Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 69
Cancer Control Opportunities in Low- and Middle-Income Countries 3 The Cancer Burden in Low- and Middle-Income Countries and How It Is Measured Each year, 5 million people in low- and middle-income countries (LMCs) die from cancer, about 10 percent of the 50 million deaths in those countries. This proportion, and the total burden of cancer,1 will continue to grow as the tobacco-induced cancer epidemic accelerates, and as the world population ages. Looked at another way today, of the 7 million cancer deaths in the world, 5 million are in LMCs. Despite this fact, cancer is not recognized as a high-priority health problem in most of these countries. Where children are dying from malaria and other infectious diseases and suffering the many consequences of malnutrition, where women die in childbirth, and where young adults are dying of AIDS, people with cancer—many dying slowly in their homes—attract less attention. However, the latter half of the 20th century witnessed major reductions in infant and childhood deaths even in the poorest countries, making cancer and the major noncommunicable diseases more prominent in the burden of disease. They now co-exist with the still heavy burden of common infectious diseases and are destined to continue growing in relative importance. This chapter briefly reviews the major shifts in mortality during the latter half of the 20th century, and then describes what is known of the current cancer burden in LMCs. The final section describes the sources of 1 The “burden of disease” ideally measures the full impact of a disease on a population. It goes beyond cases and deaths to include functional limitations imposed by the disease and the disability associated with those limitations, and non-health wellbeing (e.g., financial impacts). Quantification of risk factors known to be associated with specific diseases is also part of the measurement of burden of disease (Lopez et al., 2006). The phrase is used in this chapter both specifically and generally to describe how cancer affects populations in LMCs.
OCR for page 70
Cancer Control Opportunities in Low- and Middle-Income Countries data on the global cancer burden and discusses priorities for improving the information base. Cancer mortality, incidence, survival, and risk factor surveillance for cancer and other chronic diseases are highlighted (see Box 3-1 for definitions). WORLDWIDE CHANGES IN MORTALITY PATTERNS The decline of childhood mortality in developing countries is one of the most significant public health achievements of the 20th century. In 1950, nearly one-quarter of all children born died before their fifth birthday, most in infancy (Table 3-1). Today, about 8 percent will die before age 5. Much of the improvement is due to vaccinations against childhood infections, antibiotics against a wide range of bacterial infections, oral rehydration therapy for diarrhea, and in some places, generally improved living conditions. The upshot is that many more people are surviving to adulthood and old age—even in developing countries. This means more will eventually die from cancer, cardiovascular disease, chronic respiratory disease, diabetes, or another chronic condition of adulthood. Today’s infants—about 130 million born each year—will experience a much different pattern of deaths than those born in the past century (Table 3-2)—assuming current patterns in risk factors. More than half will live to age 70 and older. Nearly one-third (40 million of each year’s birth cohort), however, will die in middle age, between ages 35 and 69. Most of these deaths will be from chronic, noncommunicable diseases. As many as half of these “premature” deaths could be prevented if patterns in major risk factors were modified, allowing people to live longer and die in old age. Specifically for cancer, the most practicable measures involve reversing the increases in smoking prevalence; preventing liver cancer through infant vaccination against hepatitis B virus (HBV); and preventing cervical cancer through a combination of vaccination against the cancer-causing human papillomaviruses, or HPVs, (as vaccines become available and affordable) and screening for precancerous or early stages. The importance of cancer as a cause of illness and death will continue to grow, even with effective preventive measures. Appropriate cancer management—diagnosis and treatment—can extend the lives of many, particularly if diagnosed early. For those eventually dying from cancer (and other causes of death that involve chronic pain and other symptoms), at whatever age, palliative care with good pain control can vastly improve the quality of life of the patient and his or her family.
OCR for page 71
Cancer Control Opportunities in Low- and Middle-Income Countries BOX 3-1 Basic Cancer Population Statistics Defined Cancer Incidence Cancer incidence is the number of new cases occurring in a population, expressed either as an absolute number or as a rate. The incidence rate per 100,000 people per year approximates the average risk of developing cancer in a given year and is often used to compare rates over time or across populations. Measurement of incidence rates requires the identification of all new cases of cancer in a defined population, usually in a defined geographic region. The most basic cancer incidence reporting includes information about the person (age, sex, ethnicity) and about the cancer (the date of detection, anatomic site, histology, and the most valid diagnostic method used). The stage of disease at diagnosis (i.e., the extent of disease according to standard definitions) is also a valuable piece of information. Cancer Mortality Cancer mortality refers to the number of deaths attributed to cancer in the population, and the cancer mortality rate is the number of deaths per 100,000 people per year. These statistics are usually reported as rates, relating the number of deaths to the underlying population (i.e., the census population). As with incidence, reports can be more and less detailed, but information about the person (age, sex, ethnicity) and the cancer (anatomic site) is very useful. Mortality is the product of the incidence and fatality from cancer. Fatality is the proportion of people with cancer who die in a given time period, usually a year. It conveys the risk that an individual with cancer will die, while the mortality rate describes the average risk of dying from cancer in the population. Mortality rates are often used as proxies for cancer incidence, especially where incidence data are not available. For cancers with a poor prognosis everywhere, mortality may, indeed, mirror incidence, and comparisons made across time and place may be valid. In places where people are unlikely to receive curative treatment, even for cancers otherwise considered “curable,” mortality may also be a surrogate for incidence, but comparing across areas may not be straightforward. A statistic derived from cancer mortality is person-years of life lost (PYLL), which weights deaths at different ages: Death at a young age results in more PYLLs than death in old age. PYLL can be modified further by adding aspects of quality of life, such as a year spent in extreme pain would result in loss at a fraction of a “quality-adjusted life-year” or QALY.
