A review and comparison of medical expenditures models: two neural networks versus two‐part models

Abstract

This paper compares the two-part model (TPM) that distinguishes between users and non-users of health care, with two neural networks (TNN) that distinguish users by frequency. In the model comparisons using data from the National Health Research Institute (NHRI) in Taiwan, we find strong evidence in favor of the neural networks approach. This paper shows that the individuals in the selforganizing map (SOM) network clusters can be described as several different forms of frequency distributions. The integration model of SOM and back propagation network (BPN) proposed by this paper not only permits policymakers to easily include more risk adjusters besides the demographics in the traditional capitation formula through the adaptation and calculation power of neural networks, but also reduces the incentives for cream skimming by decreasing estimation biases.

Constructing “Behavioral” Comparison Groups

“…The hope was that the copayment reform would contain the total outpatient expenditures from the demand side (Manning, Newhouse, & Duan, 1987; Zweifel & Manning, 2000). The literature on the demand for medical care analyzes either discrete measures, such as the number of doctor visits (Chiappori et al, 1998; Cockx & Brasseur, 2003; Riphahn et al, 2003; Winkelmann, 2004a; 2004b), or continuous measures, such as expenditures (Hsu, Lin, & Yang, 2008; Kim, Ko, & Yang, 2005; Lin, Hsu, & Takao, 2008; Manning et al, 1987). However, it is well known that these two measures are derived from two different decision-making processes based on a principal–agent model, where the physician (the agent) determines utilization on behalf of the patient (the principal) once initial contact is made (Deb & Trivedi, 2002; Pohlmeier & Ulrich, 1995).…”