How to Combat Heart Disease Worldwide
by David S. Jones
Every day all over America, ambulances whisk people with chest pain into emergency rooms. Doctors take a history, perform a physical exam, order diagnostic tests, and, when suspicion of a heart attack is high, send the patient to coronary angiography. Once the results are available, the doctor and patient can review clinical trials, practice guidelines and other tools of evidence-based medicine. Such knowledge should enable good decisions about aspirin, thrombolytic therapy, angioplasty and bypass surgery. If medicine were nothing more than facts and data, the best decision would then be clear to the fully informed patient and doctor. But something else intervenes. Medical care, for instance, depends on where you happen to live and what doctor you happen to see. Consider rates of coronary angioplasty. Doctors in one Ohio town in 2003 performed angioplasty at a rate 10-times that in Honolulu, and even 3-times that in Cleveland, a mere thirty miles away. Such variation is ubiquitous in medicine. Practice varies between doctors within a hospital, between hospitals within a city, between cities within a state, and between different states or countries.
For centuries patients and doctors have had to decide whether to try a particular therapy for a disease. This is usually seen as a pragmatic question for the science and art of medicine. But it is also an epistemological question: what counts as a good decision, and how would you know? The history of coronary artery bypass surgery and angioplasty reveals the range of factors, some appropriate and others less so, that influence how medical decisions get made. This history raises important questions for health policy and social justice. There are now enormous disparities in access to cardiac care, with patients in much of the world unable to receive what would be considered the standard of care in the United States. What counts as an acceptable medical decision clearly depends as much on wealth and access to care as it does on medical science.
Coronary artery disease (CAD) provides a useful case study. Long the leading cause of death in the United States, and now in most countries worldwide, CAD has been the object of intense study by physicians. How can they tell if treatments work? One approach, which has developed into the field of evidence based medicine, involves comparing a new treatment to the existing standard of care and seeing if it improves outcomes. If it does, then the treatment is a good one and should be adopted. But this empirical approach is only one part of doctors’ thinking. They also try to understand the mechanisms of disease and to develop treatments that intervene on the mechanism to fix the problem. For most of the twentieth century, doctors believed that CAD was caused by the progressive growth of atherosclerotic plaques. A large plaque limits the flow of blood through an artery, depriving the heart’s muscle of needed oxygen and causing ischemia and then infarction – a heart attack. This model of blocked pipes inspired plumbing-based therapeutics. Cardiac surgeons learned how to use vein and artery grafts to bypass the obstructive plaques, while cardiologists learned how to use catheter-based balloons to compress the plaques and stents to prop the arteries open. With well over 1,000,000 of these procedures performed each year, they form a $100 billion industry.
Do they work? If you believe the plumbing model, they certainly should. Many patients, for instance, believe that angioplasty will reduce their risk of heart attack and death, and extend their life expectancy by as much as ten years. Such a dramatic benefit, however, has never been shown by the clinical trials of evidence based medicine. While revascularization extends the lives of the sickest patients, in many others it provides no benefit beyond what can be achieved with medications and lifestyle changes alone. What accounts for the mismatch between expectation and outcome? History provides several clues. Even as bypass surgery and angioplasty rose to prominence from the 1960s into the 1990s, pathologists and cardiologists abandoned the old plumbing model of progressive obstruction. They now believe that heart attacks are caused by the rupture of “fragile plaques,” often ones that are too small to be seen with angiography or treated with revascularization. The old model, despite its simple, intuitive logic, proved to be misleading basis for clinical decisions. Nonetheless it maintains a firm grip on the thinking of patients and doctors, who sometimes feel compelled to intervene even when they suspect that the procedure will provide little benefit.
Figuring out the safety of a medical treatment can be just as difficult. When coronary artery bypass surgery was introduced in the 1960s, everyone wanted to know whether it would be safe. After all, surgeons had to open the rib cage and stop the heart (in most cases) to sew in the bypass grafts. Surgeons quickly reassured themselves that the operation had an acceptable risk profile. This assessment, however, was based on a narrow vision: they focused on the risk of dying or having a heart attack during or soon after the surgery. While these were certainly the most important risks, they were not the only ones. When surgeons began doing open heart surgery in the 1950s and 1960s, they realized that the heart-lung machines that they used to keep the body alive while they stopped the heart could damage the patient’s brain, sometimes significantly. Patients suffered strokes, seizures, delirium and subtler cognitive deficits. The well-known risks, however, were largely ignored in the early years of bypass surgery. Of the first 200 studies published about bypass surgery, only four included serious discussions of these complications.
What had happened? It turns out that it is far harder to collect good data about the adverse effects of a treatment than about its desire outcomes. Sometimes the side effects cannot be predicted. But even when the complications are foreseen, many factors direct researchers’ attention away from them. Surgeons struggled to keep up with the demand for bypass surgery and had little time for thorough post-operative assessments of their patients. When they had time, they lacked the expertise needed to conduct elaborate neuropsychiatric assessments. Neurologists and psychiatrists, meanwhile, were more interested in developing treatments for their own diseases than in documenting the side effects of surgery. As a result, the full risk profile of a new treatment is often recognized slowly, only years after a procedure has been introduced into practice. Solutions to this problem are easy to design but difficult to implement. Researchers, clinicians and regulators could all take responsibility for studying complications more seriously than they do, but none yet have made it a priority.
If one set of factors leads doctors and patients to exaggerate the efficacy of interventions, and another set leads them to under-estimate risk, together they cause a serious problem. This paired error introduces an asymmetry into medical decision making that skews decisions in favor of intervention, something that drives the overuse of medical technology in the United States.
These examples show how hard it can be to generate the kind of information that doctors should use to make decisions, information about the efficacy and safety of procedures. Another set of examples reveals an equally concerning problem: evidence that a wide array of non-clinical factors influence medical decisions. The simplest way to illustrate this problem is with the geographic variations in medical care. If you keep track of the millions of medical decisions made each year by patients and their doctors, you can estimate the utilization rate of any particular procedure in different hospitals, cities, states, or countries. If evidence based medicine produced clear guidelines for medical decisions, and doctors followed them, then procedure rates should simply follow the prevalence of disease. But this is not what happens.
The problem was first recognized in 1938 when a British physician, J. Alison Glover, tabulated tonsillectomy rates and learned that they varied 27-fold across London neighborhoods, a pattern that could not be explained by underlying differences in the prevalence of disease. Although this finding was largely ignored for thirty years, researchers have now found geographical variation in surgical practice and in most other areas of medicine, everywhere that they have looked. The Dartmouth Atlas of Health Care, for instance, studied 2003 Medicare data and found that bypass surgery varied 4-fold between Mobile, Alabama, and Grand Junction, Colorado. Angioplasty varied more than 10-fold, from a high in Elyria, Ohio, to a low in Honolulu, Hawaii. Since the variation does not parallel the burden of disease, researchers have concluded that the variation is “unwarranted.”
How can this be explained? Many observers have been quick to suspect financial conflicts of interest. In traditional reimbursement systems, surgeons get paid for every procedure they do. This gives them an incentive to operate whenever it can be justified. As Boston surgeon Francis Moore observed in 1970, “For the pecuniary-minded physician and surgeon alike, or for psychiatrist or pediatrician, the American population is a happy hunting ground.” Economists have described a related phenomenon, supplier-induced demand. Physicians are more likely to use a particular procedure if that procedure is readily available. Increase the supply of surgeons in an area, and the number of referrals to surgeons increases. Some of the variation also reflects differences in medical judgment and opinion. Differences exist in the medical culture of specific institutions, and possibly between countries. Charismatic and prominent advocates for a particular operation can influence physicians in their community. Peer pressure also plays a role, with physicians conforming their practice to that of those around them.
As more and more evidence of unwarranted variation accumulated, it posed a challenge to the medical profession and its commitment to evidence based medicine. On one hand, evidence of practice variation strengthened the position of therapeutic reformers: evidence of 10-fold variation showed how much need there was for improved rationality and discipline in medical decisions. On the other hand, as the variation persisted decade after decade, it became an affront to the ambitions of evidence-based medicine, a testimony to how far medicine remains from being a fully rational enterprise.
Although the problem has been most carefully studied in the United States, it is of course not limited to the United States. British surgeons perform bypass surgery at a fraction of the rate of their colleagues in Australia (one-fourth) or the United States (one-seventh). Angioplasty varies 7-fold between Germany and Ireland. Whose rate is the right rate? No one knows. The variations have not been shown to correlate either with the burden of CAD or with the survival of patients with CAD. The disparities are even more pronounced when middle- and low-income countries are included. Even though Mexico and Germany have similar rates of heart disease, the angioplasty rate in Germany is 300-times higher than in Mexico. Fewer than 20,000 people received angioplasty in China in 2010, with a population of well over 1 billion. Contrast this with 1 million procedures on 300 million Americans.
