Review of health care access duration among cancer patients at a large health plan.

Abstract

e18758 Background: Cancer patients have unique, long-term health care needs and access to timely and affordable care is essential. Little is known about cancer patients’ propensity to leave their health plan during and after treatment. This study aimed to examine health care access duration among cancer survivors. Methods: Adult patients diagnosed with cancer between 11/1/2015 and 10/31/2016 were identified and followed until 11/1/2021. Control patients were identified as adults with no cancer diagnosis between 11/1/2015 and 11/1/2021. Patients with skin cancer were excluded due to limited care requirements. The primary outcomes were annual patient churn and median length of health plan enrollment duration. Propensity score (PS) matching was conducted between the groups utilizing patient demographics to balance comorbidities between groups. Comorbidity was described with the National Cancer Institute comorbidity index (NCI CI). Appropriate descriptive statistics were calculated and Cox proportional hazards (CPH) models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). Results: Overall, 3.2 million patients met the eligibility criteria, with 18% having cancer. Significant differences in the unmatched population were found, with the cancer group being older and having more comorbidities (p < 0.001). Following PS matching, the differences in underlying demographic variables were minimal, despite being statistically significant. Unadjusted annual churn rate, defined as the percent of cancer patients leaving the health plan per year, averaged 25.44% overall, with the non-cancer group experiencing more churn most years (Table). Median length of health plan enrollment duration was 716 days (Interquartile range (IQR) = 819) overall, with the cancer group having less median duration than the non-cancer group in the matched analysis (713 [IQR 808] vs. 717 [IQR 822], p < 0.001). After adjusting for the variables in the model, having a cancer diagnosis was associated with decreased risk of health plan disenrollment in multivariate CPH regression (HR 0.987 [95% CI 0.983 – 0.991]). Additional factors that increased the risk of disenrollment included: male gender (HR 1.011 [1.008-1.015]), Medicare eligibility (HR 1.290 [1.279-1.301]) and increasing NCI CI (1.124 [1.118-1.129]). Conclusions: Cancer patients experienced less churn from their health plan over a five-year follow-up. They were less at risk for disenrollment at any time than non-cancer patients when confounders are accounted for in multivariate modeling.[Table: see text]