The Financial Impact of Genetic Diseases in a Pediatric Accountable Care Organization

Abstract

Background: Previous studies revealed patients with genetic disease have more frequent and longer hospitalizations and therefore higher healthcare costs. To understand the financial impact of genetic disease on a pediatric accountable care organization (ACO), we analyzed medical claims from 2014 provided by Partners for Kids, an ACO in partnership with Nationwide Children's Hospital (NCH; Columbus, OH, USA). Methods: Study population included insurance claims from 258,399 children. We assigned patients to four different categories (1-A, 1-B, 2, & 3) based on the strength of genetic basis of disease. Results: We identified 22.7% of patients as category 1A or 1B- having a disease with a "strong genetic basis" (e.g., single gene diseases, chromosomal abnormalities). Total ACO paid claims in 2014 were $379M, of which $161M (42.5%) was attributed to category 1 patients. Furthermore, we identified 23.3% of patients as category 2- having a disease with a suspected genetic component or predisposition (e.g., asthma, type 1 diabetes)- whom accounted for an additional 28.6% of 2014 costs. Category 1 patients were more likely to experience at least one hospitalization compared to category 3 patients- those without genetic disease [odds ratio [OR] = 4.12; 95% confidence interval [CI] = 3.86-4.39; p < 0.0001]. Overall, category 1 patients experienced nearly five times the number of inpatient (IP) admissions and twice the number of outpatient (OP) visits compared to category 3 patients (p < 0.0001). Conclusion: Nearly half (42.5%) of healthcare paid claims cost in 2014 for this study population were accounted for by patients with single-gene diseases or chromosomal abnormalities. These findings precede and support a need for an ACO to plan for effective healthcare strategies and capitation models for children with genetic disease.

Keywords: accountable care organization; computational phenotyping; genetic disease; insurance claims; pediatrics.

Figure 1

Partners for Kids Flow of Funds: PFK receives funds to pay for each child's medical care, the amount of which is defined as a capitation rate per child. A cost-saving model can achieve a surplus of funds, which are reinvested into programs that lead to improved health for children, such as school-based clinics and neighborhood programs.

Figure 2

Categorical Distribution of Visits: (A) Frequency of IP visits among categories in 2014; 2 × 2 contingency table with odds ratio of category 1 (A&B, genetic) vs. category 3 (non-genetic); (B) Frequency of OP-office visits among categories in 2014; (C) Frequency of OP-ER visits among categories in 2014.

Figure 3

Confusion Matrix of Genetic Categorization Classification Accuracy: A confusion matrix of genetic categorization classifications from a random sampling of 100 patients. Rows correspond to the “predicted” genetic categorization for each patient based on insurance claims data. Columns correspond to the “true” categorization based on a manual review of patient electronic health charts. The diagonal cells shaded in green correspond to observations that are correctly classified, while the non-diagonal cells shaded in red correspond to incorrect observations. The column on the far right shows percentages of claims-based categorizations that are correctly (green text; positive predictive value) and incorrectly (red text; false discovery rate) classified. The row at the bottom shows percentages of chart-based categorizations that are correctly (green text; true positive rate) and incorrectly (red text; false negative rate) classified. The cell in the bottom right shows the overall accuracy (81%) of our categorization method.