04 McGregor article

Using data from the Northern Ireland Household Panel survey we demonstrate that attendance at outpatients is determined solely by respondent health. This is consistent with the GP acting as a gatekeeper to other services. Attendance thus provides valuable information on illness severity. Splitting the sample into two statistically distinct groups, we estimate ordered probit regressions of GP utilisation with and without sample selection. The results indicate that ignoring outpatient attendance may result in misspecification. Further, if health is more fully measured, age becomes redundant as a determinant of utilisation and service supply, particularly accident and emergency use, is seen as a significant determinant in explaining GP visits.