Oncology Accessibility

Home

Oncology Accessibility is a web application based on the methods published in our paper: "Disparities in accessibility to oncology care centers in France". The preprint is available on medrxiv.

Cancer is a leading cause of death worldwide, accounting for nearly 10 million deaths in 2020. Access to health services plays a key role in cancer survival, and spatial accessibility methods have been successfully used for measuring access to healthcare providers. We propose a method to i) group the care centers based on their oncology specialization; ii) compute an oncology accessibility score for each municipality in metropolitan France; iii) im-prove this accessibility by identifying which care centers to grow and by how much.
Our method will make it possible for health professionals and administrations to monitor the accessibility to oncology care. Since the accessibility optimization is very dependent on the region and local constraints, we packaged our algorithms into a web application that will let the user tune every parameter and preview the activity change.

Optimization setup

We use Linear Programming to optimize the care centers capacities \( S \) to improve the accessibility \( A \) at a given region. We are interested in maximizing the total accessibility, i.e the sum of all the municipalities' accessibility. The form below allows to pick the optimization hyper parameters:

  • Region: The region where the optimization will be run on. The optimization is ran on metropolitan France, but care centers that are not from the given region are not allowed to grow/decrease. Only the care centers and municipalities from the given region and the surrounding departments will be displayed.
  • Supply variable: The variable to use as capacity \( S \). This encodes the care center supply, vs. the population demand. Choices can be:
    • Oncology activity: Number of medical or surgery stays related to cancer + number of chemotherapy and radiotherapy patients
    • MCO activity: Number of medicine, surgery and obstetric stays
    • Chemotherapy activity: Number of chemotherapy patients
    • Radiotherapy activity: Number of radiotherapy patients
    • Oncology medical and surgery activity: Number of medical or surgery stays related to cancer
  • Additional supply: The activity to be added to the current overall activity. Setting this parameter to 0 will lead to an optimization constraint with "constant" activity, meaning that a care center will have to decrease to let another one grow. If this number is set between 0 and 1, the corresponding percentage of the current activity is added. e.g: 0.03 will add 3% of the current activity.
  • Max growth percentage: The maximum growth percentage of a care center. If set to 20%, the care center will not be allowed to grow by more of 20% of its current activity.
  • Max decrease percentage: The maximum decrease percentage of a care center. If set to 20%, the care center will not be allowed to decrease by more of 20% of its current activity. If set to 0, the care centers activity can't decrease.
  • Low cluster max capacity: The maximum capacity that the care centers from the least specialized cluster can reach. If set to 0, these care centers can't receive any activity and will be emptied if they originally had some.
  • High cluster max decrease: This is similar to the "max decrease percentage" parameter, but only applied to the care centers from the most specialized cluster. If set to 0, these care centers won't be allowed to decrease.
  • Maximum new capacity: The maximum capacity that the care centers with 0 activity can receive, unless they are within the least specialized cluster. In this case, this parameter will be ignored and "low cluster max capacity" will be used.

Analytics

Plots to better understand the accessibility and oncology activity, distribution, before running the optimization algorithm.