It might once have been acceptable to dismiss these disparities as appropriate. CAD, after all, was long seen as a disease of the industrialized west. But health officials have known for the past two decades that this is not the case. In its 1993 World Development Report, the World Bank reported that CAD caused over 10% of all deaths worldwide, making it the single leading cause of death. Since that time CAD has further tightened its hold on low- and middle-income countries. India now has more heart attack deaths than any other country. Investments in cardiac care have not increased to match the rising tide of deaths. Although it may be true that coronary revascularization is over-used in Germany and the United States, most physicians would agree that it is under-used in China, India, and elsewhere. Millions of deaths could be prevented each year if more patients had access to coronary revascularization.
The attention of global health experts remains focused, instead, on prevention. In 1993 the World Bank acknowledged the value of aspirin and blood pressure medicine but held the line at heart surgery and similar treatments with prohibitively low cost-effectiveness. Five years later, in its report Control of Cardiovascular Diseases in Developing Countries, the Institute of Medicine recommended prevention and low-cost medications and advised poor countries to avoid “sophisticated, expensive technologies,” including angiography, PCI, and CABG. When world leaders met in New York City in 2011 for the General Assembly on the Prevention and Control of Non-communicable Diseases, they focused their attention on tobacco, diet and lifestyle: “Prevention must be the cornerstone of the global response to noncommunicable diseases.” There is a clear logic here: resources are limited and prevention will always be more cost effective than treatment.
The prevention consensus, however, was tried and abandoned for the case of HIV. Just ten years ago, international health experts asserted that antiretroviral therapy was too expensive and too difficult to deploy in developing countries. Prevention promised to be a better use of limited health care resources. Treatment activists rejected this logic. First, they argued that the rhetoric of appropriate technology provided an excuse for denying medical care to people in poor countries. Second, they asserted that access to life-saving therapy was a human right: prevention did nothing to help the tens of millions of people already infected with HIV. Third, they demonstrated that the cost-benefit calculus could be transformed by reducing the cost of therapy. Fourth, they highlighted treatment’s many collateral benefits: it can reduce viral load and suppress transmission; it offered hope to demoralized communities; it captured the public imagination and mobilized resources on a scale that could scarcely have been imagined in 2002. While great challenges remain, both the Global Fund to Fight AIDS, Tuberculosis, and Malaria and the President’s Emergency Fund for AIDS Relief have made dramatic progress. Prevention may still be more cost effective, but activists, physicians and policy makers agree that treatment is the right thing to do.
The parallels of AIDS and CAD are not perfect, but they are informative. Prevention, whether through smoking cessation, weight loss or increased physical activity, does offer the best long term solution for global heart disease. But what about the tens of millions of people who already have advanced coronary atherosclerosis? They too would benefit from lifestyle change, but millions of them will still die from heart attacks each year. Amid rising support for global health equity, no one has made a serious call for global access to cardiac surgery or interventional cardiology. There is no Global Fund or President’s Emergency Program for Heart Attacks and Cardiovascular Disease. The magnitude of the need is great: heart disease kills more people each year than AIDS, tuberculosis, and malaria combined. Technology exists that might save the lives of millions. And yet no one has identified access to angioplasty as a human rights crisis.
My point here is not to argue that physicians and funders committed to global health should rush out and build angioplasty suites on every street corner in Mexico City, Mumbai or Shanghai. Instead, I want everyone to think seriously about the moral implications of health care disparities. Decisions about revascularization worldwide do not simply reflect the balance of likely risk and benefit. Instead, they also reflect other values, such as how societies choose to invest scarce health care resources. Even though the United States and other wealthy countries invest heavily in cardiac care, they have not made a commitment to fund such care elsewhere. The discrepancy might mean that we are committed to coronary revascularization but do not think people in resource-poor settings deserve comparable access (or that it is not our responsibility to provide it). Or it might mean that we are committed to providing essential medical care to resource-poor settings, but do not consider angioplasty and bypass surgery to be essential treatments. Neither options holds tremendous appeal. What rate is the right rate for coronary revascularization? Social justice demands that one target be set for all people, not one rate for the wealthy citizens of the United States, and a lower rate for poorer people worldwide. It remains to be seen whether that rate will be high or low.
About the Author:
David S. Jones is A. Bernard Ackerman Professor of the Culture of Medicine at Harvard University. His research interests include health inequalities between populations, particularly the history of explanations that have been given for health inequalities since the seventeenth century, and medical decision making, focusing on the history of cardiac therapeutics, with particular reference to the relationship between changing disease models of coronary artery disease and the various strategies used to treat it. His most recent book is Broken Hearts: The Tangled History of Cardiac Care.