Problem solving for breast health care delivery in low and middle resource countries (LMCs): consensus statement from the Breast Health Global Initiative

Problem solving for breast health care delivery in low and middle resourcecountries (LMCs): consensus statement from the Breast Health Global Initiative

Joe B. Harford a, Isabel V. Otero a, Benjamin O. Anderson b,c,*, Eduardo Cazap d, William J. Gradishar e,Julie R. Gralowb,f, Gabrielle M. Kane b, Laurens M. Niëns g, Peggy L. Porter h, Anne V. Reeler i,Paula T. Rieger j, Lillie D. Shockney k, Lawrence N. Shulman l, Tanya Soldakm, David B. Thomas h,Beti Thompson h, David P. Winchester n, Sten G. Zelle o, Rajendra A. Badwe p,1

aOffice of International Affairs, National Cancer Institute, National Institute of Health, Bethesda, MD, USAbUniversity of Washington School of Medicine, Seattle, WA, USAcBreast Health Global Initiative, Fred Hutchinson Cancer Research Center, Seattle, WA, USAd SLACOM-Sociedad Latinoamericana y del Caribe de Oncología Médica, Buenos Aires, ArgentinaeNorthwestern University Feinberg School of Medicine, Northwestern Memorial Hospital, Evanston, IL, USAf Seattle Cancer Care Alliance, Seattle, WA, USAg Institute for Medical Technology Assessment, Erasmus University, Rotterdam, the Netherlandsh Fred Hutchinson Cancer Research Center, Seattle, WA, USAiAxios International, Paris, FrancejOncology Nursing Society, Pittsburgh, PA, USAk Johns Hopkins University School of Nursing, Baltimore, MD, USAlDana-Farber Cancer Institute, Boston, MA, USAmResource and Policy Exchange, Delhi, NY, USAnAmerican College of Surgeons, National Accreditation Program for Breast Centers, Chicago, IL, USAoDepartment of Primary and Community Care, Radboud University Nijmegen Medical Centre, Nijmegen, the NetherlandspDepartment of Surgery, Tata Memorial Hospital, Mumbai, India.

Keywords:Implementation scienceCost-effectiveness researchSituational analysisMortality-to-incidence ratioPatient navigationAdvocacy

a b s t r a c t

International collaborations like the Breast Health Global Initiative (BHGI) can help low and middleincome countries (LMCs) to establish or improve breast cancer control programs by providing evidence-based, resource-stratified guidelines for the management and control of breast cancer. The ProblemSolving Working Group of the BHGI 2010 Global Summit met to develop a consensus statement onproblem-solving strategies addressing breast cancer in LMCs. To better assess breast cancer burden inpoorly studied populations, countries require accurate statistics regarding breast cancer incidence andmortality. To better identify health care system strengths and weaknesses, countries require reasonableindicators of true health system quality and capacity. Using qualitative and quantitative researchmethods, countries should formulate cancer control strategies to identify both system inefficiencies andpatient barriers. Patient navigation programs linked to public advocacy efforts feed and strengthenfunctional early detection and treatment programs. Cost-effectiveness research and implementationscience are tools that can guide and expand successful pilot programs.

� 2011 Elsevier Ltd. All rights reserved.

Introduction

Breast cancer is the most common cancer among womenworldwide with more industrialized, high income countries havinghigher incidence rates than low and middle income countries(LMCs).1 However, increases in population longevity due toimprovements in sanitation, infectious disease management, andhealth care delivery in LMCs have made breast cancer a growing

* Corresponding author. Breast Health Global Initiative, Fred Hutchinson CancerResearch Center, Box 356410, Seattle, WA 98195, USA. Tel.: þ1 206 543 6352;fax: þ1 206 543 8136.

E-mail address: [email protected] (B.O. Anderson).1 for the BHGI 2010 Summit Problem Solving Consensus Working Group. See

Appendix for the list of members included in this group.

