MediSeek Recaster

MediSeek Recaster is one big, refined, highly specific Extract Transform Load (ETL) operation you run on your computers.

Why does MediSeek have you run the ETL operation? Why doesn't MediSeek just provide the data? There are a few reasons...

We could distribute 8 CSV files that fit our database schema and let you import them and build the indexes. This is a big time-suck for expensive database administrators. who are already busy.
If we distribute a backup of the database, the size of the download would be over 60GB. Our prices would increase to pay for the storage and bandwidth, and you still need a database administrator to restore the database.

We view either option as too much of an ask and go out of our way to simplify your organizations workflow and lower your expense.

Limit The Database Administrator Duties

Using our competitors, your database administrator fees may likely be more than the price we ask for Recaster.

With Recaster you need a database administrator only one time and for one purpose - to create a user account that has privileges in SQL Server.

Recaster Requirements

MediSeek Recaster is written using Microsoft C# and runs on Windows 10 or 11, and most Windows Server versions. The installer you download consumes about 15MB and the installed program consumes about 40MB for a total of about 55MB.

At this time, Recaster uses only the Microsoft SQL Server database, it will not work with the 'Express' edition. Our first release of Recaster used the PostgreSQL database and releasing a PostgreSQL version of Recaster is a high priority.

Expect the database for each time period you process to consume about 65GB and you will be maintaining two databases at least; the current database running and the period you are processing. So on the machine where SQL Server runs, expect to consume about 130GB of space for the SQL Server MDB and transaction files.

Installer

MediSeek Recaster comes with an installer. Documentation, an uninstaller, everything it needs is included - the possible exception being the Microsoft .Net 8 framework that you *may* have to install. However, if you keep your computers updated, Microsoft likely has installed .Net 8 for you already.

Visit Download Page

Download and install Recaster, evaluate at no cost.

$245

For one month license,
no subscription required.

Competitors charge your card $4,000 up front for
12 months, amortized to
$333 per month.

In limited situations, we do offer the source code to Recaster, discussed here.

Using MediSeek Recaster

Recaster is all point-and-click and guided by a workflow. Typing in your database password is all the typing you will do.

From within Recaster, the operator signs into SQL Server with a user account - or a domain account - depending on how you set up SQL Server. This account gaining access to SQL Server does need permission to create a database, tables, indexes and views, plus this account needs CRUD permissions. With these permissions granted, decent computer skills can run MediSeek Recaster.

There is no need for SQL Server Management Studio (SSMS). Internally Recaster uses SQL Server Management Objects (SMO) to run DDL statements that create databases, tables, views and indexes as needed. Certainly, when the database is hydrated, you can use SSMS all you want.

Running Recaster

In just a minute or two, anyone can start the entire transformation process from the CSV file - to a hydrated relational database.

Click a button to ...

Start Recaster.
Signin.
Select the month and year you want to process.
Recaster downloads the NPI zip file from the CMS.GOV website. No URL or bookmark or remember. We know exactly how to construct the URL to download the NPI data file.
Recaster creates a folder for the CMS zip file and unzips it for you.
Recaster creates the database for each time-period you run, we've already named the database for you.
Recaster runs DDL statements to create all the tables in our schema.
Then click a button to start the MediSeek transformation process.
Go do something else while Recaster runs.

Recaster provides real-time analytics as it progresses through the CMS file and hydrates your brand new database. Testing on one of our slowest computers, it can take 12 hours from start to finish.

Post Hydration Actions

When Recaster is finished, it will let you know.

There are three files that CMS provides. Just click a button and each file will process.

The main NPPES file. This is the original file.
The 'other names' file is a newer CSV file that CMS recently incorporated. It contains other names that may be used by an NPI. An NPI may have several names associated with it.
The 'other locations' file. Again a newer CSV file that maintains other addresses a provider may practice at.

When the -3- files have completed, click a button to ...

Create indexes on the hydrated database. This may take a little time.
Rename the database containing this months NPI data to the name of the production database. No database administrator needed.

Renaming The Database

This new database is not the production database yet but will be. Your existing programs run using a 'current database' that we name and whose name never changes. When you have completed a new time-period database, you will want this new database to become current.

The programmers will now protest, claiming their program 'connection strings' to the database will have to change.

Not with Recaster.

Clicking the 'Rename' database button will make the database you created for the time period - the 'current' database. Your programmers do not need to change the connection string.

How?

Microsoft SQL Server stores the database in a MDB file and a transaction file. This file is where the database is stored, but SQL Server allows you to name the database using this file to any name you want. And we do this. Whatever database is current now, its name reverts back to the name of the file. Then the database you just created is re-named to 'Current'. Recaster can do this rename operation in about 5 seconds, so updating from an old database to a new database means your applications are down for about 5 seconds per month.

In many organizations, you have to schedule time with a database administrator or programmer. There can be delays. Not with MediSeek Recaster. You got it done without involving a database administrator and the programmers are out of the loop. That's it. Your new NPI database built with MediSeek Recaster is ready to use.

Database Schema

The CSV files, as we've mentioned contain just under 400 fields all in one record in one table. We reverse engineered this one record in one table into 8 tables, the goal being to normalize the data and form relationships. One of our goals was search-ability.

An NPI is issued to either an individual or an organization, which are two distinct types of entities with entirely different properties. When we segregate away from an NPI the individuals and organizations, the there is not a lot of information about the NPI, in the table named 'provider'.

There are five different name types a healthcare providers NPI may use.

An individuals name.
An individuals AKA.
An organizations name.
An organizations DBA.
And organizations may have a parent organization that is named.

This is obviously a one-to-many relationship between the NPI and the names this NPI may use. The NPI also has one-to-many relationships with the phone, addresses, taxonomy codes and payors (insurers). We end up with an NPI having 5 different one-to-many relationships.

For every provider NPI written into the 'providers' table, one record is written in to the 'providerIndividual' or the 'providerOrganization' table. A record in providerIndividual table will have exactly one record in the provider table. A record in the providerOrganization table will have exactly one record in the provider table.

Purchasing Source Code

Ooh I've worked at corporations as a Senior Software Engineer and Database Administrator. There are times - if a software is important to the line-of-business, the organization will want the ability to modify the programs. That means the corporation wants the source code - we understand the reasoning.

We do offer source code with non-exclusive, non-distribution, non-compete clauses. The pricing is based on a 2.5 year payback compared to purchasing 2.5 years of our nearest competitor pricing. These terms may change at will.

We can email qualified individuals a copy of our Source Code License Agreement for your review.

Our current source code price with non-exclusive, non-distribution and non-compete clauses is $10,000. Exclusive license, with unrestricted distribution rights is $60,000. Yet again, we are confident our quality and pricing are very competitive.