Cancer control—A global perspective
Abstract
Disparities in cancer control exist in low- and middle-income countries (LMICs). Many countries do not have cancer registries to record incidence, mortality and prevalence and are reliant on Globocan estimates of their cancer burden. Poorer cancer control within and between countries occurs in those living remotely from urban centres, those in a low socioeconomic group and some ethnic groups who have lifestyle and belief systems which impact on cancer control. High-income countries generally have population screening programmes for cervix, breast and bowel cancer. However, simpler forms of screening for cancer of the cervix like visual inspection with acetic acid have been shown to be feasible in developing nations. The widespread use of vaccines to prevent cancer has been achieved with the Hepatitis B vaccine but the human papilloma virus vaccine to prevent cancer of the cervix is largely only available in high-income countries. Access to and training of oncological surgeons in LMICs is limited, while 70% of patients in these countries cannot access radiotherapy. The World Health Organization has developed a list of essential medicines although access remains poor in LMICs. The United Nations has set targets for the control of non-communicable diseases to improve global cancer control.
1 Introduction
Disparities in cancer control exist between developing nations and higher income countries and within countries. It was estimated that in 2012, 57% of the incidence and 65% of cancer deaths were in developing countries (Globocan 2012). Many potential factors account for poorer outcomes in developing nations. There may be higher rates of lifestyle risk factors such as smoking tobacco or obesity and a later diagnosis because of poorer awareness of symptoms and late presentation. Access to preventive vaccines or to screening for cancer may be limited. There is also often poor availability of centres with appropriate diagnostic and treatment facilities (Olver, Marine, & Grogan, 2011).
Other barriers to achieving optimal cancer control include living remotely which can apply between countries and within countries at all levels of development. A lower socioeconomic status (SES) of a population correlates with poorer cancer outcomes and people of lower SES are over-represented in more remote areas (The Garvan Research Foundation, 2015). Cancer control outcomes can vary between ethnic groups where the uptake of cancer services will be dependent on their cultural appropriateness (Brach & Fraser, 2002).
In setting policy to improve cancer control, it is essential to have an accurate record of cancer incidence, prevalence and survival and their changes over time. This determines the extent of the cancer problem and where resources are most needed and allows monitoring of the impact of screening programmes and new treatment regimens. Cancer registries are essential and the data collected need to be as complete as possible.
2 Population Cancer Registries
Population cancer registries aim to capture all new cases of cancer in a geographically defined population. Their source of data can be pathology laboratories, hospital cancer registries and death certificates and reporting can be voluntary or mandatory (Leal, Fernandez-Garrote, & Mohar-Betancourt, 2016). The data are used to determine public health priorities and assess the efficacy of public health initiatives such as screening programmes (Bray et al., 2014). Unfortunately, it has been estimated that 85% of the world lack quality cancer registries (Bray et al., 2015).
The World Health Organisation's (WHO) International Agency for Research on Cancer (IARC) has established the Globocan project (Globocan 2012). This aims to provide national incidence, prevalence and mortality data or estimates for major cancers covering 184 countries. It covers periods up to 2012. It provides fact sheet maps and predictions and is a very useful source of comparative data. An analysis of Globocan methods reports that the data are strongest if national data are available (Antoni, Soerjomataram, Møller, Bray, & Ferlay, 2016). However, for some low- and middle-income countries (LMICs), estimates must be made from regional collections within a country or simply estimated from the neighbouring countries. Papua New Guinea, for example, does not have a cancer registry and so the numbers are only estimates.
A further initiative for IARC is to help countries to establish cancer registries by providing suitable software, currently CanReg 5 (Ervick, 2016). There are also guidelines on how to manage a registry, and minimum data sets have been defined along with networks and regional hubs (Bray et al., 2014). Again these are very helpful initiatives, but there are ongoing resource implications and follow-up can be problematic. For example, a review of follow-up of Kaposi's sarcoma in five countries from Sub-Saharan Africa showed that almost half were lost to follow-up by 2 years (Freeman et al., 2016). An analysis of the cost of cancer registration in four LMICs showed that the cost per case ranged between $US 3.70 and $US 15.62. More than half the cost was labour and costs were greater in smaller populations (Tangka et al., 2016).
