Data Mining FAQs

We’ve compiled answers to some of the common questions we get from health plans as they look to build or expand data mining capabilities for their organizations.

Is there a certain dollar threshold for claims you’ll review?

Discovery works closely with clients to mutually agree on an overpayment threshold. The majority of our clients set the threshold at $50-$100.

Do you rely only on algorithms or only on manual review for data mining?

Discovery relies on both algorithms and our experienced Data Mining team to identify overpayments. The algorithms drive the sample of claims from which the auditors will filter and sort to identify/validate the overpayments that have the highest scored accuracy, dollar value, time sensitivity, etc. Every overpayment we identify is 100% validated in the client system. Overpayments are never sent out the door without “human touch,” which minimizes the amount of oversight required by the client.

How long before an identified data mining concept can be moved in house?

This varies depending upon the root cause of the overpayment, but we partner with our client’s in-house team every step of the way. For example:

  • A contract load error can typically be resolved in a shorter period of time
  • Decimal point error on surgical case rate ($50,000.00 vs. $5,000.00)
  • A claims processing error that is contrary to policy design and intent may require a longer period of time for resolution
  • Claims processing allowing ungrouped surgical procedures to pay at total claim percentage of billed charges vs. line item percentage of billed charges

Is there a standard integration process for Data Mining services?

Discovery does not use a standard integration process; we customize the process based on each client’s specific requirements. Our flexible integration approach minimizes our client’s time and resources—we configure our workflows and file transfers based on the client’s custom rules and codes, utilize the client’s existing specifications and data feeds, and accepts the client’s data in its existing format.

What’s the best way to approach data mining without harming our provider relationships?

Through our work with dozens of health plans, Discovery has found that the most effective way to introduce data mining is through a phased approach. This approach allows us to help health plans balance overpayment identification while maintaining positive provider relationships. Discovery uses a three-phased approach.

Phase 1: Global concepts

These are “black and white” overpayments with little to no room for contract or regulatory interpretation. Global concepts are applicable to all lines of business. The most common examples include:

  • Duplicate payments
  • Excessive charges
  • Excessive units

While all adjudication systems deploy edits to prevent these global concept overpayments from occurring, they are not always simple to catch and prevent. As an example, duplicate payment errors are more than just “the same claim paid twice.” Duplicates can occur across a subscriber and dependent, two different provider NPIs under the same tax ID, or multiple interim claims with overlapping dates of service. Discovery deploys multiple queries to identify all possible scenarios at both the header-claim level and the detail-line level.

Phase 2: Contract & policy concepts

These concepts are based off billing guidelines and require an analysis of contract terms to develop and deploy. Contract and policy concepts include:

  • Medicare pricing for all claim types (inpatient, outpatient, rehab, etc.), including any retroactive updates from CMS
  • Medicare readmissions and transfers
  • Modifier reductions, including assistant surgeon/non-physician practitioner reviews and practitioner/surgery validation
  • Multiple procedure reduction, including surgery and imaging services

Phase 3: Contract deep-dive

The last phase introduces custom concepts based on a client’s specific provider and plan contractual language. Below are some example targets:

  • Correct reimbursement for combinations of observation, emergency room, and surgery
  • Stop-loss provisions
  • Implant and high-cost drug thresholds
  • Carve-out validation

Not all our clients move through all these phases. Some decide to stay in Phase 1 and may approach Discovery if they have a specific need. Others will give Discovery full access to contracts and policies. There is no right or wrong approach—we are flexible and tailor our Data Mining services to the exact needs of each client.

Discovery Health PartnersData Mining FAQs

Related resources