The Politics of Priority Setting for Reproductive Health: Breast and Cervical Cancer in Ghana

Abstract

Priority setting for reproductive health is affected by health sector reform policies, the often politically charged nature of issues such as abortion, decreasing funding for reproductive health activities and the broad agenda of the ICPD Programme of Action. This paper examines the influence of political and organizational factors on national priority setting for reproductive health and argues that existing priority setting tools such as disability-adjusted life years and cost-effectiveness analysis do not consider the influence of politics on the priority setting process or account for the interpretation of evidence in priority setting. It suggests that priority setting tools can be strengthened by incorporating empirical measures of political and organizational attention to an issue, and through a new measure – policy priority. The paper applies this new measure to a case study of attention to breast and cervical cancer in Ghana from 1990–97, illustrating how traditional priority setting methods cannot explain the priority given to breast cancer in Ghana. It demonstrates how local politics can trump scientific and economic evidence and suggests that the priority setting process can have unforeseen equity and social implications. It concludes by arguing that the policy priority measure provides a more complete picture of reproductive health priorities and is useful for better understanding the implications of the priority setting process for reproductive health.

Résumé

Les réformes du secteur sanitaire, la nature souvent politique de questions comme l’avortement, la baisse du financement des activités de santé génésique et la vaste portée du Programme d’action de la CIPD agissent sur les priorités en santé génésique. Cet article examine l’effet des facteurs politiques et organisationnels sur les priorités nationales et avance que les outils actuels comme l’année de vie ajustée sur l’incapacité et les analyses de coût-efficacité ne tiennent pas compte de l’influence des politiques sur la formulation des priorités et qu’ils n’expliquent pas l’interprétation des faits dans ce processus. On peut renforcer ces outils avec des mesures empiriques de l’attention politique et organisationnelle à une question, et avec une nouvelle mesure – la priorité politique. L’article applique cette mesure à une étude sur l’attention portée au cancer du sein et de l’utérus au Ghana pendant les années 90, exposant que les méthodes traditionnelles de définition des priorités n’expliquent pas la priorité accordée au cancer du sein au Ghana. Les politiques locales peuvent utiliser les données scientifiques et économiques, et la définition des priorités a parfois des conséquences sociales et sur l’équité imprévues. Selon l’auteur, la mesure de la priorité politique donne une image plus compléte des priorités et permet de mieux comprendre les conséquences de la définition des priorités en santé génésique.

Resumen

Múltiples factores afectan el establecimiento de prioridades de la salud reproductiva, tales como las polı́ticas de reforma del sector salud, las controversias que a menudo envuelven temas tales como el aborto, recortes en el financiamiento de las actividades de la salud reproductiva, y la agenda amplia del Programa de Acción de la CIPD. Este artı́culo examina la incidencia de los factores polı́ticos y organizativos en el establecimiento de prioridades de la salud reproductiva a nivel nacional, y plantea que las herramientas disponibles para establecer prioridades, tales como años de vida saludable perdidos, y el análisis de costos y beneficios, no toman en cuenta el peso de la polı́tica en el proceso de establecer prioridades ni explican la interpretación de datos y hechos en dicho proceso. Se sugiere que se pueden fortalecer dichas herramientas al incorporar medidas empı́ricas del interés polı́tico y organizativo por un tema determinado, además de una medida nueva – la prioridad polı́tica. Aplica esta medida nueva a un estudio de caso de atención a los cánceres de mama y cérvico-uterino en Ghana. Demuestra cómo la polı́tica local puede sobreponerse a los datos cientı́ficos y económicos, y sugiere que el proceso de establecer prioridades puede tener consecuencias sociales y equitativas no previstas. Concluye postulando que la medida de la prioridad polı́tica ofrece una visión más completa de las prioridades de salud reproductiva y que es más útil para comprender las consecuencias del proceso de establecer prioridades para la salud reproductiva.

Key Words:

• priority setting measures
• women’s health policy and programmes
• advocacy and political process
• global burden of disease
• cost-effectiveness analysis
• breast cancer
• cervical cancer
• Ghana

Priority setting has to do with decisions about how to allocate limited resources and is a fundamental part of any health policy development process. Within the health sector, reproductive health presents particular challenges in terms of priority setting. Reproductive health is affected by health sector reforms such as decentralization and sector-wide approaches, the often politically charged nature of issues such as abortion, decreased funding, and the broad agenda outlined in the ICPD Programme of Action. Better understanding of how reproductive health priorities are set is therefore an important area of analysis, particularly as the outcomes have public health, social and equity implications at community and individual levels.

