by Atul Gawande
IN 1971, a thirty-three-year-old internist named Harris Berman decided to do things a little differently. He and a friend who had just completed his general-surgery training moved back to their home state of New Hampshire, to the town of Nashua. They joined up with a pediatrician, a family practitioner, and an obstetrician. Together they offered health care to patients for a fixed annual fee, without any bills to insurance companies. It was a radical experiment. They paid themselves fixed salaries of thirty thousand dollars a year, a modest income for a doctor at the time, with no differences between specialties. They also bought reinsurance coverage to pay for costs that exceeded fifty thousand dollars, as Berman remembers it, in case a patient developed a catastrophic illness.
The scheme worked. Berman, who is now sixty-eight years old, told me the tale. They called themselves the Matthew Thornton Health Plan, after a physician who was one of New Hampshire's three signers of the Declaration of Independence. They were essentially an HMO, though a very tiny one. Within a short time, about five thousand patients had signed up. The doctors thrived, and there were remarkably few hassles. In the beginning, they didn't have any subspecialists, so when patients were sent to an ophthalmologist or an orthopedist the Thornton doctors had to pay for the visits. Eventually, they asked the specialists to accept a flat fee each month and dispense with the paperwork.
"Some accepted," Berman said. "And the effect on care was remarkable. The urologists, for example, suddenly became interested in having us understand which patients they really needed to see and which ones we could take care of without them. They came down and gave us talks--how to work up patients with blood in the urine and decide which ones you had to worry about. The ophthalmologists came down and told us how to take care of itchy eyes and runny eyes. They weren't going to make more money seeing these unnecessary patients, and they found a way to make sure we became more efficient."
After a few years, the Matthew Thornton Health Plan started to be cheaper than other insurers. Employers caught on and enrollment soared. Berman had to bring in more doctors. That's when things got more complicated. "In the beginning, we were all committed," he said. "We worked hard--long hours, a lot of dedication, young and hungry. Then, as we started to get bigger and bring in more staff, we found that others joined for other reasons. They liked the salaried lifestyle--the idea that being a doc could be a job rather than a day-and-night commitment. Some were part-timers. We began to see people looking at their watches as five o'clock approached. It became clear that we had a productivity problem." Also, when they tried to bring in specialists to work full-time with the group, the specialists refused to accept the same salary as the others. In order to get an orthopedic surgeon to join, Berman had to pay him considerably more than what everyone else got. It was the first of many adjustments he had to make in how and what to pay his fellow physicians.
Over the course of thirty years, Berman told me, he ended up trying to pay physicians in almost every conceivable way. He'd paid low salaries and high salaries and still watched them go home at three in the afternoon. He'd paid fee-for-service and watched the paperwork accumulate and the doctors run up the bills to make more money. He'd come up with complicated bonus schemes for productivity and given doctors budgets to oversee. He'd given patients cash accounts to pay their doctors themselves. But no system was able to provide both simplicity and the right balance of thriftiness and reward for good patient care.
By the mid-1980s, sixty thousand patients had joined the Matthew Thornton Health Plan, mainly because it had controlled its costs more successfully than other plans. It had become the second-largest insurer in New Hampshire. And now it was Berman and his rules and his contracts that the physicians complained about. In 1986, Berman left Matthew Thornton, and it was later taken over by Blue Cross. He went on to become the chief executive officer of Tufts Health Plan, one of New England's largest health insurers (where he earned a CEO's income himself). The radical experiment was over.
IN 2005, the United States spent more than two trillion dollars--one-sixth of all the money we have--on health care. This amounted to $7,110 per person. Government and private insurance split about 80 percent of those costs, and the rest largely came out of patients' pockets. Hospitals took about a third of the money; clinicians took another third; and the rest went for other things--nursing homes, prescription drugs, and the costs of administering our insurance system. Americans seem to be reasonably happy with their care, but they haven't liked the prices--insurance premiums increased by 9.2 percent in 2005.
