Random sample audit methodology misses
90 percent or more of claim errors, study finds

In our June 2009 newsletter, we reported that Healthcare Data Management, Inc. (HDM) had commissioned an independent study of the relative efficacy of the 100-percent-of-health-plan-claims audit versus the audit methodology that relies on a random sample of between 200 and 400 claims to identify claim errors…wasted expense.

The findings are in, and they are resoundingly conclusive: multiple disparate data sets were analyzed and the results compared. The analysis revealed that random sample audits would have missed a minimum of 90 percent or more of the errors identified by HDM’s 100-percent-of claims-audit.

Ronald K. Klimberg, Ph.D., Professor, Department of Decision and System Sciences at the Haub School of Business at St. Joseph’s University in Philadelphia, has completed exhaustive simulations and comparisons of these data sets. Now he and his colleague in the study, George P. Sillup, Ph.D., M.S. Assistant Professor, Department of Pharmaceutical Marketing at St. Joseph’s, are detailing the study and the findings in a white paper to be published soon.

Dr. Klimberg ran 100 simulations on 300 and 400 random claim samples each from discreet audit data sets used by HDM to perform retrospective audits for multiple different types of clients – both large organizations. In each case, HDM employed its 100-percent-of-claims methodology. Many types of benefit plans were represented.

For example, the value of overpayments identified by the 100-percent-of-claims methodology ranged from nearly $220,000 to nearly $700,000 for the multiple data sets. Randomly selected samples of 300 and 400 claims all failed to pick up 90 percent or more of these errors.

“Based on our findings, the underlying premise of the random sample method of auditing health plan expenses is flawed,” Dr. Klimberg said. “We embarked on this study with no pre-conceived notions,” Dr. Sillup added, “but we never envisioned such a dramatic result.”

“Relying on a random sample audit of health plan expenses is like panning for gold with a very coarse sieve. You are going to miss valuable nuggets both small and large. This is exactly what happened in our study. All of our random samples missed errors small and large,” Professor Klimberg concluded.

“We’ve always believed that our 100 percent approach to auditing is equivalent to the zero defects business management strategy. We urge all self-funded companies who are relying on random sample auditing, to take a long, hard look at what they’re missing,” Bill Conlan, CEO of HDM commented.

“The Klimberg/Sillup work has many implications,” David McSweeney, COO of HDM said. “It means that corporations, governmental entities and Taft-Hartley plans that rely on the random sample audit methodology are failing to identify large amounts of wasted expense.”

“For public companies, the findings have implications for Sarbanes-Oxley, which requires better corporate oversight and vigilance than the random sample health plan audit can provide.”

“We consistently find high levels of expense errors in most of the retrospective audits we do. The really amazing thing is that, with technology today, we can identify the same errors almost concurrently with the payment of claims. Our BenefitsWatch service can do this, and put plan sponsors on a level playing field with the insurance companies,” McSweeney concluded.

HDM will soon issue an alert about when the Klimberg/Sillup study will be available at www.hdminc.com, www.cfo.com and other distribution points.

For preview of the methodology and findings, contact David McSweeney at 800-859-5119 ext. 225 — dmcsweeney@hdminc.com or James Herrington at 800-859-5119 ext. 280 — jherrington@hdminc.com.