While the need for priority setting in the context of limited resources is not questioned, there are both theoretical and practical debates as to the most appropriate way to determine priorities. Quantitative methods such as cost-effectiveness, cost-benefit and burden of disease analyses differ in their methodologies, but each uses what is considered relevant data, e.g. epidemiological or economic evidence, to determine priority. Cost-effectiveness analysis and burden of disease analysis using disability-adjusted life years (DALYs) are reviewed here because they have been more frequently used by international institutions in recent years and are recommended to national governments for determining health policy. It can be argued that these are normative approaches because they refer to prescribed standards or expectations for what should be a priority, based on a summary measure. However, although they appear to be objective, there are two problems with these approaches. First, they do not account for the interpretation of evidence in the priority-setting process. Despite the current push for “evidence-based” priority setting, evidence must still be interpreted. While the data that provide evidence may be objective, interpretation of the data is not [1][2]. Second, they do not account for the political setting in which priorities are determined. Yet it is possible that national or local politics may trump scientific or economic evidence, and hence cannot be ignored.

This paper argues that normative measures for priority setting, like DALYs and cost-effectiveness analysis, are incomplete, but not that they should be abandoned. It suggests that they can be strengthened by incorporating empirical measures of the extent of political and organizational attention to a health issue. It applies a new form of measurement – policy priority – in a case study of policy attention to breast and cervical cancer in Ghana from 1990–97, to show that a better understanding of the priority setting process can be achieved.

Cost-effectiveness analysis and DALYs

Cost-effectiveness analysis uses ratios used to rank priorities based on “differences in outcomes divided by the difference in costs observed among two alternative programs” [3]. League tables, a rank-ordered listing of these comparisons, are often used by policymakers to allocate limited resources between various health problems. The literature describes several weaknesses of cost-effectiveness analysis. One is that it is not adequately standardized and leads to inappropriate comparisons. Another is that it does not improve the efficiency of resource allocation decisions [4]. Despite these criticisms, cost-effectiveness analysis has been and continues to be widely used to set health priorities. Its relatively standard methodology makes it accessible to a wide audience of stakeholders and its use of a common metric upon which to compare health problems and issues is appealing to policymakers facing difficult priority setting decisions.

A powerfully promoted method for setting international health priorities is burden of disease analysis, developed by researchers at the Harvard School of Public Health and the World Health Organization in 1991. It was the basis for the World Bank’s 1993 World Development Report: Investing in Health and is being recommended for use at the national level to carry out national burden of disease analyses. Burden of disease analysis and DALYs have generated various critiques, ranging from methodological to philosophical. Methodologically, it has been criticized for basing estimates on data that do not include socio-economic or environmental factors or measure unmet health needs [5]. The process for determining preference weights has been questioned for being unrepresentative and not transparent [6]. Others have argued that the authors have not provided philosophical justification for the methodology [7] and that the measure “obscures too much” of the process [8]. Although the creators of the burden of disease method mention the inclusion of social values, they have been criticized for not fully accounting for them. Sayers and Fliedner have criticized DALYs because they do not account for “psychological and cultural perceptions of health and disease” [8]. Sadana [9] argues for the inclusion of community values in the allocation of resources. I would extend the criticisms of DALYs to argue that political values have not been adequately considered.

In spite of these critiques, DALYs and cost-effectiveness analysis are generally viewed as rational approaches to setting national health priorities. However, while it can be argued that they are ‘rational’ because they employ a methodological approach based on analysis of epidemiological and economic evidence, they do not account for the influence of advocacy and political process in priority setting.

Advocacy and political process: the missing perspective

While the importance of the socio-political environment on priority setting in public health has been examined [10], there are few analyses of how micro-level politics influences health policymaking in developing countries [11], e.g. in how a policy problem is defined, the policy formulation process and the implementation of policy. Although the importance of politics in priority-setting decisions and health planning has been described [12][13], it has not been operationalized within the context of priority setting, including for reproductive health.

