^*See Appendix I for schematics of the heart.

^*Training for “cardiothoracic” surgery entails working on both heart and lungs, since they are properly viewed as a unified respiratory/circulatory system. The recent trend has been to specialize in one or the other after training, and the accreditation authorities are now moving to begin the specialization from the very start of training. I focus almost exclusively on the cardiac side, in part because the practices are really quite different—lung surgeons, for example, are much more involved with cancers—and I decided it was too much to absorb. In addition, as one of the lung surgeons conceded, although the lung is amazingly complex, it is the much less glamorous organ.

^*At dinner with a doctor friend from another Columbia division, he groused about the way his division split up fee income. I mentioned that in Cardiothoracic Smith made the splits himself. He said, “Sure. But everybody knows he’s so fair.” Characteristically, Smith is not Cardiothoracic’s highest-paid surgeon.

^*Coronary arteries are usually on the outside of the myocardium, which makes them easily accessible without opening the heart. Otherwise, as in virtually all valve work, the use of the heart-lung machine is unavoidable. And occasionally, some arteries in the back of the heart, next to the rear chest wall, can’t be reached without going on-pump to drain the heart. Smith does about 70 percent of his bypasses off-pump, sometimes executing the first several off-pump, then using the pump for the last one or two.

^*Although the several anesthesiologists I got to know seemed really to enjoy their work—and it is among the best-paid medical specialties—they sometimes conveyed a rueful sense of inferiority vis-à-vis the surgeons. “The surgeon always gets the case of wine, even though we may have been the one who saved the patient,” one told me. Surgeons allegedly tell anesthesiologist jokes, the way violinists tell violist jokes. My favorite was the description of anesthesiologists as “the half-asleep looking after the half-awake,” but I heard it from an anesthesiologist, not a surgeon.

^*The broader pre-and postoperative role of the Japanese surgeon, while it sounds highly desirable from the standpoint of continuity of care, is probably not in patients’ best interest. There is a massive body of data showing a high correlation between heart surgery volumes and outcomes. The same operation can evolve so differently from one patient to the next that it takes a lot of cases before a surgeon can react confidently to each new confluence of events. According to Hirata, forward thinkers in Japan would prefer to move toward the Western model, but it would create an instant surplus of surgeons.

^*In 1999, two urgency statuses, 1 and 2, were refined to 1A, 1B, and 2. 1Bs from the donor region, which involves several OPOs, will get priority over a 1A from an outside region, in part because of cold ischemic time limitations. The 2 status is becoming controversial. As cardiology improves, waiting-list death rates are dropping, and some 2s have been on the list for years, some for more than ten years. When the survival rate on the list exceeds that of a successful transplant, Naka points out, that person prima facie shouldn’t be listed—“A transplant is a dangerous operation.”

^*No joke. Fentanyl is sometimes supplied in berry-flavored lozenges, prescribed especially for patients with advanced cancers. There is inevitably an active street market.

^*Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure. Clinical trial managers have an apparently incurable weakness for clever acronyms.

^†The cost of LVADs compares favorably with that of cancer drugs that are approved based on the basis of additional survival times of only weeks or months. Accumulating examples of high-cost, low-survival treatments doesn’t justify them, of course.

^*International Center for Heart Outcomes and Innovation Research. (Cleverer than most, I admit.)

^†Allan Stewart is involved in a research project on the physiological and chemical consequences of continuous-flow devices. Although it is at an early stage, it appears that the change to continuous flow causes some degree of organ shock. Interestingly, if a patient with a continuous-flow device gets a transplanted heart after several weeks on the device, there appears to be a second shock as the body adjusts back to pulsatile mode. That suggests that true bridge-to-transplant patients may be better off with pulsatile devices, while destination patients would use continuous-flow devices for longevity.

^*In theory, Erika’s problem might also have been solved with a transplant of a new heart and lungs. Heart-lung transplants in adults had been common for a while in the 1980s, but are quite rare now—one recipient is using up a lot of organs. The number of pediatric heart-lung transplants, ever, is fewer than a dozen, and results were generally poor. Doctors mentioned it from time to time, but it was never considered a realistic option.

^*Children who get transplants therefore need one or more additional transplants to survive into adulthood. It is not because the new heart doesn’t grow. A baby who gets another baby’s heart will reach his teens with a teenage heart, but one that is already failing.

^*The final placements are determined by a national algorithmic-based system that makes a best-fit match between applicants’ rankings of schools and vice versa. The algorithm rewards realism, so both applicants and schools are punished for overreaching. That is a classic game-theory paradigm—it’s safer to aim low to ensure you get something, because if you reach too high, you might get nothing at all. The same rules apply to the schools. If they pick only top candidates, and those candidates pick a different school, they may not get anyone. It made for extremely intense selection meetings. The selection cycle I witnessed was targeted at a future class, and Columbia once again got its first choices.

