AIDS Between Science and Politics

Peter Piot, Laurence Garey

• Publisher: Columbia University Press
• Year: 2015
• ISBN:
  9780231538770
• Topic: Medicine
• Subtopic: Public Health, Administration & Care

Peter Piot, founding executive director of the Joint United Nations Programme on HIV/AIDS (UNAIDS), recounts his experience as a clinician, scientist, and activist fighting the disease from its earliest manifestation to today. The AIDS pandemic was not only catastrophic to the health of millions worldwide but also fractured international relations, global access to new technologies, and public health policies in nations across the globe. As he struggled to get ahead of the disease, Piot found science does little good when it operates independently of politics and economics, and politics is worthless if it rejects scientific evidence and respect for human rights.

Piot describes how the epidemic altered global attitudes toward sexuality, the character of the doctor-patient relationship, the influence of civil society in international relations, and traditional partisan divides. AIDS thrust health into national and international politics where, he argues, it rightly belongs. The global reaction to AIDS over the past decade is the positive result of this partnership, showing what can be achieved when science, politics, and policy converge on the ground. Yet it remains a fragile achievement, and Piot warns against complacency and the consequences of reduced investments. He refuses to accept a world in which high levels of HIV infection are the norm. Instead, he explains how to continue to reduce the incidence of the disease to minute levels through both prevention and treatment, until a vaccine is discovered.

1

A HETEROGENEOUS AND STILL-EVOLVING EPIDEMIC

RAPID GLOBAL SPREAD

AIDS was first described in 1981, which makes it a recent historic phenomenon. Following its discovery in North America and then in Western Europe, around 1985 the epicenter was considered to be in Central Africa where the HIV infection rate was 4 or 5 percent in countries such as Zaire, Rwanda, Burundi, Uganda, and Zambia. Southern Africa, which would later experience the highest HIV infection rates in the world, was hardly affected at the time. Ten years later, in 1995, the virus had spread throughout the world with an important focus in Southeast Asia, mainly connected with commercial sex. At the beginning of the twenty-first century the epidemic spread through the former Soviet republics driven by injecting drug use. Today we are still in an unstable dynamic phase of the epidemic in certain parts of the world, such as in Eastern Europe, while it has peaked in others, such as in many African countries. In Papua New Guinea, for example, an epidemic is developing with a mainly heterosexual mode of transmission. In several sub-Saharan African countries injecting drug users and men who have sex with men are now emerging as being at high risk for HIV infection, and in Asia and Western countries there is continuing high incidence among men who have sex with men.¹

FIGURE 1.1 WORLD MAP OF PREVALENCE OF HIV INFECTION.

DATA FROM UNAIDS (http://www.unaids.org), UNICEF (www.unicef.org), AND THE WORLD BANK (http://www.worldbank.org).

Three elements are important when considering the evolution of an epidemic like AIDS: the number of people living with HIV (prevalence), the incidence of new infections, and the number of deaths it causes. Prevalence means the total number or percentage of people living with HIV (combining old and new infections) at any given time in a population. Incidence refers to the number or rate of new infections over time, most commonly one year. The number of people infected with HIV reflects the sum of new infections over the years, minus the number of deaths. In 2013 the number of people with HIV was estimated at just over thirty-five million.² Overall, AIDS was the fifth leading cause of DALYs (Disability Adjusted Life Years lost) in the world in 2010, and the first cause of DALYs in twenty-one countries.³ Increased access to antiretrovirals since the mid-2000s avoided millions of deaths and thus contributed to maintaining the number of infected people at a high level. This factor thus masked the considerable drop in new infections. New infections in adults and children in 2013 were estimated at 2.1 million, a decline of 38% from 2001. This drop is spectacular, even if the figure remains high. Sub-Saharan Africa saw 1.5 million new infections in 2013, which represents about one million fewer infections than in 2001. In South and Southeast Asia there were approximately two hundred seventy thousand new cases in 2013, a decline of about one third since 2001. Western and Central Europe experienced an increase of HIV incidence, with twenty-nine thousand new infections, but Eastern Europe and Central Asia had the largest relative and absolute increase in HIV incidence in the world. For example, in Russia new infections rose by 50 percent between 2006 and 2011, from forty thousand to sixty thousand. In the United States new infections have remained relatively stable at around fifty thousand per year since the mid-1990s, though with a consistent increase in incidence among gay and bisexual men. The most pronounced decline in new infections since 2001 (52 percent) has occurred in the Caribbean.

