Q&A with Dr. John Abramson: Overdosed America
I wrote a post earlier this week about a Nieman Reports article by Dr. John Abramson, a clinical instructor at Harvard and outspoken critic of the pharmaceutical industry. After serving in the National Health Service Corps, Abramson worked as a family physician for 20 years in Massachusetts. He wrote the book, Overdosed America, which was published by HarperCollins in 2004. I spoke with him about his work, his view of journalism and what he recommends for patients trying to sort marketing myth from medical truth.
Here is a recap of our conversation, which has been edited for clarity and space:
Q: You think journalists don't get it right very often when reporting on drugs. You must understand that journalists, unlike academics, don't have the time or the space to dig into all the nuances of a topic. What do you expect from writers who are always crashing on deadline?
A: It's a tremendously difficult problem. What is reasonably expected of a reporter trying to translate extremely complex studies for readers? The percent of trials that are commercially funded has grown to about 85%. The locus of control has changed from the universities to the pharmaceutical companies. And journalists will tell you that they have not had an increase in the resources necessary to vet the integrity of these trials.
Q: But why should we trust researchers who work for pharmaceutical companies any less than academic researchers? They want to come up with the best possible medicine, right?
A: The bottom line is the odds are five times greater that commercially-sponsored trials will say the sponsors' drugs are the drugs of choice.
Q: Reporters read a study like the JUPITER trial, which shows Crestor decreasing the chance of someone dying by 20%. That sounds pretty good. But you say that they should focus on whether there is an overall benefit from the drug. Isn't living an overall benefit?
A: In that case, the primary outcome measure was reduction in cardiovascular events. And if you look at the decrease in cardiovascular events there was a nearly 50% decrease. That sounds good. But there was no decrease in serious illness overall, no difference between Crestor and the placebo. In fact, the reduction in cardiovascular illness was counterbalanced by an increase in other illnesses. The cardiovascular system is very important, but so is the rest of our body. In just cold hard terms, they had to spend more than $5,000 just in drugs to prevent one death. That doesn't include the cost of screening and the cost of the follow-up doctor visits or the laboratory tests that are necessary. The finally tally would be significantly more than $5,000 per death prevented.
Q: I'm sure a lot of people would be happy to spend $5,000 or more if they knew it would prevent them from dying.
A: But we don't know that it would prevent you from dying for sure because the study was stopped so soon. If you look at the death rates in both groups, there was a dramatic increase in the number of deaths that occurred in the placebo group right before the study was stopped. Had the study been allowed to continue, it is very unlikely that there would have been a statistical difference between the two groups.
Q: You're saying they should have continued the study to see what the trend would do?
A: At the very least the study's authors should have said this reduction in the number of deaths has to be taken with a grain of salt. You have a study that is sponsored by the drug company and the members of the data safety monitoring board are hired by the company and you get this unusual random fluctuation of deaths in the placebo group. The committee decides to stop the study, and suddenly it looks like there was a big difference between the two groups. If the death rate went down in the treatment group, you might argue that it was the drugs making a difference. Instead, you have an increase in the deaths in the placebo group and that has nothing to do with the drugs.
Q: Let's say my father has a heart problem. He takes Crestor and lives. Other people he knows don't take Crestor and die. You say the money spent on Crestor should have been spent on other healthcare needs, but my family is thrilled to have dad around. Explain.
A: Maybe your dad did live because he took Crestor. But in this study for every 400 people treated for a year there was 1 less death. The odds aren't that great that it's going to help your dad. And you are getting at the other major problem. Is the purpose of this study to help doctors and policy makers learn how to prevent cardiovascular disease or is the purpose of the study to sell more drugs so Astra Zeneca makes more money? The people in this study are overfed and under-exercised. Everyone agrees the best way to prevent them from having a cardiovascular event is to help them adopt a healthier lifestyle.
Q: This is a big issue that you have with the way scientific trials are conducted. They start with the assumption that drugs are always the answer. Lifestyle choices, exercise, healthy eating are not usually part of the equation. Is that right?
A: They're not part of the equation. The studies are done to maximize investors' return on investment. They're not done for philanthropic reasons.
Q: But drugs are a little like an on/off switch. Easy to measure. Lifestyle choices are much more complicated and subjective. I know how many pills Andy took and David did not take, but I'm not sure how many miles Dan really walked or how much he ate?
A: There's a much easier way to do it. We know from studies like the diabetes prevention study that when people are randomized to lifestyle interventions that has a positive effect. People in Lyon, France, who had heart attacks were randomized to eat a Mediterranean diet. That counseling was 2.5 to 3 times more effective at reducing future heart problems than taking a statin. So let Astra Zeneca pay for the study but have the NIH add a lifestyle arm to the study so you could have a two-by-two table. Drugs vs. no-drugs. Lifestyle modification vs. no lifestyle modification.
Q: You compare the formulas for drugs and the trials that study those drugs to the formula for the world's most famous soft drink. Isn't that a little specious? Nobody is claiming that Coke can cure heart disease, but these drug companies are truly making products that attack illnesses.
A: It's not specious. The problem here now is that almost all of the studies that inform doctors how to treat patients are commercially sponsored and the data from those trials is not released even to the researchers who write the article. Rarely do all if any have free access to all of the raw data. That's where it is similar to the secrecy surrounding the recipe for Coke.
Q: But don't they deserve to keep some of these product making secrets secret?
A: The problem is that the knowledge about the drugs is the most important product of the drug companies. And that knowledge has been commoditized. Doctors are taught to accept at face value the validity of medical and scientific information that comes from certain sources in certain formats. But there is no verification process because the companies control the data. It's very analogous to what's going on in the financial world.
Q: And how should we verify these studies?
A: You need a Sarbanes-Oxley type law for the integrity of medical science. We are in the kind of situation now we were in with Enron. It was clear from Enron that there were too many conflicts of interest. So we changed the law to say that the CEO is civilly and criminally liable for the accuracy of a company's financial reports. If CEOs of drug companies were held liable for withholding evidence from the medical review board, the problem would stop quickly. Hundreds of millions and even billions are being paid by the drug industry to settle litigation for alleged fraud. But it's very unlikely that any of the perpetrators will go to jail. Even with Vioxx, you have 47,000 personal injury cases. About 37,000 people died. $10 billion worth of the drugs were sold, and they had to pay back $4.85 billion in a global settlement. But no one is personally accountable.
Q: Can I ask you a personal question? Do you or your family members take any prescription medications regularly?
Q: OK. Well, how do you decide what to prescribe for your patients?
A: I haven't practiced since 2002. But the advice I give to people is if your doctor recommends a medication, ask if there is a lifestyle change alternative. If not, ask if there is a generic medication that will do the job. Number one, it's cheaper. Number two, a generic version means the drug has been on the market at least seven years. It's unlikely that there will be any big surprises. Number three, you will k now that your doctor is prescribing that drug because it's the best drug for you not because a drug company has marketed to that doctor through an article in a medical journal, through a continuing medical education course, through a pizza at lunch or through free prescription pads.