We use the Heckman’s two-step approach. In estimating the equation for the positive number of physician visits, we include the inverse mills ratio (lambda) which comes from a probit regression for the decision to seek a physician using all observations. We use as an exclusion restriction the choice to “opt out” of an employer provided health insurance plan – so the indicator takes a one if an individual is eligible for employer provided health insurance but not covered, and 0 otherwise. We believe this reflects a choice or preference for participation in the health care services market, but not a significant determinant for frequency of physician visits.5