Data Collection and Data Mining

back to products

We collect patient, provider, billing, admission and insurance data from a number of existing systems and organize the information onto a single high powered database server.   We then develop paper reports, spreadsheets and dashboards to present summarized data to users, and finally data cubes to help team leaders analyze and mine the data.

We negotiate with hospital IT groups to optimize the transfer of data and to overcome security and network issues.

Where information is missing or the interval between data feeds is too long we develop screens to allow hospitalists to enter, process and audit their own data.  

Processing the data feeds  

Raw data feeds are imported into database tables with just enough processing to remove erroneous data, and then the feeds are time-stamped and backed up into archive tables.  

The new data in these tables is then validated, processed with various algorithms and merged into the main database structure.  Data  integrity is maintained by allocating internal ID’s to the new information.

Patients are matched between feeds by using complex logic and a points system based on the first and last names, the DOB, Gender, SSN and address information.   The system can detect married women and twins.   If any ambiguity is found the patient is referred for an administrator for review.  Providers are matched in a similar fashion.

We have also developed many screens to maintain the base tables and track the import feeds.   The Emergency Room screen is an example of a a user interface developed to bridge gaps in the data.   It displays a list of current ER patients, along with valuable time-stamp information.

The most important of these screens  is the Event Log, which displays useful information about daily activity, such as the status of import feeds, new providers and patient name changes.  A "New Patients" screen is used to match new patients to existing patients and look for potential matches.

back to products