OCR for page 72
Cancer Control Opportunities in Low- and Middle-Income Countries Cancer Survival Cancer survival describes the proportion of individuals with cancer who are still living for defined periods after diagnosis, often aggregated by type of cancer, age group, sex, and place of residence. This statistic is often referred to as the “survival rate,” although it actually describes an individual’s probability of being alive, and not actually a rate. Cancer survival statistics are derived by calculating the proportion of people originally diagnosed with a type of cancer who are still alive at specified points after diagnosis, such as 1-year survival. For many cancers, 5-year survival is synonymous with “cure,” because relatively few of those surviving 5 years go on to die from the cancer (breast cancer is the most important exception). Survival is influenced strongly by the stage of disease at diagnosis and the availability of effective treatment. If no adjustment for stage at diagnosis is made in calculating survival, people diagnosed at earlier stages will appear to have better survival than those diagnosed with later stage disease, regardless of treatment, but this is simply a statistical artifact. But for cancers for which effective treatments exist, early detection and treatment means a real survival advantage. If existing treatments are not very effective (e.g., for pancreatic cancer, lung cancer, and stomach cancer) or if the person does not have access to medical services, the stage of disease makes little difference in survival. Survival is, therefore, a crude measure of the effectiveness and/or availability of cancer treatment. Cancer Prevalence Cancer prevalence indicates the number of people alive with cancer in a population. Unlike incidence and mortality, there is no standard definition of a prevalent cancer case. The most appropriate definition may depend on how the information is going to be used. One approach is to count people with cancer who are in active treatment or follow-up, which has strong economic effects. As a practical matter, this has been interpreted by some as cases within 5 years of diagnosis. However, many cancer survivors live with long-term effects of the disease itself and the treatments, some of which require further management, so an argument could also be made for a more inclusive definition. Cancer prevalence can be calculated from cancer registries with good long-term follow-up, or the more usual, estimated from incidence and survival data.
OCR for page 73
Cancer Control Opportunities in Low- and Middle-Income Countries TABLE 3-1 Worldwide Childhood Mortality, 1950–2000 Year of Birth Percentage Dying Before Age 5 1950–1954 23 1970–1974 14 1990–1994 9 2000–2004 8 (about 10 million of 130 million born each year) TABLE 3-2 Approximate Distribution of Deaths by Age Group of Those Dying in the Early 21st Century and Anticipated for Those Born in the Early 21st Century Age Range Deaths Each Year in Early 21st Century Future Deaths of Those Born in Early 21st Century 0–34 (children and young adults) 20 milliona (33%) 20 million (15%) 35–69 (middle age) 20 million (33%) 40 millionb (30%) 70+ (old age) 20 million (33%) 70 million (54%) TOTAL 60 million 130 million aIn 2001, there were 7 million deaths (out of 56 million), but this number is increasing because of deaths from AIDS. bDeaths at ages 35–69 (in 2035–2069) will be mainly from noncommunicable disease: cardiovascular diseases, cancer, chronic respiratory diseases, etc. SOURCE: Personal Communication, R. Peto, University of Oxford, June 2006. BASIC CANCER STATISTICS In 2002, about 11 million new cases of cancer occurred and about 7 million people died of cancer worldwide. LMCs account for more than 80 percent of the world’s population, 72 percent of the world’s cancer deaths, 78 percent of years of life lost (YLL), and 77 percent of disability-adjusted life-years (DALYs) (Table 3-3). Cancer becomes relatively more important as other causes of premature death decline. This leads to substantial variation in the proportion of deaths attributable to cancer in different parts of the world and at different income levels. In 2002, cancer accounted for 12.5 percent of deaths worldwide, but just over 25 percent of all deaths in the low-mortality countries of Europe. In contrast, among the highest mortality countries of Africa, 3.6 percent of deaths were from cancer, and in the highest mortality countries of Southeast Asia, 7.1 percent (World Health Organization, 2003). Deaths from
