Secondary Logo

Journal Logo

Letters

Evaluating the Impact of Increasing Allowable Inpatient Diagnoses in Medicare Claims Data

Connolly, John G.; Gagne, Joshua J.; Lin, Kueiyu Joshua

Author Information
doi: 10.1097/EDE.0000000000001138

To the Editor:

In January 2011, the Centers for Medicare & Medicaid Services (CMS) increased the number of allowable inpatient diagnosis and procedure codes per claim from nine and six, respectively, to 25 for each.1 A sudden expansion in the number of allowable inpatient codes could affect covariate or outcome measurement in claims-based studies.2 Specifically, it might lead to differential measurement of the covariate or outcome in the time periods before and after the policy change, resulting in confounding by calendar time in a comparative study or biased time trends in time series analyses. To quantify the potential differences in measurement, we used Medicare claims data to estimate the prevalence of several commonly researched medical conditions before and after code expansion.

Methods

We identified Medicare beneficiaries between 2008 and 2013 among patients receiving care from the largest healthcare provider system in Massachusetts, which comprises seven hospitals and more than 30 ambulatory centers. Eligible patients entered the study cohort after 180 days of continuous Medicare enrollment, during which they were required to have at least one medical encounter of any kind in any care setting recorded in the provider system. We assessed the mean yearly prevalence of 27 medical conditions across all major organ systems. We divided the 27 conditions into two categories: (1) 18 covariates based on both outpatient and inpatient codes (mimicking high-sensitivity variables; e.g., chronic conditions) and (2) nine covariates based on validated algorithms or published pharmacoepidemiologic studies that used only inpatient codes (mimicking high-specificity variables; e.g., acute outcomes).3–10 We classified the assessment year of the covariates based on the year of cohort entry, but the assessment was done in the 365 days following the cohort entry date. We then compared the yearly prevalence of each condition in the 3 years before (2008–2010) and 3 years after (2011–2013) code expansion.

We identified 119,766 eligible patients over the study period. The number of total monthly inpatient diagnoses increased by 45% from approximately 14,500 in December 2010 to approximately 21,000 in January 2011, although the corresponding numbers remained similar for outpatient diagnoses in December 2010 and January 2011 (eFigures 1; http://links.lww.com/EDE/B614 and 2; http://links.lww.com/EDE/B614). The yearly prevalence of each covariate over the study period is displayed in the Figure. None of the nine covariates defined using only inpatient codes increased in mean prevalence after the code expansion. Among the 18 covariates based on both outpatient and inpatient codes, we noted small increases in the prevalence of 4 covariates after the policy change, with the largest increase for all cancers, which rose from 46% in the period before code expansion to 48% after.

Figure
Figure:
Covariate prevalence by care setting and calendar year. ACUTE MI indicates acute myocardial infarction; AFIB, atrial fibrillation; AKI, acute kidney injury; CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; CRI, chronic renal insufficiency; DIAB, diabetes; DVT, deep vein thrombosis; HEM STROKE, hemorrhagic stroke; HEPATOTOX, hepatotoxicity; HF, heart failure; HIP FRACT, hip fracture; HYPERLIP, hyperlipidemia; HYPERTEN, hypertension; ISCH STROKE, ischemic stroke; LIVER, liver disease; PE, pulmonary embolism; PNEUM, pneumonia; PUD, peptic ulcer disease; UGIB, upper gastrointestinal bleed.

Despite a substantial increase in the number of total inpatient diagnoses following code expansion, there were no large increases in prevalence for the selected covariates as typically defined in pharmacoepidemiologic studies after the policy change. It is possible that we did not find an increase in prevalence based on inpatient codes because these selected medical conditions were serious health events that warranted coding under both the old and new code limits or because they were defined using the primary diagnosis position. The covariates based on outpatient and inpatient codes did not increase in prevalence possibly because these conditions tended to be recorded during an outpatient encounter that was unaffected by the policy change. Although the policy appeared to have little impact on the selected conditions, our results cannot rule out potential changes in prevalence for conditions not tested. It is also possible that true changes in the occurrence of these conditions over time could have modestly affected the observed trends. Nevertheless, these findings are reassuring for epidemiologists using CMS claims data, especially when study periods include time before and after 1 January 2011.

John G. Connolly

Joshua J. Gagne

Kueiyu Joshua Lin

Division of Pharmacoepidemiology and Pharmacoeconomics

Department of Medicine

Brigham and Women’s Hospital and Harvard Medical School

Boston, MA

joc121@mail.harvard.edu

REFERENCES

1. Centers for Medicare and Medicaid Services. Details for title: R2028CP. https://www.cms.gov/Regulations-and-Guidance/Guidance/Transmittals/2010-Transmittals-Items/CMS1238530.html?DLPage=1&DLEntries=10&DLFilter=processing%20additional&DLSort=1&DLSortDir=descending. Accessed May 16, 2019.
2. Tsugawa Y, Figueroa JF, Papanicolas I, Orav EJ, Jha AK. Assessment of strategies for managing expansion of diagnosis coding using risk adjustment methods for Medicare data. JAMA Intern Med. 2019
3. Wahl PM, Rodgers K, Schneeweiss S, et al. Validation of claims-based diagnostic and procedure codes for cardiovascular and gastrointestinal serious adverse events in a commercially-insured population. Pharmacoepidemiol Drug Saf. 2010;19:596–603.
4. Andrade SE, Harrold LR, Tjia J, et al. A systematic review of validated methods for identifying cerebrovascular accident or transient ischemic attack using administrative data. Pharmacoepidemiol Drug Saf. 2012;21(suppl 1):100–128.
5. Kiyota Y, Schneeweiss S, Glynn RJ, Cannuscio CC, Avorn J, Solomon DH. Accuracy of Medicare claims-based diagnosis of acute myocardial infarction: estimating positive predictive value on the basis of review of hospital records. Am Heart J. 2004;148:99–104.
6. Birman-Deych E, Waterman AD, Yan Y, Nilasena DS, Radford MJ, Gage BF. Accuracy of ICD-9-CM codes for identifying cardiovascular and stroke risk factors. Med Care. 2005;43:480–485.
7. Cushman M, Tsai AW, White RH, et al. Deep vein thrombosis and pulmonary embolism in two cohorts: the longitudinal investigation of thromboembolism etiology. Am J Med. 2004;117:19–25.
8. Myers RP, Leung Y, Shaheen AA, Li B. Validation of ICD-9-CM/ICD-10 coding algorithms for the identification of patients with acetaminophen overdose and hepatotoxicity using administrative data. BMC Health Serv Res. 2007;7:159.
9. Solomon DH, Rassen JA, Glynn RJ, Lee J, Levin R, Schneeweiss S. The comparative safety of analgesics in older adults with arthritis. Arch Intern Med. 2010;170:1968–1976.
10. Waikar SS, Wald R, Chertow GM, et al. Validity of international classification of diseases, ninth revision, clinical modification codes for acute renal failure. J Am Soc Nephrol. 2006;17:1688–1694.

Supplemental Digital Content

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.