Jeff was ready to die. An Oregon man in his late 50s with kidney failure, he had cleared all the legal and practical hurdles that can keep someone from pursuing a medically assisted death. All that was left for him to do was ingest the lethal medication. Or so he thought.
For much of his life, Jeff had worked as a long-haul trucker, sleeping in the cab of his truck and going fishing in between jobs. When his kidneys started to fail and he went on dialysis, he gave up trucking and became homeless. Jeff eventually stopped dialysis. He knew he wasn’t eligible for a transplant, and he didn’t want to spend the rest of his life hooked up to a machine.
By the time Jeff was admitted to a Portland nursing home and enrolled in hospice, he had come to accept that he was dying. And he wanted to go out on his own terms. When Jeff first made his request to use Oregon’s assisted dying law, no one was sure whether he would make it through the 15-day waiting period imposed by the law. He barely ate anymore and constantly threw up blood and bile. He lumbered around the facility bent over his cane, his skin so pale he looked like a ghost. But Jeff pulled through and at the end of the two weeks he secured his prescription for life-ending medication.
Early one Saturday morning, Jeff’s social worker Ada checked him out of his nursing home and drove him to her apartment. He had nowhere else to go. His facility prohibited assisted deaths on their premises and he didn’t want to die in a motel. When Jeff reclined on Ada’s L-shaped sofa later that morning, he requested that she join him in prayer. Jeff had recently returned to his Christian faith and he was anxious to have God on his side. “Do you think he’s going to forgive me?” he kept asking Ada. “Please, God, forgive me. Jesus, please welcome me.” Jeff drank the lethal cocktail through a straw. Within 20 minutes he was asleep. Twelve hours later, Ada saw a shadow out of the corner of her eye and looked over. Jeff was sitting up and awake. Then, he started vomiting profusely.
* * *
Today, 22 percent of Americans live in a state with access to a medically assisted death. Most people probably imagine this process, while emotionally difficult, to be technically straightforward: ingest the medication and die. The reality, however, is far messier. There is no magic pill that will end a person’s life, and physicians aren’t taught how to end someone’s life in medical school. Jeff was the eighth Oregonian to regain consciousness during the course of an assisted death. Others have undergone protracted deaths, sometimes up to four days. And though Jeff’s experience is rare—to date 1,657 people in Oregon have died under their state’s assisted dying law—it points to a larger issue: the complicated science of life-ending medications.
Dr. Lonny Shavelson, a Berkeley physician, knows that the pharmacology of dying is still in its infancy. He runs Bay Area End of Life Options, a practice specializing in medically assisted deaths. Over the past three years, Shavelson has been at the bedside of over 200 patients as they ingested the lethal medications. And he has faced his fair share of riddles. “Why did Prince die in an elevator after taking a small dose of fentanyl, when we are giving our patients huge doses of medications and they’re not dying in an hour? They’re dying in five or six hours.” Shavelson and his colleagues across the country are getting closer to unlocking the science of dying, he says, but they still have a way to go.
The letter of the law
When a committee of lawyers, physicians, and activists sat down to craft Oregon’s Death with Dignity Act in 1993, similar ballot initiatives had just failed in Washington and California. To appease strong opposition, lawmakers laced the Oregon statute with a long list of restrictions and protections. In contrast to all previous proposals, the Oregon measure no longer allowed for euthanasia, the act of a physician actively ending a person’s life with a lethal injection. Under the new law, patients would have to administer the life-ending medications themselves—a final safeguard meant to ensure the voluntary nature of their act.
The Oregon statute and other statutes that came after it use verbs like “ingest” or “take” to describe the process of self-administration. Many physicians have interpreted this language to mean any method that involves a person’s digestive system. In addition to taking the medications orally, patients can ingest them through a feeding tube or a rectal catheter—as long as they are the ones pushing down on the syringe.
Some physicians are unhappy about these limitations. They point to the experience of other countries with assisted dying laws, which shows that administering the medications through the bloodstream is much faster, safer, and more effective. In Canada, where qualified patients can choose between a self-administered death and a lethal injection by a clinician, more than 99 percent opted for an injection between January and October of 2018. For injections, the average time to death is 10 minutes.
That’s not the only wrinkle. Assisted dying laws in the United States don’t specify which medications to use, leaving the decision up to the prescribing physician. When Oregon became the first state to enact its assisted dying law in 1997, the Oregon Health Authority didn’t provide a list of approved medications. And pharmacy textbooks don’t offer any direction. In the absence of trials or studies, physicians have had to devise their own protocols.
