Skip Navigation

Statistical Sampling Processes

Topics on This Page:

 


Instructions:

II. Describe the statistical sampling process conducted to estimate the improper payment rate for each program identified.


A. Medicare Fee-For-Service

The Medicare fee-for-service (FFS) improper payment estimate is derived from two programs: the Comprehensive Error Rate Testing (CERT) Program and the Hospital Payment Monitoring Program (HPMP). The CERT program reviews claims that account for approximately 60 percent of the total Medicare FFS payments. HPMP reviews claims that comprise the remaining 40 percent. The CERT Program calculates the error rate for Carriers, Medicare Administrative Contractors (MACs), Durable Medical Equipment Medicare Administrative Contractors, and non-Prospective Payment System (PPS) inpatient Hospital claims submitted to Fiscal Intermediaries (FIs). The HPMP calculates the error rate for PPS inpatient hospital claims submitted to the FIs. The Medicare FFS improper payment methodology includes:

  • Randomly selecting approximately 140,000 claims;
  • Requesting medical records from providers on these claims;
  • Reviewing the claims and medical records for compliance with Medicare coverage, coding and billing rules; and
  • Treating non-response by a provider as an error.

B. Medicare Advantage

A methodology to estimate improper payments in the Medicare Advantage (MA) program is currently being developed. During FY 2007, HHS prepared a comprehensive project plan to develop error rates for the MA program and a comprehensive risk assessment to determine potential areas vulnerable to payment error in the MA program. These efforts led to the completion of a measurement project on the payment systems calculation.

Payment System Calculation Discrepancies (PSCD) is one of the areas identified in the risk assessment. The PSCD measures monthly discrepancies between the payment processing system and a simulation of monthly prospective payments that are calculated independently. The simulated payment amounts are generated from a series of Statistical Analysis Software (SAS) programs that use 100 percent of the monthly beneficiary-level payments in MA from the Monthly Membership Reports to independently calculate the monthly prospective payments. The simulated payments are used to validate the monthly prospective payments. Discrepancies identified could contribute to future improper payments if not resolved. Most PSCDs are adjusted in the multiple reconciliation processes and systems in place at HHS. It is important to note that since MA payments are made prospectively and reconciled at the end of the year, the PSCDs are not improper payments. In the MA program, a payment is considered improper if the amount paid was incorrect after final reconciliation.

C. Medicare Prescription Drug Benefit

A methodology to estimate improper payments in the Medicare Prescription Drug Benefit (MPDB) program is currently being developed. During FY 2007, HHS prepared a comprehensive project plan to develop error rates for the MPDB program and a comprehensive risk assessment to determine potential areas vulnerable to payment error in the MPDB program. These efforts led to the completion of a measurement project on the payment systems calculation.

Payment System Calculation Discrepancies (PSCD) is one of the areas identified in the risk assessment. The PSCD measures monthly discrepancies between the payment processing system and a simulation of monthly prospective payments that are calculated independently. The simulated payment amounts are generated from a series of Statistical Analysis Software (SAS) programs that use 100 percent of the monthly beneficiary-level payments in MPDB from the Monthly Membership Reports to independently calculate the monthly prospective payments. The simulated payments are designed to validate the monthly prospective payments. Discrepancies identified could contribute to future improper payments if not resolved. Most PSCDs are adjusted in the multiple reconciliation processes and systems in place at HHS. It is important to note that since MPDB payments are made prospectively and reconciled at the end of the year, the PSCDs are not improper payments. In the MPDB program, a payment is considered improper if the amount paid was incorrect after final reconciliation.

D. Medicaid

To measure Medicaid improper payments, seventeen states, from a total of 50 states plus the District of Columbia, were selected each year to create a three year rotation cycle. To select the 17 states for each year of the 3-year cycle, states were ranked by size based on their past Federal FFS expenditures and grouped into three major strata with 17 states in each stratum. The expenditure data showed that nine states represented a substantial portion (approximately 50%) of the total Federal FFS expenditures. To get a precise estimate for the national rate, it was important to group these nine high-expenditure states into their own stratum. Therefore, the 17 states in Strata 1 were further divided into two substrata – Stratum1A (consisting of the nine states with the highest federal FFS expenditures) and Strata 1B (consisting of the eight remaining highest-expenditure states). The states were sampled such that three states were selected from Strata 1A each year. Given the criterion that each state would be selected once over a three-year cycle, for each stratum there is one year in which only 5 states are sampled. The sample distribution over the three year period, by strata, is illustrated in Table 1 below.

Table 1: Number of States to be Selected from Each Stratum in Each Year

Strata Year 1 Year 2 Year 3
1A 333
1B 332
2 656
3 566

Each state’s sample size is determined based on annual expenditures. The average FFS annual sample size for each state included in the FY 2007 rate is 1,000 claims. States submit quarterly adjudicated claims data from which a randomly selected sample of approximately 250 FFS claims, stratified by service type, is drawn each quarter. Each selected FFS claim is subjected to a medical and data processing review.In FY 2008, HHS expects to report a comprehensive national Medicaid error rate that includes a FFS,managed care, and eligibility component.