OCR for page 74
Cancer Control Opportunities in Low- and Middle-Income Countries TABLE 3-3 Deaths, Years of Life Lost (YLL), and Disability-Adjusted Life Years (DALYs), All Causes and Cancers by Country Income Level, 2002 (all figures in millions) All Causes of Death Cancer Deaths YLLa DALYsb Deaths YLLa DALYsb Population Country Income Level Low 28.5 606.4 877.7 1.8 20.8 21.4 2,560,762 Lower middle 17.2 221.0 402.6 2.7 29.4 30.3 2,214,697 Upper middle 3.4 42.2 90.4 0.6 5.9 6.3 513,406 High 7.9 52.4 118.7 2.1 15.4 17.4 933,917 All LMCs 49.1 869.6 1,370.7 5.1 56.1 58.0 5,288,865 World 57 922.5 1,490.1 7.1 71.6 75.5 6,224,985 LMC share of global burden 86% 94% 92% 72% 78% 77% 85% aThe component of the DALY that measures years of life lost by a population due to premature mortality. b A measure of the gap in healthy years of life lived by a population as compared with a normative standard. SOURCE: Data from World Health Organization (2006). communicable diseases, maternal and perinatal conditions, and nutritional deficiencies are of greater importance as the income level is lower, but cancer still occupies a prominent place in the overall statistics (Table 3-4). Clearly, cancer is not rare anywhere, even where other health problems are more pressing, but significant variations exist (Figure 3-1). For men, cancer incidence is highest in North America, with an age-standardized rate of about 400 per 100,000, or an 18 percent risk of developing cancer by age 65. The risk of dying from cancer is highest for men in Eastern Europe, at about 200 per 100,000, and the cumulative risk of dying before age 65 is about 10 percent. For women, incidence is also highest in North America, at about 300 per 100,000, while cancer mortality is highest in East Africa, at about 120 per 100,000. As with all age-standardized international comparisons, the data are adjusted for differences in population age distribution—which is heavily influenced by birth rates and mortality from other diseases—by applying age-specific rates from the country in question to a “standard” population in order to focus on comparable cancer risks and rates for individuals within a population. Patterns of Cancer in LMCs Knowing approximately how many cancers are occurring in a population, the distribution of types, and who is being affected are essential for
OCR for page 75
Cancer Control Opportunities in Low- and Middle-Income Countries TABLE 3-4 Deaths by Cause, Sex, and Age, 2001 (thousands) Low- and Middle-Income Countries World Males Females M + F M + F Males Females Cause 0–29 30–69 0–29 30–69 Total Total 0–29 30–69 0–29 30–69 All causes 7,975 10,853 7,437 7,312 48,351 56,242 8,117 12,263 7,515 8,088 Communicable, maternal, perinatal, and nutritional conditions 5,797 2,532 5,956 1,829 17,613 18,166 5,822 2,598 5,977 1,862 Noncommunicable diseases 871 6,699 827 4,810 26,023 32,891 915 7,871 859 5,495 Malignant neoplasms 111 1,691 95 1,255 4,955 7,021 121 2,186 101 1,597 Injuries 1,307 1,623 654 673 4,715 5,186 1,379 1,792 678 730 Percentage of deaths from cancer 1.4 15.6 1.3 17.2 10.2 12.5 1.5 17.8 1.3 19.7 SOURCE: Lopez et al. (2006).
OCR for page 76
Cancer Control Opportunities in Low- and Middle-Income Countries FIGURE 3-1 Cancer incidence and mortality by geographic area, 2002. LAC = Latin America and the Caribbean. SOURCE: Reprinted, with permission, from Parkin et al. (2005). Copyright 2005 by Lippincott Williams & Wilkins. understanding the burden that cancer imposes on society. Any attempt to assess needs and priorities in health logically starts with an examination of the extent of the problem, and cancer is no exception. How important is cancer? Who in the population is most affected? Is the burden of disease from cancer increasing or decreasing? How does it compare with other health problems along these dimensions? Unfortunately, health statistics are poor where health problems are most severe.
