Demographic Characteristics Influence Treatment Costs of Invasive Melanoma in Florida

Abstract

Background Demographic characteristics are known to influence the treatment and outcomes of patients with invasive melanoma. Whether these characteristics influence treatment costs is unknown. We aimed to analyze whether patient demographics and tumor characteristics influence treatment costs for patients with invasive cutaneous melanoma in Florida. Methods This was a cross-sectional study in which the Florida Inpatient and Outpatient Dataset of the Agency for Health Care Administration was analyzed for patients with a diagnosis of invasive melanoma between January 1, 2013 and December 31, 2018. Categorical variables were assessed using Pearson χ2 tests, and continuous variables were evaluated using Kruskal-Wallis tests. Logistic regression analysis was conducted to identify the association between patient demographics and total costs. All analyses were done using SAS 9.4 statistical software (SAS Institute, Inc). Results Multivariate analysis showed that sex (P < 0.001), hospital setting (P < 0.001), race/ethnicity (P < 0.01), patient region (P < 0.01), Elixhauser Comorbidity Index score (P < 0.001), presence of metastasis (P < 0.01), total number of procedures (P < 0.001), and length of stay (P < 0.001) were correlated with the cost of treatment of invasive cutaneous melanoma. After stratification, the association between cost and race/ethnicity disappeared for inpatients but remained for Black patients in the outpatient setting (P < 0.001). The association between cost and patient residence regions also differed when the cohort was stratified. Conclusions Strategies addressing disparities in treatment cost of invasive melanoma should differ, depending on the hospital setting where the patient is being treated.

Emergency Department Visits Before Cancer Diagnosis Among Women at Mayo Clinic

“… 18 An Elixhauser score was used based on the investigator’s previous experience with the score at our institution and previous literature showing advantages in the score. 19 , 20 , 21 The score was calculated for each patient by summing the individual weights of comorbidities, except for solid tumors, metastatic cancer, and obesity. 22 Cancers were excluded from the score to eliminate the possibility that the score included newly diagnosed primary cancer.…”

Challenges of Deep Learning in Cancers

“…Commonly used ML methods for cancer analysis include linear regression, logistic regression, 7 decision trees, random forests, support vector machines, 8 neural networks, k-means clustering, principal component analysis, 9 naïve Bayes, gradient boosting, 10 and so on. However, ML methods suffer several shortcomings.…”