2.1 The cause of the disparities
Cancer registries help identify disparities which span the spectrum of cancer prevention, early detection and treatment. Poorer outcomes can be associated with living in more remote areas, lower SES, riskier lifestyles and ethnicity.
2.1.1 Remoteness
Many of the studies focusing on the disparities between rural and urban patients come from high-income countries, but provide a guide to the factors which can be applied to LMICs. In Australia, 27% live in regional areas and 2.3% in remote or very remote areas (Australian Bureau of Statistics 2013). From data spanning 2004–2008, cancer deaths were higher in inner regional areas than major cities, outer regional or remote areas (Australian Institute of Health and Welfare, 2014).
Further complicating the survival outcomes is that although Indigenous Australians constitute 3% of the population, they are over-represented in remote areas. Cancer causes one in five deaths in Indigenous people, and although a larger percentage of non-Indigenous Australians (30%) die of cancer than Indigenous Australians (20%), mortality is improving in the non-Indigenous group but not in the Indigenous group (The Garvan Research Foundation, 2015).
People in lower SES groups have poorer cancer outcomes and represent 39% of those living in remote areas, compared to 24% in rural areas and only 17% in major cities. Moreover, those living in rural and remote areas, particularly men, are less likely to have heeded cancer prevention messages and they engage in lifestyle behaviours associated with a greater risk of cancer. Daily tobacco smoking rates were reported to be 22% in outer regional and remote areas of Australia compared to 15% in cities. Consuming harmful levels of alcohol over time occurred in 24% in outer regional and remote areas compared to 19% in cities. Rates of obesity or being overweight were 10% higher in absolute terms in regional and rural areas (70% vs. 60%), whereas sedentary behaviour applied to 60% compared to 54% in major cities (The Garvan Research Foundation, 2015).
Multiple factors interact to impact on cancer control. For example, a population-based cohort study using data from the New South Wales Cancer Registry in Australia showed that there was a higher rate of deaths from cancer and more presentations with later stage disease in Indigenous Australians compared to non-Indigenous people across SES and remote categories (Tervonen et al., 2016).
Furthermore, an impact of living in rural locations can be found in long-term survivors of cancer when health-related quality of life is assessed (Pedro & Schmiege, 2014). Rurality also impacts on social function and financial difficulty.
One solution to delivering healthcare to rural and remote communities harnesses technologies which enhance connectivity. Telemedicine can be used to export multidisciplinary opinions from large urban centres to isolated practitioners or used to deliver personal care. In both cases, this reduces the need for costly and time-consuming travel, by either patients or medical staff (Bradford, Caffery, & Smith, 2016). Web-based support resources and information which can be tailored to patient needs and is easily accessed by rural and remote communities (Fennell et al., 2016).
2.1.2 Ethnicity
Different cultures may have lifestyle and belief systems that impact on cancer control. Non-biological explanations for cancer, such as it being a punishment for past transgressions would render screening and medical treatments less relevant. The reluctance to travel away from communities in rural and remote locations may compromise care (Cunningham, Rumbold, Zhang, & Condon, 2008). Family and community responsibilities in some cultures may also outweigh the individual requirement for treatment at a particular time, which can compromise treatment schedules.
Alternatively, diets such as the Mediterranean diet or vegetarian diets dictated by cultural prohibitions against meat, or diets based on hunting and gathering food from surrounding bushland, may have a favourable impact on cancer risk compared to an energy dense Western diet.
Various ethnic communities may demonstrate different patterns of cancer. In New Zealand, Maoris have a higher incidence of lung and cervical cancer, while Pacific Islanders like the Tongans have more liver cancer, yet lower incidences of prostate cancer, bowel cancer and melanoma than the non-Maori population (Foliaki, Jeffreys, Wright, Blakey, & Pearce, 2004). Some of this can be explained by lifestyle factors such as a high incidence of smoking or obesity. Other factors include a high rate of HBsAg positivity in Maoris (76.6%) and Pacific Islander (80%) of patients with liver cancer. Differences in survival between ethnic groups are also apparent. Continuing the New Zealand example, Maori and Pacific Islander women have been found to have poorer survival from breast cancer. When the characteristics of the cancers were analysed, they were younger at presentation, and lower SES. Their cancers were larger and they had a higher percentage of ductal to lobular cancers than other women. They also had less triple negative breast cancer and more HER2-positive cancers. When adjusting for other known risk factors such as location, grade and nodal status, Maori women still had a higher death rate (Campbell et al., 2015).