OCR for page 77
Cancer Control Opportunities in Low- and Middle-Income Countries More so than for other health conditions, however, enormous effort goes into constructing best estimates for cancer incidence and mortality for every country on the globe. The International Agency for Research on Cancer (IARC, an agency of the World Health Organization, or WHO) is the source of the most widely respected global cancer database, GLOBOCAN. WHO compiles vital and health statistics for all countries and all causes of death. The statistics presented in this chapter come from these two sources. They are considered broadly accurate and indicate the magnitude of the cancer burden in LMCs, though they are mainly estimates, made in the absence of directly collected data, particularly in the lowest income countries. The status of direct data collection in LMCs is reviewed later in this chapter. The mix of common cancers varies between high-income and low- and middle-income countries (Figure 3-2 and Tables 3-5 and 3-6), and among LMCs in different parts of the world. The patterns vary by geography and economic status, which correlate roughly with the causes of cancer in the “environment” in its broadest sense. Genetic variation plays a lesser role overall. The majority of cancers in more developed countries are those associated with more affluent lifestyles—cancers of the lung, colon and rectum, breast, and prostate. All except lung cancer have a reasonably good prognosis where comprehensive cancer management is available. In contrast, cancers of the liver, stomach, esophagus, and cervix—all related to infectious agents—are relatively more common in developing countries. Where treatment is largely unavailable, all cancers have a poor prognosis, but in this group, all but cervical cancer have poor outcomes everywhere (Parkin et al., 2005). Differences in rankings between developed and developing countries in both incidence and mortality are broadly explicable by differences in exposures, both to infectious and environmental agents, and the availability of medical care. Stage Distribution of Cancers at Diagnosis Most cancers in LMCs are detected at later stages than in high-income countries. Although this is the common wisdom and too logical to dispute, the actual evidence on which to judge this is sparse. Population-based data are not available, but a number of hospital-based studies have been published that report the cancer stage distribution of patients at those hospitals. Table 3-7 is a compilation of these studies for breast cancer. The percentage of advanced cancers ranges from 30 to 98 percent. CANCER TRENDS AND CURRENT STATUS OF COMMON CANCERS Over time—mostly over fairly long periods of years and decades—certain cancers become more or less common. This is known largely from
OCR for page 78
Cancer Control Opportunities in Low- and Middle-Income Countries
OCR for page 79
Cancer Control Opportunities in Low- and Middle-Income Countries FIGURE 3-2 Estimated global cancer incidence and mortality, 2002. SOURCE: Reprinted, with permission, from Parkin et al. (2005). Copyright 2005 by Lippincott Williams & Wilkins.
OCR for page 95
Cancer Control Opportunities in Low- and Middle-Income Countries TABLE 1 Five-Year Relative Survival (in %) of Ugandan Patients with Cancer, Kampala, Uganda, 1993–1997 Cancer Site 5-Year Relative Survival (%) Number of Patients Contributing to Analysisa Nasopharynx 0.0 50 Esophagus 4.5±2.3 182 Stomach 0.0 91 Colon/rectum 8.3±3.9 104 Liver 3.2±2.2 117 Lung 0.0 50 Breast 45.4±7.1 174 Cervix 18.2±4.8 285 Ovary 16.2±8.1 69 Prostate 46.9±7.7 161 Eye 34.2±9.8 88 Thyroid 13.4±11.4 41 Lymphomas 35.4±5.8 199 Kaposi’s sarcoma, HIV positive 9.1±3.6 188 Kaposi’s sarcoma, HIV negative 65.7±14.2 32 aThis includes patients with complete and incomplete follow-up; overall, 27 percent were lost to follow-up before 5 years. SOURCE: Gondos et al. (2005).Gondos et al. (2005). survived 5 years, compared with only 12 percent of the Ugandans. Large gaps between the two populations also were apparent for nasopharyngeal, colorectal, cervical, ovarian, and prostate cancers. Discussion Cancer patients in Uganda have very poor survival odds—lower than the few other developing countries where survival has been documented (Sankaranarayanan et al., 1998). This is the outcome that must be expected where annual per capita incomes are less than $300 (World Bank, 2003) and health care spending is less than $50 per capita. An estimated 5 percent of the population has access to the meager cancer treatment facilities, all of which are centered in Kampala. Data on cancer stage at presentation are lacking, but the evidence points to the majority being in late stages. Sixty percent of the deaths occurred in the first year after diagnosis, and 80 percent by the end of 2 years. Those presenting at earlier stages would have a better chance of finding life-saving treatment, but with treatment so scarce, earlier diagnosis may make little difference.
OCR for page 96
Cancer Control Opportunities in Low- and Middle-Income Countries only once that area is sufficiently developed will the task of following up on registered cases commence. The 10 registries included in Cancer Survival in Developing Countries are likely to include what are among the best survival experiences in the developing world. These countries were among those developing rapidly, with better cancer services than other countries. The existence of the registries also signals an urban catchment area with above-average cancer services. Findings from Cancer Survival in Developing Countries The 10 registries and the time periods represented in the volume are listed in Table 3-10. To the extent possible, the data were made comparable (Sankaranarayanan et al., 1998). Even among these developing countries, wide variations in survival from some cancers was reported. The analysis also includes comparisons with registries in the United States (white population) and Europe, so differences were also noted between the higher- and lower income areas. Figure 3-4 broadly summarizes the relationships found. These are not surprising and are, in fact, intuitive, but it is useful to see them drawn on the basis of evidence. Three major patterns are apparent: Cancers with poor prognosis: These cancers have the smallest survival differential between low- and high-income countries, and include cancers of the esophagus, liver, lung, and pancreas. They are often detected at advanced stages in both low- and high-income countries, because no TABLE 3-10 Registries Included in Cancer Survival in Developing Countries Registry Area Cancer Registration Period (diagnosis) Closing Date of Follow-Up China Qidong 1982–1991 31 Dec 94 Shanghai 1988–1991 31 Dec 94 Cuba 1988–1989 31 Dec 94 India Bangalore 1982–1989 31 Dec 93 Barshi 1988–1992 31 Dec 95 Bombay 1982–1986 31 Dec 93 Madras 1984–1989 31 Dec 93 Philippines: Rizal 1987 31 Dec 93 Thailand Chiang Mai 1983–1992 30 Jun 94 Khon Kaen 1985–1992 31 Dec 95 SOURCE: Sankaranarayanan et al. (1998).