Physicians' after-expense incomes are a fairly small percentage of medical costs. But we're responsible for most of the spending. For the patients I see in the office in a single day, I prescribe somewhere around thirty thousand dollars' worth of medical care--in the form of specialist consultations, surgical procedures, hospital stays, X-ray imaging, and medicines. And how well these services are reimbursed inevitably affects how lavish I can be in dispensing them. This is where money-mindedness becomes inescapable--and likewise the struggle between doing right and doing well.
I remember, twelve years ago, getting the bill for the heart surgery that saved my son Walker's life. The total cost, it said, was almost a quarter-million dollars. My payment? Five dollars--the cost of the copay for the initial visit to the emergency room and the doctor who figured out that our pale and struggling boy was suffering from heart failure. I was an intern then and in no position to pay for any significant part of his medical expenses. If my wife and I had had to, we would have bankrupted ourselves for him. But insurance meant that all anyone--either us or his doctors and nurses--had to consider was his needs. It was a beautiful thing. Yet it's also the source of what economists call "moral hazard": with other people paying the bills, I did not care how much was spent or charged to save my child. To me, all the members of the team deserved a million dollars for what they did. Others were footing the bill--so it's left to them to question the price. Hence the adversarial relationship patients and doctors have with insurers. Whether insurance is provided by the government or by corporations, there is no reason to think that the battles--over the fees charged, the bills rejected, the preapproval contortions--will ever end.
Given the struggles over payment, what's striking is how substantial medical reimbursements have continued to be. Physicians in the United States today remain better compensated than physicians anywhere else in the world. Our earnings are more than seven times those of the average American employee, and that gap has grown over time. (In most industrialized countries, the ratio is under three.) This has allowed American medicine to attract enormous talent to its ranks and kept doctors willing to work harder than members of almost any other profession. At the same time, we as a country have shown little concern for the uninsured. One American in seven has no coverage, and one in three younger than sixty-five will lose coverage at some point in the next two years. These are people who aren't poor or old enough to qualify for government programs but whose jobs aren't good enough to provide benefits, either. They face difficulty finding doctors who will treat them, unconscionable rates of bankruptcy from health care bills, and a proven increased likelihood that problems such as high blood pressure, heart disease, appendicitis, and cancer will go undetected or inadequately treated. Our byzantine insurance system leaves gaps at every turn. Some day soon that must change.
A FEW DAYS after the chairman of surgery offered me the job, I returned to his office and named my figure.
"That'll do fine," he said, and we shook hands. Now I am the one who's too embarrassed to say what I earn. We talked for a while afterward: about how to fit research in, about how many nights I'd have to be on call, about how to keep time for my family. The prospect of my new responsibilities filled me with both exhilaration and dread.
As the meeting was ending, though, I realized that there was one final important question I had not brought up.
"What are the health insurance benefits like?" I asked.
The Docto
rs of the Death Chamber
On February 14, 2006, a United States district court issued an unprecedented ruling concerning the California execution by lethal injection of murderer Michael Morales. The ruling ordered the state to have a physician, specifically an anesthesiologist, personally supervise the execution or else to drastically change the standard protocol for lethal injections. Under that protocol, the anesthetic sodium thiopental is given at massive doses that are expected to halt breathing and extinguish consciousness within one minute after administration; then the paralytic agent pancuronium is given, followed by a fatal dose of potassium chloride. The judge found, however, that evidence from execution logs showed that six of the previous eight prisoners put to death in California had not stopped breathing before technicians gave the paralytic agent; the finding raised a serious possibility that the prisoners had experienced suffocation from the paralytic, a feeling much like being buried alive, and felt intense pain from the potassium bolus. This experience would be unacceptable under the Constitution's Eighth Amendment protections against cruel and unusual punishment. So the judge ordered the state to have an anesthesiologist present in the death chamber to determine when the prisoner was unconscious enough for the second and third injections to be given--or to have a general physician supervise an execution performed with sodium thiopental alone.
