12.2 PEPPER, Forecasting, Hospital Financial Data, and Public Reporting

12.2 PEPPER, Forecasting, Hospital Financial Data, and Public Reporting

Quick Answer: The Program for Evaluating Payment Patterns Electronic Report (PEPPER) is a Centers for Medicare & Medicaid Services (CMS) contractor comparative billing report: a hospital-specific Microsoft Excel file of Medicare statistics for CMS-chosen target areas. Compare your percents with the nation, Medicare Administrative Contractor (MAC) jurisdiction, and state. High outlier status is commonly at or above the 80th percentile; coding-focused areas also use a 20th percentile low-outlier lens. An outlier is not a payment error. Forecast CMI and query impact with spreadsheets and current Inpatient Prospective Payment System (IPPS) relative weights. Documentation and coded claims also feed public sites such as Care Compare (formerly Hospital Compare), Leapfrog, and Healthgrades. Always use the current PEPPER user guide for target definitions; do not memorize an unpublished or outdated list.

The ACDIS inpatient CCDS outline asks for a working knowledge of PEPPER data, the ability to create forecasting data, basic computer and spreadsheet skills, hospital-specific financial data, and an explanation of how physician documentation affects publicly reported data (the handbook examples include Leapfrog and Healthgrades). This independent OpenExamPrep section teaches those analysis tools for inpatient IPPS CDI. It is not an ACDIS product.

What PEPPER is—and what it is not

PEPPER summarizes provider-specific Medicare claims statistics for discharges and services that CMS treats as vulnerable to improper payment because of billing, Medicare Severity Diagnosis Related Group (MS-DRG) coding, and/or admission-necessity issues. Target areas are determined by CMS. Each target is a ratio: a numerator of discharges CMS wants watched, over a denominator that is the larger related group. The file typically lets you review several federal fiscal quarters—often described as about three years of statistics—so you can see movement, not one isolated percent.

Authorized users download the hospital's report from the CMS PEPPER portal. CDI specialists usually receive a copy from compliance, health information management, or finance rather than administering portal passwords. Short-term acute care (ST) PEPPER is the flavor that matches IPPS hospitals on this exam. Critical access, long-term, and post-acute PEPPERs exist; do not mix their target lists into an inpatient CCDS answer.

PEPPER does not identify the presence of payment errors. It does not recoup money. It does not replace a Recovery Audit Contractor review. Formal tests of statistical significance are not used to assign outlier status. Treat it as a prioritization tool for auditing and monitoring.

Percentiles, comparison groups, and the method

For each target area, PEPPER compares the hospital's target percent with other hospitals in the nation, the MAC jurisdiction, and the state. Current short-term user guides describe a Compare Targets view in which a high outlier (at or above the 80th percentile) is emphasized (commonly red bold type) and a low outlier on coding-focused targets (at or below the 20th percentile) is emphasized (commonly green italics). If fewer than 11 hospitals in the state or jurisdiction have reportable data for that target, those percentiles may display as zero. Reports display target-area data only when the release's minimum case-count threshold is met. No reportable data for a quarter may appear as unavailable, not as a hidden passing score.

Method, not memorized appendix. CMS has paused, limited, and later expanded short-term PEPPER releases; target-area sets change. A slide deck from a prior year is not the current report. The CCDS skill is:

1. Obtain the current Excel PEPPER and that same release's user guide.
2. Read each target's numerator and denominator for that release.
3. Note whether the target is coding-focused or admission-necessity-focused; the audit sample differs.
4. Compare your percent with nation, jurisdiction, and state, and look at the trend across quarters.
5. If you are an outlier, sample records. For coding-focused targets, review documentation, code assignment, and CC/MCC logic. For admission-necessity targets, review medical necessity of the inpatient stay and setting. Do not "fix" a percentile by changing codes without a record that supports the change.
6. Record the sample findings in a spreadsheet and remeasure on the next PEPPER.

You may see coding-focused families such as pneumonia, septicemia, stroke-related groupings, ventilator support, or CC/MCC concentration on a given release, and admission-necessity families such as short stays or selected procedures. Those examples illustrate types. They are not a promise that every named family appears on the sitting or on next quarter's file. If a question asks for the current national target list, the correct move is to consult the current user guide, not to recite a frozen inventory.

Forecasting CMI and query impact

Forecasting in CDI is a planning range, not a promise to the chief financial officer. Start from cases, not from a slogan.

Hospital Medicare CMI for a period is the sum of MS-DRG relative weights for the included discharges, divided by the number of those discharges. Relative weights come from the current fiscal year IPPS tables CMS publishes. Do not invent a weight, and do not reuse last year's table after the October update without checking.

A simple query-impact sketch uses case-level rows:

• Working MS-DRG and its relative weight.
• Plausible alternative MS-DRG if a pending query is answered and coded as expected.
• Delta = alternative weight minus working weight.
• Expected delta = delta × historical response rate × historical agreement rate for that query type (or a conservative range).
• Forecast CMI movement ≈ sum of expected deltas ÷ expected discharges.