At first, they looked in some unlikely places. They spent hours scouring pharmacological reports on animal deaths, accidental poisonings, and drug-induced suicides. Eventually, they settled on a handful of possible medications, knowing they would have to refine them later on. Pentobarbital (Nembutal) and Secobarbital (Seconal), both powerful and fast-acting barbiturates, soon became the gold standard in lethal medications for an assisted death. Physicians gave their patients 2.5 times the fatal dose. The medication put them to sleep, the sleep progressed to a coma, and the coma stopped the brain’s drive to breathe. Between 2001 and 2018, about half of all Oregon patients who used Seconal died within 25 minutes (20 minutes for Nembutal). But there were sporadic, unexplained outliers that took significantly longer, and five patients who ingested Seconal regained consciousness. Overall, though, the medications were working reasonably well—until they disappeared.
The pharmacology of dying
In the past decade, pharmaceutical companies have taken the most effective and affordable life-ending medications off the market. In 2011, Lundbeck, a Danish pharmaceutical company, stopped selling Pentobarbital in the United States, to curtail its use for capital punishment. That same year, the European Union instituted a general export ban on the medication. Soon after, the price of Seconal skyrocketed. Up until 2009, a lethal prescription of Seconal had cost less than $200. Over the next few years, Marathon Pharmaceuticals hiked the price up to $1,500 per 100 capsules. In 2015 Valeant, a Canadian pharmaceutical company now known as Bausch Health, acquired the rights to the medication and doubled the price to $3,000, one month after California proposed a bill to become the fifth state to legalize assisted dying.
It was around that time that physicians in Washington State started looking into alternatives. They formed a committee that included a cardiologist, a pharmacist, an anesthesiologist, and some veterinarians. Together, they set out to create their own formula. According to Dr. Carol Parrot, a retired anesthesiologist who has assisted over 300 patients in ending their lives, they needed the new protocol to accomplish four things: it had to be safe to handle, effective, cheap, and reliably available through compounding pharmacies. The group eventually settled on a mixture of four medications by the name DDMP. The compound was made up of diazepam and morphine—meant to sedate the person and depress their respiratory drive—and digoxin and propranolol to stop the heart. After the first 30 deaths, the group increased the individual doses and named the new formula DDMP2, which cost around $750.
On average, patients who ingested DDMP2 took much longer to die than those who could still afford Seconal. Until the end of 2019, the mean time to death for DDMP2 was over four hours in Oregon, and one death lasted as long as 47 hours. But Seconal had moved out of reach for many, and in 2018 fewer and fewer pharmacies were stocking it, knowing the current batch was set to expire by the end of the year. Since then, Seconal has become unavailable, forcing physicians across the country to rely exclusively on DDMP2.
Physicians started to prepare families for the possibility of a delayed death. Dr. Charles Blanke, an Oregon oncologist, saw how difficult long deaths could be on families. “The first hour or two, families are holding their loved one’s hand, they get in bed with them, they don’t want them to die. But by hour five, by hour eighteen, they are ready for their loved one to leave this earth. And it’s hard on everybody when they don’t.” It wasn’t necessarily that families minded longer deaths, but they minded not knowing about that prospect in advance. “The main difficulty is the family expecting them to die quickly. They get uncomfortable waiting for the last breath,” says Dr. Terry Law, a physician who was part of the committee that developed DDMP2. With preparation, some families have said they appreciated the chance to spend the extra time saying goodbye.
The search for better medications has been driven by what physicians perceive as unacceptably long deaths. Yet there’s no medical standard on how long a death is meant to last. For Shavelson, anything over two hours qualifies as a protracted death, anything over four as problematic. But some of his colleagues disagree with the assumption that a good death must always be rapid. In a standard hospice setting where the process of dying can easily take days, an 18-hour death might be considered quick. A former emergency physician, Shavelson is something of a renegade in the assisted dying world. When California enacted the End of Life Option Act in 2016, he launched his own private practice, focused only on patients considering an assisted death. Wiry and energetic, he taught himself the science of dying. Shavelson requests that all his patients wear a heart and oxygen monitor when they ingest the lethal cocktail. Having these data has allowed him to differentiate, for the first time, between a respiratory death, which usually happens within an hour as a result of someone’s system being flooded with sedatives, and a cardiac death, which takes longer.
Shavelson hypothesized that prolonged deaths had something to do with the cardiac medications not working as well as they should. He began isolating the digoxin and giving it to patients half an hour before the rest of the dose. Shavelson suspected that the tiny quantity of digoxin was getting lost in a sludge in the gut, and that the bowels took a long time to absorb it. With D-DMP2, he reported a reduced mean time to death of 1.3 hours. Jeff’s physician had followed Shavelson’s new technique, having Jeff ingest the digoxin in advance of the other medications.
Over the past year, Shavelson has fine-tuned the protocol, swapping the propranolol out for amitriptyline, an anti-depressant that can cause a fatal change in heart rhythm at high doses. With this change, 90 percent of his patients have died in under two hours. Physicians in other states have adopted the new protocol, reporting similar times to death.
