Figure 2: Shows the difference between insured and uninsured patients with respect to likelihood of attending a GHE in the near future.

The two graphs are constructed using conditional probabilities calculated from the estimated coefficients presented in Table 1. The method of calculation was as described by Agresti²⁷. From (a,b) the shifting trends of empirical probabilities are similar for both insured and uninsured patients. However, the changes in numerical probabilities are significantly different, and in (b), the two probability lines intersect, at about 50%. So 50% can be seen as a probability threshold where uninsured patients are indifferent in their decisions to have a GHE in the near future or not. There exists no such threshold for insured patients as seen in (a).