The Determination and Management of Risk
by Siegfried Othmer | December 31st, 2004I had chosen this topic to end the year well before the recent tsunami brought home to us the disconnect between risks that we face–as societies, as a global community, as a species, and as individuals–and how we actually live our lives. This is a follow-up to the previous newsletter on Triage.
What got me started down this track was the December 5 issue of Parade Magazine, which featured an article by Michael Crichton that lampooned our tendency to overstate risk of future catastrophe and to over-react to it. First of all, this seemed strange coming from the person who has made a living by exaggerating risk: “The Andromeda Strain”, “Jurassic Park”, and “Prey.” But that’s science fiction, and this is real life.
Crichton mocks our concern about global warming, for example. We can’t seem to get our story straight. First it was global cooling, now it’s global warming that is scaring us. Then it was the population explosion that was supposed to have millions at the edge of starvation by the end of the century. Then there was the scare about power lines killing us through its huge electric and magnetic fields. And of course there was that famous non-crisis, the Y2K scare. And the fear promulgated by the Club of Rome about running out of resources. Then in passing he mentions saccharine, swine flu, endocrine disruptors, fluorescent lights, and just for good measure, killer bees.
Hold on. Global cooling and global warming are really part of the same story. It is that we are living through a lengthy period of extraordinary stability in climate, a warm period that occasionally punctuates what on the larger time-scale is still an ice age. So if we regard our current climatic history on almost any timescale larger than centuries, our current era is at minimum unusual in terms of its stability. Now that we are reaching the point of having our economic activity influence climate directly, it is certainly high time to worry about the consequences. Even those who worry about global warming know that this may be a prelude to–even a trigger for–another sudden deep freeze.
And as for the population explosion, was it not partly averted by the fact that action was taken: the one-child policy in China; the promotion of birth control in India. And is it irrelevant that we currently have more than one Billion children in the world significantly under-nourished? Was the Y2K scare not averted by the fact that billions were expended to clean up old software? Who knows by what margin of safety we skinnied by on January 1, 2000? As for swine flu, the matter of anticipating which of the emerging flu viruses may in practice constitute an epidemic hazard for humanity was not only a guessing game in President Johnson’s era, but remains so to this day. Was SARS anything to dismiss lightly?
As for saccharine, why not also include aspartame in the discussion while we are at it? We still have a problem here. And some people clearly do react badly to fluorescent lights. But these issues affect a subset of the population, and are not at all in the same league. The same goes for the power line scare, which was never anything that could threaten the society at large. And I will grant Michael Crichton’s dismissal of killer bees as a threat. The killer bees in our government are far scarier.
The Economist magazine took on apocalyptic visions in its December 18th edition, complete with a bolide crashing into the ocean on its cover. But it allows for a positive side effect of apocalyptic thinking in the simple fact that it suggests a progression through time as opposed to a cyclical character in history, even if things do end badly. “The apocalyptic narrative may have helped to start the motor of capitalism.” Both the sense of potential progress in humankind, and the radical division into good and evil, into light and darkness, started with Zoroastrianism in what is now Iraq, and has been carried forward to our day through the Judeo-Christian tradition. But the secular narrative of environmental catastrophe is alive and well also. We need to make sense of the evidence we have.
An article in Science in June of 2003 starts off: “The risk of asteroid and/or comet impacts is real.” It was finally determined that the mysterious flattening of a Siberian forest in 1908 was due to an asteroid thought to be 50-70m in diameter. No crater was ever found, so the object must have disintegrated in the atmosphere. The likelihood of such impacts is calculated to be one in a thousand years. An object only 100 times bigger can take us all out as surely as it once took out the dinosaurs.
The risk of bolide impacts is great enough that NASA has set a goal of detecting 90% of all objects greater than 1km in size that could potentially do us harm. More than that, it must do so at sufficient range that we can mobilize resources in time to deflect such an object. The fact that we are not there yet was illustrated late last year when an asteroid came closer to the earth than the moon. Had it been on an impact trajectory there would just have been time to ring the church bells and if necessary to get the President out of the White House.
It is in the matter of a water impact of a bolide that our concern merges with the earthquake/tsunami that just happened near Sumatra. A 1km bolide impacting on the ocean can create water height as much as 65 meters. That’s thirty to sixty times taller than the tsunami we just witnessed. These swells could carry inland as much as four kilometers. As with the tsunami, an impact anywhere in the ocean would be communicated with little attenuation to its periphery everywhere. If there is a problem here it is in not taking this threat seriously enough, rather than its opposite.
With regard to another influenza pandemic, the World Health Organization projects a potential world-wide mortality risk of some 7 million. However, it has been taken to task for this estimate, with other models showing potential mortality up to 50 million, and even 100 million. In its own defense, the WHO has argued that it has been deliberately cautious in its estimates in order not to damage its credibility when the worst possibilities laid out don’t actually occur. This, of course, has it entirely backwards. Nobody wants these estimates to be validated by reality. The point is to give policymakers a realistic estimate of risk in order that steps might be urgently taken so that the events may be avoided. A bureaucrat simply has to be able to tolerate the mocking of a Michael Crichton. Success lies in the failure of the prediction of catastrophe.
The current issue of Science supports this concern. In an article entitled “Will Vaccines Be Available for the Next Influenza Pandemic,” the authors state boldly: “The events leading to influenza pandemics are recurring biological phenomena that cannot be prevented.” They can only be managed, and that only if anticipated. With a range of 6-8 months for a global spread, there is a small window for new vaccines to have an impact.
What may also be at play here is a general aspect of “noblesse oblige” on the part of government: “Don’t scare the horses.” Government officials have a dreadful fear of panicking the public, even though the public has rarely shown any propensity to do so. Who knows how much we are being protected from adverse information because of how we might react? The most obvious case where prior knowledge might actually affect behavior is in the area of economics. The theme of financial instability has been a recurring one in these newsletters, by virtue of its analogy to CNS instability. On this occasion, I am dealing with it as a concern in its own right. My own fears were fanned by two experts in financial instability: Paul Krugman, the economist, and George Soros, the investor. [A third book for this list has just been published: Benoit Mandelbrot’s “The (Mis)Behavior of Markets,” a book that directly reconnects us with our thoughts about brain instability.]
Krugman reflects on the situation in 1998 when the economy of Russia collapsed, and the problem was exported to the United States by virtue of the investments of Long-Term Capital Management, an investment consortium involved in arbitrage, and led by eminent Nobel-winning economists, econometricians, and mathematicians. By virtue of superior and more comprehensive financial analysis, they could eke out small but consistent profits in international arbitrage. Since the risks were always balanced, they were small. Hence the investors were willing to be leveraged to the hilt, putting billions in borrowed funds to work in order to arbitrage hundreds of billions in financial instruments. The system was based on traditional equilibrium economic models, however, and the Russian default on its international obligations took them outside of the range of their assumptions. The collapse of LTCM might well have brought the web of international finance into a depression-scale crisis. It had to be avoided at all costs.
So Paul Krugman found himself at a briefing at the Federal Reserve at the time, and when officials were asked what they could do, the answer was: “Pray.” Now this was not coming from some Bush appointee recommending faith-based economics. It was just that there were not a lot of policy options. In fact the Federal Reserve had intervened massively already, but at some point things are just beyond anybody’s control. Krugmann’s book is titled “The Great Unraveling.” We have lots of reasons to hope that he is excessively pessimistic, but it would be better simply to be about the business of building more stability into the system.
What conclusions can we draw from the recent tsunami. Here we have an event that is much more predictable than either bolide impacts, financial collapses, or influenza epidemics. We know the incidence of large earthquakes fairly well. There have been five 9+ earthquakes and accompanying tsunamis since 1960. Tsunamis are reasonably predictable in magnitude. There is warning time because of the slow progress of the wavefront. Yet our global community appeared to be completely unprepared for the event. What might one expect of our institutions when matters are less certain than in this case?