^*New York has adopted rules limiting resident working hours to not more than eighty hours a week. I had seen Eric Rose, in a surgery recruiting talk for medical students, tell them that the new rules would let them lead a normal personal life. (By comparison, eighty hours is only four hours a week shorter than the twelve-hour–seven-day factory work schedules that outraged nineteenth-century labor reformers.) The State, cleverly, enforces the rules through a private contractor that is reimbursed with the fines it collects from violations. Individual hospital violation rates are not published, but total violations have fallen sharply since the law went into effect. My limited sample suggests that almost all residents, regardless of the intent of the hospitals, do their best to circumvent the rules. This is one of the crucial learning periods of their lives, and they don’t want to miss anything that might be important. Smith once commented that they might have to make them wear radio tags.

^*Outsiders are rarely allowed in such meetings. Had the question been posed directly, I doubt if I would have been. Smith had introduced me at one of the regular staff meetings, albeit amid the early buzz of people assembling, and by the time of my first M&M, I was already part of the wallpaper. After the doctors were used to seeing me at M&Ms, most would take the time to explain case details, and I almost always met with Smith afterward to review my understanding of what had transpired.

^*The numbers are also reflected in the State outcomes tracking system, an innovation in which New York has led the nation. See Appendix II for a brief description and the Columbia-Presbyterian rankings.

^*Six Sigma is statistical shorthand for an error rate of one in less than 3.4 million (or 99.9997 percent accuracy). It evolved originally from high-tech manufacturing process control. Since there are thousands of processes in making a memory chip, they all must be managed to something like Six Sigma levels to produce final yields of 90 percent or so. Nothing that happens in a hospital is controlled at anywhere near that level of accuracy. But even at GE the concept is often employed more as a useful slogan (in financial services operations, for instance) than as a real-world standard.

^*And may be impossible within the current lotterylike system of redress through tort suit. Accessible and standardized error documentation—who can believe it won’t be leaked?—could feed endless lawsuits. Some kind of workman’s compensation–type arrangement with expert reviews and an appeal process may be a prerequisite to a completely transparent quality control and fault correction regime.

^*Another much-less-publicized study using similar methods was published at about the same time. In comparison with the alternative drugs, it found adverse renal effects related to aprotinin, but not cardiac or cerebral effects, and did not show mortality differences.

^*Fibrin is an important blood-clotting agent. “Lysis” is a kind of molecular destruction. So “fibrinolytic” drugs attack fibrin and prevent clotting, while “antifibrinolytics” help preserve fibrin, enhance clotting, and reduce bleeding.

^*That is a scary finding, for it suggests that researchers may be exercising unconscious bias in trial patient selection within the pool of those who are officially eligible. Or, conceivably, trial patients get much better care than normal patients, or both selection and treatment factors may be at work.

^*The institute’s affiliates include thirty universities, ten government agencies, two of which are NIH units, and fifteen large companies, five of which are pharmaceutical companies. Bayer is not one of them.

^*Conceivably, of course, aprotinin may be the cause of the high German morbidity and mortality rates. But practices differ so much from country to country that it would require a substantial research effort to pinpoint the causes of the German variance.

^*The notes to this chapter include an extended excerpt from the hearing exchanges on this point. See Notes.

^*An NEJM spokesperson told me that they never give interviews on editorial decisions. I also made several requests to Mangano for interviews, and sent him written questions, but he did not reply.

^*During a similar display at the 2004 TCT, a procedure went seriously wrong, and the patient died later that day. But Leon has stuck to the same live format. As I’ll show later in the chapter, he is seriously committed to showing everything. Rather than sugarcoat difficult operations, he stresses the dangers.

^*Notice that in measuring pulmonary hypertension, the backpressure into the blocked pulmonary artery is being used as a proxy for the pressure coming out of the lungs into the left atrium, which necessarily introduces some degree of error. To add to the uncertainty, the hypertensive index that transplant specialists use, the PVR (pulmonary vascular resistance), is a function of three variables, the (unblocked) pulmonary artery pressure, the so-called wedge pressure from the Swanz-Ganz described above, and cardiac output, which is based on yet another estimate. In the Maynard case, the surgeons later expressed considerable skepticism about Erika’s near-miraculous PVR drop from close to 12 to only 1.7. Linda Addonizio, however, argues for its accuracy, based on the exceptional reactivity of pediatric cases, and the special skills in her section. The entire discussion is a good example of the fundamental imprecision of physiological indices, and the importance of experience and intuition in interpreting ostensibly precise data.

^*The initial market penetration of the da Vinci, it appears, has been in areas where laparascopic procedures are well established. (They are small-incision, camera-assisted surgeries, usually in the belly.) Surgeons used to laparoscopic techniques usually find that the 3-D, highly intuitive character of the da Vinci is a big improvement over laparoscopy, and requires minimum adaptation on their part. It is also becoming the technology of choice for prostate surgery, since the fine-grained control substantially enhances preservation of sexual and urinary functions. The gas pressure within the abdomen also reduces bleeding. Unlike most prostatectomies, a urologist told me, da Vinci prostatectomies rarely require blood transfusions. Amid all this good news, the company, unfortunately, is increasing its prices to well over a million dollars for a basic system. If the history of computer hardware is any guide, the victor in robotic surgery will be the company that produces a highly functional, compact, low-priced system that becomes standard in every OR in the world, not the one selling a “mainframe” solution. Finally, the da Vinci is not in any sense robotic—it is a servo-mechanism—but the “robotic” tag has stuck, and is a great attention-getter.

