CFR > Title 42 - Public Health > CHAPTER IV—CENTERS FOR MEDICARE & MEDICAID SERVICES, DEPARTMENT OF HEALTH AND HUMAN SERVICES > SUBCHAPTER C—MEDICAL ASSISTANCE PROGRAMS (parts 430 to 456) > PART 431—STATE ORGANIZATION AND GENERAL ADMINISTRATION > SUBPART P—Quality Control (§431.800 to §431.865) > 431.814—Sampling plan and procedures.

431.814—Sampling plan and procedures.

(a) Plan approval. The agency must submit a basic MEQC sampling plan (or revisions to a current plan) that meets the requirements of this section to the appropriate CMS regional office for approval at least 60 days before the beginning of the review period in which it is to be implemented. If a plan is unchanged from a previous period, the agency is not required to resubmit the entire plan. Universe estimates and sampling intervals are required 2 weeks before the first monthly sample selection for each review period. The agency must receive approval for a plan before it can be implemented.
(b) Plan requirements. The agency must have an approved sampling plan in effect for the full 6-month sampling period that includes the following:
(1) The population to be sampled;
(2) The list(s) from which the sample is selected and the following characteristics of the list(s):
(i) Sources;
(ii) All types of cases in the selection lists;
(iii) Accuracy and completeness of sample lists in reference to the population(s) of interest;
(iv) Whether or not the selection list was constructed by combining more than one list;
(v) The form of the selection list (whether the list or part of the list is automated);
(vi) Frequency and length of delays in updating the selection lists or their sources;
(vii) Number of items on the lists and proportion of listed-in-error items:
(viii) Methods of deleting unwanted items from the selection lists; and
(ix) Structure of the selection lists.
(3) The sample size, including the minimum number of reviews to be completed and the expected number of cases to be selected. Minimum sample sizes are based on the State's relative level of Medicaid annual expenditures for services for active cases, and on the total number of negative case actions in the universe for negative cases. When the sample is substratified, there can be no fewer than 75 cases in each substratum, except as provided in paragraph (c) of this section or as provided in an exception documented in an approved sampling plan which contains a statement accepting the precision and reliability of the reduced sample.
(4) The sample selection procedure. Systematic random sampling is recommended. Alternative procedures must provide a representative sample, conform to principles of probability sampling, and yield estimates with the same or better precision than achieved in systematic random sampling.
(5) Procedures used to identify amounts paid for services received in the review month.
(6) Specification as to whether the agency chooses to—
(i) Use billed amounts to offset recipient liability toward cost of care (No indication will be interpreted to mean that the agency will use paid claims); and
(ii) Use denied claims to offset recipient liability toward cost of care in the payment review. (No indication will be interpreted to mean denied claims will not be used.)
(7) Indication of whether the agency opts to drop or complete cases selected more than once in a sample period. (No indication will be interpreted to mean that the agency will complete cases selected more than once.)
(c) Eligibility universe—active cases. The MEQC universe for active cases must be divided into two strata, the Aid to Families with Dependent Children (AFDC) stratum and the Medical Assistance Only (MAO) stratum.
(1) All States must use the AFDC quality control sample for the AFDC stratum.
(2) States must include in the MAO stratum all cases certified as eligible for Medicaid that are not in the AFDC stratum, excluding individuals specified in paragraph (c)(4) of this section.
(3) States that do not have an agreement with the Social Security Administration under section 1634 of the Act and do not have more restrictive eligibility criteria under section 1902(f) of the Act but require a separate Medicaid application for recipients of SSI and determine Medicaid eligibility using SSI criteria must divide the MAO stratum into two substrata: MAO cases and SSI cash cases for the first review period beginning after July 1, 1990 and for review periods thereafter. The SSI substratum sample size must be 75 cases or one-half of the total MAO sample, whichever is smaller. The non-SSI MAO substratum sample will be the remainder of the MAO stratum cases.
States may be exempt from this requirement when implementing an approved sampling option that does not accommodate this stratification method.
(4) States must exclude from the MEQC universe all of the following:
(i) SSI beneficiaries whose eligibility determinations were made exclusively by the Social Security Administration under an agreement under section 1634 of the Act.
(ii) Individuals in foster care or receiving adoption assistance whose eligibility is determined under Title IV-E of the Act.
(iii) Individuals receiving Medicaid under programs that are 100 percent Federally-funded.
(iv) Individuals whose eligibility was determined under a State's option for Express Lane Eligibility under section 1902(e)(13) of the Act.
(d) Eligibility universe—negative cases. Unless the agency has an approved superior system under § 431.812(c) that provides otherwise, the universe for negative Medicaid eligibility cases must consist of all denied applications, suspensions, and terminations occurring during the review month except transfers between counties without any break in eligibility, cases in which eligibility is exclusively determined by SSA under a section 1634 contract, cases determined eligible for foster care and adoption assistance under title IV-E of the Act, and cases under programs that are 100 percent federally funded.
(e) Sampling procedures. The agency must document all sampling procedures used by the State agency, including 98 percent accuracy of program identifier codes used in the sampling frame to separate listed-in-error cases from those in the population of interest, must make them available for review by CMS, and must be able to demonstrate the integrity of its sampling procedures in accordance with this section.
(f) Sampling periods. The agency must use 6-month sampling periods, from April through September and from October through March.
(g) Statistical samples. The agency must select statistically valid samples of both active and negative case actions.
(h) Sample selection lists. The agency must submit to CMS monthly a list of cases selected in the sample to be reviewed, after the State's sample selection and before commencing MEQC reviews on the cases in the sample.
(i) Universe estimates and sampling intervals. The agency must submit detailed universe estimates and sampling intervals to CMS for approval at least 2 weeks before the first sample selection of the review period if the estimates differ from the previous period. The sampling intervals must be used continuously throughout the sampling period unless otherwise specified in an approved sampling plan. Final universe counts based on the actual sampling universe must be determined and reported to CMS for each stratum/substratum designated in the sampling plan.

Code of Federal Regulations

The agency also must submit universe counts for cases eligible for foster care and adoption assistance under title IV-E of the Act, and, for States with an agreement under section 1634 of the Act, for cases found eligible by the Social Security Administration.
(j) Sample size and methodology options. The agency may select a sample size in accordance with the minimum established under paragraph (b)(3) of this section or use one of the methodologies specified in paragraph (j)(1) or (2) of this section.
(1) Increase in size. The agency may, at its option, increase its sample size for a sampling period above the federally prescribed minimum sample size provided for under paragraph (b)(3) of this section, and receive FFP for any increased administrative costs the agency incurs by exercising this option.
(2) Retrospective sampling. The agency may, at its option, implement retrospective sampling in which cases are stratified by dollar value of claims paid. If the agency selects retrospective sampling, it must—
(i) Draw an initial case sample size each month that is no less than 5 times the required sample size. The sample will be selected from the universe of cases that were certified eligible in the fourth month prior to the month of case selection;
(ii) Identify claims paid for services furnished to all individuals during the review month (and, if indicated, any months prior to the review month in the agency's selected spenddown period) for these cases;
(iii) Stratify the cases by dollar value of the claims into three strata; and
(iv) Select a second statistically valid sample within each group subject to the sample size requirements specified in paragraph (b)(3) or (j)(1) of this section.
[55 FR 22166, May 31, 1990, as amended at 75 FR 48847, Aug. 11, 2010]