and tissue banks have provided biological specimens and some limited clinical and pathological data to multiple laboratories and each laboratory has worked independently to generate and maintain data specific to its specific aims. This model is effective in helping address specific and focused hypotheses, but it is not an effective mechanism for enhancing collaboration between laboratories and maximizing mining of data derived from independent sources. This is an issue that is particularly relevant to autism brain research given the scarcity of well characterized tissue specimens. Although the principal focus of the Mount Sinai/Bronx VA Department of Psychiatry brain bank is Alzheimer disease and schizophrenia, some specimens from patients with autism are banked and the need for data sharing and data warehousing is similar. To address this need, we are in the process of developing a centralized data warehouse to support not only tissue and donor specific data, but to also maintain data derived from utilization of research specimens by different projects and laboratories. A data warehouse structure that addresses these needs must accommodate data from a multitude of diverse sources, strictly preserve donor confidentiality, preserve research data confidentiality, respect the proprietary nature of research data generated by different investigators, while at the same time enhancing data exchange and data mining by diverse investigators from multiple disciplines and multiple research groups. These needs for absolute adherence to patient and investigator confidentiality on the one hand, and cross-laboratory data sharing and data mining on the other hand, are at face value mutually exclusive. The challenge of a maximally useful data warehouse is in its ability to accommodate these apparently diametrically opposed aims. The Mount Sinai/Bronx VA Department of Psychiatry data warehouse project seeks to reconcile these aims by a) adherence to strict but minimal data normalization rules and criteria; b) identity stripping and encoding; c) internet and web driven interfaces; d) predefined and "on the fly" query capacities; e) a system of data-tags that define "belongingness" to different laboratories; and f) password protected limits on ranges of accessible data sets. These systems permit each laboratory access to their own data and access to a set of common variables. Each laboratory with access to the warehouse can also view the class of data variables that other laboratories have contributed to the warehouse, but they can only gain access to the underlying data by permission of the laboratory that generated that particular data set. Since each laboratory controls access to its own data, they can grant access to the underlying data to other investigators by collaborative agreements. This protocol safeguards tissue donor confidentiality and investigator confidentiality, while permitting multiple investigators a view of all of the data classes available on any given specimen and the capacity to establish collaborations for in dept data mining and data sharing.
Back to Agenda