E. State Children’s Health Insurance Program

The SCHIP program did not measure an improper payment rate in FY 2007.In FY 2008, HHS expects to report a comprehensive national SCHIP error rate that includes a FFS,managed care, and eligibility component.

F. Temporary Assistance for Needy Families

HHS’ Office of the Inspector General (OIG) has developed a methodology to measure improper payments in the TANF program. In FY 2007, pilot reviews were conducted in three states. The OIG randomly selected 150 cash assistance cases in each state and reviewed the eligibility and payment status of the sampled cases based on Federal and state requirements.In FY 2008, HHS expects to report a national TANF error rate.

G. Foster Care

Foster Care Eligibility Reviews are conducted systematically in each state (the 50 states, the District of Columbia and Puerto Rico) every three years. During these primary reviews, a team comprised of Federal and state staff review 80cases selected from the state’s title IV-E foster care population during a six month period, the Period Under Review (PUR). The reviews determine a state’s level of compliance in meeting the Federal regulatory eligibility requirements for the Foster Care Program and validate the accuracy of a state’s claim for Federal reimbursement of Foster Care payments.

Each regulatory review specifies the number of error cases and the amount of payment errors. An error case is defined as a case in which a payment is made on behalf of an ineligible child during the PUR. Payment errors may include payments for error cases, “ineligible” payments made to non-error cases which failed to meet an eligibility criterion outside the PUR, and “unallowable” payments for services not covered by Title IV-E or its regulatory provisions (e.g. therapy). The information gathered in the regulatory monitoring review is used to correct underpayments as well as overpayments.

HHS employs a 10 percent error threshold to determine the level of state compliance in meeting the Federal requirements in the Foster Care program. If a state exceeds the error threshold for both the case and payment error rates in the primary review, the state will receive a secondary review. During the secondary review, 150cases are selected. If a state exceeds the error threshold for both the case and payment error rates in a secondary review, the state is assessed an additional extrapolated disallowance, which is equal to the lower limit of a 90percent confidence interval for the state foster care population’s total dollars in error during the six-month PUR. The extrapolation increases geometrically the resulting disallowance. Since FY 2000, HHS has systematically conducted more than 110 regulatory Foster Care reviews, with over 8,000 Foster Care cases reviewed.

H. Head Start

HHS is legislatively required to perform reviews of each Head Start program every three years. The design of the sample for the Erroneous Payments Study of Head Start programs is a three-stage element sample. Since each program is reviewed once every three years, the first stage of the sample is to identify the programs up for review.The second stage of the sample is to select the programs to be reviewed. As was done in the FY 2006 Erroneous Payments study, the FY 2007 study selected 50 programs and 10 alternates. Programs were selected through a stratified random sample, where programs were divided into five stratum by enrollment. The number of programs sampled within each stratum is roughly proportional to the number of children represented in each stratum, based on the most recent Program Information Report funded enrollment data. The third stage of the sample is to select the records to be reviewed in each selected program, using a systematic sampling scheme.

In the FY 2007 Erroneous Payments Study, 50 Head Start programs from 31 states were reviewed. A total of 11,083records were examined. The purpose of the reviews was to determine whether documentation demonstrated that a Head Start child was income eligible. A payment error in the Head Start program is defined as a payment for an enrolled child from a family whose income exceeds the allowable limit (in excess of the 10 percent program allowance for families above the income limit). To make this determination, reviewers were required to look at each sample child’s folder and determine if the child was ineligible. A child was deemed ineligible if (1) there was not, as required by 45 CFR Part 1305.4(e), a signed statement by a Head Start employee stating the child was eligible to participate or (2) there was income documentation in the child’s folder that, in the reviewer’s judgment, suggested the child was not Head Start eligible. In FY 2007, reviewers were asked to review income documentation regardless of whether there was a signed statement from the staff in the file.

I. Child Care and Development Fund

During FY 2004, HHS initiated an improper payment pilot project to measure improper payments and to assess the efforts of states to prevent and reduce improper payments in its Child Care program. The project was implemented in two phases with a total of nine states participating in the measurement portion of the project. HHS reported the results of the first phase of the pilot project, which included four states, in its FY 2005 PAR. The second phase of the pilot project was completed by the remaining five states in FY 2007.

In addition to completing the remaining pilot projects, in FY 2007, HHS promulgated a Final Rule revising the ChildCare and Development Fund (CCDF) regulations to provide for the measuring and reporting of error rates in the fifty states, District of Columbia and Puerto Rico.

In FY 2008, HHS expects to report a national Child Care improper authorization for payment error rate.

Other sections of the Improper Payments Information Act Report

 

Report Date: November 15, 2007


AFR Section III Links: