Appellant, Quality Clinical Laboratories, Inc., appeals from a judgment rendered in the Wayne County Circuit Court affirming a decision of appellee, Michigan Department of Social Services, that required appellant to reimburse the DSS for over $63,000 in medicaid overpayments already made to the lab.
Appellant lab is an independent provider of medical services to medicaid recipients. It provides services to those individuals at the request of their physicians and then files claims for reimbursement with the DSS. The administrative hearing in this case resulted when the DSS determined, after an audit, that it had overpaid the lab during the *599years 1973 through 1975. The DSS claimed that the lab owed it for overpayments which fell into some 14 categories. The case was fully heard by an administrative law examiner, who made findings of law and fact leading to his conclusion that appellant need reimburse appellee DSS for over $63,000.
On appeal, appellant raises two issues. First, appellant claims that the decision of the administrative law examiner was not supported by competent, material and substantial evidence on the whole record, as required under MCL 24.306(l)(d); MSA 3.560(206)(l)(d).
The substantial evidence test applies to review of social services cases; that test requires evidence which a reasoning mind would accept as sufficient to support a conclusion and consisting of more than a mere scintilla of evidence, though it may be somewhat less than a preponderance of the evidence.1 Thus, on appeal we examine the whole record before deciding, but our examination avoids invasion of the agency’s fact-finding province, while affording meaningful review of the administrative decision.2 Review of this record reveals that far more than a scintilla of evidence acceptable to reasoning minds was produced. The trial court’s decision was fully supported by the evidence.
Second, appellant argues that the DSS determination of the refund was arbitrary and capricious. This claim is directed at the DSS sampling procedures used to calculate the refund amount in this case. The lab claims that the statistical extrapolation formula used by the DSS had no rational and *600supportable basis. The lab claims that the sampling technique was not accurate. The audit consisted of taking a random sample of the claims made by plaintiff for 150 recipients of laboratory services for the period audited. The amount of overpayment was arrived at by determining the percentage of overpayment in the sample of 150 files (error rate) times the total dollars paid the appellant for all services during that period from January 1, 1973, through December 31, 1975. The department listed 14 separate categories as the basis for considering certain claims to be erroneous. They were as follows:
"1. Missing: No documentation to verify that recipient had services performed on that date.
"2. No Results: Either no values printed across from test name or no test at all.
“3. No Request: Either referring physician request does not indicate request or profile ordered does not include such testing.
"4. Culture Only: Either lab or physician documentation confirms non-performance of sensitivity testing or; culture results indicate non-performance of sensitivity.
"5. Included in 8891: Culture count is billed through both codes 8888 and 8891 resulting in a duplicate payment.
"6. Normal Flora Sensitivity: Sensitivity performed on non-pathogenic growth and therefore not reimbursable.
"7. Hemoglobin Coding Error: Hemoglobin Electrophoresis billed while only simple Hemoglobin test performed.
"8. 8224 Results: T7 test billed but PBI performed.
"9. 8508 Results: Australian Antigen test performed but Immuno Electrophoresis billed.
"10. Derived from T3 , T4: Values for this test are obtained by a simple calculation of T3, T4 values and is therefore not reimbursable.
"11. Chiropractic Referrals: Testing referred to laboratory by chiropractor.
"12. Credit to Provider: Provider not correctly reim*601bursed as a result of Optical Character Reader (OCR) error. Refund amount is expressed negatively.
"13. 8305 Overlap: Values found in 8307, Total Protein and A/G Ratio, are present in 8305, Protein Electrophoresis, and thus is not reimbursable.
"14. Included in 8297: Values for codes 8112, Cholesterol, and 8340, Triglycerides, are present in code 8297, Lipoprotein Electrophoresis and Phenotyping.”
While the initial amount of overpayment indicated by the audit was established at $104,093.45, as a result of meetings and a preliminary conference that amount was reduced to $68,195.42. The trial court responded to this claim as follows:
"Thus, the recognition of statistical methods by the courts of several jurisdictions lends support to the statistical results of the department’s audit. Had the department desired an actual reimbursement figure, rather than a calculated one, it would have had to peruse through thousands of files. This method has been found to be ineffective due to its time-consuming nature. In answer to appellant’s argument that this sampling method is invalid, this court does not find that the statistical survey was self-serving or improper.”
In Illinois Physicians Union v Miller,3 a federal court has decided this precise issue. The court in that case determined that the use by the Illinois Department of Public Aid of a sampling extrapolation formula for determining overpayments to physicians under the medicaid program was not arbitrary, capricious, or invidiously discriminatory, where there was opportunity to rebut the initial determination of overpayment.
Our review leads us to conclude that no error was created by the use of the statistical extrapolation formula in this case.
Affirmed.
Russo v Dep’t of Licensing & Regulation, 119 Mich App 624, 631; 326 NW2d 583 (1982); Tompkins v Dep’t of Social Services, 97 Mich App 218, 222; 293 NW2d 771 (1980); Soto v Director of the Michigan Dep’t of Social Services, 73 Mich App 263, 271; 251 NW2d 292 (1977).
Freiberg v Big Bay De Noc School Dist, 91 Mich App 462, 468; 283 NW2d 775 (1979), lv den 407 Mich 964 (1980).
675 F2d 151, 156 (CA 7, 1982).