Treatment Effect Bounds: An Application to Swan-Ganz Catherization∗

We implement alternative bounding strategies to reanalyze data from the observational study by Connors et al. (1996) on the impact of Swan-Ganz catheterization on mortality outcomes. While the Connors et al. (1996) study assumes that there are no unobserved differences between patients who are catheterized and patients who are not, we employ a bounding strategy that allows for such unobserved differences. We implement both the traditional bounds that exploit access to an instrumental variable but impose no other assumptions (Manski, 1990) and the bounds of Shaikh and Vytlacil (2004) that impose additional relatively mild nonparametric structural assumptions. Both of these approaches require an instrumental variable that shifts the probability of catheterization, but which does not alter mortality risks. We propose and justify using indicators of weekday admission as an instrumental variable in this context. We find that, while the traditional instrumental variable bounds are almost entirely uninformative in our application, the Shaikh and Vytlacil (2004) bounds that do impose some nonparametric structure often produce a clear answer—catheterization reduces mortality at 7 days, and increases it at 30 days and after. Applying a further nonparametric structural assumption—that doctors catheterize individuals with systematically worse latent health—further narrows these bounds and strengthens these conclusions. Our findings suggest an explanation for the fact that many ICU doctors are deeply committed to the use of the Swan-Ganz catheter. Since most ICU patients leave the ICU well before 30 days after admission have elapsed, ICU doctors never observe the increase in mortality. They do, however, observe the decline in mortality at 7 days. ∗We would like to thank seminar participants at Brigham Young University, University of Chicago, Harvard/MIT, Michigan Ann-Arbor, Michigan State, and at the ZEW 2nd Conference on Evaluation Research in Mannheim. Correspondence: Landau Economics Building, 579 Serra Mall, Stanford CA 94305; Email: [email protected]; Phone: 650-725-7836; Fax: 650-725-5702.