Contents lists available at ScienceDirect

The Breast

journal homepage: www.elsevier .com/brst

0960-9776/$ e see front matter � 2011 Elsevier Ltd. All rights reserved.doi:10.1016/j.breast.2011.02.007

The Breast 20 (2011) S20eS29

concern in LMCs. The adoptionof “Western”behaviors have resultedin women beginning childbearing later in life and engaging in lessbreastfeeding, which will likely result in higher risk for breastcancer.2 Although significant reductions in breast cancer mortalityhave been accomplished in high income countries, primarily due tomajor investments in research and advances in early detection,treatment, and management, the same cannot be said of poorereconomies such as the continent of Africa, Southeast Asia, and partsof Latin America and the Caribbean. In Africa, for example, where70e90% of womenwith breast cancer present with stage III (locallyadvanced) or IV (metastatic) diseasewith tumors averaging 10 cm indiameter,3 the mortality-to-incidence ratio is approximately 0.7,meaning that in a given year there are 7 deaths for every 10 newlydiagnosed cases of breast cancer.4 The comparable mortality-to-incidence ratio in most Latin American countries is approximately0.4 and in North America is approximately 0.2.4 The significantdifferences in mortality-to-incidence ratios suggest that trans-ferring successful practices used in high-resourced environments tolow-resourced environments may be difficult, but has the potentialto save many lives and decrease global health disparities.

The purpose of this consensus manuscript is to summarize thefindings and practical considerations of the Breast Health GlobalInitiative (BHGI) Problem-Solving Working Group. This paper isbased on the presentations made at the 4th BHGI Global Summitheld in association with the SLACOM-Sociedad Latinoamericana ydel Caribe de Oncología Médica, in Chicago, Illinois on June 9e11,2010, and the discussions that resulted from them. The BHGImethods for guideline and consensus development have beenpreviously described.5 Definitions and more detailed descriptionsof the state of breast cancer in low income countries and in middleincome countries can be found in the two other BGHI 2010consensus reviews.6,7

Consensus findings

Baseline cancer incidence and mortality data

Definitive statistics regarding breast cancer incidence andmortality may not be available in LMCs due to the paucity of pop-ulation-based cancer registries and inaccurate and incompletedeath registration. However, estimates can be provided usingstatistical models, like those used to produce GLOBOCAN by theInternational Agency for Research on Cancer (IARC) of the WorldHealth Organization (WHO).1 GLOBOCAN provides estimates ofcancer incidence and mortality rates based upon the best dataavailable from registries around the world. Quality breast cancerdata from LMCs are important because data helps policymakers andhospital administrators define the breast cancer problem in a givenpopulation, or subpopulation, often definable by factors such asage, socioeconomic status, race, ethnicity and place of residence.

Population-based cancer registries provide an importantresource for epidemiological research and the planning and eval-uation of breast health care programs for early diagnosis andtreatment.8 Quality and reliable data on breast cancer incidence areavailable from selected LMCs with long-standing population-basedregistries, which have observed increases in breast cancer inci-dence.9 However, data from most LMCs in Africa, Asia and theMiddle East are lacking. Of the data included in Volume IX of CancerIncidence in Five Continents also produced by IARC, more than 70%are derived from North America and Western Europe.10

Some experts argue that a population-based cancer registryshould only be established if the information needed for cancercontrol cannot be provided more efficiently by other means.A cancer registry is not a project with a beginning, middle, and end,but rather an activity that requires substantial resources to be

sustained. The IARC, within the framework of the WHO, providestechnical support to LMCs to establish cancer registries through itstraining, education and research programs.11

Practical considerations for cancer registries in LMCs

The often referenced cancer incidence and mortality report,GLOBOCAN 2008, provides estimates for all countries of the worldwhether registries are present or not.1 GLOBOCAN 2008 statisticsare calculated using the provisional estimates of the age- and sex-specific deaths from cancer (of all types) in each country of theworld, extracting national population estimates from the UnitedNation population division. Although GLOBOCAN uses the mostrelevant data available, it may need to extrapolate from neigh-boring countries when an LMC lacks its own population-basedregistry. This extrapolation could compromise accuracy.

Alternate approaches exist to estimate the numbers of breastcancer deaths and deaths of all-causes among women in a givenLMC. Data can be collected from formal death certificate reviews ina defined region of a country for a specific number of years. Alsouseful are reviews of hospital records to capture those cancer casespassing through one or more hospitals. Because many cancer casesnever make it to the hospital, hospital record registries or caseseries cannot be generalized to describe incidence rates for a givenpopulation. Nonetheless, they can provide very important infor-mation regarding median tumor size and stage of disease atdiagnosis, data which is particularly valuable for assessing thesuccess of early detection activities by providing a measure, albeitimperfect, of cancer down-staging over time. Hospital registriescan also be used to collect data on breast cancer treatments thatcan then be used to identify gaps in cancer treatment and training.Institution-based pilot programs can also create data collectionsystems.

