Validity of Diagnostic Codes to Identify Metastatic Prostate Cancer in Medicare Claims Database
Author(s)
Bo Zhou, PhD1, Amit D. Raval, PhD2, Yifan Zhang, MPH, PHD1, Nethra Sambamoorthi, PhD1, Usha Sambamoorthi, MA, PhD1;
1University of North Texas Health Science Center, Fort Worth, TX, USA, 2Bayer Healthcare Pharmaceuticals, Piscataway, NJ, USA
1University of North Texas Health Science Center, Fort Worth, TX, USA, 2Bayer Healthcare Pharmaceuticals, Piscataway, NJ, USA
OBJECTIVES: Claims are a common source of real-world data and often rely on diagnostic codes to identify medical conditions. The validity of codes to identify conditions is critical to general real-world evidence. Prostate cancer (PC) is the most common cancer in men in the United States, with a rise in the incidence of metastases at diagnosis. With the linkage of Medicare claims and the Surveillance, Epidemiology, and End-Results (SEER) registry, this study aimed to examine the validity of diagnostic codes for metastatic PC (mPC) diagnosis in Medicare claims.
METHODS: The study utilized a retrospective cohort of older men (age≥ 66 years) diagnosed with PC between 2016 and 2019 using the SEER-Medicare data. Men with continuous enrollment in Medicare fee-for-service Parts A and B for ≥ 2 months pre- and post-PC diagnosis were included. Claims-based diagnosis of mPC was confirmed using the International Classification of Disease (ICD)-10 diagnosis codes for metastases within 2 months pre and post- (registry-based) PC diagnosis date. Using registry-based mPC as a gold standard measure, validity parameters were estimated for claims-based mPC. Logistic regression was used to examine factors (social determinants of health: age, race, marital status, county-level socioeconomic index) associated with the discordance in mPC diagnosis.
RESULTS: The study cohort comprised 72,840 men diagnosed with PC. A total of 6,684 (9.2%) had registry-based mPC, and 5,899 (8.1%) had claims-based mPC at diagnosis. The sensitivity, specificity, positive-predictive value, negative predictive value, and accuracy for claims-based mPC were 68.9%, 98.0%, 78.1%, 96.8%. and 95.3%, respectively. Older men (85+ years), African Americans, and low socio-economic index were key factors for discordance in mPC diagnosis.
CONCLUSIONS: Findings highlight the claims-based measure (using ICD-10-based codes) can accurately identify mPC at diagnosis among older men with prostate cancer, albeit a bit under-coding in claims, resulting in moderate sensitivity.
METHODS: The study utilized a retrospective cohort of older men (age≥ 66 years) diagnosed with PC between 2016 and 2019 using the SEER-Medicare data. Men with continuous enrollment in Medicare fee-for-service Parts A and B for ≥ 2 months pre- and post-PC diagnosis were included. Claims-based diagnosis of mPC was confirmed using the International Classification of Disease (ICD)-10 diagnosis codes for metastases within 2 months pre and post- (registry-based) PC diagnosis date. Using registry-based mPC as a gold standard measure, validity parameters were estimated for claims-based mPC. Logistic regression was used to examine factors (social determinants of health: age, race, marital status, county-level socioeconomic index) associated with the discordance in mPC diagnosis.
RESULTS: The study cohort comprised 72,840 men diagnosed with PC. A total of 6,684 (9.2%) had registry-based mPC, and 5,899 (8.1%) had claims-based mPC at diagnosis. The sensitivity, specificity, positive-predictive value, negative predictive value, and accuracy for claims-based mPC were 68.9%, 98.0%, 78.1%, 96.8%. and 95.3%, respectively. Older men (85+ years), African Americans, and low socio-economic index were key factors for discordance in mPC diagnosis.
CONCLUSIONS: Findings highlight the claims-based measure (using ICD-10-based codes) can accurately identify mPC at diagnosis among older men with prostate cancer, albeit a bit under-coding in claims, resulting in moderate sensitivity.
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
RWD140
Topic
Real World Data & Information Systems
Topic Subcategory
Health & Insurance Records Systems
Disease
SDC: Oncology, SDC: Urinary/Kidney Disorders