OCR for page 97
Cancer Control Opportunities in Low- and Middle-Income Countries FIGURE 3-4 Cancer survival differences between developed and developing countries and implications or control measures. SOURCE: Reprinted, with permission, from Sanakaranarayanan et al. (1998). Copyright 2006 by the International Agency for research on Cancer. effective screening or early diagnosis techniques are available. Even with the best available treatment, most people who develop these cancers anywhere in the world do not survive for an extended period of time. Cancers with early detection, diagnosis, and treatment options that are relatively easy to implement: For a second group of cancers, including melanoma and cancers of the head and neck, large bowel, breast, cervix, ovary, urinary bladder, and thyroid, there is greater variation in survival between developing and developed countries, and probably between low-income and middle-income countries, at least in some cases. For these cancers, early detection, diagnosis, and treatment that, in principle, can be delivered through basic health care facilities is effective and improves survival. To the extent they are implemented in low-resource settings, survival will be improved. Cancers with effective diagnostic and treatment interventions that require improved logistics: A third group of cancers, including leukemia, lymphoma, and testicular cancer, are marked by an even greater variability in survival between developing and developed countries. Effective treatments are available for these cancers, but they are multimodal treatments requir-
OCR for page 98
Cancer Control Opportunities in Low- and Middle-Income Countries ing a greater degree of medical resources, a good health care infrastructure, and sophisticated knowledge. The cancers in this group are relatively less common in developing countries. GLOBOCAN: Cancer Incidence, Mortality, and Survival for All Countries Given that neither mortality nor cancer incidence data are recorded directly in most LMCs, how are we able to include numbers representing global cancer mortality and incidence, such as those at the beginning of this chapter? How do we know there were 11 million cases and 7 million deaths from cancer in 2002? As referenced earlier, these are estimates from GLOBOCAN, a database created by IARC, part of WHO, that includes estimates of the incidence and prevalence of, and mortality from, 27 cancers for all countries in the world, most recently for 2002. GLOBOCAN data reflect all reliable information from cancer registries and mortality reporting, and where these sources of information are missing or incomplete, estimates of incidence, mortality, population, and prevalence made following explicit rules (see http://www-dep.iarc.fr/). GLOBOCAN is accessible through the IARC CancerMondial webpage (see http://www-dep.iarc.fr/) and on CD-ROM, and is widely cited in the global cancer literature. Surveillance of Risk Factors In addition to knowing how many cancers are occurring and how many people are dying from cancer, knowing the distribution of risk factors—in particular, those that are modifiable—can be extremely useful. As noted in a WHO report, “the risk factors of today are the diseases of tomorrow” (Bonita et al., 2001). Cancer shares risk factors with other major causes of death from noncommunicable diseases, and it is surveillance of these major, shared risk factors that are the basis of two approaches described in the next sections. The first is WHO’s approach, cross-sectional sample surveys that can be adapted for use in every country, even in low-resource areas. The second is longitudinal studies that involve following large cohorts of people over decades; these studies are generally more appropriate for middle-income countries. WHO “STEPwise” Approach WHO has developed an initial three-step approach to population surveys for surveillance of risk factors in response to a resolution on the prevention and control of noncommunicable diseases, passed by the World Health
OCR for page 99
Cancer Control Opportunities in Low- and Middle-Income Countries Assembly in 2000. The steps provide flexibility in the level of effort that can be made initially, while allowing for expansion when resources permit. The use of standard survey instruments across time and countries allows for more valid comparisons on those dimensions. Countries may also choose to add more detailed questions or tests, depending on their situation. As with all surveys, the sampling frames must be carefully defined and the numbers surveyed sufficient to provide reliable estimates of the actual population rates. (WHO provides step-by-step guidance for all aspects of the survey through a series of documents available at http://www.who.int/chp/steps/en/, and carries out training programs all over the world.) The major noncommunicable diseases are cardiovascular diseases, cancers, diabetes, and chronic respiratory diseases. The most important common risk factors that are amenable to intervention are identified as smoking, unhealthy alcohol consumption, unhealthy diet (specifically, low intake of fruits and vegetables), physical inactivity, overweight and obesity, raised blood pressure, raised blood glucose, and raised blood lipids. Of these, smoking has the greatest impact on cancer, but all except raised blood pressure are considered by WHO to have some relevance to cancer (Bonita et al., 2001). The steps are the following: Questionnaire-based assessment: Reports from respondents on socioeconomic data, data on tobacco and alcohol use, and some measurements of nutrient status and physical inactivity. Step 1 plus physical measurements: Simple physical measurements are added to step 1, including at least blood pressure, height, weight, and waist circumference. Step 2 plus biochemical measurements: Measurements on blood samples, including at least fasting blood sugar and total cholesterol. At each step, a “core” set of data is defined, and an “expanded core” and optional items are suggested. The exact measures should be tailored to specific country needs. Steps 1 and 2 are considered “desirable and appropriate for most countries,” while step 3 is not recommended by WHO in “less well-resourced settings unless low-cost technology is used” (Bonita et al., 2001). The STEPwise approach is meeting with success in terms of training and initial surveys. WHO has provided training to 82 countries, including every country in Africa, and 23 countries have completed initial reports (Personal communication, J. Lippe, WHO, May 2006). It will be at least a few years before sufficient data are built up from different countries to appreciate the full value, but this tool appears to be a good choice for most countries, regardless of other data collection efforts.