As expected, these patterns of cancer change over time as seen in the Hawaiian Islands, where lung cancer is decreasing in Hawaiian men, while breast cancer is on the increase in women (Johnson, Oyama, LeMarchand, & Wilkens, 2004).
There may also be differing genetic factors, which requires culturally sensitive research, and yet these groups are under-represented in research (Dachs et al., 2008). One example of an attempt to remedy this was made by the Morehouse School of Medicine supported by the National Cancer Institute of the United States when in 2010 they started a Cancer Disparities Research Training Program, recruiting from their Master of Public Health and Master of Science in Clinical Research Programs. Those who enrolled were predominantly African Americans who it is hoped will be motivated to focus on research into disparities (Caplan et al., 2016).
2.1.3 Socioeconomic status
The impact of SES on cancer incidence and outcomes has been more difficult to isolate from other factors. Some studies like the analysis from the British Columbia cancer registry from 1981 to 2009 show low SES to be a significant determinant of stage and survival of oral cancer, or in South Australia that SES had an impact on colorectal cancer survival (Auluck et al., 2016; Beckmann et al., 2016).
No impact of SES on survival was found in Northern Portugal when a sensitivity analysis accounted for socioeconomic impact on mortality from other causes (Antunes, Mendoca, Bento, & Rachet, 2016). The methodology of studying the impact of SES can impact on the results, as was found in a study from Australia where when larger sizes of the geographic unit used to demonstrate the impact of SES on survival resulted in underestimation (Stanbury, Baade, Yu, & Yu, 2016).
Another issue is the differing definitions of low SES. Often there is a composite score such as described in the study from British Columbia which variably weighted variables around income, employment status, highest educational level and single parenting (Auluck et al., 2016). In a different example, a Finnish study examining breast cancer mortality simply compared income levels. There was higher mortality in lower income groups who presented with later stage disease and had more comorbidities. This did not explain all of the differences and it was suggested that there may be differences in care (Vehko, Arffman, Manderbacka, Pukkala, & Keskimääki, 2016). Indeed, a New Zealand study found that breast cancer survival was better for patients treated in private compared to those in the public nationally funded health system (Tin et al., 2016). Similarly, in the population data linkage study previously quoted from South Australia, socio-demographic disparities existed for survival from colorectal cancer despite the State having a universal health system (Beckmann et al., 2016).
The complexity of assessing the impact of socioeconomic factors was highlighted in a qualitative study using a Welsh database of low SES individuals who had previously participated in a cancer awareness survey. The study looked at barriers to presenting with cancer symptoms. It found a poor awareness of non-specific symptoms in this group, and often a fatalistic view of cancer which could be reinforced by discussing symptoms with family or friends. Poverty, often meant that essentials like daily food took priority over seeking medical advice (McCutchan, Wood, Smits, Edwards, & Brain, 2016).
3 Access to Cancer Prevention and Screening
Prevention and early detection should be the mainstays of cancer control. Population screening of asymptomatic individuals allows early detection when a cancer is localised and potentially curable, or detection of precancerous lesions which can be prevented from becoming invasive cancer. There is evidence of the efficacy of this approach in screening for cancer of the cervix, breast and bowel but benefits must be weighed against the negative consequences of over-diagnosis and overtreatment.
Organised population screening programmes which are designed to capture everyone in the target group occur in high-income or upper middle-income countries such as Argentina, Australia, Britain, Canada, Finland, Korea and Singapore and Uruguay, while more opportunistic screening occurs in other high-income countries such as France, Germany, Japan and the United States. Within those countries, people from lower SES groups and various ethnic groups have lower rates of participation in screening (Department of Health and Ageing & AIHW, 2008). LMICs (Sankaranarayanan, 2014) have no programmes or sporadic programmes that lack population-wide impact.
