Prakash Nadkarni, M.D.
Lecturer, Yale Center for Medical Informatics
Research Interests Web Presentations Selected Publications Database Course (BIS560) Materials Microsoft Access Lecture materials
Email: pmnadkarni @ geisinger.edu
Phone:
Fax:
Mailing Address:
Senior Research Investigator II
Geisinger Center for Healthcare Research
100 North Academy Avenue
Danvile, PA 17821
Research Interests
My hobby (which also happens to be my job) is working with biomedical databases of all kinds.
At the moment, I'm playing with Entity-Attribute-Value (EAV) databases, which are used in domains where the number of potential descriptors (attributes) describing an object is a couple of orders of magnitude greater than the actual number of descriptors for a given object. For example, when dealing with patient data across all clinical specialties, the number of history elements,symptoms, clinical examination findings, lab tests and so on ranges in several tens of thousands, and this number is constantly growing. Yet, for a given patient, not more than a few dozen types of positive or significant negative findings are actually relevant. That is, the data is highly sparse, and a set of conventional relational tables, with one finding per column, would result in much wasted space, because most columns would be null. In the EAV approach, one stores only non-null findings in a table containing three types of information: the Entity (the patient, the date/time the finding was recorded), the Attribute (i.e., the name of the finding) and the Value of the finding.
TrialDB, an EAV database for management of Clinical Studies Data that is copyrighted by myself and my colleague Cindy Brandt (though it is open-source freeware), is described on the TrialDB Home Page. This page has an FAQ, and links to the ftp site, online documentation (also ftp-able) and the demo site where you can try it out..
I also dabble in information retrieval (a fancy phrase for text processing). This is an offshoot of my database interests: a large component of biomedical databases consists of narrative text (which captures nuances that coded text cannot): examples are discharge summaries and operative notes. I'm looking at ways to optimize the searching process by indexing the content based on recognition of concepts in controlled biomedical vocabularies (I mainly play with the National Library of Medicine's Unified Medical Language System), and at ways of integrating text search with conventional database search.
In the past, I've worked in the area of genome informatics as well as parallel computation in molecular biology and genetics.
The following links point to the contents of presentations (organized as HTML framesets) that should be of general interest to medical informaticians.
Clinical Data Warehousing presented at AMIA Fall Symposium, Orlando, FL, Nov 8 1998
ACT/DB: An Infrastructure for Clinical Trials Data Management Columbia University, Jan 21 1999
The EAV/CR Physical Data Model for Heterogeneous Scientific Databases Human Brain Project Annual Meeting, NIH, Jun 5,1999
Understanding and Implementing the EAV Database in the General Clinical Research Center. National GCRC Meeting, Baltimore, MD, April 13, 2002. The URL above is a converted PowerPoint presentation. A detailed explanatory paper can be found below.
An Introduction to EAV systems: National GCRC Meeting, Baltimore, MD, April 13,2002.
Informatics Support of Data management for multi-centric clinical studies: Integrating clinical and genetics/genomic data American College of Medical Informatics, Fort Lauderdale, Florida, March 2003.
Database Representation of Phenotype Data: Issues and Challenges Human Genome Variation Society, American Society for Human Genetics Meeting, Los Angeles, CA , November 4, 2003.
Selected Publications:
Here is a list of selected recent publications. (Some of the papers are downloadable as MS-Word files plus figures, compressed into zip files. See the hyperlinks at the bottom of the publications list..)
Book Chapters
Shepherd, G.M., M.D. Healy, M.S. Singer, B.E. Peterson, J.S. Mirsky, L. Wright, J.E. Smith, P.M. Nadkarni, & P.L. Miller. Senselab: a project in multidisciplinary, multilevel sensory integration. pp. 21-56 of Neuroinformatics: An Overview of the Human Brain Project, ed. S.H. Koslow & M.F. Huerta. Lawrence Erlbaum Associates, Inc. Mahwah, NJ: 1997.
Prakash Nadkarni., Jason Mirsky, Emmanouil Skoufos, Matthew Healy, Michael Hines, Perry Miller and Gordon Shepherd. Modeling Heterogeneous Data on the Nervous System (Book Chapter) in Bioinformatics Databases ed. Stanley I. Letovsky. Kluwer Academic Publishers, Dordrecht , Netherlands. pp. 38-51
Books
Prakash Nadkarni: Parallel Programming with Linda: An Advanced Introduction.
Linda, conceived by David Gelernter and initially implemented by Nick Carriero, both of Yale University, is a "mini-language" constructing of just 4 constructs that is embedded in a conventional language (e.g., C or FORTRAN) to give it parallel capabilities. While conceptually very simple, its reliance on a pre-processor (that must be hand-sculpted for the language in which it is to be embedded) has limited its widespread use, by contrast with MPI, which though somewhat more difficult to use, depends purely on a subroutine library. This book was written back in 1992, and also gives an introduction to parallel programming. It can be downloaded by clicking here.
Downloadable Publications
The downloadable zip files linked to below typically contain more than one publication. Refer to the numbers in the list above. Please note: figures are generally bundled with the MS-word file, or separately, but some figures may be missing). Many of the publications have originally appeared in the Journal of the American Medical Informatics Association, from where you can get excellent content related to Medical Informatics. JAMIA publications that are more than three years old are also freely downloadable (as PDFs) from NCBI's PubMed Central Site
Click on the following links to download full text of
publications of interest :
publications 1, 3 and 4;
publications 7 and 8
publication 9
publications 10, 12 and 13
publication 14
publications 16 and 17
.
publication 18
publication 19
publication 20
publication 21
publication 22
publication 23
publication 24
publication 25
publication 26
publication 28
publication 29
publication 30
publication 32