The California Medical Association, the American Medical Association (AMA), and the American Society of Anesthesiologists (ASA) immediately and loudly opposed such physician participation as a clear violation of medical ethics codes. "Physicians are healers, not executioners," the ASA's president told reporters. Nonetheless, in just two days, prison officials announced that they had found two willing anesthesiologists. The court agreed to maintain their anonymity and to allow them to shield their identities from witnesses. Both withdrew the day before the execution, however, after the Court of Appeals for the Ninth Circuit added a further stipulation requiring them to personally administer additional medication if the prisoner remained conscious or exhibited pain. This they would not accept. The execution was then postponed (Morales remained on death row as of January 2007), but federal courts have since continued to require that medical professionals assist with the administration of any execution by lethal injection.
Execution has become a medical procedure in the United States. That fact has forced a few doctors and nurses, asked to participate in executions, to choose between the ethical codes of their professions and the desires of broader society. The codes of medical societies are not always right and neither are the laws of society. There are vital but sometimes murky differences between acting skillfully, acting lawfully, and acting ethically. So how individual doctors and nurses have sorted these out and made their choices interested me.
The Morales ruling is the culmination of a steady evolution in methods of execution in the United States. On July 2, 1976, in deciding the case of Gregg v. Georgia, the Supreme Court legalized capital punishment after a decadelong moratorium on executions. Executions resumed six months later, on January 17, 1977, in Utah, with the death by firing squad of Gary Gilmore for the killing of Ben Bushnell, a Provo motel manager.
Death by firing squad, however, came to be regarded as too bloody and uncontrolled. (Gilmore's heart, for example, did not stop until two minutes after he was shot, and shooters have sometimes weakened at the trigger, as famously happened in 1951 in Utah when the five riflemen fired away from the target over Elisio Mares's heart, only to hit his right chest and cause him to bleed slowly to death.)
Hanging came to be regarded as still more inhumane. Under the best of circumstances, the cervical spine is broken at the second vertebra, the diaphragm is paralyzed, and the prisoner suffocates to death, a minutes-long process.
Gas chambers proved no better: asphyxiation from cyanide gas, which prevents cells from using oxygen by inactivating a vital enzyme known as cytochrome oxidase, took even longer than death by hanging, and the public revolted at the vision of suffocating prisoners fighting for air and then seizing as their ability to use oxygen shut down. In Arizona, in 1992, for example, the asphyxiation of triple murderer Donald Harding took eleven minutes, and the sight was so horrifying that reporters cried, the attorney general vomited, and the prison warden announced he would resign if forced to conduct another such execution. Since 1976, only two prisoners have been executed by firing squad, three by hanging, and eleven by gas chamber.
Many more executions, 74 of the first hundred after Gregg and 153 in all, were by electrocution, which was thought to cause a swifter, more acceptable death. But officials found that the electrical flow frequently arced, cooking flesh and sometimes igniting prisoners--postmortem examinations often had to be delayed for the bodies to cool--and yet in the case of some prisoners, it took repeated jolts to kill them. In Alabama, in 1979, for example, John Louis Evans III was still alive after two cycles of 2,600 volts; the warden called Governor George Wallace, who told him to keep going, and only after a third cycle, with witnesses screaming in the gallery, and almost twenty minutes of suffering, did Evans finally die. Only Florida, Virginia, and Alabama persisted with electrocutions with any frequency, and under threat of Supreme Court review, they too abandoned the method.
Lethal injection now appears to be the sole method of execution accepted by courts as humane enough to satisfy Eighth Amendment requirements--largely because it medicalizes the process. The prisoner is laid supine on a hospital gurney. A white bedsheet is drawn to his chest. An intravenous line flows into his arm. Under the protocol devised in 1977 by Dr. Stanley Deutsch, the chairman of anesthesiology at the University of Oklahoma, prisoners are first given 2,500 to 5,000 milligrams of sodium thiopental (five to ten times the recommended maximum for ordinary therapeutic use), which can produce death all by itself by causing complete cessation of the brain's electrical activity, followed by respiratory arrest and circulatory collapse. Death, however, can take fifteen minutes or longer with thiopental alone, and the prisoner may appear to gasp, struggle, or convulse. So 60 to 100 milligrams of pancuronium (ten times the usual dose) is injected one minute or so after the thiopental to paralyze the prisoner's muscles. Finally, 120 to 240 milliequivalents of potassium is given to produce rapid cardiac arrest.