Another recent book is helpful in assessing risk. Jared Diamond’s: “Collapse: How Societies Choose to Succeed or Fail.” In looking at the ultimate failure of a number of societies in the past, Diamond came up with some decisive factors: 1) environmental degradation, usually man-caused; 2) climate change; 3) hostile neighbors; 4) disruption of trade relations. The Vikings, for example, may have been done in by an unfortunate confluence of environmental degradation (their own) and an adverse change in climate. Diamond concludes: “The modern world faces environmental problems that are serious enough to do us in.” The same goes for his second category, climate change.
So resource depletion is a realistic concern, even if the Club of Rome was premature in its alarums. Climate change is a clear threat, even if the Bush Administration is trying to sweep it under the carpet. Globalization prepares the ground for the spread of contagion as well as democracy. And even if killer bees are not the problem they were thought to be, there are many real concerns about invasive species (e.g., zebra mussels in the Great Lakes) that cannot upset the juggernaut by themselves, but do gnaw at the economic basis of our prosperity. Then there are the potential economic shocks. If there is disinclination to take these seriously, one should just consider the example of Russia. There life expectancy among males is now back down to 58 years, less than in Bangladesh, with more than 50% of Russian males not making it to retirement age. And AIDS is still just on the horizon.
I conclude from all of these observations that our society is woefully under-reporting major risk factors, and hence is not devoting the necessary resources to ameliorating risk. Curiously, this parallels what happens at the level of an individual, a topic to which we now turn.
The story begins with a recent article in the December 6 issue of the New Yorker Magazine by Atul Gawande, MD, titled “The Bell Curve.” The topic is cystic fibrosis and the progress that has been made in dealing with it. Since the condition is genetic, and there is no cure, the remedies are largely behavioral. As thickened mucus slowly fills the small airways in the lung, regular measures must be taken to restore and maintain lung function. With such measures, a gradual but significant increase in life expectancy has been achieved over the decades. By 1966 life expectancy was ten years, already a vast improvement over earlier status. By 1972 it was up to 18 years, a gain of one year of life expectancy per calendar year. In 2003, life expectancy nationwide was up to 33 years, a further gain of half a year in life expectancy per calendar year.
However, in the best centers for CF treatment the life expectancy was already up to 47 years in 2003. One key point of the article was the existence of a broad bell curve in treatment efficacy among CF treatment centers, and secondly the persistence of a broad bell curve even as everyone learned the lessons of successful care. One might have expected a significant narrowing of the spread as laggards caught up with the standard. This was the point of Stephen Jay Gould’s analysis of the “co-evolution” of hitters and pitchers in professional baseball over the decades. A tightening of the distributions eventually made the appearance of another 400-hitter such as Joe DiMaggio very unlikely. But this same thing was not happening in CF care despite the huge upward migration in efficacy across the board. On the contrary, the best hospitals were continuing to outdistance their peers from the middle of the bell curve by a good deal, and in fact were threatening to become a separable, non-overlapping distribution entirely.
When these very successful treatment centers were more closely evaluated, it was found that they did not in fact have any secrets they had not shared. They were simply more insistent on patient compliance with their health maintenance routines, and they were very intolerant of lapses in self-care. As one of the pioneers describes it, there is ordinarily a chance of 0.5% per day of a CF patient acquiring a major lung ailment. With proper self-care that probably can be reduced by an order of magnitude, to 0.05%. This makes for a difference between 83% chance of avoiding major illness over a year versus 17%.
Unfortunately, this is not the perception of the patients, who are in fact also trying to assess their own situation and optimize their own quality of life. They may choose to neglect their odious and burdensome daily care at some point, and they find that they do not get punished immediately by getting sick. So, the preventive care regimen is progressively neglected, and eventually the probability of overt illness rises substantially. In the best treatment programs, there is sufficient involvement with the medical staff that such back-sliding is caught early and remedied. The patients are kept well by force of will of the medical staff, and by virtue of the organization of care that has been put in place. Here we have a clear instance in which appropriate medical care, and a large reliance on self-care, results in substantial improvements in quality of life and in life expectancy. Such is not always the case.
In a review of the benefits of mammography, Malcolm Gladwell (writing in The New Yorker, December 13), cautions that we may have set our expectations much too high for such an imprecise art as the reading of mammograms. In one comparison, ten board-certified radiologists were asked to look at 150 mammograms, of which it was known that 27 had developed breast cancer and the rest had remained healthy. One radiologist indeed identified 85% of cancers successfully, but at the expense of 65% false positives. For every cancer case correctly identified, there were two false positives. Another identified only 37% of the cancer cases successfully, and yet another flagged 85% of all 150 cases as suspicious. If we make the assumption that this 85% included all 27 of the actual cancer cases, the false positive rate was nearly 80%–four false positives for every correct identification.
As in the case of cystic fibrosis, there is a broad distribution of outcomes among practitioners. Only in this case it is very difficult to tell where good medicine lies and where the boundaries should be. As it is, radiologists are subject to many malpractice claims because of the unsupportable assumptions of rigor on the part of the consuming public. In addition to the trauma that all the false positives cause, there is the more fundamental problem that all the mammography campaign put together reduces cancer mortality at most ten percent, and that in turn adds mere days to the life expectancy of a 60-year-old. Mammography tends to catch the slow-growing tumors, and in these cases thorough breast exams may have comparable efficacy. We have here the opposite of what was the case for cystic fibrosis–a modest increment in health outcome for a high expenditure of resources.
What is clearly needed, just as in the society at large, is a much better calibration of risk, and to go along with that, a determination of the incremental benefit of certain procedures, be they diagnostic, preventive, or remedial. From that could follow a rational policy for the application of the resources for care–ok, let’s not mince words: rationing. We effectively have the rationing of health care already. Clearly lacking is visibility into the implicit decisions being made, and any kind of rational framework for making such decisions.
This is a broad topic that cannot be comprehensively addressed in this article. Let me just illustrate with an example given by Ivor Shapiro, writing in The Walrus under the Title “The End of Health Care.” A Pap smear every third year is sufficient to catch nearly all cervical cancers in time for treatment. The incremental benefit of an annual Pap smear is so marginal that the cost comes to $600,000 per additional life extended. Is that worth it? The answer cannot be given in isolation. There are many implicit judgments being made in our society that attach a value greater than $600,000 to a life extended. (One estimate, by David Cutler, Harvard economist, assigns a $100,000 value to each year of added life.) Perhaps a two-year cycle for the Pap smear would give us a better balance of costs and risk. In any event, this illustrates the kind of calculation that can go into the “rationalization” of health care.
One significant issue that is almost always missing from cost-benefit analyses is the fact that complicated medical procedures are not without their own risk. The cumulative risk is such that medical care by itself constitutes the largest cause of death in this country, and the management of iatrogenic illness is one of the major cost drivers in the US medical enterprise. One estimate puts the cost impact of medical errors at nearly $300B per year, which exceeds the cost of all prescription medications by a third, and amounts to the equivalent of 20% of all medical costs.
More important for our purposes is the realization that the whole issue of allocation of medical resources is freighted with great emotional weight, particularly as the end of life is approached. We observe that people will endure the most horrific tortures just for the sake of the possibility of a restored quality of life at some point in the future. We observe people surviving with very low “apparent” quality of life with chronic pain and severe disabilities–even locked-in syndrome–and they clearly choose life. And we observe that people are content–even happy–to be alive although constrained by Alzheimer’s. It is incumbent on relatives and caregivers to respect those value judgments, and not to supersede them with their own appraisals. But is this also an obligation for the government?
The answer is yes in the sense that the government implements our collective, communal values. But in many cases it will be the will of both the patient and the family to “spare no cost to extend this life,” and to that the government may not be able to say yes. We are back to needing a rational process to allocate end-of-life care that is publicly financed and supported. The more significant message for present purposes is that a small amount of preventive care, as in cystic fibrosis, can either displace far into the future, or even displace entirely, major medical expenditures. Yet both on the level of the individual, and on the level of society at large, we short-change the amelioration of long-term risk while piling on expensive end-of-life care.
