The Price We Pay

by Marty Makary

  In the opening of the book, I described my experience observing the predatory screening practices of doctors at church health fairs. At that time, I had shared my observations with Dr. Jim Black and Dr. Caitlin Hicks, my vascular surgery colleagues. Both were markedly disturbed by what was happening in the community. They told me about a surgeon in the area who they believed was often doing unnecessary stent procedures, generating as much as $100,000 per day.

  Susan Hutfless, a PhD epidemiologist on my team, hadn’t forgotten about that issue. We ran our new appropriateness metric in the national Medicare data and found an alarming trend. The expected procedure rate for claudication—the leg pain resulting from poor circulation—should be less than 10%. Also, surgeons who do the procedure should not routinely separate it into stages on different days. Despite this, the data revealed that some doctors operated on more than half of the patients they evaluated with claudication and routinely strung out the procedure over multiple days. One of the top people on the list in the United States was the very doctor Dr. Black had told me about. He alone was milking millions of dollars from the system.

  Susan ran the data several times, looking to see if there was any mathematical or clinical reason why this one doctor would be such an outlier. But every time she reviewed the data, it was clear that this finding was not a statistical error. Plus, it fit with what Dr. Black had heard. Months later, Susan asked to meet with me. She walked into my office, initially calm, but as she spoke, her voice cracked with emotion.

  “Marty, isn’t there something we should do about this doctor?”

  I understood her frustration. Doing this data analysis had big implications and put the people who saw the results in an awkward position. We had tried to get the professional association involved with the vascular procedures but hit bureaucratic roadblocks. It takes time for large organizations to obtain consensus within their members.

  Susan is highly professional and stoic. But as we discussed the dilemma, tears welled up in her eyes. “I just don’t understand. This person is hurting a lot of people.”

  Susan raised a good point. She confronted me with a moral decision: Now that we can identify extreme outliers in the data, do we have a duty to let them know?

  Susan knew the politics involved and the importance of engaging specialty societies rather than working around them. The societies were sometimes quick to respond, but other times they moved like molasses. In one instance, a specialty society’s leadership told me that letting outliers know they were outliers was a great idea but it wasn’t a priority of theirs.

  Susan’s ethical concern had a big impact on me. Thinking about the opioid epidemic alone—another manifestation of overtreatment—makes me wish something could be done immediately. Now that data on extreme outliers is available and experts agree on which practice patterns are inappropriate and even downright wrong, what should our society do about it? I believe we have a duty to act.

  CHAPTER 9

  Opioids like Candy

  For most of my surgical career, I gave out opioids like candy. I was unaware that about 1 in 16 patients become chronic users, according to the recent research by doctors at the University of Michigan. My colleagues and I didn’t realize we were fueling a national crisis. But today opioids are the leading cause of death in America of people under 50 years of age.1

  As a medical student and surgical resident, I spent thousands of hours learning how to diagnose breast cancer, how to cut out breast cancer, and how to recommend chemo and radiation after breast surgery. But at no point was I taught that the way we liberally prescribed opioids was creating an epidemic of addiction that would eventually kill more people in the United States than breast cancer itself.2

  My “aha” moment came when I watched my father recover from an operation. Coincidentally, it happened to be an operation I routinely perform. I customarily prescribed 60 opioid tablets when I did it. But that night after dad’s surgery, I watched him recover comfortably at home with a single tablet of ibuprofen.

  Wow. It directly contradicted my residency training, in which I was taught to give every patient a boatload of opioid tablets upon discharge. The medical community at large ingrained in all of us that opioids were not addictive and urged us to prescribe generously. And that’s exactly what we did.

  Coming Clean

  The hundreds of excessive opioid prescriptions I wrote in 2015 alone (the last year for which national data is available) were only a tiny fraction of the country’s 249 million prescriptions filled that year. That means physicians gave out the equivalent of one opioid prescription for every American adult. And in that year, the U.S. pharmaceutical industry produced 14 billion opioid pills. That’s about 40 pills for every American citizen. These were all funded by you, the American public, through tax dollars, health insurance premiums, and cash payments.

  Now that I see headline after headline about tragedies stemming from prescription narcotics, I feel horrible about how copiously I prescribed opioid pills. Each news story makes me angry that I didn’t know more about the bondage that can arise from a prescription. Each headline also makes me wish I had known then what I know now—that non-opioid alternatives can be extremely effective in managing pain.

  My dad practiced hematology: the science of blood disorders, leukemia, and lymphoma. He explained to me how we got into this mess. For decades, cancer patients were undertreated for pain. But then things swung to the other extreme when pharmaceutical companies sold us on the lie that opioids were not addictive. Then pain became a vital measurement in medicine. The consumerist pain rating system elevated pain as a leading quality measure, overshadowing actual medical quality. As the race to measure quality took off, things that were easy and cheap to measure rose to the top: readmission rates, patient satisfaction, and pain scores. The question “How often did the hospital staff do everything they could to help you with your pain?” became a measuring stick by which all U.S. hospitals were rated. This created a perverse incentive to distribute excessive opioids.

