AAPC-CPMA Certified Professional Medical Auditor - Set 5 - Part 1

Correct Answer: Convenience sampling

Convenience sampling is non-random and uses results that are readily available. Choosing the charts easiest to pull fits this method, and its results may be highly biased because the selection favors certain outcomes. Option a is incorrect because it requires dividing the population into strata and picking the same number of values from each stratum, which the auditor did not do. Option b is incorrect because it randomly selects whole groups and includes every member of those groups, but here no groups were chosen at random. Option c is incorrect because it follows an established rule, such as every nth entry from a random starting point, rather than taking whatever is easiest to reach.

Correct Answer: The result of sampling from a population

Data are the result of sampling from a population. A portion, or subset, of the population is selected and studied, and the information collected from that sample is the data, from which a statistic can be calculated. Option b is incorrect because that defines a statistic, which is calculated from the sample data after collection and so is derived from the data rather than being the data. Option c is incorrect because that describes a parameter, which belongs to the entire population and is estimated by a statistic, not the results gathered through sampling. Option d is incorrect because that is the definition of a population, the larger group from which a sample is selected, not what sampling produces.

Correct Answer: Validity

Validity refers to how much a measure or conclusion accurately reflects the real world. The note describes exactly that quality, so validity is the term, as distinct from terms describing sample selection or sample variation. Option a is incorrect because it is a numerical characteristic of a sample that estimates the corresponding population parameter, not a judgment of how accurately a conclusion reflects reality. Option b is incorrect because it is the natural variation that comes from selecting a sample to represent a larger population, and it shrinks as sample size grows. Option c is incorrect because it describes a sample in which not all members of the population are equally likely to be selected, which is a selection problem rather than accuracy.

Correct Answer: A statistician or equivalent expert

The manual requires the sampling methodology to be reviewed and approved by a statistician or a person with equivalent expertise in probability sampling and estimation methods, ensuring a statistically appropriate sample and sound estimation methods. Option a is incorrect because the provider is the party from whom medical records are requested; the manual gives it no role in approving the methodology. Option c is incorrect because the COR is consulted, with the BFL, about using extrapolation when the criteria are not met, but does not approve the sampling methodology itself. Option d is incorrect because OIG audits or evaluations are one listed indicator of a sustained or high payment error level, not a reviewer of the methodology.

Correct Answer: Reviewing sampling results and extrapolation before release

Before a findings letter or overpayment demand letter goes out, the contractor has the statistical expert review the sampling results and any later overpayment estimation or extrapolation, and verifies the contractor agrees with those statistical findings. Option a is incorrect because the contractor releases the findings letter or demand letter; the statistical expert's task is reviewing the statistical results beforehand, not issuing the letter. Option b is incorrect because CMS approves the time period for RACs and the SMRC; the statistical expert only helps the contractor determine the period. Option c is incorrect because the contractor, not the statistical expert, consults its BFL with a copy to the COR when no sustained or high error criterion is met.

Correct Answer: A cluster of claims associated with a patient

Medicare program integrity guidance says the sampling unit may be a claim line or a higher-level unit, including a claim, a cluster of claims or lines associated with a patient, or a treatment day. Option a is incorrect because this is the result produced after sampling, an estimate of total overpayment for units in the frame, rather than something chosen for review. Option b is incorrect because the period for review sets the time boundary of the universe and must be documented consistently, but it is not a selectable element. Option d is incorrect because the universe is the target population from which the sampling frame is built, not an element that the sampling software selects.

Correct Answer: Sampling units in the sampling frame

The Medicare program integrity guidance states that the extrapolation estimate of total overpayments covers the sampling units in the sampling frame. Exclusions made while building the frame, such as prior-review or no-payment claims, are outside its scope. Option a is incorrect because the universe is the target population used to build the frame; exclusions made during frame construction remove claims from the estimate's scope. Option b is incorrect because the sample is the subset actually reviewed, and the estimate projects from it to the larger frame rather than stopping at those reviewed units. Option c is incorrect because these are an example of claims the auditor may exclude when building the frame, so they fall outside the estimate.

Correct Answer: statistical expert

The Medicare manual states that after the provider or supplier is selected, the contractor, in conjunction with a statistical expert, determines the time period and the number of days, weeks, months, or years reviewed. Option a is incorrect because the Medicare manual does not give this role any part in choosing the review period; the contractor works with a statistical expert instead. Option b is incorrect because CMS approves the time period only for RACs and SMRC, so it acts as an approving body rather than as the expert consulted on the period. Option d is incorrect because medical reviewers are named in the self-disclosure guidance among qualified individuals who conduct a review, but the Medicare manual does not assign them the period-setting role.

Correct Answer: nonoverlapping

RAT-STATS describes a stratified plan as dividing the universe into two or more nonoverlapping categories (strata), then drawing a random sample from each stratum. Each universe item therefore belongs to exactly one stratum. Option b is incorrect because strata that overlap would let one item sit in several groups, whereas RAT-STATS describes strata as separate divisions of the universe. Option c is incorrect because the program asks for the number of universe items in each stratum, and these values must be known before estimates are made. Option d is incorrect because randomness applies to the selection of items within each stratum; the strata themselves are defined categories, and no such label is given to them.

Correct Answer: Cluster sample

A cluster sample divides the population into clusters, randomly selects some of the clusters, and includes all members of the selected clusters. The policy describes exactly this process of random group selection followed by full review of each chosen group. Option a is incorrect because a stratified sample draws a random sample from every stratum, so no stratum is left out, whereas this policy reviews only some groups. Option c is incorrect because convenience sampling is non-random and uses readily available results, but here the groups were chosen by random selection. Option d is incorrect because a systematic sample follows a rule, such as taking every nth entry from a random starting point, rather than selecting whole groups.