Cancer registration methods in LMCs are very different frommostly passive methods used in high-resourced registries of highincome countries. Twinning (two countries partnering to achievea common goal) between high income countries and LMCs or moreestablished LMCs and less established LMCs (southesouth collab-orations) allow for the sharing of breast cancer data registry bestpractices and tools. Twinning or cooperation between a wellfunctioning established registry and a new LMC registry couldcatalyze effective program development. International cancerregistry organizations like the IARC12 can integrate with regionalcancer registry networks such as the European Network of CancerRegistries,13 the North American Association of Central CancerRegistries,14 the Asian Cancer Registry Network, and the Group ofLatin Language Registries.

The 2008 BHGI consensus guidelines recommend recordkeeping for countries at the “basic resource level” that includesindividual medical records and service-based patient registration(Table 1). The guidelines recommend facility-basedmedical recordsand centralized patient registration and hospital level cancerregistration for countries at the “limited resource level”, and pop-ulation-based cancer registries for countries at the “enhancedresource level”.15 There is a tendency to think that a national cancerregistry is one that covers the entire country. While completepopulation coverage is practical (and even necessary) in countrieswith smaller populations, it is impractical in more populousnations. In larger countries, sentinel registries that providereasonable population sampling can yield statistics that areadequate for situation analysis and assessment of cancer controlactivities. In countries where total health spending defines limita-tions in health care delivery, the costs associated with covering theentire population with a cancer registry is unlikely to be identifiedas the highest priority activity.

J.B. Harford et al. / The Breast 20 (2011) S20eS29 S21

The U.S. Institute of Medicine (IOM) has recommended thatcancer registration in LMCs should begin in urban areas where thelogistical advantages of data acquisition are significant and thecosts per case registered are lower than in rural areas.16 Costsassociated with a cancer registry include costs related to ascer-tainment of cases, and the cost of training epidemiologists toanalyze and interpret the data, preferably in association witha health department, university, or research institution, to ensurethat problems related to public health are competently addressed.

Situational analysis for program planning in LMCs

To develop breast health care programs in any setting, it isnecessary to have information on common program capacityindicators, such as the number of radiotherapy, ultrasound andmammography machines (along with their distribution) andinformation on the number of health care professionals and breastcancer specialists (along with their distribution). Collecting thistype of health care system data can help develop metrics andidentify gaps and opportunities for equitable allocation ofresources. The American Cancer Society has initiated a score cardproject, involving four Latin American countries, that includes thecreation of data collection tools to measure patient outcomes anddetermine health care policy and government structures relevantto cancer control, including the existence and roles of a nationalcancer institute, and the inclusion of breast cancer in nationalcancer control plans.17

Data collection should be tailored to the key situation analysisquestions being asked. For example, an analysis of screeningmammography capacity requires more than an enumeration ofmammography units in a region. In addition, it is necessary to knowwhether the units are being used, how they are being used, if theresulting images have adequate diagnostic quality, and if there aresufficient trained personnel to support best practices. In Brazil, theNational Cancer Institute (INCA), working in partnership with theBrazilian Department of Informatics for the Unified Health System(DATASUS,) has developed an Information System for the Control ofBreast Cancer (known by its Portuguese abbreviation SISMAMA) tobe used as a management tool to capture, organize and makeavailable data about the population tested, test results, follow-up of

abnormal cases, the quality of the services, and other essentialinformation generated in the course of providing screening exams.18

Practical considerations for situational analysis

Mortality statistics can also be used, but accurate informationmay be difficult to collect in LMCs.15 WHO collects country-leveloverall mortality statistics, but the accuracy of these statistics arequestionable. To calculate accurate mortality rates in a population,the numbers of males and females of different agesmust be known,and information from death certificates on the age and gender of allpersons who die, and the cause of death, must be available orobtainable using reasonable methodologies.

The accuracy and completeness of death certificates varies fromcountry to country. