Every push has a pull, every to has a fro, and every surgical mortality study has an offsetting counterpart. The medical literature is littered with well-intentioned but flawed epidemiologic studies of surgical mortality that attempt to demonstrate the capabilities of one type of anesthesia provider compared to another. Surely our efforts and resources can be better directed toward improving the outcomes of our patients through the application of valid studies focused on patient care issues.

The April 2003 American Association of Nurse Anesthetists Journal brought us another peer-reviewed (and, therefore, valid?) study of surgical mortality. This particular article, “Surgical Mortality and Type of Anesthesia Provider,”¹ was authored by Michael Pine, M.D., President of Michael Pine and Associates, Inc, a Chicago firm hired by the American Association of Nurse Anesthetists (AANA) Foundation for this study, and two of his company’s associates. The study reviewed surgical inpatient (prior-to-discharge) mortality rates in 404,194 Medicare patients from 22 states undergoing eight nonemergent surgical procedures from 1995-97. The authors concluded, “… for the surgical procedures included in this study, the type of anesthesia provider does not affect inpatient surgical mortality.”

I suppose that as anesthesiologists we should be alarmed at this study’s conclusion. However, the fact of the matter is that in-hospital surgical mortality is a very bad way to define anesthesia outcomes — or hospital outcomes at all. Curiously, Dr. Pine previously noted this problem: “… mortality rates based on administrative data alone may be biased and misleading, even when statistics show excellent predictive accuracy.”² From a statistical standpoint, the primary conclusions of this recent study really are that 1) the administratively derived Medicare billing data are too inadequate to allow risk-adjusted outcome comparisons and 2) valid comparisons between anesthesia providers require analyses of much larger numbers of patients. The complex risk adjustments needed to discern the impact of patient-related compared to anesthesia provider-related factors require information on millions, not thousands, of patients if large administrative databases are used.

Really, who can blame AANA for trying to produce a paper to counteract the earlier study of anesthesia provider differences and their impact on Medicare patient outcomes published by Silber et al.³ in Anesthesiology? Ironically, the current Pine study and the earlier Silber study validate three important observations, none of which serve the AANA’s primary purpose for commissioning an epidemiologic investigation:

A knowledgeable physician is important when things go wrong. Both the Silber and Pine studies show that “two heads are better than one” when unexpected problems arise. In these studies, the anesthesia care team (defined as an anesthesiologist working with a nurse anesthetist) had lower failure-to-rescue and mortality rates. Anesthesiologists are vital in resolving medical emergencies and implementing life-saving medical care.

Medicare databases do not provide good enough information to use in outcome studies. Multiple studies have concluded that Medicare’s billing and other administrative data do not provide valid information for outcome studies.^2,4-6 Until the databases improve or, by a miracle, someone learns how to glean the wheat from the chaff within the data tapes, authors should be dissuaded from using this information for outcome studies.

Type of in-room anesthesia provider is an interesting but relatively unimportant issue. Please recall that ASA worked with the Centers for Disease Control (CDC) in the early and mid-1980s to develop a prospective study that would clarify the importance of anesthesia providers on the outcomes of surgical patients. Based on preliminary data, the definitive study was projected to be large and very expensive. Although the potential price tag dampened enthusiasm for the project, it was the advent of HTLV-III infection (later known as HIV, or AIDS) that diverted the CDC’s attention to this more important and pressing issue.

You might argue that while the issue of type of anesthesia provider continues to be interesting, its value is quite insignificant compared to the various patient safety and care issues that we should be addressing. Would our attention and resources be better directed toward understanding disease-specific pain mechanisms, individualized treatment of acute and chronic pain with pharmacogenetically determined opioid profiles or uniquely tailored approaches to the management of sepsis and acute respiratory failure?

The Pine study has provided little political clout to AANA in its attempt to provide “safety” data that support a rationale for eliminating physician supervision by state governor waiver from Medicare requirements. Given the apparent political impasse from the Pine and Silber studies, let us take a deep breath, re-evaluate the most important research needs of our specialty and vigorously pursue studies that will improve the care of our patients. Both the Foundation for Anesthesia Education and Research and the Anesthesia Patient Safety Foundation stand ready to help young investigators begin to address these questions and advance the science of anesthesiology.

Mark A. Warner, M.D., is Professor and Chair, Department of Anesthesiology, Mayo Clinic, Rochester, Minnesota.