Correct Answer: Single-Stage Random Numbers

In the stratified attribute appraisal example, both random samples were obtained with the Single-Stage Random Numbers program, which generated 100 random numbers within the range of each stratum's universe size. Option b is incorrect because it is the appraisal program that estimates error rates for each stratum and the universe, not the program that generated the random numbers. Option c is incorrect because it randomly selects whole groups and keeps all their members, while the example drew random numbers within each stratum. Option d is incorrect because it is a method of choosing every nth entry from a random starting point, and the example used random numbers instead.

Correct Answer: Median

The median is generally the better measure of the center when extreme values or outliers exist, because it is not affected by the precise numerical values of the outliers. The mean is affected by them. Option a is incorrect because it is a parameter describing the whole population and is also an average of all values, so extreme values still influence it. Option b is incorrect because it is pulled toward extreme values, so one very large claim would distort it and give a misleading sense of the middle. Option c is incorrect because this is simply the symbol for the sample mean, a statistic calculated from every value, so the outlier still shifts it.

Correct Answer: The center stays about the same, and the spread of the means gets smaller.

Under the central limit theorem, as sample size increases the mean of the sample means remains approximately the same, while their standard deviation shrinks. The distribution of means tends toward normal and looks steeper and thinner. Option a is incorrect because larger samples make the sample means tend toward a normal distribution, with a graph that appears steeper and thinner rather than flatter and wider. Option b is incorrect because the central limit theorem shows the standard deviation of the sample means getting smaller with larger n, so a widening spread is the opposite of what occurs. Option d is incorrect because the mean of the sample means remains approximately the same as n grows, so no upward shift in the center is expected.

Correct Answer: It is two standard deviations below the mean

A z-score states how many standard deviations a value lies from the mean. Values smaller than the mean have negative z-scores, so negative two places the claim two standard deviations below the mean. Option a is incorrect because a value equal to the mean has a z-score of zero, so a score of negative two cannot describe it. Option b is incorrect because positive z-scores mark values larger than the mean, while a negative z-score places the value to the left of the mean. Option c is incorrect because z-scores count standard deviations, not raw units, so the distance depends on the standard deviation, which is not given.

Correct Answer: Sampling Frame

The Self-Disclosure Protocol defines the Sampling Frame as the totality of the sampling units from which the sample was selected. The report must also explain how the audit population differs from the frame and what effect that difference has on conclusions. Option b is incorrect because it is a report item describing the group of claims needing information, plus the method and basis used to develop that group. Option c is incorrect because it states how many items were reviewed to reach the damages estimate, and the protocol requires at least 100 claims. Option d is incorrect because it is defined as any of the designated elements that constitute the population of interest, not the full set from which the sample was drawn.

Correct Answer: Mean weekly hours statistics students spend watching television

Mean weekly television hours calls for a confidence interval for the mean, as in the statistics students example. Proportion problems instead involve categorical data with two categories, such as Success or Failure, Yes or No. Option a is incorrect because voting for a candidate is cited as a proportion question, and election polls report confidence intervals as proportions or percentages. Option c is incorrect because it appears among the standard examples of estimating a true population proportion, so it belongs to the set the stem names. Option d is incorrect because college-level education is listed beside smoking and voting as a situation where the true population proportion is estimated, so it fits the set.

Correct Answer: Exact hypergeometric distribution underlies the procedure

RAT-STATS derives an exact confidence interval based on the hypergeometric distribution rather than the normal approximation. Because that interval is not symmetric, the point estimate is not in the center of the interval. Option b is incorrect because zeroing a negative lower limit appears in the stratified discussion and has nothing to do with the off-center interval in the unrestricted example. Option c is incorrect because the standard error formulas are not used by RAT-STATS when deriving the interval; they belong to approximate methods in other software packages. Option d is incorrect because RAT-STATS does not use the normal approximation; other software packages use it, and it would produce a symmetric interval centered on the estimate.

Correct Answer: Unrestricted using expected error rate

The Variable Unrestricted procedure using the expected error rate assumes the percentage of nonzero errors equals the expected error rate. The expected zeros in the error population are therefore factored into the sample size calculation. Option a is incorrect because that is Situation 2, which allocates an overall sample already determined across strata, favoring larger or more varied strata, rather than estimating sample size. Option b is incorrect because the manual says the error-rate procedure often gives more reliable estimates than this module, because the error-rate procedure builds in the zero values expected in the error population. Option d is incorrect because it obtains an estimate of the universe mean and standard deviation from a probe sample file or from direct entry, not from an error rate.

Correct Answer: I, II, and III

The module serves two situations: choosing sample size for an unrestricted or stratified variable appraisal, and distributing an already determined overall sample among strata. Picking a random number seed is not described as one of its purposes. Option a is incorrect because leaves out the second situation the module supports, which is allocating an already determined overall sample among the strata. Option c is incorrect because includes random number seed selection, which is not a function of this module, and omits the unrestricted appraisal it also supports. Option d is incorrect because drops the stratified appraisal, which the module covers, and adds seed selection, a task the module does not describe.

Correct Answer: 1128(b)(7) exclusion authority

OIG often concludes exclusion is unnecessary when a person agrees to appropriate integrity obligations. In those cases OIG requires integrity obligations in exchange for a release of its 1128(b)(7) exclusion authority. Option a is incorrect because it is a point on the risk spectrum between exclusion and integrity obligations, not an authority that OIG releases in return for obligations. Option b is incorrect because referrals to CMS contractors are a monitoring step OIG takes where integrity provisions are absent; the CIA does not exchange them for a release. Option d is incorrect because it is the authority OIG uses for unilateral monitoring, such as audits and investigations, when a person refuses a CIA, so it is not released.