How Are You Using Physician Data Mining?

The pay for performance model of payment has been the preference of health insurance companies for several decades. Previous models made it too easy for money to be paid for services that were not truly needed, were poorly documented or in some cases, were entirely fabricated for the purpose of payment. Pay for performance measures are meant to validate payments to physicians in accordance with The Affordable Care Act by using data mining techniques to keep tabs on physician costs.

Tracking physician payments, however, requires new datasets that physician practices, hospitals and healthcare systems may not have previously had the means to track. Metrics included in this required tracking might be Medicare measures, guidelines and performance comparison between physicians within a healthcare system; fostering a healthy sense of competition may prove to help keep physicians on track to meet the goals of the ACA.

Data collection on physician performance is not just valuable for pay out: quality improvement and patient safety rank high among the concerns of the executives orchestrating such tracking measures. By tracking performance measures across the broad, systemic problems within a healthcare system could potentially appear. What at first seemed to be an issue only with one practice or specialty may in fact exist across the board.

This becomes actionable data for quality improvement programs that will want to target the most troubled areas with fine-tuned solutions. By mining healthcare system wide data, they can also acquire metrics on the biggest costs to their particularly healthcare system; whatever the specific high cost, cost-saving measures can be designed to target that area.

While this data is potentially fuel for hospital reform, it could also usher in a slew of aggressive, disciplinary actions against physicians and their practices that may be unwarranted: some hospitals may use the data as justification for punishment, rather than using the resources for finding solutions.