There’s another issue that no one has yet been able to address: Life-ending medications taste awful. For the medications to work, patients must drink the entire 4oz cocktail in less than two minutes—otherwise they risk falling asleep on a partial dose. Sweet juices, sorbets, and hard liquors can reduce the bitterness and burning sensation, and in Oregon patients have experimented with swishing a flat Coke around in their mouth. While some of these techniques help, they don’t completely alleviate the problem. For many patients, life’s final taste is still repulsive.
When things go wrong
One puzzle has plagued prescribing physicians since the early days of the law: Why do the sickest-looking people sometimes have the longest deaths? It turns out that whatever illness is wreaking havoc on a patient’s overall health also compromises their digestive system. The medications can sit in the gut for hours before being absorbed. Shavelson and his colleagues came up with a red-flag checklist for these and other high-risk cases. Patients with strong hearts—those under 55 and extreme athletes—are on that list, as are people suffering from gastro-intestinal diseases such as pancreatic cancer, those with a high tolerance to opioids and alcohol, and very large patients. Physicians finally had a theory for why certain of their patients took so long to die, or failed to die.
In retrospect, Jeff ticked many of these boxes. He had been on oxycodone for his chronic pain and on diazepam for his anxiety. Perhaps he had built up such a high tolerance that the sedatives in D-DMP2 couldn’t keep him asleep. It’s also possible that Jeff suffered from gastroparesis, a condition that slows the passage of food from the stomach into the small intestine. He had been eating little and throwing up frequently. Or perhaps not all the digoxin had entered his body, clinging to the glass or straw. Before he woke up, Jeff had shown signs of agitation, causing Ada to worry that he might regain consciousness. She informed Jeff’s physician, who immediately called Jeff’s hospice provider to request medications to keep him sedated, buying more time for the medications to work. As a religious organization that opposes medically assisted dying, they refused to help, saying they don’t participate in the law.
Jeff’s death raises another question: Should the time of ingestion for an assisted death be treated as a medical event? Physicians like Shavelson highly recommend the presence of a clinician during an assisted death. Too many things can go wrong, he says, and families shouldn’t be left to fend for themselves. Mixing up the four medications can be stressful, as can administering the medications at the correct time intervals. Most families have no idea how to tell when someone has died, and some are alarmed when a person takes loud and disturbing breaths after they have slipped into a coma. Shavelson considers it good medical practice to be at the bedside for this process. It gives him the chance to reassess a person’s mental and physical capacity prior to their death, and to do any last-minute problem-solving. He typically recommends that the pharmacy hold a patient’s medications until they are ready to use them, under his or another clinician’s supervision.
Parrot believes that it’s her role to empower families to navigate an assisted death on their own, if that’s their preference. She provides them with all the resources they need and she makes herself available in case they call. Plus, she knows that families are rarely alone. In the Pacific Northwest, trained volunteers from End of Life Washington and End of Life Choices Oregon attend the majority of all assisted deaths. They offer support to patients and their loved ones, acting primarily in an advisory and non-medical capacity (they will, however, help with mixing up the medications so that families can focus on being with their loved ones).
An assisted death isn’t just science, Parrot says, it’s also an art. “I try to allow each patient to have input on how they want it to be. I had one patient who was estranged from his family and wanted to take the medicine alone. I told him that his family might feel badly for not being able to say goodbye. I was very happy to hear that he was surrounded by family when he died.”
One step to improving the science of assisted dying requires better and more standardized data collection. Dr. Matthew Wynia, a bioethicist at the University of Colorado, says that formalizing assisted dying research and data sharing is imperative. He is slated to help lead the research wing of the newly founded American Clinicians Academy on Medical Aid in Dying, a national organization meant to serve as a resource portal for physicians and patients. Another step is to enhance cooperation with hospice and palliative care providers, whose training and expertise in caring for the dying will be a vital asset.
As new states are rolling out assisted dying laws, a new cohort of physicians will be joining efforts to tackle the science of dying, building on a rapidly growing body of knowledge. Though extremely uncommon, Jeff’s death is a painful reminder that, under current assisted dying laws in the United States, an entirely predictable death isn’t yet realistic.
After he awoke, scared and confused, Jeff was admitted to the hospital. When Ada visited Jeff the next day, she found him in the “comfort care” ward. Nurses had given him a mild tranquilizer to help with his nausea and vomiting. Over the next four days he declined rapidly. Ada was with Jeff when he died. As he lay flailing in his hospital bed, yanking at the sheets, she held his hand. “He didn’t die peacefully,” she said, her voice almost breaking. “It was a terrible death.”
Anita Hannig is associate professor of anthropology at Brandeis University, where she teaches classes on the cultural dimensions of death and medicine. Over the past four years, she has studied how access to a medically assisted death is transforming the ways Americans die. She is the author of Beyond Surgery, and her work has appeared in Cognoscenti, the Conversation, and Undark Magazine. You can follow her on Twitter @AnitaHannig