Nothing illustrates this better than a cost-benefit analysis on a tsunami warning system in the Indian Ocean. The cost of implementing such a system would have been trivial in comparison with the current cost of rescue efforts, never mind any appraisal of the actual loss that has been suffered. But it is difficult for us to live life in the contemplation of risk–both on the personal and the state and international levels. At the state level, the problem can be solved by instituting a formal process of risk assessment. But at the level of the individual, things are much tougher. We are biologically geared toward near-term feedback loops. The long-term impacts must be rationally appraised; yet our own rationality has been much over-sold.
At the level of the society at large, survival as a functioning entity is much more important than the survival of any individual within it. Indeed, the death of the individual is programmed in. If we obtain the society’s “will to live” just by adding up the individual “will to live” of all the stake holders, we get a collective “will to live” that is far larger than anything that has found political expression. We are treating the survival of a functioning society much more casually than this calculation would suggest. As a society we are leading a most profligate life, and charging our credit card to boot. There is no care for the morrow. Indeed, one has to dismiss a whole raft of real problems in order to have Social Security, missile defense, and new nuclear weapons come to the top as risks to be attended to urgently. Ultimately the problem of inattention does get solved, in that we will just have one disaster after another to remind us of our neglect. We have just had one.
We have also placed our fortunes in the hands of entities–the corporations–which are known for both short-term and narrow-gauge thinking. Corporations, it has been pointed out, are sociopaths when characterized in mental health terms. They are fundamentally antithetical to the concept of socialization of risk, which is one of the principal purposes of government. And they both live and die by the sword: Fully one-third of Fortune 500 companies in 1970 no longer existed in 1983. The half-life of corporations in the twentieth century was 75 years. We don’t have a model here for societal survival.
Finally, what are the implications for us in neurofeedback? Our principal functions in neurofeedback are two-fold: 1) The amelioration of brain instabilities; and 2) the “tuning” of the system into better self-regulation in the steady-state and under ordinary challenge. Of these the more profound impact is obtained through the management of brain instabilities. These hold the most serious risk for the person, and they have the lowest likelihood of being well-managed medically.
On the other hand, our remedy also has its shortcomings. Instabilities often survive our best ministrations. In these cases we may need to move in the direction of continuous monitoring in order that an excursion into instability can be detected in real-time, and behavioral measures taken. Even a few seconds of warning might be helpful in cases of narcolepsy, hot flashes, vertigo, or absence seizure. Instabilities with longer cycle times, such as manic-depressive illness, may give us considerably longer warning time. What if we could anticipate a person’s episodic relapse into alcohol abuse by some hours or days? Just as we now wonder why a tsunami warning system was not in place in the Indian Ocean, one must ask why this obvious next step for neurofeedback is not high on our agenda.
Secondly, there is the issue of the appropriate place of neurofeedback in a rationally based, optimized cost-driven healthcare system. I hold out great hope that neurofeedback can be a key cost-reducer in elder-care, and will be a powerful instrument in the displacement of much expensive medical care at the end of life. Thus many of its benefits will be in the far-off future for those who undertake the training presently. But that rationale will not be sufficient as an enticement. Neurofeedback must establish and sustain itself on the basis of near-term positive results, not just distant hopes. In the case of clinical applications, such near-term results are plentiful. But with respect to the person who is not suffering or disabled, the challenge is very different. I suspect that matters are similar in the field of physical fitness. There may be a long-term objective with respect to longevity and maintenance of function, but those who actually stick with the regimen are probably those who derive a tangible near-term benefit as well.
In the case of neurofeedback, the near-term payoff must be in terms of reliably better performance, better emotional regulation, and better quality of life through enhanced awareness and depth of feeling–a chance to live life more fully, with a larger range of choices. It’s even better if such improvements are quantifiable. But in any event, to be accepted at the level we aspire to, progress through neurofeedback must impact quality of life in the near-term in an obvious way. It cannot simply look good on an actuary’s ledger. People will not work hard for a statistical likelihood. At the same time, doing neurofeedback just for “quality of life” must be minimally intrusive and minimally disruptive. We have good systems in place for remedial care, but the answer is not in yet for broad utilization within the society for non-clinical purposes. So we still face two major challenges: “always-on” monitoring and neurofeedback for emergency care, and neurofeedback in support of “positive psychology” and optimum performance.