There is a theoretical need to consider politics and its influence on the priority setting process in addition to the epidemiological and economic evidence considered by cost-effectiveness analysis and DALYs. Two theories from political science, social construction and agenda setting, inform this analysis and provide useful frameworks to better explain which issues end up becoming priorities. “Social construction” refers to the idea that broad societal concerns influence how particular issues or problems are portrayed and, as a result, how those issues are perceived [14][15]. “Agenda setting” refers to the process by which an issue is placed on a policy or political agenda. Agenda setting is an important part of the policy process, because it influences the type of politics that affect an issue, the chances that an issue will become part of a particular policy agenda, and the ability of advocates to influence the development of a desired policy [16]. Although agenda-setting theory has not previously been applied to reproductive health policy, it has been used to explain the policy process for other international health policies such as child health, polio, tuberculosis and malaria [17][18]. Both social construction and agenda-setting theories suggest that the way in which a health problem is defined can affect how it is perceived by decision-makers and influence the level of priority it is accorded.

By incorporating measures of the social and political construction of health issues, one can better distinguish between the normative aspects of the priority-setting process (determining what should be a priority using quantitative measures) and the empirical aspects (what actually happens). Policy priorities are not static but can change over time due to influences such as mortality data or media attention. Examining the empirical aspects of policy priority enables the tracking of the priority-setting process as it occurs. International donors and agencies and national groups such as medical associations or women’s organizations, can also influence policy priority. Empirical measures of the attention and influence of these groups is therefore important.

Methodological issues: defining and measuring policy priority

In this paper, I define a “policy priority” as a specific health issue that is receiving attention or consideration on the policy agenda [19]. An issue is considered to have obtained policy attention if policymakers (or those closely associated with them) explicitly decide to address an issue or not. Policy attention to a health issue can be measured in three ways:

•	direct attention – commonly used systematic measures such as incidence data, mortality and morbidity data, DALYs, results of cost-effectiveness analysis and actual costs;
•	process attention – direct and indirect measures of social and organizational capacity to address a particular health issue, which vary in scope and include not only physical resources such as drugs, equipment and other commodities and supplies, but also technical guidelines and recommendations, treatment protocols, and the number of training courses and workshops organized for clinicians and other service providers to develop capacity to address a health issue;
•	political attention – indicators that groups or individuals in a position of influence, including politicians, civil servants and Ministers, non-governmental organizations such as academic research institutions, women’s organizations, women’s health advocates, health and medical professionals and the media are engaged in advocacy and policymaking, raising the issue publicly and publishing information.

Direct attention reflects a normative framework for priority setting according to traditional priority setting methods, while process attention and political attention describe empirical frameworks because they attempt to measure the priority actually accorded an issue.

Each of these three measures of policy attention includes indicators and draws on data regarding resources and services for a particular health issue, as well as information on local policy and political contexts. The indicators used are based on several types of observations and analysis, including resource flows, content analysis of policy and programme documents and public statements, analysis of media attention and interviews with policymakers. This approach allows the combination of qualitative and quantitative data, providing a more complete analysis than a single or summary measure. The policy priority measure is widely applicable and can be used to view an issue over a period of time or at a single point in time. It is also presented in such a way that two or more health issues or problems can easily be compared. Furthermore, it is transparent and clearly presents several types of measures for the same health issue.

During the course of an institutional affiliation with the Ghana School of Public Health in Legon, I became interested in the amount of policy attention being paid to breast versus cervical cancer in Ghana. In 1996 and 1997, I collected data based on key informant interviews, analysis of international and national reproductive health policy documents; and secondary epidemiological and economic data related to breast and cervical cancer for the years 1990 to 1997. I carried out 115 key informant interviews with senior policymakers, programme managers and field personnel associated with the academic, scientific and NGO communities working on reproductive health in Ghana, and senior policymakers and programme managers of international technical agencies and the international women’s health policy community working on reproductive health. Further, I reviewed Parliamentary proceedings, policy and programme documents of international technical agencies and NGOs working on reproductive health and government agencies and NGOs working on reproductive health in Ghana, international and local medical and scientific literature, and studied media attention to breast and cervical cancer. Analysis involved content analysis of interviews and documents, media analysis, and an in-depth comparative case study of breast and cervical cancer.

