RESUMO
Breast cancer is one of the most common cancers in women all over the world. Due to the improvement of medical treatments, most of the breast cancer patients would be in remission. However, the patients have to face the next challenge, the recurrence of breast cancer which may cause more severe effects, and even death. The prediction of breast cancer recurrence is crucial for reducing mortality. This paper proposes a prediction model for the recurrence of breast cancer based on clinical nominal and numeric features. In this study, our data consist of 1,061 patients from Breast Cancer Registry from Shin Kong Wu Ho-Su Memorial Hospital between 2011 and 2016, in which 37 records are denoted as breast cancer recurrence. Each record has 85 features. Our approach consists of three stages. First, we perform data preprocessing and feature selection techniques to consolidate the dataset. Among all features, six features are identified for further processing in the following stages. Next, we apply resampling techniques to resolve the issue of class imbalance. Finally, we construct two classifiers, AdaBoost and cost-sensitive learning, to predict the risk of recurrence and carry out the performance evaluation in three-fold cross-validation. By applying the AdaBoost method, we achieve accuracy of 0.973 and sensitivity of 0.675. By combining the AdaBoost and cost-sensitive method of our model, we achieve a reasonable accuracy of 0.468 and substantially high sensitivity of 0.947 which guarantee almost no false dismissal. Our model can be used as a supporting tool in the setting and evaluation of the follow-up visit for early intervention and more advanced treatments to lower cancer mortality.
RESUMO
BACKGROUND: Previous studies in Taiwan utilizing the Taiwan's National Health Insurance Database (NHIRD) have estimated the direct healthcare costs of RA patients, but they have not focused on patients on bDMARDs, or considered patients' response to therapy. OBJECTIVES: The objective of this study was to estimate the rate of inadequate response for patients newly treated with biologic disease-modifying antirheumatic drugs (bDMARDs) as well as their costs and resource use. METHODS: Data were from the catastrophic illness file within the NHIRD from 1/1/2009 to 12/31/2013. Patients with RA, which was categorized by the presence of a catastrophic illness card, that were previously bDMARD-naïve, were included in this study if they initiated their first bDMARD during the index period. The index period included all of 2010, a pre-index period consisting of the index date- 365 days, and a follow-up period including the index date to 365 days post-index, were also included. Previously biologically-naïve patients were indexed into the study on the date of their first claim for a bDMARD. A validated algorithm was used to examine the rate of inadequate response (IR) in the biologically-naïve cohort of patients. Inadequate responders met one or more of the following criteria during their year of follow-up: low adherence (proportion of days covered <0.80); switched to or added a second bDMARD; added a new conventional synthetic DMARD (csDMARD); received ≥1 glucocorticoid injection; or increased oral glucocorticoid dosing. All-cause mean annual direct costs and resource use were measured in the year of follow-up. Costs were converted from NT$ to USD using 1 NT$ = 0.033 USD. RESULTS: A total of 818 patients with RA initiated their first bDMARD (54% etanercept and 46% adalimumab) in 2010. After one year of follow-up, 32% (n = 258) were classified as stable, 66% (n = 540) had an IR, and 2% (n = 20) were lost to follow-up. During the follow-up period mean annual total direct costs were $16,136 for stable patients compared to $14,154 for patients with IR. Mean annual non-medication direct costs were $937 for stable patients and $1,574 for patients with IR. Mean annual hospitalizations were higher for patients with IR (0.46) compared to stable patients (0.10) during the one year follow-up period. CONCLUSIONS: The majority of patients that were previously naïve to bDMARDs had an IR to their first bDMARD during the year of follow-up. Patients with an IR had numerically increased all-cause resource utilization and non-medication costs during the follow-up period compared to patients with stable disease. This level of IR suggests an unmet need in the RA treatment paradigm.