Taking the example of cervical cancer screening, developed countries offering regular Papanicolaou testing in Europe, North America and Asia have reduced the incidence and mortality from cancer of the cervix by as much as 80% (International Agency for Research on Cancer, 2005). Cheaper alternative strategies such as VIA (visual inspection with acetic acid) which can be followed up by immediate cryotherapy is being introduced in low- and middle-income countries (Sankaranarayanan, Nessa, Esmy, & Dangou, 2012). Mammographic screening which in high-income countries has been found to be associated with a reduce mortality from breast cancer (International Agency for Research on Cancer, 2002). Most LMICs do not have the infrastructure to provide a mammographic screening programme and so breast self-examination (BSE) or clinical breast examination (CBE) is promoted. However, a randomised trial of BSE showed no impact on reducing mortality and trials of CBE are pending (Thomas et al., 2002). Colorectal screening programmes using faecal immunochemical testing are just ramping up in high-income countries but there are little data on their feasibility in low- and middle-income countries (Khuhaprema et al., 2014).
Oral inspection for leucoplakia or early oral cancer has been shown to reduce mortality in India. People who smoke, drink alcohol or use betel nut are at high risk (Sankaranarayanan et al., 2013).
The greatest advances in prevention have seen the development of the human papilloma virus (HPV) vaccine which can prevent at least 70% of cancer of the cervix, and a vaccination programme against hepatitis B which will reduce the incidence of hepatocellular cancer. In the 8 years after the introduction of the HPV vaccine, 64 countries have it available nationally, 4 countries regionally and 13 overseas territories have developed vaccination programmes. This targets 118 million women but only 1% were from low- or lower middle-income countries and these are the countries where the death rates from cancer of the cervix are highest (Bruni et al., 2016). The GAVI initiative has improved access to the vaccine in eligible low-income countries (Campos, Sharma, Clark, Kim, & Resch, 2016). The hepatitis B vaccination programme has been more effective with 184 member states of the WHO vaccinating up to 82% of their infants (WHO 2016).
4 Access to Treatment
One of the major causes of disparate outcomes in cancer control is variable access to the treatment modalities of surgery, radiotherapy and systemic therapies. Surgery is often the primary treatment for solid tumours and yet LMICs to do no have sufficiently trained surgeons or resources to treat curable cancers like early stage breast, bowel and head and neck cancers (Are, Wong, & Leon, 2016). Unlike in high-income countries, LMIC surgeons do not have access to the imaging facilities for accurate staging, multimodality treatment or specialist surgical techniques such as sentinel node biopsy or endoscopic surgery (Hoekstra et al., 2016). There is organ-based surgery rather than oncologic surgery and variable training (Sullivan et al., 2015). A solution is that surgical societies and individual surgeons can provide training programmes and fellowships for their colleagues in LMICs. Teaching can be enhanced by using eHealth (electronic communicators) and mHealth (using mobile technology) (Hoekstra et al., 2016).
It is estimated that between 50% and 60% of patients will require radiotherapy at some time during their treatment (Atun et al., 2015), but it is estimated by the International Atomic Energy Agency (IAEA) that in developing countries because of a shortage of 5,000 machines, 70% of cancer patients cannot benefit from curative or palliative radiotherapy (International Atomic Energy Agency, 2016). The IAEA are working to introduce radiotherapy to more of these countries. Moreover, it has been shown that investment in radiotherapy would not only save lives but also have positive economic benefits.
A great challenge for LMICs is being able to afford increasingly high-cost cancer drugs. In Southeast Asia for example, in a high-income country, Singapore, 55% of patients had access to a range of drugs and targeted therapies in common usage but only 15% of patients in LMICs had the same access (Lopes, 2011).
The WHO has attempted to construct an essential medicines list as a guide and in 2014 had UICC (Union for International Cancer Control) convened groups of oncologists to update the list. They looked at the potential for long-term remissions, the contribution of the drugs over other modalities and whether the therapies required advanced pathology testing such as genomic analysis. The 2015 WHO list includes 46 medicines for 26 different cancers (Lopes, de Souza, & Barrios, 2013).
There is a range of measures suggested where LMICs may be able to access these medicines for more of their population. The creation of some form of universal healthcare may help, but more taxes in a low-resourced population is a challenge. The use of generics and biosimilars for drugs off patent may eventually increase availability but any compulsory licensing of essential medicines may contravene trade agreements and be a disincentive for innovation by the pharmaceutical industry. Participation in research can improve access to drugs. There may be industry-led access schemes with subsidised drugs or funding schemes where governments partner with private industry or philanthropic organisations to make drugs available (Lopes et al., 2013).
Of course the incentives to make cancer drugs available may depend on whether other priorities such as food security or countering resistant infectious diseases need to take precedence in low-income countries. I have also visited countries, for example Papua New Guinea, where the funding for cancer medicines dictated identifying six or seven “essential” chemotherapy drugs. Global co-operation will be needed to reduce the cancer burden in LMICs.
5 Conclusions
It is encouraging that the worldwide burden from non-communicable diseases such as cancer has been recognised by the United Nations (UN) who adopted resolution 66/2 as a result of a high-level summit in September 2011 and have created targets for member countries to achieve to improving cancer prevention and control (WHO, 2011).
The initial step in most countries will be to establish a cancer registry to define the incidence of cancer and measure the mortality from cancer and to enable the assessment of the burden of the disease and be able to evaluate the success of cancer control strategies. Even in low-income countries that still have problems with infectious diseases or food security, the cancer burden may be shown to significantly reduce productivity as the prevalence of cancer increases and therefore warrants cancer control measures.
Each country must then implement programmes to meet the UN cancer prevention targets in order to improve global cancer control. The risk factor with the most widely accepted framework for its control is reducing tobacco use with the Framework Convention on Tobacco Control of the World Health Organization. Countries can learn strategies for implementation from others who have achieved success. For example, Australia first legislated for plain packaging of tobacco products to eliminate one of the last advertising strategies, successfully defended this policy from tobacco industry legal challenges, and other countries then followed with similar policies.
More problematic is implementing strategies to reduce obesity and alcohol consumption. Cultural practices around food and alcohol consumption can be strongly entrenched in both low- and high-income countries. In high-income countries, the food and alcohol industry and advertising industries can form strong political lobby groups which dissuade governments from acting in the public good, by labelling such action as paternalism which stifles free choice. This defence of legislative inaction should not apply to governments' protecting a nation's children from multi-million dollar advertising campaigns which encourage overconsumption of nutrient poor, energy dense foods and beverages.
The HPV vaccine represents a breakthrough in the prevention of cancer of the cervix, which is only limited by the ability to disseminate the vaccine. Despite subsidies, some low-income countries cannot or are unwilling to allocate the resources to fund the vaccine and lack the infrastructure to administer it. This leads to disturbing disparities. Papua New Guinea, for example is 4 km from the closest island off Australia. Australia has a very low incidence of mortality from cancer of the cervix because of decades of screening for cancer of the cervix, added to which boys and girls are now both vaccinated, and there is a real possibility of eliminating the disease. Yet, it is the greatest cause of cancer deaths in women in PNG where there is neither screening nor HPV vaccination programmes. Solving such disparities requires international collaborative action.
Finally in terms of treatment, we must identify what minimum essential cancer treatments are needed, be it surgery, radiotherapy or systemic treatments. For example, a minimum essential list of drugs to provide a benchmark for developing nations and will ensure that at least curable and highly responsive cancers can be treated. This can be expanded as a country's economy strengthens.
Just as new vaccines and drugs improve the prevention and treatment of cancer, other developing technologies can help overcome disparities to improve cancer control. Telemedicine, for example, can allow the export of medical expert advice to patients in remote communities in a cost-effective manner. Other disparities have solutions well beyond health, for example the more equitable distribution of wealth, or the design of our built environments to reduce population exposure to carcinogens.
Improving cancer control is a multifaceted exercise but the setting of achievable international targets for reducing the burden of non-communicable diseases is a welcome first step.