Cervical and breast cancer in developing countries

Epidemiological data on cervical cancer show it to be a significant public health problem for women living in developing countries [20], where 80% of the estimated 466,000 cases in the world occur annually, and where it is the leading cause of cancer deaths in women [21]. One reason for the higher incidence of cervical cancer in developing countries may be that the risk factors are prevalent. These include early age at intercourse, more than one sexual partner and low socio-economic status [22]. Importantly, cervical cancer is associated with infection with human papillomavirus (HPV), a sexually transmitted virus [23], which is often transmitted at the same time as other sexually transmitted diseases (STDs). In many parts of the developing world, especially sub-Saharan Africa, the level of STDs has reached epidemic proportions. The main reason for the high incidence in developing countries, however, is the lack of screening and treatment for pre-invasive cervical abnormalities, which can develop into invasive cancer if left untreated [24].

Breast cancer is more prevalent in developed countries, where it is the most common cause of cancer deaths among women and where over half of the 1,000,000 new cases in the world occur each year [25]. Although significant information on breast cancer in the developed world exists, there are fewer data on breast cancer incidence and mortality in developing countries. Based on risk factors of early menarche, low parity, late age at first pregnancy and short duration of breastfeeding, breast cancer is considered to have a lower incidence in developing as compared to developed countries [26]. However, this does not mean cases are negligible, and as women in developing countries begin to exhibit these risk factors more often, breast cancer incidence is likely to increase. Some of this information was not readily available, especially in developing countries, in the early 1990s.

Activities relating to cervical cancer in Ghana

Attention to cervical cancer in Ghana during the period 1970 to 1997 was episodic and primarily came from the medical community, especially obstetricians and gynaecologists (ob-gyns). Very little screening for cervical cancer is done in Ghana and there are no screening guidelines in place. As a result, only doctors with a particular interest in cervical cancer carry out screening or women who specifically request it are screened. The lack of screening is due in part to limited equipment and resources. Specula, which are required to perform a pelvic examination, and other basic equipment required to perform Pap smears, are in short supply. Only those ob-gyns educated overseas were trained in the treatment of cervical cancer.

Besides individual clinicians, there were no identifiable groups advocating for cervical cancer screening in any organized way in Ghana during the case study period of 1990–1997. There have been television and radio programmes to educate the public about cervical cancer but these have been intermittent and the result of the efforts of individual physicians. However, it was portrayed in the media as being associated with early and illicit sexual activity and poor genital hygiene.

There is also evidence of episodic Ministry of Health (MOH) attention to cervical cancer. A former head of the Maternal and Child Health (MCH) Unit proposed a study to determine the prevalence of cervical cancer for the development of a national screening programme. However, this study did not take place because of lack of funding. In 1993, the Johns Hopkins Program for International Education in Reproductive Health (JHPIEGO) sent literature to the MCH unit proposing to start a cervical cancer screening study. The proposed study “died a natural death”, according to the MCH Unit, again because of lack of funding. The issue of cervical cancer was revived again when the MOH drafted the Reproductive Health Strategy and Protocol for Ghana in 1995. This policy document was intended specifically to outline strategies and activities relating to reproductive health in Ghana. The National and Reproductive Health Service Policy and Standards, published by the MOH in April 1996, does discuss the prevention and management of cervical cancer at the sub-district and higher levels. It does not provide screening or treatment guidelines, but mentions prevention of cervical cancer as one of the objectives of the policy. It also discusses breast cancer, under the prevention and management of cancers of the reproductive tract, but does not provide screening or treatment protocols for breast cancer either. In the fall of that same year, two medical schools developed a proposal to screen for cervical cancer in Ghana. As of 1998, however, were they were still waiting for MOH funding.

In 1995, the head of the MCH unit created a Cervical Cancer Working Group, which included clinicians and Ministry of Health staff, to organize a pilot study of cervical cancer in Ghana. However, the working group disagreed about the type of screening method to use. Some members advocated visual inspection, a method that was and is still the subject of clinical trials being conducted in developing countries under the auspices of the Alliance for Cervical Cancer Prevention, a consortium of US NGOs (including JHPIEGO), which is less expensive, less reliable and requires less training than properly done Pap smears. Others felt Pap smears should be used. As of July 1998, the Cervical Cancer Working Group in Ghana had held two meetings. According to the head of the MCH Unit, however, there had been little actual organized activity related to cervical cancer screening up to that time.