Work a numeric skeleton without inventing a published weight. Suppose 1,000 Medicare discharges and a current CMI of 1.6200. If a set of heart-failure queries is expected, after response and agreement, to add a total of 8.0 relative-weight points across the quarter, the sketched CMI lift is 8.0 / 1,000 = 0.0080, to about 1.6280. If response is weaker than last quarter, rerun the same sheet at a lower response assumption. Report a range, and list what would falsify it: CC exclusion, a coder who cannot assign the diagnosis, a clinical disagreement, or a later denial.

Never multiply "queries sent" by an unpublished national conversion factor. Never substitute OpenExamPrep practice-bank size for CMI. Never treat PEPPER outlier status as a dollar figure.

Basic spreadsheet skills that the outline actually tests

The content outline names basic Excel-style skills because PEPPER arrives as a workbook and because physician and CMI trends live in tables. You are not sitting an Excel certification. You should be able to:

• Keep one row per query or per discharge with dates, provider, response type, working and final MS-DRG, relative weight, and an unspecified-code flag.
• Compute response rate with counts (for example, count of non-blank response dates divided by count of queries).
• Use a PivotTable (or equivalent) to slice by provider, service, and month.
• Calculate percent change: (new − old) / old.
• Chart CMI, non-response volume, and a PEPPER target percent across quarters.
• Filter and sort; freeze the header row; avoid hard-coding a target-area name that a later PEPPER release retired.

Those functions turn a dump of queries into a trend. A screenshot of last month's dashboard is not forecasting.

Hospital-specific financial data

"Hospital-specific" means this facility's numbers, not a national blog average. IPPS operating payment for a case is conceptually a standardized amount (base rate) times the MS-DRG relative weight, with geographic and other adjustments CMS publishes. Finance may share a blended average payment or contribution figure; use that figure if you are allowed to see it. Do not guess a base rate.

Other financial files CDI actually uses include documentation- or coding-related denial dollars, length of stay in PEPPER-flagged MS-DRG families, and the share of Medicare volume in high-weight families you are educating on. The point of applying financial data is to prioritize education and audit time, and to explain why complete documentation changes both payment and the coded dataset that quality reports consume. It is not a license to write reimbursement language into a query.

Public reporting: Care Compare, Leapfrog, Healthgrades

Physician documentation becomes coded claims (and, for some measures, abstracted events). Public sites then republish hospital performance, often on a lagged measurement period. Always read the dates on the public page.

CMS Care Compare on Medicare.gov is the current consumer hospital-quality site. CMS first reported many of these measures on Hospital Compare; CDI materials and the CCDS outline's sibling credential still use that older name. Care Compare displays complications, mortality, readmissions, infection measures, patient experience, and overall star ratings, among other topics. Many of those measures use Medicare claims or other datasets that start with diagnoses, procedures, and POA indicators.

The Leapfrog Group publishes a Hospital Safety Grade (letter grades, issued on a regular cycle) and a separate hospital survey. Safety Grade methodology mixes process and outcome measures; several inputs draw on CMS data such as infections and patient-safety indicators. Incomplete or unspecified documentation, missing POA, and complication coding can move claims-based inputs even when bedside care did not change.

Healthgrades is a commercial rating and comparison publisher. It uses Medicare claims among other sources for many hospital rating products. CDI does not control Healthgrades' proprietary formulas, but CDI does influence the documented conditions that appear on the claims those formulas ingest.

Public reporting is not identical to CMS pay-for-performance programs such as Hospital Value-Based Purchasing or the Hospital-Acquired Condition Reduction Program; those formulas belong with Domain VIII. The Domain IV skill is that documentation and coded data are visible outside the hospital, so physician performance and PEPPER-style coding patterns have reputational as well as payment consequences.

Exam-style scenarios

Scenario: red bold, no denial letter. Compare Targets shows a coding-focused target in red bold at the 82nd national percentile. A colleague says CMS has already decided the hospital overcoded. The accurate reading is high-outlier status versus the comparison group, which is a reason to sample records. It is not a recoupment and not proof of error. If the sample shows supported CC/MCC diagnoses, the "fix" is not to strip codes to chase a lower percentile.

Scenario: last year's target list. A binder lists PEPPER targets from a release that predated a program pause and a later limited rebuild. Using that list to design this year's audits skips the method. Open the current user guide, confirm numerators, and only then pull the internal cases that match this definition.

Scenario: forecast as a guarantee. An analyst multiplies 200 queries by a homemade "$4,000 per MCC" factor and presents a single revenue number to finance. That is not CCDS-style forecasting. Use case-level relative-weight deltas from the current IPPS table, haircut by actual response and agreement, and present a range with assumptions.

Scenario: public grade shock. Leapfrog issues a lower Safety Grade while Care Compare still shows older measurement dates. The CDI question is which coded and abstracted inputs moved, and whether documentation of severity, complications, and POA in the period those sites actually used explains part of the change—not whether last week's queries have already appeared on Medicare.gov.

How to study this section

Recall: PEPPER is a CMS contractor comparative Excel report; 80th/20th percentile flags; outlier ≠ error; Care Compare is the current CMS hospital site formerly called Hospital Compare; Leapfrog and Healthgrades are additional public publishers influenced by documentation and claims. Application: given a percentile color or a CMI sketch, choose sample-and-trend rather than recoup-or-guarantee. Analysis: choose the current user guide over a memorized target list, and choose hospital-specific weights and rates over invented multipliers.