^*The “purse-string suture,” which is the standard method of inserting a tube into an organ, is just one of dozens of high-efficiency surgical techniques that have evolved from long practice. The surgeon places a suture in a circle around the area of incision, with the two ends coming together, as in the top of a sack. The incision is made after the suture is in place. In almost a single motion, a skilled surgeon can make the incision, insert the tube, and pull the purse-string shut, so there is almost no loss of blood. When it’s time to extubate, he pulls out the tube, tightens the purse-string, and ties it shut.

^*Nonprofit tax returns are available to the public and must list the entity’s top five earners. Previous years’ returns show that Jan Quaegebeur’s earnings were usually right behind Smith’s. Jeffrey Moses, the director of the interventional cardiology division came to Columbia in 2004, however, and, at $2.2 million, pushed Quaegebeur off the list. The top-earning Columbia physician in recent years has been a dermatologist, David Silvers, who earned an impressive $3.6 million in 2004.

^*In the framework of a widely used medical ethics textbook, Stewart described a conflict of obligations, the obligation to give value to his sponsor and to give the best information to his audience. Conflicts of obligations are almost inescapable—doctors’ financial obligations to their employers often conflict with their obligations to their patients. Ethical conflicts arise when an obligation is transposed into a self-interest, like making money or advancing academically. Since Stewart expected that Bayer would not be happy if he shared his unvarnished views of aprotinin, he anticipated being caught in a conflict of interests and was right to turn down the opportunity.

^*The macabre joke underlying the government’s successful multibillion-dollar lawsuit against the tobacco companies, ostensibly to recover the excess health care costs of smokers, was that neither side could admit that smokers’ high death rates saved the government money. Health care economists often try to offset the costs of medical interventions by the economic benefits of extended lives—if, say, an extra year of a “quality-adjusted” life is worth $100,000 (a common low-end assumption), a $300,000 heart transplant that extends life by ten years, on average, is well worth it. I’ve never found such arguments especially convincing. In real life cost-benefit analyses, the benefits are usually paid into the same coffers that the costs come from. But the assumed returns from Grandma’s extra years of life are priced in theoretical dollars that you can’t find on anyone’s income statement.

^*The 67 ten-year periods since 1929 comprise the sequence, 1929–1939, 1930–1940, 1931–1941…1995–2005.

^*The original idea was that the new channels would increase blood flow, but autopsy evidence suggested that the channels mostly closed up. An FDA review speculated that the injury to the muscle tissue stimulated angiogenesis (new vessel growth), or that it may have destroyed nerves that were sensitive to pain, or that, since the end point was essentially subjective—rate your pain on a scale—and it’s hard to blind surgical interventions, the reported improvements may have been mostly a placebo effect.

^*An important subsidiary question is whether patients actually stick with the anticlotting regimens. Anecdotal evidence, at least, suggests that depending on their insurance, many probably don’t, which is dangerous. The rights to Plavix (clopidogrel) are owned by Sanofi-Aventis (in Europe) and Bristol-Myers Squibb (BMS). As of early 2007, there is a complicated patent dispute between BMS and several generic drug makers as to whether the clopidogrel patents expired in 2003, or are good until 2011, as BMS maintains. In addition, Sanofi-Aventis and BMS are under investigation for possible antitrust violations for attempting to forestall potential generic competition. Annual sales of Plavix exceed $6 billion.

^*As this is written, there is great enthusiasm for “pay for performance,” or P4P, systems. There are in fact some protocols so widely accepted—almost all cardiac patients should be on aspirin regimens—that it makes sense to penalize or reward practices for their compliance. But the number of instances where such rules can be definitively enforced, relative to the totality of health care services, will be very small. To speak of P4P as a major step toward higher quality and less expensive health care strikes me as a counsel of desperation.

^*There is a direct, and successful, precedent on this point. In the 1970s, so-called Medigap insurance programs to supplement Medicare coverage sprang up by the thousands. Provisions were so complicated that even the most capable seniors had difficulty choosing among plans and frequently found themselves overcharged. Congress stepped in in 1981 and mandated ten basic plans. All Medigap vendors had to market all ten plans, and benefit provision had to be identical from vendor to vendor. The market quickly stabilized around three or four of the plans, companies competed hard on price and service quality, and the vast majority of seniors purchased the coverage. It is a pity that a similar approach was not used for the new Medicare Part D pharmaceutical plan.

^*The formulas set the statewide actual, expected, and risk-adjusted mortality rates as identical.