Activities relating to breast cancer in Ghana

During the 1970s and 80s, breast cancer in Ghana, much like cervical cancer, was addressed as a clinical issue by the medical community. This is evidenced by the fact that attention to breast cancer primarily came in the form of episodic clinical studies. Data from one study in 1972–77 showed that breast cancer accounted for 7.45% of all cancers treated at Korle Bu Teaching Hospital. It ranked fourth after liver cancer, cancer of the cervix and Burkitt’s lymphoma [27]. General surgeons are responsible for the diagnosis and treatment of breast disease, and clinical breast examination has been a part of the medical school curriculum since the 1970s. During the 1990s, however, awareness about breast cancer in Ghana increased dramatically due to the fact that an organized public awareness campaign about breast cancer took place, during the same years as women’s health advocates in the developed world, particularly the United States and Canada, were conducting similar national-level campaigns. As was happening elsewhere, calls were made by women’s organizations in Ghana for breast cancer screening using breast self-examination and mammography to be available.

Political attention to breast cancer as a public health issue was spearheaded by two powerful women’s groups in the country – the National Council on Women in Development (NCWD) and the 31st December Women’s Movement (31DWM). The NCWD, established in 1975 in response to the United Nations International Women’s Year, is the official national body in Ghana to work with national and international organizations on issues that affect the status of women. The 31DWM organizes for social, political, and economic justice for urban and rural women in Ghana. Their charter includes “mobilizing ‘ordinary’ women throughout the country” and any Ghanaian woman over the age of 18 is eligible to join. In 1996, 31DWM had an estimated membership of 1.5 million women. It was headed at that time by the then First Lady of Ghana, Nana Konadu Rawlings, which provided the 31DWM close and powerful political connections.

The NCWD holds member meetings on the first Thursday of each month. It was in March 1991, at one of its regional meetings in Accra, that the issue of breast cancer was raised as a topic for NCWD attention. The selection of breast cancer that year also coincided with the release of a medical professor’s book, Breast Cancer in Ghana, which was launched by a female justice and received front-page news coverage in May of 1991. NCWD invited a general surgeon from Korle Bu Teaching Hospital to speak about breast cancer at this particular meeting. He performed clinical breast examinations on volunteers from among the more than 60 attendees. 20 women volunteered and received a clinical breast examination in front of the other women attending the meeting. He palpated breast lumps in four of the 20 volunteers, which created understandable alarm among the women attending the meeting.

Following this meeting, NCWD decided to address breast cancer as an important health issue for women in Ghana. They set two main goals: to create awareness programmes and to raise funds to obtain mammography equipment for breast cancer screening for the entire country. As part of their awareness campaign, they invited survivors of breast cancer – including the US ambassador’s wife at the time – to their meetings to talk about their experiences of breast cancer. Together with other women’s groups, NCWD began a fundraising campaign for mammography equipment and sent fundraising letters to many countries through Ghanaian embassies. The executive secretary of NCWD at the time these activities started was a Foreign Ministry diplomat and she used her connections to contact Ghanaians in other countries to raise money. Formal dinner dances that included speeches on breast cancer were another fundraising technique used. It was the invitation and programme to the first of these dances that the NCWD slogan, “Save a Breast, Save a Loved One” was introduced. The campaign was later referred to as the “Save a Breast” campaign. One NGO activist convinced women’s groups in Ghana to recognize October as “Breast Cancer Awareness Month” as it is also in North America.

Media coverage of breast cancer and mammography also helped to generate and sustain public attention. One nationally televised programme in particular got women’s attention and raised public awareness. It was a programme made in the US featuring the actress Felicia Rashad from the “Cosby Show”, extremely popular in Ghana at the time, describing her experience with breast cancer. NCWD and 31DWM arranged to have copies of this video shown to women at beauty parlours and hair salons.

The First Lady and 31DWM became actively involved in breast cancer awareness in 1994. 31DWM participated in fundraising efforts and worked with the MOH at the district and community levels. Their members held regional rallies to raise awareness about breast cancer and to instruct village women to perform breast self-examination. This was their comparative advantage as they were already very active with women at the grassroots level throughout the country. 31DWM was also keen to acquire mammography equipment and set a goal of putting mammography machines in every region. Both NCWD and 31DWM were well placed to acquire these expensive machines, and mammography equipment was both donated and purchased. For example, the charity group Soroptomist International in Denmark donated one machine, and the MOH purchased another from Japan, that began to be used at Korle Bu Teaching Hospital in 1994. By 1996, four mammography units had been acquired and installed in three regions of the country (two in the capital Accra, one in Kumasi and one in Tamale, a city in the north). In addition to purchasing the equipment, the MOH employed the technicians required to maintain the units, purchased the film and paid the associated recurrent costs.

The National and Reproductive Health Service Policy and Standards mentioned above also discussed breast cancer, under the prevention and management of cancers of the reproductive tract, though it did not provide screening or treatment protocols. Even without these, however, breast cancer became an important policy priority in Ghana in the 1990s while cervical cancer did not, in spite of the epidemiological and economic evidence that suggests cervical cancer is a greater public health problem for women in Ghana and should have been given greater policy priority. This case study therefore illustrates the importance of including policy priority measures to accepted normative measures in order to understand why a particular health problem may get priority.

Policy priority measure applied to the Ghana case study

In this section, I apply both normative measures and empirical measures, which include the three types of policy priority measures outlined earlier, to cervical and breast cancer in Ghana. The data for the 1990s are presented in three tables, with data for breast cancer and cervical cancer in separate columns for purposes of comparison. For each variable, a cell is shaded to suggest which issue is likely to be a greater priority. For some of the variables this requires a value judgement based on the interpretation of the data. In the case where there is no difference between the two types of cancer or a determination cannot be made, both cells remain unshaded.

The variables included in Table 1

Table 1 Direct attention measures: breast and cervical cancer, Ghana, 1990–1997

Variables	Breast cancer		Cervical cancer
Regional incidence rates 1990 [28]
Gambia	2.9		8.2
Ghana	19.0		26.0
Mali	9.8		24.4
Nigeria	19.2		25.3
Regional mortality rates 1990 [28]
Gambia	1.3		4.1
Ghana	8.6		14.6
Mali	4.5		12.9
Nigeria	8.7		14.4
DALYs [29]
(all ages, sub-Saharan Africa)	238		434
Cost effectiveness analysis results –	Physical exam	Physical exam+mammography screening	Pap Smear
% GNP per capital per year life gained [30]	12	97–120^a	26^b
		135^c	26
			26
Cost data^d	Screening	Treatment	Screening	Treatment
Private sector	US$ 50	n.a.	US$ 8	n.a.
Public sector	US$ 29	US$ 1522–1539	US$ 4.30	US$ 416–426

a Data based on physical exam and mammography at one-year intervals for women aged 50+. Discount rate .03.

b Based on screening every three years of women aged 40–70. Discount rate .03.

c Data based on five-year screening intervals of women aged 35–59 in China. Discount rate .03.

d Collected by the author, 1997. Public sector data from Korle Bu Teaching Hospital in Accra. Private sector data from SSNIT Hospital, Accra, in 1997 figures.

are frequently used to determine health priorities and to provide evidence of a health problem. These are usually based on economic and epidemiological evidence and include mortality and incidence rates, DALYs, cost-effectiveness or cost data. All but one of the direct attention measures included in Table 1 suggest that cervical cancer should be a greater priority than breast cancer in Ghana. Incidence and mortality rates of cervical cancer in Ghana and neighbouring countries are consistently higher than for breast cancer during the period studied. The DALY results for sub-Saharan Africa also suggest cervical cancer constitutes a greater burden of disease than breast cancer. The cost-effectiveness analysis suggests that screening women aged 35–59 at five-yearly intervals using Pap smears is more cost-effective than mammography screening with physical examination for breast cancer. Finally, the actual costs of screening and treating each form of cancer in Ghana in 1997 suggest that it is cheaper to screen for cervical cancer than for breast cancer in both the public and private sector, and cheaper to treat cervical cancer than breast cancer in the public sector.

The process attention measures shown in Table 2

Table 2 Process attention measures: breast and cervical cancer, Ghana, 1990–1997

Variables	Breast cancer	Cervical cancer
Budgetary measure
Allocated	n.a.^a	n.a.
Expended	n.a.	n.a.
Resource measures	Yes	Yes
Facilities	1 Radiotherapy Unit - Korle Bu	1 Radiotherapy Unit - Kumasi
	1 Radiotherapy Unit - Korle Bu	1 Radiotherapy Unit - Kumasi
Resource measure	Yes	Yes
Equipment	Mammography units - 2 Accra, 1 Tamale, 1 Kumasi Mammography film and maintenance (MOH responsible)	Pap smear equipment (Papkits) Speculum – 3 Colposcopes – limited 1 intra-cavity ceasium unit but it is not functional
Formal programme goals	No	No
National coordinating body	No	Yes (at MOH level - MCH/FP Unit)
Technical guidelines	Yes	Yes
If yes, have they been adapted to local context?	National Reproductive Health Guidelines and Standards	National Reproductive Health Guidelines and Standards
Screening/treatment protocols and standards	No	No
Trainings and level of staff being trained	Yes 1 radiographer+3 assistants	None specific to cervical cancer Yes for radiotherapy - 4 currently training in S. Africa
Workshops	Yes	No
International level	None	None
National level	Yes	None
Regional level	Yes	None
Number of technical and evaluation reports	None	Yes 2 – JHPIEGO, MOH
Regular system of service statistics exists and is functional	No	No

