May 2002

Data Mining for Dollars
The More You Know, the More You Can Get Paid
By Jeff Pasternack, MSM

Are you ready to learn more about how to “find” more money? Not in a get-rich-quick sort of way, but in a steady, methodical fashion. That’s what “data mining” is all about, and you’re going to learn how you or your office manager can use it to your advantage.

Let’s make the assumption that you have some sort of practice management software that tracks your patients by insurance company, i.e., which patients are insured by which payor. And let’s assume this same software package also handles your billing and accounts receivable. Finally, let’s assume you can export data from your system into a spreadsheet, or, failing that, assume your software can at least print any report you wish to create. Armed with these capabilities and a little bit of know-how, you have what you need at your fingertips to improve your practice’s cash flow.

First, identify your 10 most frequently performed procedures. Now generate a report that shows the average reimbursement rate paid by all insurers (excluding Medicare, Medicaid, and cash) for each of these 10 procedures. By compiling this information, you have performed a basic type of data mining and generated an internal benchmark that shows your practice’s average reimbursement rates.

Next, for each of the 10 procedures, run a report that shows what each individual insurer pays per procedure. In your spreadsheet program, make a matrix for each procedure. On one axis, list the payor, and on the other axis, list the dollar range. Rank each payor with its position on the matrix, i.e., Payor 1 pays the most for Procedure 1, thus it is ranked #1.

Once all 10 matrices are created, add up the payor ranks. The best score possible would be 10. The worst is dependent upon how many payors you have, i.e., if you have 10 payors, the worst possible score would be 100, if the lowest-ranked payor finished in 10th place on all 10 procedures.

Make one matrix that combines the 10 procedures with the ranking. Now you have a single view of your practice’s payor rankings as determined by rate of reimbursement for the top 10 procedures.

Next, you will want to find out what the patient mix is and how it is spread out among the payors. Run a report on your patient population, so it identifies the percentage of patients covered by an insurer but excludes Medicare, Medicaid, and cash. Add a column to your matrix and fill it with that data.

If you were ranking your top five procedures and five payors, the matrix would look like this:

You can extend the matrix out based on how many procedures and payors you are ranking. By looking at this simple matrix, you now know which insurers, on average, pay more than others. In addition, you can relate that information back to the percentage of your patients who are with each company.

So now that you’ve mined all this great data, what should you do with it? The answer, of course, is “it depends.”Now that you know who pays more, you have a couple of different options. First, you can launch a marketing campaign to attract those patients who are covered by the top payors. As the practice grows with those patients, you can begin to phase out your acceptance of insurance from those companies that pay the least.

Second, you can perform similar data-mining tasks on procedures that you don’t perform as often, but appear to be more profitable. Once identified, and depending on the types of procedures identified, your practice could launch a marketing effort to attract those patients who would benefit from more profitable procedures (in keeping with the insurance companies’ guidelines, as well as state and federal laws, of course).

Third, depending on your practice’s situation (i.e., whether you are close to retirement), you could choose to hire an associate doctor to manage the cases of lesser reimbursing insurers so you can focus on other interests. Or, if you’re not in a position to retire or would rather concentrate on your work, you could focus on highly targeted marketing activities to bring in the patients who fit the first and second options previously noted.

Fourth, you could bring on a coding specialist to analyze your coding decisions. There is a fine line here, but in many instances your billing staff may not be totally aware of all coding nuances. A specialist may be able to identify some areas suitable for “right coding” that can legitimately increase your billing rates.

Fifth, you could choose to do nothing and simply monitor these statistics on a monthly or quarterly basis. By analyzing trends and being aware of the changes, your practice may be in a better position to negotiate a more favorable contract than if you were unaware of the statistics.

In and of itself, data mining is not an answer. It does, however, provide access to data that, when combined with knowledge about the data’s origins, becomes transformed into information.

When information is used in a decision-making process, it becomes knowledge. In today’s competitive environment, knowledge can make the difference between simply surviving and thriving.

Mr. Pasternack is the president of Dynamic Physicians Management, a web services and technology consulting firm based in Maryland. He is a member of the Medical Society of Washington, D.C.'s Information Technology Taskforce and lectures around the country about the role of the Internet in healthcare and business. Mr. Pasternack holds a master’s of science in management degree in health-care administration. He can be reached at jeff@dynamicpm.com or 301-610-6666.