(298) Discrepancies in ICD Codes Used to Identify Three Common Diseases in Cancer Patients in Real-World Settings and Their Implications for Disease Classification in Patients Without Cancer: A Literature Review and Descriptive Study

Publication Number: 972

Background: International Classification of Diseases (ICD) codes are frequently used in the United States by health insurers and disease registries, and are often recorded in electronic medical records. Due to their widespread use, ICD codes are a valuable source of data for epidemiology studies, but there are challenges related to the accuracy and reliability of their use.

Objectives: This study aims to 1) identify ICD-9/ICD-10 codes reported in the literature/other online sources used to identify three common diseases in elderly patients with cancer (anemia, hypertension, arthritis), 2) compare ICD codes identified in the literature/web search to a code list provided by SEER-Medicare (considered the “gold-standard” in this study) to determine their discordance, and 3) evaluate numerical differences in disease classification in a cohort of non-cancer SEER-Medicare patients using the literature/web search codes compared to the SEER-Medicare codes.

Methods: A literature search was performed in Ovid (no time limits) to find sources that reported ICD codes for at least one disease of interest. Articles were screened in two levels (title/abstract; full text). Comparisons between ICD code lists and evaluation of discrepancies in patient disease classification were performed in SAS Version 9.4.

Results: Of 106 references identified, 29 reported ICD codes and were included. Overall, 884 codes were reported (180 ICD-9; 704 ICD-10): 155 anemia, 80 hypertension, and 649 arthritis codes. A high degree of discordance was seen between the SEER-Medicare and literature/web search code lists. Overall discordance was 32.9% (22.2% for ICD-9; 35.7% for ICD-10). The SEER-Medicare list contained several codes not found in literature/web sources, including certain codes for hypertensive retinopathy/encephalopathy, Page Kidney, spondylosis/spondylitis, juvenile arthritis, thalassemia, sickle cell disorder, autoimmune anemias, and erythroblastopenia. When classifying a cohort of SEER-Medicare non-cancer patients (N=684,376) with the literature/web search versus the SEER-Medicare codes, the latter classified an additional 129 patients with anemia, 33,683 with arthritis, and 510 with hypertension compared to the former.

Conclusions: There were large discrepancies in ICD codes used to identify three common diseases in real-world settings that resulted in differences in disease classification. In all cases, the SEER-Medicare code list captured patients who were missed (and potentially misclassified) using the literature/web search codes. Researchers should use standardized and/or validated coding algorithms when available to increase consistency in research and reduce the risk of patient misclassification, which can significantly alter the findings of a study.