ISER Discussion Paper, Institute of Social and Economic Research, Osaka University 547
Objective The treatments of ischaemic heart disease have drastically changed due to the developments in coronary-anterial bypass graft operations and coronary intervention procedures. The treatment outcomes of these procedures are compared using observational data. Method The data used are microdata from Surveys of Medical Care (SMC) in Japan from 1992 to 1998, which are claim data, not using a randomized controlled trial. Hence, endogeneity in the choice of treatment should be important and thus both the instrumental variables method and the nonparametric propensity score matching method are adopted for estimation. Result Although the instrumental variables method and nonparametric propensity score matching method produce different results in many cases, predominance of the latter over the instrumental method is confirmed. Although claimed with some reservations, patients treated with precutaneous transluminal coronary angioplasty (PTCA) have significantly lower possibility of severe heart failure than those treated with coronary artery by-pass grafting (CABG). The results of comparing invasive procedures with preserving treatment depend on the existence of constraints. In data for more acute situations, invasive procedures have a significantly higher possibility of severe heart failure than preserving treatment.