  When I reflect on the enormity of the opioid crisis, I’m amazed that politicians still talk about solving it almost exclusively in terms of how much money they can spend on treatment and putting naloxone, the overdose antidote, in every McDonald’s and Starbucks in America. Yes, treatment is vital. But the root cause of the problem is overprescribing. Ironically, as our country debates how we can get out of the opioid crisis, we continue to flood communities with these powerfully addictive, sometimes fatal pills. The opioid crisis is another manifestation of the broader problem of overtreatment in medicine. It is essentially an issue of appropriateness.

  Even now that the opioid crisis has blown up and become a top priority for many health care leaders, far too many Americans still walk out of a hospital, surgery center, or dental clinic with a prescription for opioids they don’t need. Take C-section, for example, one of the most common operations paid for by Medicaid. Some doctors appropriately prescribe 5 to 10 opioid tablets to certain patients after the procedure (in combination with non-opioid medication, as recommended by the American Pain Society). Meanwhile, other doctors are still doing what I did for years: prescribing every patient a bottle of 60 highly addictive narcotic painkillers.

  Perhaps the most embarrassing study that was released at the height of the opioid epidemic looked at surgical patients who stayed in the hospital after their procedure. The study showed that half of patients who did not take opioids on their last day in the hospital were still given a prescription for the potentially deadly painkillers when they went home. That was the situation for my patient Suzette Morgan.

  Suzette works at Johns Hopkins in the office of research administration. She knew me because she personally helped me compile and mail several research grant applications. As I’ve come to know her professionally during my years at the university, she came up to me one day and asked me to perform an operation she needed: a laparoscopic gallbladder removal. I agreed and did the procedure, which went well. Bu
t weeks later she came to my office to give me an earful about the 30 opioid tablets she was prescribed after surgery.

  “I didn’t take them while in the hospital after surgery, so why would I need them after leaving the hospital?” she complained. Humiliated, I apologized, realizing that someone on my team had given her the opioid prescription.

  I called an urgent meeting of my clinical and research teams to discuss opioid prescribing at our hospital. As a scientist of health care quality, I was perplexed. This epidemic had been going on for about a decade but I was unaware of any national best practices for what should be prescribed to a patient after a standard operation.

  Building Consensus

  For this urgent meeting, I invited our local expert in pain management at Johns Hopkins, and a few residents and nurses. About a dozen of us crowded around a table in my office. I asked the million-dollar question: What should we be prescribing when patients are discharged after a standard operation, like a gallbladder surgery, assuming the patient is an average person who does not have a preexisting pain issue? When the experts in the room began talking, instead of directly answering the question, they said, in varying ways, that it’s really a matter of surgeon style or preference. I rephrased the question: What should a prescription read for an average adult patient going home after a laparoscopic gallbladder surgery?

  I put our head of pain management on the spot. She responded by saying what is done rather than what should be done. “Some surgeons like to give a few, others like to give a lot,” she said. I responded by repeating my original question. She turned it back on me: “Marty, you’re a surgeon, you might know better than me what they should be prescribing.”

  “I don’t know,” I admitted. “That’s why I called this meeting. And to be very frank, I don’t actually write the prescriptions for my patients. My residents do.”

  I turned to my residents in the room. One of them pointed out that our electronic health record, EPIC, had an e-prescribing default that recommended a 30-day supply. An intern also explained, “Some chief residents will yell at us if we prescribe anything below 30 pills.” A different resident said prescribing too few could result in calls during off hours when it’s inconvenient to prescribe more. These reasons were frustrating to hear, but they were honest. I knew these unwritten rules from my “residency survival manual.”

  “This is pathetic,” I said. “We’ve been the number one hospital in the United States for 22 of the last 28 years in the U.S. News & World Report rankings. The experts in this room are at the mountaintop of academic medicine. But we can’t even agree on what patients should go home with?”

  Again, I pressed them on the number of pills we should be prescribing rather than what we are prescribing. One resident said 30, another said 50, and another said 60, then one of them got bold. “We prescribe based on what our last resident taught us,” he admitted. “To be honest, we don’t know because we don’t follow the patients after they go home like the nurse practitioners do.”

  Finally, I turned to my nurse practitioner, Christi Walsh. She calls every patient at home after surgery. She knew more about this than any of us. Christi silenced the room by answering my question in her usual matter-of-fact manner: “Marty, they don’t need any opioids.”

  We were all stunned, as well as embarrassed. “Most patients don’t need the opioids we’ve been prescribing after a laparoscopic gallbladder operation,” she explained. “Good patient education and non-opioid alternatives keep the vast majority of them comfortable.”

  Christi had said what no one else was willing to say. At that moment, it became clear to me. The field of surgery needed thoughtful guidelines on what we should be prescribing after surgery.

