Evaluating Data Sources for Monitoring Guideline-Concordant Endometrial Cancer Treatment Among Older Adults
Caitlin E. Meade, Jennifer A. Sinnott, Simran A. Kanal, Allison M. Quick, Kemi M. Doll, Ashley S. FelixBackground:
Guideline-concordant endometrial cancer treatment – care consistent with recommended treatment based on tumor characteristics – is critical for survival. However, cancer registry-based assessments typically determine concordance based on whether a recommended treatment modality was received, without clinical details needed to assess treatment adequacy, including appropriate timing, amount, and duration. These details are observable in claims, raising the possibility that registry-based measures of guideline-concordance may misclassify guideline-concordant treatment. We compared guideline-concordant treatment classifications based on claims data, which capture both treatment receipt and adequacy, with classifications based on SEER registry data, which capture treatment receipt but not adequacy.
Methods:
Older patients diagnosed with endometrial cancer from 2000-2019 were identified from the Surveillance, Epidemiology, End Results (SEER)-Medicare linked dataset. For Medicare claims, observed surgery and adjuvant treatment were classified using treatment receipt and adequacy criteria, including timing, number of chemotherapy cycles or radiation fractions, and duration. For SEER, observed treatment was classified using registry-coded treatment receipt without adequacy criteria. For each data source, observed treatment was then compared with treatment recommended by the National Comprehensive Cancer Network (NCCN) according to year of diagnosis, stage, histology, and grade to determine guideline concordance. Agreement between data sources was assessed using kappa statistics.
Results:
Among 29,209 patients for surgery and 14,372 patients for adjuvant treatment, guideline-concordant care was more frequent in SEER than Medicare claims (surgery: 88.2% vs. 79.8%; adjuvant treatment: 85.4% vs. 71.6%). Agreement between SEER and Medicare was moderate for both surgery (kappa=0.46) and adjuvant treatment (kappa=0.47).
Conclusions:
SEER-based measures that rely solely on treatment fields may overestimate guideline-concordant treatment for older patients with endometrial cancer by failing to account for treatment adequacy.