Processing and validation of inpatient Medicare Advantage data for use in hospital outcome measures.

OBJECTIVE: To determine the feasibility of integrating Medicare Advantage (MA) admissions into the Centers for Medicare & Medicaid Services (CMS) hospital outcome measures through combining Medicare Advantage Organization (MAO) encounter- and hospital-submitted inpatient claims. DATA SOURCES AND STUDY SETTING: Beneficiary enrollment data and inpatient claims from the Integrated Data Repository for 2018 Medicare discharges. STUDY DESIGN: We examined timeliness of MA claims, compared diagnosis and procedure codes for admissions with claims submitted both by the hospital and the MAO (overlapping claims), and compared demographic characteristics and principal diagnosis codes for admissions with overlapping claims versus admissions with a single claim. DATA COLLECTION/EXTRACTION METHODS: We combined hospital- and MAO-submitted claims to capture MA admissions from all hospitals and identified overlapping claims. For admissions with only an MAO-submitted claim, we used provider history data to match the National Provider Identifier on the claim to the CMS Certification Number used for reporting purposes in CMS outcome measures. PRINCIPAL FINDINGS: After removing void and duplicate claims, identifying overlapped claims between the hospital- and MAO-submitted datasets, restricting claims to acute care and critical access hospitals, and bundling same admission claims, we identified 5,078,611 MA admissions. Of these, 76.1% were submitted by both the hospital and MAO, 14.2% were submitted only by MAOs, and 9.7% were submitted only by hospitals. Nearly all (96.6%) hospital-submitted claims were submitted within 3 months after a one-year performance period, versus 85.2% of MAO-submitted claims. Among the 3,864,524 admissions with overlapping claims, 98.9% shared the same principal diagnosis code between the two datasets, and 97.5% shared the same first procedure code. CONCLUSIONS: Inpatient MA data are feasible for use in CMS claims-based hospital outcome measures. We recommend prioritizing hospital-submitted over MAO-submitted claims for analyses. Monitoring, data audits, and ongoing policies to improve the quality of MA data are important approaches to address potential missing data and errors.

Incorporating Medicare Advantage Admissions Into the CMS Hospital-Wide Readmission Measure.

Importance: Medicare Advantage (MA) enrollment is rapidly expanding, yet Centers for Medicare & Medicaid Services (CMS) claims-based hospital outcome measures, including readmission rates, have historically included only fee-for-service (FFS) beneficiaries. Objective: To assess the outcomes of incorporating MA data into the CMS claims-based FFS Hospital-Wide All-Cause Unplanned Readmission (HWR) measure. Design, Setting, and Participants: This cohort study assessed differences in 30-day unadjusted readmission rates and demographic and risk adjustment variables for MA vs FFS admissions. Inpatient FFS and MA administrative claims data were extracted from the Integrated Data Repository for all admissions for Medicare beneficiaries from July 1, 2018, to June 30, 2019. Measure reliability and risk-standardized readmission rates were calculated for the FFS and MA cohort vs the FFS-only cohort, overall and within specialty subgroups (cardiorespiratory, cardiovascular, medicine, surgery, neurology), then changes in hospital performance quintiles were assessed after adding MA admissions. Main Outcome and Measure: Risk-standardized readmission rates. Results: The cohort included 11 029 470 admissions (4 077 633 [37.0%] MA; 6 044 060 [54.8%] female; mean [SD] age, 77.7 [8.2] years). Unadjusted readmission rates were slightly higher for MA vs FFS admissions (15.7% vs 15.4%), yet comorbidities were generally lower among MA beneficiaries. Test-retest reliability for the FFS and MA cohort was higher than for the FFS-only cohort (0.78 vs 0.73) and signal-to-noise reliability increased in each specialty subgroup. Mean hospital risk-standardized readmission rates were similar for the FFS and MA cohort and FFS-only cohorts (15.5% vs 15.3%); this trend was consistent across the 5 specialty subgroups. After adding MA admissions to the FFS-only HWR measure, 1489 hospitals (33.1%) had their performance quintile ranking changed. As their proportion of MA admissions increased, more hospitals experienced a change in their performance quintile ranking (147 hospitals [16.3%] in the lowest quintile of percentage MA admissions; 408 [45.3%] in the highest). The combined cohort added 63 hospitals eligible for public reporting and more than 4 million admissions to the measure. Conclusions and Relevance: In this cohort study, adding MA admissions to the HWR measure was associated with improved measure reliability and precision and enabled the inclusion of more hospitals and beneficiaries. After MA admissions were included, 1 in 3 hospitals had their performance quintile changed, with the greatest shifts among hospitals with a high percentage of MA admissions.