OCR for page 100
Cancer Control Opportunities in Low- and Middle-Income Countries Large Prospective Studies of Risk Factors and Deaths Surveying for well-known risk factors for cancer and other noncommunicable diseases is of clear benefit for disease control planning and monitoring. In order to extend our understanding of these risk factors in populations not yet studied, and identify and characterize risk factors not yet well studied, a different approach is needed. Large prospective (longitudinal) studies serve this function, as well as providing surveillance for known risk factors. In each such study, hundreds of thousands of adults are interviewed briefly about major risk factors (e.g., smoking, diet) and have basic physical measurements taken (and blood samples stored, in some cases). Households are revisited periodically to record household members’ vital status, and the participants are resurveyed periodically (e.g., every 3 to 5 years) for changes in risk factors. When cohort members die, the cause of death is ascertained. Such studies, involving more than 2 million people, are under way in a handful of countries. A modest number of additional studies should be started periodically to capture unstudied populations, expand the information base on known risk factors, add measures based on new science, and exploit new technologies (e.g., genetic and information) to gather information on large cohorts economically. This is already the case in the current studies, which have an estimated cost of follow-up of $1 per person per year (Personal communication, R. Peto, University of Oxford, June 2006). A leader in large prospective studies to date has been the Clinical Trial Service Unit and Epidemiological Studies Unit (CTSU) at the University of Oxford, England (Clinical Trial Service Unit, 2006), in collaboration with researchers and governmental and academic organizations in the survey areas. The most mature of the CTSU collaborative studies is a nationally representative study of men only, in China. In 1991, information was recorded on smoking, drinking, weight, height, blood pressure, lung function, medical history, and various social factors in about 225,000 adult men throughout China. Reliable systems of follow-up were put into place, and by 1996 the dates and causes of 12,000 deaths had been recorded. Of these deaths, vascular, neoplastic, respiratory, and all other causes each accounted for about one-quarter. Preliminary analyses were consistent with the conclusion from other studies that, during the 1990s, at least 0.6 million deaths each year in China were directly attributable to smoking (Niu et al., 1998). Long-term follow-up of the initial Chinese study is continuing, with periodic resurveys of all middle-aged adults living in the study areas. In addition, it is proposed that blood samples be collected at the next resurvey so that nested case-control analyses can be conducted subsequently, as is being done already in some other sites, which include women as well as men. This will allow the age-specific relevance of established risk factors and of newer risk factors (e.g., various details of the lipid profile, coagulation factors, antioxidants, micronutrients, antibodies, genetic variants, etc.) for vascular
OCR for page 101
Cancer Control Opportunities in Low- and Middle-Income Countries and other diseases to be studied reliably in “nested” case-control studies. Additional studies are under way in Mexico City (200,000 adults over age 40); Russia (four regions: 100,000 adults); Egypt (150–200,000 adults); Trivandrum, south India (200,000 adults); Bombay, west India (200,000 adults); Madras, east India (300,000 adults); and Cuba (200,000 adults). Many of CTSU’s prospective studies have been paired with large retrospective studies that can provide a reliable snapshot of tobacco-attributable deaths at about the time the prospective studies begin. This was done in southern India, where the smoking habits of 43,000 men who had died of various diseases in the late 1990s were compared with the habits of 35,000 living men (the study was restricted to men because few women smoked). About one-quarter of the smokers studied died at ages 25–69, with those dying at these ages losing, on average, 20 years of life. The expected results for cancer and vascular diseases were confirmed, but an unexpected finding also emerged: about half of the deaths from tuberculosis, which causes more than 10 percent of deaths in this population, were attributable to smoking (Gajalakshmi et al., 2003). Had these people not smoked, they would not have died of tuberculosis. This is not a finding about cancer, but it demonstrates that even in a study of smoking—about which quite a lot is known, mainly from wealthy countries—new and surprising (and potentially life saving) information can come to light. There may be a misperception that little more can be learned from further studies of this kind, particularly for established risk factors (such as smoking and blood lipids). But the effects of such factors can vary enormously from one population to another, and there is still substantial uncertainty as to how important they are in different settings and how their importance is changing with time. CTSU’s studies, for example, have defined the outlines of the future epidemic of deaths due to tobacco in developing countries: If current smoking patterns persist, worldwide deaths from tobacco will increase from about 3 million a year now to about 10 million a year by 2025. SUMMARY AND RECOMMENDATIONS There should be no doubt that cancer imposes a substantial burden on all countries, even though the proportion of mortality (and other burden measures) is less in LMCs than in high-income countries. Infectious diseases are still the biggest killers, particularly in low-income countries and some middle-income countries. What lies behind the cancer burden figures, however, is a very thin veneer of data and a great deal of estimation, mainly of cancer incidence and mortality. A small amount of data on survival also exists. Conceptually, burden would also include disability, loss of productivity, caregiving burden, out-of-pocket expenditures, etc., but such information is largely unavailable for these countries.
