1973. Estimating influenza vaccination status among Medicare beneficiaries: Comparison of billing claims and vaccination data obtained from medical records and self-report
Session: Poster Abstract Session: Clinical: Respiratory Track
Saturday, October 7, 2017
Room: Poster Hall CD
Posters
  • CMS_HAIVEN_IDWeek_02OCT17.pdf (673.9 kB)
  • Background: Accurate ascertainment of vaccination status is needed to evaluate influenza vaccine effectiveness (VE). Few studies have compared methods of identifying vaccination among hospitalized adults. We assessed accuracy of self-reported influenza vaccination status among adults hospitalized with acute respiratory illness with influenza vaccination billing claims submitted to the Centers for Medicare & Medicaid Services (CMS) and other vaccination records. 

    Methods: For Medicare enrollees ≥65 years old recruited for the 2015-16 US Hospitalized Adult Influenza Vaccine Effectiveness Network (HAIVEN) study, we compared self-reported influenza vaccination status with CMS claims and vaccination data collected from inpatient and outpatient medical records, pharmacies, and immunization registries (“HAIVEN vaccination records”). Measures were compared by kappa, sensitivity (Sn.) and specificity (Sp.).

    Results: Among 391 HAIVEN enrollees with continuous Medicare enrollment, influenza vaccine coverage was 75% by self-report, 66% by HAIVEN vaccination records and 51% by CMS claims [Tbl 1]. Of 293 patients self-reporting vaccination, 107 (37%) lacked CMS claims (Sn. = 63%) and 49 (17%) lacked HAIVEN records (Sn. = 83%). Of 259 patients with HAIVEN vaccination records, 87 (34%) lacked CMS claims (Sn. = 63%) [Fig 1]. Agreement between measures was moderate, with kappas of 0.38-0.61.

    Conclusion: In this sample of hospitalized older adults, influenza vaccination coverage estimates from CMS claims and other documented sources underestimated coverage compared to patient self-report. Reliance on billing claims could lead to substantial misclassification of vaccination status, suggesting that a lack of vaccination claims data should not be taken as evidence of lack of vaccination. Further improvement in understanding the accuracy of vaccination status data is needed for VE studies.

    Table 1. CMS and HAIVEN influenza vaccination status

    Tbl. 1A

     

    HAIVEN Vacc. Records
    No. Subjects (%)

     

    Yes

    No

    Total

    CMS

    Yes

    172 (66)

    27 (20)

    199 (51)

     

    No

    87 (34)

    105 (80)

    192 (49)

     

    Total

    259 (66)

    132 (34)

    391

    Tbl. 1B

     

    Self-Report

     

    Y

    N

    Total

    CMS

    Y

    186 (63)

    13 (13)

    199 (51)

     

    N

    107 (37)

    85 (87)

    192 (49)

     

    Total

    293 (75)

    98 (25)

    391

     

     

     

     

     

     

     

     

     

     

    Tbl. 1C

     

    Self-Report

     

    Y

    N

    Total

    HAIVEN

    Y

    244 (83)

    15 (15)

    259 (66)

    Vacc.

    N

    49 (17)

    83 (85)

    132 (34)

    Records

    Total

    293 (75)

    98 (25)

    391

     

     

     

    Elif Alyanak, MPH1, Alicia M. Fry, MD, MPH1, Courtney Strickland, MPH1, Jeffrey Kelman, MD MMSc2, Yoganand Chillarige, MPA3, Xiyuan Wu, MS4, Michael Wernecke, BA3, David Shay, MD, MPH1 and Jill M. Ferdinands, PhD, MSc1, (1)Influenza Division, Centers for Disease Control and Prevention, Atlanta, GA, (2)Centers for Medicare and Medicaid Services, Baltimore, MD, (3)Acumen, Inc., Burlingame, CA, (4)Acumen LLC, Burlingame, CA

    Disclosures:

    E. Alyanak, None

    A. M. Fry, None

    C. Strickland, None

    J. Kelman, None

    Y. Chillarige, None

    X. Wu, None

    M. Wernecke, None

    D. Shay, None

    J. M. Ferdinands, None

    Findings in the abstracts are embargoed until 12:01 a.m. PDT, Wednesday Oct. 4th with the exception of research findings presented at the IDWeek press conferences.