Current through Register Vol. 49, No. 13, September 23, 2024
Subpart 1.
Authorization.
For the purpose of part
9505.2215, the commissioner is
authorized to calculate the amount of monetary recovery from a vendor based
upon extrapolation from a systematic random sample of claims submitted by the
vendor and paid by the program or programs. The department's random sample
extrapolation shall constitute a rebuttable presumption regarding the
calculation of monetary recovery. If the presumption is not rebutted by the
vendor in the appeal process, the department shall use the extrapolation as the
monetary recovery figure specified in subpart
3.
Subp. 1a.
Definitions.
For purposes of this part, the following terms have the
meanings given them.
A. "Cluster
sampling" means a method of sampling in which a selection of groups of items
are sampled rather than a selection of individual items.
B. "Extrapolation" means estimating an unknown
overpayment by projecting, with a calculated margin of error, the results from
a random sample to the population from which the random sample was drawn. The
form and computations for the extrapolation and its confidence interval depend
on the method of random sampling.
C. "Interval estimate" means the estimation
of a parameter, in terms of an interval, for which one can assert with a given
probability or degree of confidence that it contains the actual value of the
parameter.
D. " 90 percent
confidence interval" means an interval estimate, or estimate with a margin of
error, for a population parameter computed using a procedure that produces
intervals that contain the true parameter for 90 percent of all random
samples.
E. "Population" or
"population of claims" means a defined set of paid claims for a specified
period that exist at the time of the audit or investigation.
F. "Population parameter" means a measure
such as mean, median, standard deviation, or proportion that is calculated or
defined by using every item in the population.
G. "Probe sample" means a limited initial random
sample of at least 50 units that can be used to provide guidance in selecting
the sample size for a full random sample.
H. "Random sample of claims" means a subset of claims
chosen from a population of claims using a random sampling method including
simple random sampling, stratified sampling cluster sampling, or another
sampling method that the department determines is more likely to lead to
greater precision, or a closer approximation to the population mean.
I. "Simple random sampling" means a method of
sampling in which all individual items in the population have an equal
probability of being included in the sample.
J. "Stratified sampling" means a method of
sampling in which the population is divided into two or more parts and a simple
random sample is selected for each part. An estimate is determined separately
for each part, and these are combined to form an estimate for the entire
population.
K. "RAT-STATS" refers
to the primary statistical software used by the Office of the Inspector General
of the United States Department of Health and Human Services and the Centers
for Medicare and Medicaid Services.
L. "Sampling plan" means the combination of the
identified population, the random sampling method, the sample size, and the
technique for implementing the random sampling method on the
population.
M. "Statistical
analysis system" or "SAS" means an analytics and statistics software system
developed by the SAS Institute.
N.
"SPSS" means an analytics and statistics software system licensed by
IBM.
Subp. 2.
Decision to use samples.
The department may use sampling and extrapolation to
calculate a monetary recovery if:
A.
the claims to be reviewed represent services to 50 or more recipients;
or
B. there are more than 1,000
claims to be reviewed.
Subp.
3.
Statistical method.
The department must use the procedures in this subpart when
calculating the amount of monetary recovery by random sample extrapolation from
the audit results of a random sample of claims.
A. The sampling plan and extrapolation must
be chosen and performed according to guidance from the Centers for Medicare and
Medicaid Services including the most recent version of the Medicare Program
Integrity Manual.
(1) The sampling plan must
include a probe sample.
(2) The
department must use tools that include RAT-STATS, SAS, or SPSS.
B. The department must tell the
provider the sampling method the department is using prior to drawing the
sample.
C. The vendor must pay an
overpayment identified under this part only if the overpayment identified has a
90 percent confidence interval that does not contain zero dollars.
