How many medical treatments have you endured that may have been ineffective- or even done you more harm than good? How often are the negative side effects of medical intervention greater than the benefit they provide? In this episode, EconTalk host Russ Roberts may have met his skeptical- er, nihilist- match. Philosopher Jacob Stegenga joins Russ to talk about his new book, Medical Nihilism.
As you might guess, Stegenga is very skeptical about the efficacy of medical interventions, particularly pharmaceuticals. (Though he also tells a fascinating story about his own torn ACL toward the end of the conversation…) Stegenga argues that there are very few “magic bullets” in medicine. This does not, he asserts, make him purely anti intervention.
1- What characterizes a medical “magic bullet,” according to Stegenga? He and Roberts stress it’s not that drugs that aren’t magic bullets are useless… Why do most interventions fail to become magic bullets? How can we get more and more magic bullets in the future?
2- Why does Stegenga argue that we should be most skeptical of first-person reports of a drug’s efficacy? (Hint: You might want to revisit this episode with Gary Greenberg for part of the answer. You might also want to look at this episode with Eric Topol for a slightly different perspective…)
3- The conversation takes up side effects of medical intervention, and Roberts notes that Stegenga’s book has relatively little discussion of cancer. What are the forces that cause us to underestimate the harm of intervention and overestimate the benefits, according to Stegenga? How do you think Steven Horwitz would reply to Stegenga?
4- “… there’s an inverse correlation between trial quality and measured effect size in the trial.” What does this mean, and what does it suggest about the possibility of detecting harms? Does this constitute a “principled argument” for medical nihilism, as Stegenga suggests?
5- Stegenga states, “… this argument from economists that the FDA standard is killing people–doesn’t carry much weight… I think you could go even further and say the economists’ standard would kill orders of magnitude more people, because more harmful drugs would get through the regulatory standard. Thereby killing a lot of people.” Is the FDA- too lenient? How does this compare with Charles Hooper’s perspective in this Econlib article?
READER COMMENTS
Arde
Apr 9 2019 at 10:35am
Question 3. According to Stegenga, the incentive mechanism is the main reason why the harms of intervention are underestimated and the benefits are overestimated. The trials are done or are commissioned by the pharmaceutical companies and their aim is to get the drug approved. Consequently, there is no incentive to be honest about harms and all incentives are there to overestimate the benefits. This can be done, for example, by p-hacking and selection of an inappropriate trial population.
Steven Horwitz, when faced with cancer diagnosis, suggested to take every pill, make every appointment and learn more about how to treat the disease. Horwitz disregarded the costs side, incl. opportunity cost, of his advice. The pills can have side effects, which can lower the quality of life. Some pills can be extremely expensive and there might be other uses of the money. There is an opportunity costs of making every appointment, for example, less time spent with family and friends, less time for contemplation. The reason why Horwitz disregards the cost side could be that his own life has a very high value in his utility function and he is ready to pay any cost to extend it (or to increase the probability of extending it). I read in Wikipedia that he is awaiting for bone marrow transplantation, my best wishes for him and I hope he will have a good recovery. Placing an infinitely high value on one’s own life is a very understandable position and many people would feel the same. But not all.
The question why the costs are often disregarded in the case of cancer could be complemented with the idea of another Econtalk guest – Robert Aronowitz, who told the story about the 19th century woman suffering from cancer. She was ready to undergo an operation on kitchen table without anaesthesia to be sure that she has done everything before she dies Robert Aronowitz referred to Daniel Kahneman’s idea of anticipated regret as decision making heuristic. When faced with something bad, we want to make sure that we have done everything possible so that we don’t have regret and feelings of guilt.
In the conversation with R.Aronowitz Russ Roberts mentioned that the quality of life is often disregarded, which is another reason why costs are underestimated. Doctors typically give a much lower weight to quality of life as compared to extension of life or reduction of the probability of death. Aronowitz illustrated this with the example of prostate cancer treatment, which can cause incontinence, impotence and blood clots in legs.
Arde
Apr 9 2019 at 11:21am
Question 3 continuation.
I can add three more reasons on why the costs of medicines and treatment are underestimated. Number one – there are different time spans for benefits and harm. The benefits of treatment usually show up immediately, but harms often take more time to develop. For example, some drugs can have carcinogenic effects, and cancer takes long time to develop. As an extreme example, some drugs can affect the DNA and the negative impacts might show up only in the next generation. When doing an RCT for drug approval, the evaluators are not waiting for decades to estimate the impacts. The differences in time span affects also our decision making as patients. When we suffer from pain, we want pain relief now. We don’t care that painkillers in long term can lower the immunity and damage stomach, kidneys and liver. Another example comes from the Econtalk guest Banerjee in his book “Poor Economics”. When being sick, people in India often take steroids, which make them instantly feel better. However, they disregard (or are not aware of) the long term cost that steroids increase the speed of ageing. Banerjee writes that there are many 30 year olds in India who look like 70 year olds. Hyperbolic discounting – the tendency to give higher weight to present than future – is quite common.
Reason number two – excessive specialization of medical profession where each specialist is interested in the part of body of his expertise and not in the body as a whole. This has a tendency to overestimate the benefits of treatment that improves “his” part of body and disregard the negative effects on other parts of body. For example, many dentists take an x-ray during a regular check-up each six months even if the patient does not have any complaints. The dentists say that there can be cavities and decay, which can be seen only by x-ray, and the sooner they detect them the better. These dentists do not care about potentially excessive exposure to radiation because this will have a long-term effect and will affect other parts of body (for example, thyroid), not teeth.
Reason number three – when doing clinical trials the number of outcome variables for benefits is much smaller than the number of potential outcome variables for harm. This leads to situation when only a limited number of harms is tested, as it is impractical and prohibitively expensive to test all possible harms. For example, the benefits of a drug are tested by measuring its impact on a limited number of outcomes related to disease x it claims to cure, and usually there is only one primary outcome. However, when measuring the harms, the researcher might not be sure which harms to measure. There might not be a very well developed theory about harmful effects, and body is a very complex system. There are many ways how the new drug could affect other parts of body, and these effects are often difficult to predict. As a result the researcher might pick up some outcomes related to harm. When he estimates that that there are no effects on these outcome variablesm, he can conclude that there is no harm overall. However, it is possible that there might be a harmful effect on other outcomes which he did not measure. This leads to underestimation of harms.
Michael McEvoy
Apr 16 2019 at 7:54am
Arde – I agree with all your points . Very well put. I do think Stegenga was naive or uninformed regards the methods needed to track down harms . It is very costly as soon as you get into the 1 in 10000 users category . Yet of course, some drugs have millions of users .
Arde
Apr 17 2019 at 6:31am
Michael McEvoy.
Thank you. Yes, indeed you mentioned another reason why the harmful effects are underestimated. If the effects are rare, you need a very large sample size to detect them. Because of this some harmful effects go undetected in RCTs.
But I would not agree that Stetenga is uninformed about the methods. I had an impression he knows the methods very well, he just focused more on the incetive side of the story, which is the most important part. I disagree with him about some other things he mentioned. I am more optimistics about modern medicine and I appreciate more what it can do. I am also less dismissive about the patient reports as an evidence. He mentioned three reasons why we should not trust first-person reports, and I did not find them very convincing.
Comments are closed.