Development of a novel, objective measure of healthcare-related financial burden for U.S. families with children

Many families face increasing financial pressure due to rising healthcare costs, with implications for their future health and healthcare use. Existing studies have often relied on a “10% cut-point” that may not apply across all family types and income-levels. The goal of this study was to develop a theoretically-based and empirically-driven objective measure of financial burden using a nationally representative population-based sample of US families with children.

The measure was developed using data on 149,021 families with children from the 1999-2010 National Health Interview Survey. The measure was validated using data on 18,488 families with children from the 2002-2010 Medical Expenditure Panel Survey. We estimated the marginal probability of unmet healthcare need using bivariate, tensor product splines for family income and healthcare expenditures, while adjusting for relevant predisposing, need and enabling factors. Recursive partitioning was then performed on these marginal probabilities, as a function of income and expenditures, in order to empirically establish thresholds which demarcate different levels of predicted risk. These grouping were further collapsed to create a novel measure of financial burden (4 groups); the performance of this novel measure was evaluated and validated.

Multivariable results revealed that, compared to ‘low burden,’ increasing categories of burden demonstrated increased odds of unmet need (mid-low OR: 1.86, 95%CI: 1.71-2.01; mid-high OR: 2.70, 95%CI: 2.48-2.94; high OR: 4.46, 95%CI: 4.11-4.84); with a significant linear trend (p<0.0001). Similar results were observed using the validation sample. The 10% cut-point was associated with increased odds of unmet need in development sample but not the validation sample, and demonstrated substantially worse model fit than the novel measure. Finally, the 10% cut-point misclassified 74.5% of families identified as having ‘high burden’ with the novel measure, and 98.4% of those with ‘mid-high burden.’

This study was successful in generating and validating a novel financial burden risk measure that was strongly associated with unmet need and can easily be derived using income and expenditures. When compared against the commonly utilized alternative, we found evidence to support the superior performance of the novel measure and these comparisons revealed that substantial misclassification of financial burden are probable when utilizing the existing measure. The newly-developed measure of financial burden could be applied in future studies of healthcare utilization and expenditures (e.g., to improve identification of financial burden among families) and has implications for use in policy development and evaluation (e.g., ACA’s current use of income thresholds for premium and cost-sharing subsidies).