Validation of a Rule-Based ICD-10-CM Algorithm to Detect Fall Injuries in Medicare Data

BACKGROUND

Diagnosis-code-based algorithms to identify fall injuries in Medicare data are useful for ascertaining outcomes in interventional and observational studies. However, these algorithms have not been validated against a fully external reference standard, in ICD-10-CM, or in Medicare Advantage (MA) data.

METHODS

We linked self-reported fall injuries leading to medical attention (FIMA) from the Strategies to Reduce Injuries and Develop Confidence in Elders (STRIDE) trial (reference standard) to Medicare fee-for-service (FFS) and MA data from 2015-2019. We measured the area under the receiver operating characteristic curve (AUC) based on sensitivity and specificity of a diagnosis-code-based algorithm against the reference standard for presence or absence of ≥1 FIMA within a specified window of dates, varying the window size to obtain points on the curve. We stratified results by source (FFS versus MA), trial arm (intervention versus control), and STRIDE's ten participating healthcare systems.

RESULTS

Both reference standard data and Medicare data were available for 4941 (of 5451) participants. The reference standard and algorithm identified 2054 and 2067 FIMA, respectively. The algorithm had 45% sensitivity (95% confidence interval [CI], 43%-47%) and 99% specificity (95% CI, 99%-99%) to identify reference standard FIMA within the same calendar month. The AUC was 0.79 (95% CI, 0.78-0.81) and was similar by FFS or MA data source or trial arm, but showed variation among STRIDE healthcare systems (AUC range by healthcare system, 0.71 to 0.84).

CONCLUSIONS

An ICD-10-CM algorithm to identify fall injuries demonstrated acceptable performance against an external reference standard, in both MA and FFS data.

An assessment of completeness and medical coding of Medicare Advantage hospitalizations in two national data sets

OBJECTIVE

To compare the Encounter Data System (EDS) and Medicare Provider Analysis and Review (MedPAR) completeness and medical coding of Medicare Advantage hospitalizations.

DATA SOURCES

FY 2016-FY 2019 data limited to hospitals paid under Medicare's Inpatient Prospective Payment System.

STUDY DESIGN

Secondary data analysis.

DATA COLLECTION/EXTRACTION METHODS

Completeness of EDS and MedPAR data was estimated using the total number of unique hospitalizations in both data sources as denominator. Deriving this denominator involved matching cases in the EDS and MedPAR by MA enrollee, discharge date, and hospital. The higher the match rate, the more informative the comparison of EDS and MedPAR medical coding of the same hospitalization. EDS and MedPAR codes were assessed for similarity on six measures of Medicare Severity Diagnosis-Related Group (MS-DRG) assignment and identical diagnosis and procedure codes.

PRINCIPAL FINDINGS

EDS hospitalizations' completeness increased steadily each year from 90% to 93%, driven by the 23 largest Medicare Advantage Organizations, which account for 83% of total cases. MedPAR completeness was relatively stable (89%) and benefited from 91% completeness among the largest hospitals, which are often teaching hospitals and account for 63% of MedPAR cases. By 2019, 97% of medical cases were assigned the same MS-DRG, indicating the high consistency of the severity level coding, since 98% were assigned the same base MS-DRGs, which include all severity levels for the same condition. Without chart reviews, medical cases with identical diagnosis codes increased from 87% to 92%.

CONCLUSIONS

The EDS has a completeness advantage over MedPAR for studies of non-teaching disproportionate share (DSH) hospitals and individual hospitals generally. MedPAR is only slightly less complete for hospitalizations of teaching DSH hospitals and large hospitals in general. A highly consistent EDS and MedPAR medical coding of matched cases is an important finding since the matched cases are 88% of EDS and 90% of MedPAR cases.

Assessing the usability of MAX 2008 encounter data for comprehensive managed care

BACKGROUND

As growing numbers of Medicaid enrollees receive health benefits through comprehensive managed care, researchers and policymakers seeking to understand the service use of these enrollees must rely on encounter data.

OBJECTIVE

To assess the availability, completeness, and quality of physician, clinic, and outpatient service (OT), inpatient (IP), and prescription drug (RX) encounter claims to judge the usability of the 2008 Medicaid Analytical eXtract (MAX) encounter data.

DATA

2008 MAX encounter data, which are derived from the state-submitted Medicaid Statistical Information System (MSIS) files.

METHODS

For each basis of eligibility (BOE) group in each state that had at least ten percent participation in comprehensive managed care and submitted at least 200 encounter claims, the completeness and quality of the OT, IP, and RX encounter data were evaluated using comparison metrics created from the full-benefit, non-dual fee-for-service (FFS) population across all states with substantial FFS participation. Data that met both the completeness and quality criteria were considered usable.

RESULTS

The completeness and the quality of the encounter data were high. The encounter data were considered usable for a least one BOE category for 22 of the 25 states that submitted OT encounter data, 20 of the 24 states that submitted IP data, and 13 of the 15 states that submitted RX data.

CONCLUSIONS

Most states that have comprehensive managed care plans are reporting OT, IP, and RX encounter data. Of those data, the majority are complete and of comparable quality to FFS data for adults, children, the disabled, and aged populations.