Officials liked this method. Because it borrowed from established anesthesia techniques, it made execution more like familiar medical procedures than the grisly, backlash-inducing spectacle it had become. (In Missouri, executions were even moved to a prison-hospital procedure room.) It was less disturbing to witness. The drugs were cheap and routinely available. (Cyanide gas and 30,000-watt electrical generators, by comparison, were awfully hard to find.) And officials could turn to doctors and nurses to help with technical difficulties, attest to the painlessness and trustworthiness of the technique, and lend a more professional air to the proceedings.
But medicine balked. In 1980, when the first execution was planned using Deutsch's technique, the AMA passed a resolution against physician participation as a violation of core medical ethics. The resolution was quite general. It did not address, for example, whether pronouncing death at the scene--something doctors had done at previous executions--was acceptable or not. So the AMA clarified the ban in its 1992 Code of Medical Ethics. Article 2.06 states, "A physician, as a member of a profession dedicated to preserving life when there is hope of doing so, should not be a participant in a legally authorized execution," although an individual physician's opinion about capital punishment remains "the personal moral decision of the individual." The code further stipulates that unacceptable participation includes prescribing or administering medications as part of the execution procedure, monitoring vital signs, rendering technical advice, selecting injection sites, starting or supervising placement of intravenous lines, or simply being present as a physician. Pronouncing death is also considered unacceptable, because the physician is not permitted to revive the prisoner if he or she is found to be alive. Only two actions are acceptable: provision, at the
prisoner's request, of a sedative to calm anxiety beforehand and signing a death certificate after another person has pronounced death.
The code of ethics of the Society of Correctional Physicians establishes an even stricter ban: "The correctional health professional shall . . . not be involved in any aspect of execution of the death penalty." The American Nurses Association (ANA) has adopted a similar prohibition. Only the national pharmacists' society, the American Pharmaceutical Association, permits involvement, accepting the voluntary provision of execution medications by pharmacists as ethical conduct.
States, however, wanted a medical presence. In 1982, in Texas, the state prison medical director, Ralph Gray, and another doctor, Bascom Bentley, agreed to attend the country's first execution by lethal injection, though only to pronounce death. But once on the scene, Gray was persuaded to examine the prisoner to show the team the best injection site. Still, the doctors refused to give advice about the injection itself and simply watched as the warden prepared the chemicals. When he tried to push the syringe, however, it did not work. He had mixed all the drugs together, and they had precipitated into a clot of white sludge.
"I could have told you that," one of the doctors reportedly said, shaking his head.
After a second effort, Gray went to pronounce the prisoner dead but found him still alive. The doctors were part of the team now, though; they suggested allowing time for more drugs to run in.
Today, all thirty-eight death-penalty states rely on lethal injection. Of 1,045 murderers executed since 1976, 876 were executed by injection. Against vigorous opposition from the AMA and state medical societies, thirty-five of the thirty-eight states explicitly allow physician participation in executions. Indeed, seventeen require it: Colorado, Florida, Georgia, Idaho, Louisiana, Mississippi, Nevada, North Carolina, New Hampshire, New Jersey, New Mexico, Oklahoma, Oregon, South Dakota, Virginia, Washington, and Wyoming. To protect participating physicians from license challenges for violating ethics codes, states commonly promise anonymity and provide legal immunity from such challenges. Nonetheless, despite the promised anonymity, several states have produced the physicians in court to vouch publicly for the legitimacy and painlessness of the procedure. And despite the immunity, several physicians have faced license challenges, though none have lost as yet.