Defunding Sexual Healthcare: A Topological Investigation of Resource Accessibility

TL;DR

The paper investigates how defunding Planned Parenthood clinics could affect access to sexual and reproductive healthcare in California. It applies Topological Data Analysis, specifically persistent homology on Vietoris-Rips filtrations built from travel-time distances between facilities, to compare current (PPHC+FQHC) versus FQHC-only scenarios. The analysis reveals that removing PPHCs generally increases travel times and connectivity gaps, with statistically significant evidence, underscoring potential declines in access for Medicaid beneficiaries. The work highlights the utility of PH/TDA for evaluating spatial inequalities in public health and outlines avenues for more granular, nationwide analyses.

Abstract

Government actions, such as the Medina v. Planned Parenthood South Atlantic Supreme Court ruling and the passage of the Big Beautiful Bill Act, have aimed to restrict or prohibit Medicaid funding for Planned Parenthood Healthcare Centers (PPHCs) at both the state and national levels. These funding cuts are particularly harmful in states like California, which has a large population of Medicaid users. This analysis focuses on the distribution of Planned Parenthood clinics and Federally Qualified Health Centers (FQHCs), which offer essential reproductive healthcare services including, but not limited to, abortions, birth control, HIV services, pregnancy testing and planning, STD testing and treatment, and cancer screenings. While expanded funding for FQHCs has been proposed as a solution, it fails to address the locational accessibility of Medicaid-funded health centers that provide sexual and reproductive care. To assess this issue, we analyze the proximity of data points representing California's PPHC and FQHC locations. Topological Data Analysis (TDA)-an approach that examines the shape and structure of data -- is used to detect disparities in reproductive and sexual healthcare coverage. To conduct data collection and visualization, we utilize R and Python. We apply an n-closest neighbor algorithm to examine distances between facilities and assess changes in travel time required to reach healthcare sites. We apply persistent homology to analyze current gaps across multiple scales in healthcare coverage and compare them to potential future gaps. Our findings aim to identify areas where access to care is most vulnerable and demonstrate how TDA can be used to analyze spatial inequalities in public health.

Figures (10)

• Figure 1: Arbitrary point cloud $W$ with points $w_i$.
• Figure 2: As the radii ($r$) of each point intersects— when the distance between each point is less than $2r$— simplices form between the points. Here, edges emerge in Figures 2.c. and 2.d. simplicesgif.
• Figure 3: A VR filtration through 3 phases. Phases 1 through 3, from left to right, ranges on a scale parameter $r_i$ from $r_0$ to $r_2$. At the $0$D homology, the two vertices are born at $r_0$. There is a homology class in $r_0$ and $r_1$ as the surrounding radii have not yet intersected and there is space between the vertices. In $r_2$, the homology class dies with the edge connecting the two vertices. Therefore, the connecting edge is the death simplex of the homology class and the death value of the homology class is $r_2$.
• Figure 4: A VR filtration through 4 phases. Phases 1 through 4, from left to right, range on a scale parameter $r_i$ from $r_0$ to $r_3$. At the $1$ D homology, the homology class is born at the third phase, $r_3$. There is a homology class in $r_0$ and $r_1$ as the surrounding radii have not yet intersected and there is space between the vertices. In $r_2$, the homology class dies with the edge connecting the two vertices. Therefore, the connecting edge is the death simplex of the homology class and the death value of the homology class in $r_2$.
• Figure 5: Point cloud variations for each VR filtration.