OCR for page 102
Cancer Control Opportunities in Low- and Middle-Income Countries Few LMCs have accurate recent data about their cancer burden or major risk factors for cancer (or other chronic diseases). The lack of health status information extends well beyond cancer to the entire range of vital and health statistics. The estimates produced by IARC for each country are useful for setting initial priorities, but cannot be used to track progress— positive or negative—or to define more precisely what priorities should be in the medium to long term. Major improvements in overall vital and health statistics will take time. Over the short term, it is feasible to propose modest improvements in the information base, however. In particular, it is relatively inexpensive to gather information on the major risk factors for cancer and other noncommunicable diseases in periodic cross-sectional surveys. WHO’s STEPwise Approach to Chronic Disease Risk Factor Surveillance is well developed, and training and other assistance is available from WHO. The standardized STEPwise Approach has the advantage of producing comparable information across countries as well as over time. Measuring causes of death in a population is more ambitious, but nevertheless highly worthwhile. In low-income countries in particular, this is difficult because many people die without medical care, or at least without a diagnosis before death. Systems based on “verbal autopsies” (determinations based on interviews of family members, health care workers, and others with information about the circumstances of a person’s death) can be developed in place of medical certification, as demonstrated in India’s “Million Death Study” in a network of sample registration areas that constitute a nationally representative sample of deaths (Jha et al., 2006). Prospective (longitudinal) studies of chronic disease risk factors and causes of death involving several million people have been initiated as collaborations between researchers in LMCs and high-income countries. Results are already available from a few LMCs, such as China, India (including the “Million Death Study”), and Mexico. These studies have documented the unique patterns of diseases and their risk factors, such as the strong link between smoking and tuberculosis deaths in India or smoking and lung cancer and chronic lung disease in China. Cohort studies such as these are complex, requiring extensive planning as well as the sustained commitment of human and financial resources for data collection, processing, and analysis. The investment is a significant one on all counts, but the cost need not be prohibitive. Studies now under way cost on average $1 per person/per year to maintain (Personal communication, R. Peto, University of Oxford, April 2006). Finally, cancer registries that record cancer cases and the outcomes of those cases—at least in specific hospitals, and more usefully, in defined geographic areas—are important for understanding local conditions, at least for those who come to medical attention. Registries require sustained
OCR for page 103
Cancer Control Opportunities in Low- and Middle-Income Countries commitments and trained personnel, which are most feasible in urban areas where diagnosis and treatment are available. International assistance and collaboration should be available for all of these monitoring and surveillance activities, to take advantage of existing knowledge and experience. The recommendations from this chapter focus on better characterizing the cancer burden (along with other diseases) in LMCs to support cancer control planning and monitoring. They are based on the usefulness and feasibility (including cost, although we were not able to make specific cost estimates) following from the discussion in the chapter. RECOMMENDATION 3-1. The following should be considered: Risk factor surveillance for chronic diseases should be initiated in many countries using standardized questionnaires (e.g., STEPS). Collection of cause-specific mortality data should be a long-term goal in every country. Where vital statistics systems are weak or nonexistent, initial data collection may be sentinel sites rather than nationwide. Improved mortality reporting at a level appropriate to the country should be supported as a part of cancer control activities. Longitudinal studies of chronic disease risk factors and mortality should be initiated in a few additional middle-income countries. Cancer registries should be developed in conjunction with cancer control activities, mainly in urban areas where diagnostic and treatment services exist. Where new or existing cancer centers are developed into centers of excellence, registries in the catchment area should be a part of the development. REFERENCES Amir H, Azizi MR, Makwaya CK, Jessani S. 1997. TNM classification and breast cancer in an African population: A descriptive study. Central African Journal of Medicine 43(12):357–359. Anyanwu SN. 2000. Survival following treatment of primary breast cancer in eastern Nigeria. East African Medical Journal 77(10):539–543. Barton M, Frommer M, Shafiq J. 2005. The Role of Radiotherapy in Cancer Control in Low-and Middle-Income Countries. Commissioned by the Institute of Medicine. Typescript. Bonita R, de Courten M, Dwyer T, Jamrozik K, Winkelmann R. 2001. Surveillance of Risk Factors for Noncommunicable Diseases: The WHO STEPwise Approach. Summary. Geneva, Switzerland: World Health Organization.