More than any other figure, deaths illustrate the drama of AIDS: 1.5 million people died in 2013 in spite of the existence of antiretrovirals. The majority of these deaths occurred in sub-Saharan Africa, where AIDS is the first cause of death in about half the countries. In Europe and the United States twenty-seven thousand people died from AIDS in 2013, most of whom could have been saved if they had had access to early diagnosis and appropriate treatment.

HOW ARE HIV NUMBERS ESTIMATED?

How can we estimate the number of people infected with HIV, since tests are not implemented systematically on all populations? To estimate the extent of a disease, infection and risk are constant challenges in public health. There is a complex worldwide epidemiological survey that probably provides better public health data for AIDS than for most other conditions. Indeed, since the end of the 1980s systematic surveys of the prevalence of HIV antibody in pregnant women presenting at prenatal clinics are held, using a standardized methodology developed by the Joint United Nations Programme on HIV/AIDS (UNAIDS) and the World Health Organization (WHO). They are confidential, anonymous, and repeated every two years in the same clinics, but not with the same patients. The results allow an estimation of the infection rate in the sexually active heterosexual population. The aim is to establish trends (increase, decrease, or stability), rather than an absolute number, to facilitate organization of prevention, and anticipate needs for treatment or support for affected people.

Besides a number of logistical challenges and inadequate resources for epidemiological surveillance, estimating the level and spread of HIV infection in a large population is complicated by at least two factors. The first is the reality that HIV is not evenly distributed in a population, with some individuals and subgroups being at greater risk than others, making nationwide estimates more hazardous, as neither the burden of infection, nor the size of the groups at higher risk, are known in many countries (compounded by the fact that in some countries it is very difficult to work with certain populations as they are illegal, such as homosexual men). Specific surveys are necessary to estimate HIV prevalence in such populations. A second more recent challenge for estimating new infections is that more infected people are surviving due to antiretrovirals. The success of a program for treating AIDS is measured by a fall in mortality and therefore also by an increase in people living with HIV. Conversely, falling numbers of infected people may indicate an increase in deaths, or a long-term decrease in new infections. So, paradoxically, for at least one generation we can only hope for a decrease in new infections but an increased number of people living with HIV.

Beyond the absolute number and distribution of infections in given regions over time, national AIDS programs seek to understand which groups are affected by recent infection, as this is where efforts for HIV prevention should be focused. However, sentinel surveillance of HIV detects people whose infection may date from years earlier, which is not ideal for early prevention of infection, though it remains useful to estimate needs for antiretroviral treatment. The major problem is the absence of a reliable biological test to measure the incidence of recent HIV infections in large-scale surveys. Such a test would be particularly important to estimate whether prevention has an impact or not, or if it is targeting the groups most at risk.

HIV infection rates among young people may be a proxy for incidence in their age group, as infections through sex or injecting drug use are likely to be recent. Nevertheless extrapolating HIV prevalence in young pregnant women to the general population must be adjusted with care as we might overestimate the prevalence among other women of the same age in the general population, as well as among men, since particularly in Africa women are infected at a much younger age than men.

In spite of data from sentinel surveillance and behavioral and serological surveys it remains difficult to estimate national HIV prevalence even in countries with well-functioning systems. When new data and better methods become available estimates are updated, sometimes generating controversy or even conspiracy theories. For example, in 2008 the Centers for Disease Control (CDC) in the United States, which have access to thousands of data banks, had to adjust the HIV prevalence estimates for 2006 and recognize that 1.1 million people were HIV-positive, higher than had previously been estimated.⁴ And in 2007 UNAIDS announced a major reduction in its estimates of the number of people living with HIV in India from five to 2.7 million when new data became available from a much larger number of epidemiological surveillance sites than before.

In spite of data from sentinel surveillance and behavioral and serological surveys it remains difficult to estimate national HIV prevalence even in countries with well-functioning systems. When new data and better methods become available estimates are updated, sometimes generating controversy or even conspiracy theories.

Reliance on complex statistical models is indispensable for obtaining the best estimates and predicting future evolution. UNAIDS and WHO rely heavily on a reference group of independent experts, of whom most are not directly involved in surveys. This group developed the methodology for the estimates and the relevant software. In most countries there are committees of national and international experts who examine local data. National estimates are made about every two years and the data are integrated with a worldwide bank run by UNAIDS and WHO. An alternative methodology, the Global Burden of Disease Study 2010 by the Institute for Health Metrics and Evaluation (Cause of Death Ensemble model), which incorporates mortality for all causes, generated mortality estimates close to those published by UNAIDS.⁵ HIV estimates can be politically highly sensitive, and countries such as Russia, China, India, and South Africa at some point challenged the UNAIDS figures, though without providing alternative evidence.

