TY - JOUR AU - Bhattacharya,Jay AU - Shaikh,Azeem AU - Vytlacil,Edward TI - Treatment Effect Bounds: An Application to Swan-Ganz Catheterization JF - National Bureau of Economic Research Working Paper Series VL - No. 11263 PY - 2005 Y2 - April 2005 UR - http://www.nber.org/papers/w11263 L1 - http://www.nber.org/papers/w11263.pdf N1 - Author contact info: Jay Bhattacharya 117 Encina Commons Center for Primary Care and Outcomes Research Stanford University Stanford, CA 94305-6019 Tel: 650/736-0404 Fax: 650/723-1919 E-Mail: jay@stanford.edu Azeem Shaikh Department of Economics University of Chicago 1126 E. 59th Street Chicago IL 60637 E-Mail: amshaikh@uchicago.edu Edward J. Vytlacil Department of Economics Yale University Box 208281 New Haven, CT 06520-8281 Tel: 203/432-3244 Fax: 203/432-6167 E-Mail: edward.vytlacil@yale.edu AB - We reanalyze data from the observational study by Connors et al. (1996) on the impact of Swan-Ganz catheterization on mortality outcomes. The Connors et al. (1996) study assumes that there are no unobserved differences between patients who are catheterized and patients who are not catheterized and finds that catheterization increases patient mortality. We instead allow for such differences between patients by implementing both the bounds of Manski (1990), which only exploits an instrumental variable, and the bounds of Shaikh and Vytlacil (2004), which exploit mild nonparametric, structural assumptions in addition to an instrumental variable. We propose and justify the use of indicators of weekday admission as an instrument for catheterization in this context. We find that in our application, the Manski (1990) bounds do not indicate whether catheterization increases or decreases mortality, whereas the Shaikh and Vytlacil (2004) bounds reveal that catheterization increases mortality at 30 days and beyond. We also extend the analysis of Shaikh and Vytlacil (2004) to exploit a further nonparametric, structural assumption -- that doctors catheterize individuals with systematically worse latent health -- and find that this assumption further narrows these bounds and strengthens our conclusions. ER -