LATEBREAKER
Cancer
Pre-diagnostic Symptom Related Healthcare Utilization in Ovarian Cancer Patients: Insights from Latent Class Analysis Akalya Villenthi* Akalya Villenthi May Kuo Dawn Michelle Ritzwoller Sharon Peacock Hinton Stephanie B Wheeler Christopher Baggett Victoria L Bae-Jump Caroline A Thompson
Ovarian cancer (OCa) is the 6th leading cause of cancer-related mortality in U.S. women due to frequent late stage at diagnosis. While there is no recommended screening test, a better understanding of the pre-diagnostic symptom burden could provide evidence about how to improve earlier diagnosis. We used Latent Class Analysis (LCA) a statistical method to identify subgroups in high dimensional data, from analyzing symptom-related pre-diagnostic healthcare utilization (HCU) in a population-based sample of OCa patients residing in North Carolina. Adult OCa patients diagnosed 2009-2019 with 12+ months pre-diagnostic continuous insurance enrollment were identified from a linkage of statewide cancer registry and multiple-payer claims. We classified potential OCa symptoms using ICD codes and identified the provider type visited for the earliest symptom-related HCU. The LCA model included sociodemographic variables, comorbidity scores, OCa symptoms, and earliest provider type visited. We used logistic regression to estimate the relationship between class and stage at diagnosis. The study population included 2147 OCa patients who had symptom related pre-diagnostic HCU,64% were diagnosed with late-stage tumors, and 16% presented to the emergency department (ED). Among black patients (13%), 70% were late stage, and 24% presented to the ED. LCA identified two classes, Class 1 (N = 1165, 54%): younger women with primarily abdomen-specific symptoms and lower comorbidity scores who more often presented to primary care, and Class 2 (N = 982, 46%): older women with higher non-specific symptom burden and multiple comorbidities who more often presented to the ED. Notably, 73% of Black patients and 64% of lower income patients belong to Class 2 (vs. 42% white and 34% highest income). Class 2 patients had 90% higher odds of late-stage OCa(OR:1.9; 95% CI:1.6-2.3). Our findings provide valuable insights for tailoring healthcare interventions aimed at enhancing the early detection of OCa.