The future of the HIV epidemic depends on numerous unknowns and no model can fully determine the extent of long-term changes, either in prevention or treatment. Thus, in 1990 the worst case scenario the WHO envisaged for 2000, using the EPI Model computer program, was that fifteen to twenty-five million people would be infected with HIV throughout the world, which unfortunately proved to be millions short.⁶

GENERALIZED AND CONCENTRATED EPIDEMICS

Generalized epidemics are de facto those resulting from heterosexual transmission in the general population, as in sub-Saharan Africa, Haiti, Cambodia, or Papua New Guinea. Most countries experience concentrated epidemics, where HIV is associated with certain high-risk groups. In the Americas, Europe, and Australia these are mainly gay men, but in Eastern Europe around 50 percent are drug users. Paradoxically we sometimes have better figures for sub-Saharan Africa than Europe or other continents because it is easier to measure a phenomenon when it is common than when it is rare or very rare, such as in China.

In more than thirty countries, especially in Africa, random sample surveys of HIV infection in the general population have been performed, often in conjunction with existing demographic and health surveys. This approach should provide more representative estimates, but is onerous, costing seven to fifteen million dollars for a population of fifteen to twenty million, and cannot be repeated often. However, population surveys on HIV prevalence are also not without problems. For example, the rate of nonparticipation by men in such national surveys varies from 5 to 15 percent, making results difficult to interpret, the more so because groups most exposed to HIV, such as truck drivers, migrant works, soldiers, or sex workers, are systematically underrepresented, resulting in an underestimation of infection rates. Thus, in general, national population surveys gave infection rates slightly lower than those in pregnant women.⁷ For example, in Lesotho estimates from pregnant women were 26.5 percent but 24 percent in a national household survey. In South Africa the differences were particularly great (29 versus 17 percent). Which figure is closer to reality? Surveys based on samples from the general population often do not include people who are not regular members of a household, such as those who do not work locally. There is probably a slight underestimate of HIV prevalence obtained in national surveys and an overestimate in those involving pregnant women, except in South Africa.

To estimate the prevalence of HIV in concentrated epidemics in high-risk groups one must undertake surveys not only in antenatal clinics, but especially in the highest risk populations, such as men who have sex with men, injecting drug users, and mobile populations, depending on the specific context and country. When dealing with concentrated epidemics methods for estimation are even more problematical. Indeed, to estimate infection rate in these high-risk groups one needs first to know their number at the national level. Establishing estimates of the number of people engaged in such behavior implies working with the populations in question. Many societies condemn and even punish behavior such as illicit drug use, homosexuality, and prostitution. Because of this, people at the highest risk for HIV may be ignored by official epidemiological surveys. If the evolution of high-risk behavior and the number of infections in such groups is not monitored, the inevitable consequence is that efforts to control an epidemic will be inadequate. Most countries have made considerable progress in HIV surveillance in the last fifteen years but there remain areas where it is difficult to obtain reliable estimates of the extent of the epidemic. Population migration dynamics are important in some parts of the world: sometimes mainly internal, sometimes external, masculine or feminine, they can influence estimates of prevalence. Sometimes problems of epidemiological surveillance are political: the problem may be simply ignored, as was the case in China before 2005. It was difficult to obtain information, especially in certain provinces like Hunan where a major part of the adult population of whole villages was contaminated by blood transfusions. The situation has changed since 2005 thanks to greater openness by the central authorities. Other countries present other problems: for example the Democratic Republic of the Congo is an enormous country of which half has no proper roads and where the health services often cannot even perform HIV antibody tests for diagnostic purposes in sick patients. Furthermore in countries like the Congo, Sudan, Afghanistan, or Somalia political instability, insecurity, and war are major obstacles to surveillance of infectious diseases, including AIDS. In most of these cases a large part of differences in data is not a deliberate intention to hide results but simply that the data are just not available.

THE DIVERSIFICATION OF EPIDEMICS

When considering AIDS, it is more correct to speak of HIV epidemics rather than a single epidemic. HIV has become endemic in most countries and is seen in different forms within a country, a region,...