OSIP Newsline
Newsline No. 02-11
December 12, 2011


OSIP to improve audit procedures in 2012

The Office of Self Insurance Plans (OSIP) Audit Unit will change the method that it uses to conduct audits of claims to better establish estimated future liabilities reserves. Effective for all audits initiated on or after January 1, 2012, OSIP will begin using a random sample methodology for selecting claims file audits.

For each firm being audited, the Audit Unit will select a statistically significant number of claims for auditing based on the  firm’s total number of claims. The claims will be randomly selected for audit using a random number generator program. The audit is designed to determine if the claims administrator has established proper reserves on the medical and treatment information available.

In instances where the files are under or over-reserved, the percentage of under or over-reserving will be determined. This percentage of reserve increase or decrease will then be extrapolated across the entire universe of claims files in determining the proper level of required security deposit that must be maintained by the self-insured employer.

Previously, OSIP has used a method of identifying and auditing high value claim files that have a probability of being significantly under-reserved. Reserve adjustments were then made only to the audited files.

“Under the new system, employers will have a fairer degree of application of both over and under-reserved files being audited, with a blended rate being applied to the entire universe of claims,” said OSIP Chief Jon Wroten.

In instances where it is determined that a self-insured employer has a significant across-the-board practice of under-reserving, OSIP will either audit further, or order a special audit. A special audit is usually performed by an independent auditor chosen by the employer at their expense.

The random sample method is currently in use by the Division of Workers’ Compensation Audit Unit, which OSIP will be mirroring their best practices after.

# # #