Medical Billing Audit, Clean Claims Metrics, And the Payer-Provider Conflict
Dr. Noah Payne shook his head in disbelief: the practice reimbursements shrank instead of climbing in response to the recent hiring of Dr. Inna Ternist. The new doctor clearly additional to the total number of patients seen however overall payments did not mirror the additional charges. Perhaps the new claims were not produced, submitted, or paid? Dr. Noah remembered noticing the growing pile of rejected and denied claims accumulating dust on his desk – he never had the time to review them…How many of these claims are clean? How many of them require manual review and correction?
Dr. Noah looked at his Vericle screen and began analyzing the numbers. The system showed 58 percent clean claims (PCC). In other words, almost every second claim required manual correction. Who could be causing such a high level of problems: the practice, the billing service, or the payer? Dr. Noah’s instinctively felt that perhaps the billing service was negligent about data entry course of action and kept introducing enormous data errors. But the service manager was quick to explain a demanding quality assurance course of action for data entry. What else could be causing such a high level of manual work in a seemingly streamlined course of action?
A quick review shows that PCC varies along several dimensions:
- 19 and 70 percent for financial class
- 37 and 66 percent for month of service
- 55 and 59 percent for physician
- 29 and 70 percent for various CPT codes
Trying to discover a pattern, Dr. Noah looked for a root cause size. He drilled into 99213 – the single largest frequency CPT code for his practice. Vericle showed 3,135 claims and the above average 62 PCC carrying charges and payments for 99213 code.
Having secluded the single most frequent CPT code, Dr. Noah was thinking about other dimensions that influence PCC. He hypothesized that if all doctors in his practice had the same coding skills, and assuming uniform dispensing of errors, he should observe no PCC variance across the doctors. however, a quick click on a Vericle screen yielded a spread, confirming his suspicion that different doctors maintained slightly different coding skills:
- Dr. Ted 1,554 claims and PCC = 63%
- Dr. Lori 865 claims and PCC = 62%
- Dr. Inna 194 claims and PCC = 61%
- Dr. Noah 516 claims and PCC = 60%
Next, Dr. Noah switched his attention to dispensing of PCC across the financial classes. Again, he hypothesized that if all payers used the same rules to deny claims then there should be no difference in the average PCC for different payers, unprotected to a uniform dispensing of errors over a large sample of submitted and paid claims. however the numbers showed a meaningful (30 percent) variation of PCC for the same CPT code: UHC – 82, Blue Cross Blue protect – 73, Oxford – 64, Aetna – 59, Medicare – 59, and Cigna – 51, confirming his conclusion that various payers used various rules to deny and underpay claims.
Dr. Noah recalled reading an article about PacifiCare, a Californian insurance company being fined upon an audit. The joint Department of Managed Health Care and Insurance Department recently analyzed 1.1 million paid claims from June 2005 to May 2007 that covered about 190,000 members in PacifiCare’s HMO plans and PPO coverage [Gilbert Chan , “PacifiCare fined record $3.5 million,” http://www.sacbee.com , January 30, 2008]. They discovered 30 percent of the HMO claims wrongly denied and 29 percent of the disputes with doctors were handled incorrectly. PacifiCare paid out over $1 million and was fined additional $3.5 million. Dr. Noah’s findings approximately equaled PacifiCare audit – the insurance companies were failing anywhere between twenty to fifty percent of his claims and each insurance company showed a different failure rate, depending on a system used to fail submitted claims.
Finally, Dr. Noah thought of the billing service operation. Is his billing service systematically working to discover failed claims and enhance its response to such discoveries? Is there a pattern of an occasional drop of PCC reflecting its decline in response to various payer’s initiatives? Conversely, is there any evidence for a methodic improvement effort? A chart of the dispensing of a single CPT-code clean claim percentage over the complete year must answer his question. In his mind, PCC should iterate between drops and climbs, hopefully each time at a higher level. Vericle confirmed his expectations, showing an overall improvement of PCC over the year (46% 1-07, 39% 2-07, 52% 3-07, 55% 4-07, 63% 5-07, 67% 6-07, 72% 7-07, 69% 8-07, 72% 9-07, 68% 10-07, 74% 11-07, 73% 12-07)
In summary, Dr. Noah concluded that PCC must be a time-dependent function, which dives down and climbs up depending on four important factors. Specifically, PCC deteriorates in response to any of (a) continuous payer initiatives to obstruct billing, rejecting, losing, delaying, and underpaying claims, (b) practice missing or incorrectly submitting demographics and coding information, or (c) billing service entering data erroneously and inconsistently; and PCC improves in response to a concerted effort by both the practice and the billing service to discover, correct, and avoid demographics, coding, and data entry problems. Large-extent medical billing networks create the needed volumes and resulting economies of extent to permit the payment audits capable to discover systemic claims processing problems.