a The Ministry of Health budget does not break down expenditures beyond Unit level (e.g. MCH) so these data not available. Data for Ghana collected by the author, 1997.

indicate that breast cancer is being given a higher priority than cervical cancer. The equipment resource indicator strongly suggests that breast cancer has a greater priority because expensive mammography equipment was made available for breast cancer while cervical cancer screening was hampered by a lack of very basic and inexpensive equipment – speculums as well as other tools for taking Pap smears. While there were no technical guidelines or screening protocols for either breast or cervical cancer, there were training courses held at several levels of staff and workshops held for breast cancer, suggesting that it was a higher priority. I did not collect data on number and location of clinicians trained in carrying out breast and cervical cancer surgery, or other health workers trained in doing Pap smears, though these data would be valuable to include in similar studies in future.

Table 3

Table 3 Political attention measures: breast and cervical cancer, Ghana, 1990–1997

Variables	Breast cancer	Cervical cancer
Committee created to study the problem	No	Yes
Statements by governmental and NGO leaders	Yes^a (favourable)	No
Regional and district level activities	Yes	No
Multi-Ministry agency involvement	Yes (NCWD, MOH)	No
Involvement of private sector	Yes (donations)	No
Involvement of NGO sector	Yes	No
Funding from government sources	Yes	No
Funding from bilateral and multilateral donors	No	No
Media attention
Use of mass media to inform or motivate public	Yes (radio, TV, newspaper)	Yes (radio, TV, newspaper)
Count of articles^b	60	16
Radio and television programmes	Yes – count not available	Yes – count not available

a This determination is made through coding and analysis of primary interviews and documents.

b 3,247 health-related articles in 1991–1996 were reviewed and coded. Source: Daily Graphic, Accra.

shows that more political attention was given to breast cancer than cervical cancer. Although both forms of cancer were recognized by Ministry of Health officials as public health problems, only breast cancer received MOH attention at the regional and district levels. Evidence from interviews with MOH officials suggested an MOH preference for addressing cervical cancer over breast cancer. Favourable statements made by leaders and multi-Ministry agency involvement in promotion of breast cancer screening very likely reflects the powerful political connections of NCWD and 31DWM. Regional and district level activities for breast cancer can be explained by 31DWM’s large membership and effective outreach to the grassroots in rural areas. Ministry of Health financial support for mammography film and machine maintenance suggests increased political attention and commitment to breast cancer that may be due to pressure from the politically powerful 31DWM.

Discussion

The results of applying the policy priority measure in a case study of breast and cervical cancer in Ghana during the period 1990–97 lead to several important observations about priority setting for reproductive health.

•	Differential availability of evidence, cross-national differences in the public health importance of different diseases and the social construction of issues affect priority setting.

Breast and cervical cancer in Ghana is an instructive case in which political attention essentially trumped available scientific and economic evidence in terms of the priority given to breast cancer rather than cervical cancer. Had summary measures of disease such as the incidence and mortality data and DALYs presented in Table 1 been referred to by all stakeholders, it would very likely have led to the prioritisation of cervical cancer over breast cancer.

This raises the question of how best to incorporate scientific and economic evidence into the political process of priority setting. This case study has shown that scientific and economic evidence matter, but it also suggests that their interpretation may affect the type and amount of influence they have. Perhaps by including measures of policy priority, i.e. organizational and political attention to a health issue, the role of scientific and economic evidence in the priority setting process can be better understood.

In this case, the social construction of breast and cervical cancer affected whether or not each was prioritised. Ghanaian women’s organizations successfully drew attention to breast cancer by drawing on the powerful symbols associated with the breast. References to breastfeeding, motherhood and the nurturing of children enabled them to mobilize not only their own members, but a larger public of men and women as well. The social construction of cervical cancer was associated with more negative images. As a disease caused by a sexually transmitted infection, it was portrayed in the media as being associated with early, and therefore illicit, sexual activity and poor genital hygiene. The negative social connotations associated with cervical cancer therefore made it a more difficult issue on which to carry out a public campaign, had such a campaign even been introduced.