Calculation of Overall Hospital Quality Star Ratings With and Without Inclusion of the Peer Grouping Step.

Importance: The Centers for Medicare & Medicaid Services (CMS) Overall Star Rating is widely used by patients and consumers, and there is continued stakeholder curiosity surrounding the inclusion of a peer grouping step, implemented to the 2021 Overall Star Rating methods. Objective: To calculate hospital star rating scores with and without the peer grouping step, with the former approach stratifying hospitals into 3-, 4-, and 5-measure group peer groups based on the number of measure groups with at least 3 reported measures. Design, Setting, and Participants: This cross-sectional study used Care Compare website data from January 2023 for 3076 hospitals that received a star rating in 2023. Data were analyzed from April 2023 to December 2023. Exposure: Peer grouping vs no peer grouping. Main Outcomes and Measures: The primary outcome was the distribution of star ratings, with 1 star being the lowest-performing hospitals and 5 stars, the highest. Analyses additionally identified the number of hospitals with a higher, lower, or identical star rating with the use of the peer grouping step compared with its nonuse, stratified by certain hospital characteristics. Results: Among 3076 hospitals that received a star rating in 2023, most were nonspecialty (1994 hospitals [64.8%]), nonteaching (1807 hospitals [58.7%]), non-safety net (2326 hospitals [75.6%]), non-critical access (2826 hospitals [91.9%]) hospitals with fewer than 200 beds (1822 hospitals [59.2%]) and located in an urban geographic designations (1935 hospitals [62.9%]). The presence of the peer grouping step resulted in 585 hospitals (19.0%) being assigned a different star rating than if the peer grouping step was absent, including considerably more hospitals receiving a higher star rating (517 hospitals) rather than a lower (68 hospitals) star rating. Hospital characteristics associated with a higher star rating included urbanicity (351 hospitals [67.9%]), non-safety net status (414 hospitals [80.1%]), and fewer than 200 beds (287 hospitals [55.6%]). Collectively, the presence of the peer grouping step supports a like-to-like comparison among hospitals and supports the ability of patients to assess overall hospital quality. Conclusions and Relevance: In this cross-sectional study, inclusion of the peer grouping in the CMS star rating method resulted in modest changes in hospital star ratings compared with application of the method without peer grouping. Given improvement in face validity and the close association between the current peer grouping approach and stakeholder needs for peer-comparison, the current CMS Overall Star Rating method allows for durable comparisons in hospital performance.

Assessing the accuracy of race-and-ethnicity data in the Outcome and Assessment Information Set.

BACKGROUND: Limitations in the quality of race-and-ethnicity information in Medicare's data systems constrain efforts to assess disparities in care among older Americans. Using demographic information from standardized patient assessments may be an efficient way to enhance the accuracy and completeness of race-and-ethnicity information in Medicare's data systems, but it is critical to first establish the accuracy of these data as they may be prone to inaccurate observer-reported or third-party-based information. This study evaluates the accuracy of patient-level race-and-ethnicity information included in the Outcome and Assessment Information Set (OASIS) submitted by home health agencies. METHODS: We compared 2017-2022 OASIS-D race-and-ethnicity data to gold-standard self-reported information from the Medicare Consumer Assessment of Healthcare Providers and Systems® survey in a matched sample of 304,804 people with Medicare coverage. We also compared OASIS data to indirect estimates of race-and-ethnicity generated using the Medicare Bayesian Improved Surname and Geocoding (MBISG) 2.1.1 method and to existing Centers for Medicare & Medicaid Services (CMS) administrative records. RESULTS: Compared with existing CMS administrative data, OASIS data are far more accurate for Hispanic, Asian American and Native Hawaiian or other Pacific Islander, and White race-and-ethnicity; slightly less accurate for American Indian or Alaska Native race-and-ethnicity; and similarly accurate for Black race-and-ethnicity. However, MBISG 2.1.1 accuracy exceeds that of both OASIS and CMS administrative data for every racial-and-ethnic category. Patterns of inconsistent reporting of racial-and-ethnic information among people for whom there were multiple observations in the OASIS and Consumer Assessment of Healthcare Providers and Systems (CAHPS) datasets suggest that some of the inaccuracies in OASIS data may result from observation-based reporting that lessens correspondence with self-reported data. CONCLUSIONS: When health record data on race-and-ethnicity includes observer-reported information, it can be less accurate than both true self-report and a high-performing imputation approach. Efforts are needed to encourage collection of true self-reported data and explicit record-level data on the source of race-and-ethnicity information.