Specifically, endogeneity may arise because there might be some unobserved heterogeneity that drives both the investments in non healthcare services treatment (information) and physician services. Unobserved propensity toward good health, or marginal value of health, may increase consumers’ use of both health information and health services. It results in a biased positive effect of information without proper controls for valuation of health. Moreover, other unobserved preference parameters such as an individual’s risk preference and time preference may also affect both consumer health information-seeking as well as health care utilization in some ambiguous ways. For example, people who are risk-averse may seek both health services and consumer health information, or they may use more health services but be weary
of consumer information. In sum, in models of entry into physician service and number of physician visits, Info may be endogenous and
correlated with the residuals of both equations so that a sample
selection correction for propensity to initiate a first visit is insufficient
for correcting the potential bias. The direction of the potential
bias is unclear according to our analysis.