EXPERIMENT WITH ARTIFICIAL LARGE MEDICAL DATA-SETS WITHOUT WORRYING ABOUT PRIVACY.
It is difficult to get access to Electronic Medical Records (EMRs) due to privacy concerns and technical burdens.
I am in a process of founding a company focused on developing a new EMR management platform and I want to demonstrate to a venture capital company and to potential customers the ability of my product to handle big data. Current simulated medical databases are limited and are hard to configure.
I am a student or a researcher working at a university that does not have yet an access to EMR system and I am interested in evaluating machine learning algorithms. Tedious bureaucracy.
I am teaching a computer science course and I wish to let my 150 students to experiment with electronic medical records. Not possible due to privacy issues.
Solution: a database of artifical patients.
The data is generated according to pre-defined criteria and is not based on any human data.
The database contains the same characteristics that exist in a real medical database such as patients’ admission details, demographics, socioeconomic details, labs, medications, etc.
The database is customizable. For example, it is possible to generate a population of 100,000 patients of which 60% are male, 40% are African American, 15% are diabetic, specific lab range distributions can be set, etc.
The number of records can range from several thousands to millions, depending on the desired configuration.