Researchers can apply for data use agreements with the Centers for Medicare and Medicaid Services (CMS) to acquire sensitive Research Identifiable File (RIF) data. Encounter RIFs contain administrative records submitted to CMS by Medicare Advantage organizations. Unique to these files are chart reviews. Along with service records, Medicare Advantage Organizations (MAOs) are allowed to submit adjustments to diagnoses reported on service records in the form of chart reviews for the purpose of risk adjustment(1). The purpose of this article is to explain how to identify chart reviews in encounter RIFs, and to give some sense of their impact via a few high-level descriptive tables.
Background
Part C encounter RIF data includes two types of records: service records and chart reviews. Medicare Advantage Organizations (MAOs) perform retrospective reviews of their beneficiaries’ medical records to identify diagnoses that were not originally submitted by providers or were submitted in error. MAOs may report diagnosis changes identified because of these reviews to CMS as chart reviews; that is, MAOs use chart reviews to add or delete diagnoses to members’ records for the purpose of risk adjustment. It has been shown that chart reviews have significantly increased coding intensity and inflated payments to some MAOs(2,3).
Descriptive Statistics
Detailed specifications for identifying and categorizing chart reviews are provided in Chapter 4 of the Medicare Encounter Data File User Guide(4). Here are some descriptive figures describing chart reviews in the 2023 RIFs.
The proportion of chart reviews varies significantly by RIF type (2023).
Table 1. Percent Chart Reviews by File Type (2023)
| Encounter File Type | Percent chart reviews |
|---|---|
| Inpatient | 45.1% |
| Carrier | 18.4% |
| Outpatient | 4.4% |
| SNF (Skilled Nursing Facilities) | 0.8% |
| HHA (Home Health Agencies) | 0.4% |
| DME (Durable Medical Equipment) | 0.2% |
Table 1 caption: Percent of chart reviews by file type.
In the 2023 Inpatient file, 17.1% of service records link to at least one chart review. Of those, 35.7% link to a single chart review and 64.3% to multiple. The following table contains a detailed account of chart reviews in this file.
Table 2. Inpatient Chart Review Counts by Category (2023)
| Inpatient File | Record count | Percent | Percent of chart reviews |
|---|---|---|---|
| Chart Review Records | 6,222,149 | 45.1% | 100.0% |
| Linked Chart Reviews | 3,305,826 | 23.9% | 53.1% |
| Add diagnoses to a service record | 2,933,836 | 21.3% | 47.2% |
| Delete diagnoses from a service record | 279,603 | 2.0% | 4.5% |
| Add diagnoses to a chart review record | 3,382 | 0.0% | 0.1% |
| Delete diagnoses from a chart review record | 89,005 | 0.6% | 1.4% |
| Replace a chart review record | 0 | 0.0% | 0.0% |
| Unlinked Chart Reviews | 2,916,323 | 21.1% | 46.9% |
| Original claim number not recorded on chart review | 2,799,777 | 20.3% | 45.0% |
| Original claim number not found in dataset | 116,546 | 0.8% | 1.9% |
| Service Records | 7,586,092 | 54.9% | |
| Total | 13,808,241 | 100.0% |
Table 2 caption: Chart review counts by category in the 2023 inpatient encounter RIF.
Curiously, there were no service records found in the file for 24% (135,000) of unique BENE_IDs (571,000) identified on unlinked chart reviews.
More Characteristics of Chart Reviews
- Some instances of chart reviews linking to records in prior annual RIFs have been found (e.g., ~6,400 linkages between the 2018-2017 IP RIFs).
- There are chart reviews that do not link to any other record. Such chart reviews can only add distinct diagnoses to a member’s list for the year.
- Chart reviews in a base RIF can be joined to the detail-level file (i.e., “line” or “revenue center” file) using the key variables BENE_ID, CLM_TYPE_CD and ENC_JOIN_KEY; however, the purpose of chart reviews is to amend diagnoses and all diagnoses are found in the base file, therefore the relevance of detail-level records for chart reviews is not apparent.
- Inpatient encounter records can list up to 25 diagnoses. In the 2023 inpatient encounter RIF, the median number of diagnosis codes on a service record that linked to an "add" chart review is about 20, whereas the median number of diagnoses on the "add” chart review is about 2, indicating MAOs are generally following guidelines to only report new diagnoses on chart reviews.
- For "add diagnoses" chart reviews, DGNS_CD1 is filled 100% of the time. The fill percentage decreases with each subsequent numbered diagnosis (e.g., DGNS_CD2 at 53% and DGNS_CD3 at 37%).
Impact of Chart Reviews on Member Diagnoses
In the 2023 IP encounter RIF, 23% of patients received at least one new diagnosis code because of chart reviews. The top diagnosis categories added are listed in the table below.
Table 3. Top Ten New Diagnosis Categories Added by Inpatient RIF Chart Reviews (2023)
| Diagnosis category | Diagnosis category | Percent |
|---|---|---|
| E11 | Type 2 diabetes mellitus | 6.9% |
| I50 | Heart failure | 6.1% |
| E78 | Disorders of lipoprotein metabolism and other lipidemias | 5.6% |
| I10 | Essential (primary) hypertension | 4.3% |
| I25 | Chronic ischemic heart disease | 3.6% |
| I48 | Atrial fibrillation and flutter | 3.1% |
| Z79 | Long term (current) drug therapy | 3.1% |
| J96 | Respiratory failure, not elsewhere classified | 2.7% |
| F32 | Major depressive disorder, single episode | 2.5% |
| J44 | Other chronic obstructive pulmonary disease | 2.4% |
| All others | 59.7% | |
| Total | 100.0% |
Table 3 caption: Top ten new diagnosis categories added by inpatient RIF chart reviews in 2023.
The impact of chart reviews varies by MAO. In 2023 a single MAO was responsible for 51% of members whose unique list of diagnoses increased because of chart reviews (while insuring 29% of all MA members); the next two highest MAOs contributed 18% and 8% respectively. In 2019 the highest MAO was responsible for 63% of these cases and the next highest 7% indicating a trend towards more MAOs using this mechanism more frequently.
Given the added complexity that chart reviews bring to a research plan, researchers should consider whether it’s necessary to include chart reviews in their analysis. For research questions that would not be impacted by diagnosis reviews, encounter service records alone would suffice.
- MedPAC (Medicare Payment Advisory Commission). Payment Basics: Medicare Advantage Program Payment System. 2025 Nov 18. Available from: https://www.medpac.gov/document-type/payment-basic.
- March 2020 Report to the Congress: Medicare Payment Policy, Chapter 13: The Medicare Advantage program: Status report. Available from: https://www.medpac.gov/document/http-www-medpac-gov-docs-default-source-reports-mar20_entirereport_sec-pdf/
- Meyers DJ, Trivedi AN. Medicare Advantage Chart Reviews Are Associated With Billions in Additional Payments for Some Plans. Med Care. 2021 Feb 1;59(2):96-100. Available from: https://pubmed.ncbi.nlm.nih.gov/32925467/
- Chronic Conditions Warehouse (CCW). Medicare Encounter Data File User Guide. Centers for Medicare & Medicaid Services. 2025 June. Available from: https://www2.ccwdata.org/web/guest/user-documentation.