  Opioid prescribing policies have been introduced since the addiction crisis took center stage. But it’s mostly state governments and insurance companies setting limits that get imposed on every patient, no matter the circumstances. I knew as a surgeon that every operation carries a different level of invasiveness and a different level of pain. An open chest operation is a heck of a lot more painful than needle biopsy. An opioid prescribing guideline, in principle, needed to be procedure-specific. It also needed to exclude patients already on other pain medications, since that can change a patient’s tolerance level and pain threshold.

  How could one prescribing limit be applied to different types of operations? I asked the group in my office that day to move on and talk about another operation, an open hysterectomy. For that procedure, Christi again educated us. She said we should prescribe at most 15 opioid pills, based on the hundreds of conversations she had had with patients who had recovered from the procedure. Christi’s estimate would be affirmed months later when a study from Dartmouth found that 70% of the opioid pills surgeons prescribe are never taken.3

  In the absence of any published procedure-specific guidelines, it was time to do something. My colleagues and I decided to create guidelines for our own Johns Hopkins surgical group and, in the spirit of transparency, offer them for the world to see. We agreed that the best approach to creating new guidelines would be to invite a broad range of experts, including surgeons, anesthesiologists, pain specialists, residents, nurses, pharmacists, and patients to form an expert panel. That panel would issue a consensus statement on the maximum number of opioids a narcotic-naïve adult should be prescribed after each of 20 common surgical procedures. Recognizing that treating pain depends on many patient factors, we created a range rather than a number of opioids for each procedure. But the upper limit was a real ceiling—a limit that should not be exceeded, according to our panel’s consensus.

  To do this, we considered tapping the bureaucracy of our professional associations, but we were in a national opioid crisis and wanted to move fast. The next week, my research colleague Dr. Heidi Overton and I invited surgeons, pain specialists, nurses, residents, pharmacists and, most important, patients to an expert panel meeting.

  For three hours in a large room at the center of the Johns Hopkins Hospital, we went through 20 of the most common procedures in health care. We had a patient and a clinician discuss their opinions about what the opioid prescribing guideline should be for each operation. We used a consensus-building process called the Delphi method. Everyone on the expert panel voted, then listened to one another explain their votes, then revoted. The average of the revote would become the consensus guideline.

  The discussions were intense but productive. Patients described their pain after surgery in detail and doctors shared their experiences treating pain after surgery. Others in the room described their experiences caring for patients in the days after surgery. Interestingly, when it came time to vote for an opioid prescribing limit, patients always voted for fewer pills than the surgeons.

  For each operation, a consensus was finally achieved. For many of the procedures, we recommended between 10 and 15 opioid pills, and we never recommended exceeding 20. Of course, this did not mean that patients who had a genuine need for more opioids would not be able to get them. It applied only to the prescriptions that patients received when they were sent home.

  We didn’t want to wait for a medical journal to publish our guidelines. That would take six months if we were lucky, after submitting a formal manuscript perfectly formatted according to their stringent specifications about fonts, margins, and spacing. Instead, we created a website called SolveTheCrisis.org, where the day after the expert panel concluded we posted our opioid prescribing guidelines. The site got thousands of hits a day. Months later, the New York Times wrote a piece about the guidelines and we started getting tens of thousands of hits each day. People were eager for guidance on what doctors should be prescribing after common operations.

  Nine months later, we also published the recommendations in a formal article in the Journal of the American College of Surgeons.4 It would be the first ever procedure-specific opioid prescribing guideline. We didn’t stop. Working with a dentist in the school of public health, Dr. Owais Farooqi, we conve
ned another expert panel of dentists and dental hygienists. We used the same consensus approach to create opioid prescribing guidelines for dental procedures. And in collaboration with the organization Allied Against Opioid Abuse, we created educational videos5 for patients to watch before surgery, reminding them it’s normal for surgery to cause soreness. It’s pain that limits activity that we want to aggressively treat, beginning with non-opioid alternatives in most cases. Opioids are an option, if patients choose, to treat severe pain, but they should know that they carry a risk of addiction, which can be fatal. Usually when I mention that to patients, they beg me to prescribe something else when they leave the hospital—a nudge for good.

  During and after our work to create procedure-specific guidelines, I watched the government and some insurance companies continue to issue draconian policies that limited opioid prescriptions to a 4-, 10- or 30-day supply. How could anyone dictate hard-and-fast limits when the amount of pain resulting from every procedure was different? Drilling bone to do a hip replacement is far more of a shock to the system than a lymph node biopsy.

  Guiding Change

  The opioid crisis created an obvious opportunity to use the Improving Wisely approach to look at a surgeon’s prescribing patterns in the national data. We wanted to see which clinicians had a pattern of prescribing far too many pills after certain operations. Our analysis excluded patients with preexisting opioid use or pain syndromes. We wanted to include only patients who had not previously taken narcotics, which is an important factor that clinicians should take into account when prescribing opioids.

  The analysis uncovered a stunning range in the way doctors prescribed opioids. Some doctors fell within our Johns Hopkins expert panel guideline of zero to 10 opioid tablets after, say, a prostate surgery. Other surgeons were still routinely prescribing a whopping 50 or 60 tablets to their narcotic-naïve patients.