Furthermore, lack of data and lack of information on the part of major stakeholders can lead to incorrect perceptions of whether a health issue should be prioritised in the first place. The existence of clear epidemiological data on breast cancer in the developed world supported and influenced policy attention to breast cancer in Ghana. One form of this influence was the sharing of materials such as educational videos and pamphlets from breast cancer groups in North America. Had there been clear data for Ghana, it might have served to counter-balance this effect. Such transnational influences should not be underestimated [31]. According to both normative measures and policy priority measures, breast cancer was a far greater public health problem in the US in the 1990s than cervical cancer and has rightly received high policy priority there for that reason. In Ghana, on the other hand, cervical cancer is a greater public health problem than breast cancer, but relatively little political attention was given to cervical cancer in the 1990s globally.

•	Politics is an important factor in the determination of priorities.

The Ghana case study illustrates how both the local political situation and transnational influences can affect priority setting, sometimes in unexpected ways. The role of national women’s groups in Ghana, influenced by US and Canadian women’s groups, explains in large part why breast cancer became a higher priority than cervical cancer. Their political connections in Ghana and abroad and access to government officials help to explain their success. Cervical cancer did not have the attention or support of women’s groups to the same extent. Members, and especially office-holders, of the large women’s organizations tend to be elite women from higher socio-economic groups – who may be at greater epidemiological risk for breast cancer. Cervical cancer more frequently affects women of lower socio-economic status. They may also be organized in women’s groups but more often at community level and with few resources and little influence, and they are less able to create attention to such issues. While cervical cancer did enjoy the support of the Maternal and Child Health Unit of the Ministry of Health, it did not have the broader organizational and political attention that would have helped to make it a priority.

•	Priority setting not only has public health implications but also equity implications.

Prioritising breast cancer over cervical cancer has obvious public health implications. However, the difference between the normative and empirical findings for breast and cervical cancer in Ghana also suggests important health equity implications. Despite Ministry of Health preference for giving attention to cervical cancer, it never reached the health agenda. This suggests that the lack of comprehensive priority setting measures and data, and efforts to make all stakeholders aware of these, may reflect and perpetuate existing inequities in the health system and shift limited health resources to one socio-economic group over another, advertently or otherwise.

Conclusion

The case of breast and cervical cancer in Ghana demonstrates the value of including process and political attention measures in an analysis of policy priority as part of any priority setting exercise. In this case study, normative measures clearly showed that cervical cancer should be a higher priority in Ghana than breast cancer, while the empirical variables of process and political attention indicated that the opposite had been the case. Traditional priority setting measures could not have explained this on their own, let alone given markers on how to address and resolve such a disjuncture. The inclusion of process and political measures allowed this distinction to be made.

The interpretation of evidence should be viewed as an important but not always predictable part of the priority setting process. Theories about the social and political construction of health needs and problems can inform the discussion and provide a useful framework to examine priority setting. It is important to note that the evidence actually used in creating policy attention to a health issue may not be the most appropriate evidence, as the case of breast cancer in Ghana illustrates. The political setting within which decisions are made can have important health equity implications and lead to an inequitable allocation of resources, as this case study also shows. In the context of limited resources for reproductive health care, achieving a better understanding of these matters and their health equity implications is especially important.

Where conflicting evidence and resulting inequities are identified, policymakers and others can act together to ensure that a more comprehensive view of priority setting is taken. In bringing all stakeholders to the table and seeking a consensus on which health needs and conditions should get greater or lesser priority, more comprehensive forms of evidence will be required and may be helpful in better understanding the priority setting process.

Acknowledgements

This paper is based on doctoral dissertation research carried out in Ghana in 1996–97. The Hewlett Foundation provided financial support for data collection, analysis and write-up. The Saltonstall Population Innovation Fund provided a travel grant. The author is grateful to the Faculty of the Ghana School of Public Health in Legon for logistical and other support during fieldwork. This paper is based on comments received on a working paper published in October 2001 by the Harvard Center for Population and Development Studies entitled Priority Setting in International Health: Beyond DALYs and Cost-Effectiveness Analysis.