OCR for page 104
Cancer Control Opportunities in Low- and Middle-Income Countries Bray F, Sankila R, Ferlay J, Parkin DM. 2002. Estimates of cancer incidence and mortality in Europe in 1995. European Journal of Cancer 38(1):99–166. Chopra R. 2001. The Indian scene. Journal of Clinical Oncology 19(18 Suppl):106S–111S. Clinical Trial Service Unit, University of Oxford. 2006. Need for Large-Scale Observational Epidemiology. [Online]. Available: http://www.ctsu.ox.ac.uk/projects/observational.shtml#rp [accessed May 31, 2006]. Gajalakshmi V, Peto R, Kanaka TS, Jha P. 2003. Smoking and mortality from tuberculosis and other diseases in India: Retrospective study of 43,000 adult male deaths and 35,000 controls. Lancet 362(9383):507–515. Gondos A, Brenner H, Wabinga H, Parkin DM. 2005. Cancer survival in Kampala, Uganda. British Journal of Cancer 92(9):1808–1812. Gondos A, Chokunonga E, Brenner H, Parkin DM, Sankila R, Borok MZ, Chirenje ZM, Nyakabau AM, Bassett MT. 2004. Cancer survival in a southern African urban population. International Journal of Cancer 112(5):860–864. Jha P, Gajalakshmi V, Gupta PC, Kumar R, Mony P, Dhingra1 N, Peto R, RGI-CGHR Prospective Study Collaborators. 2006. Prospective study of one million deaths in India: Rationale, design and validation results. PLoS 3(2):e18. Lopez AD, Ahmad O, Guillot M, Ferguson B, Salomon J, Murray CJL, et al. 2002. World Mortality in 2000: Life Tables for 191 Countries. Geneva, Switzerland: World Health Organization. Lopez AD, Mathers CD, Ezzati M, Jamison DT, Murray CJL. 2006. Global Burden of Disease and Risk Factors. New York: Oxford University Press. Malik IA. 2002. Clinico-pathological features of breast cancer in Pakistan. Journal of the Pakistan Medical Association 52(3):100–104. Mathers CD, Ma Fat D, Inoue M, Rao C, Lopez AD. 2005. Counting the dead and what they died from: An assessment of the global status of cause of death data. Bulletin of the World Health Organization 83(3):171–177. Niu SR, Yang GH, Chen ZM, Wang JL, Wang GH, He XZ, Schoepff H, Boreham J, Pan HC, Peto R. 1998. Emerging tobacco hazards in China: Early mortality results from a prospective study. BMJ 317(7170):1423–1424. Omar S, Khaled H, Gaafar R, Zekry AR, Eissa S, el-Khatib O. 2003. Breast cancer in Egypt: A review of disease presentation and detection strategies. Eastern Mediterranean Health Journal 9(3):448–463. Parkin DM. 2006. The evolution of the population-based cancer registry. Nature Reviews Cancer 6:603–612. Parkin DM, Bray F, Ferlay J, Pisani P. 2005. Global cancer statistics, 2002. Ca: A Cancer Journal for Clinicians 55(2):74–108. Parkin DM, Wabinga H, Nambooze S. 2001. Completeness in an African cancer registry. Cancer Causes & Control 12(2):147–152. Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas DB. 2002. Cancer Incidence in Five Continents, Volume VIII. Lyon, France: International Agency for Research on Cancer. Sankaranarayanan R, Black RJ, Parkin DM. 1998. Cancer Survival in Developing Countries. Lyon, France: International Agency for Research on Cancer. IARC Scientific Publications No. 145. Schwartsmann G. 2001. Breast cancer in South America: Challenges to improve early detection and medical management of a public health problem. Journal of Clinical Oncology 19(18 Suppl):118S–124S. Vorobiof DA, Sitas F, Vorobiof G. 2001. Breast cancer incidence in South Africa. Journal of Clinical Oncology 19(18 Suppl):125S–127S. World Bank. 2003 World Development Indicators. Washington, DC: World Bank.
OCR for page 105
Cancer Control Opportunities in Low- and Middle-Income Countries World Health Organization. 2003. The World Health Report 2003: Shaping the Future. Geneva: World Health Organization. World Health Organization. 2006. Revised Global Burden of Disease (GBD) 2002 Estimates. [Online]. Available: http://www.who.int/healthinfo/bodgbd2002revised/en/index.html [accessed January 4, 2006]. World Health Organization. 2006. Core Health Indicators. [Online] Available: http://www3.who.int/whosis/core/core_select.cfm?strISO3_select=MYS&strIndicator_select=MortInfantBoth&intYear_select=latest&language=english [accessed 7/10/06].
Representative terms from entire chapter: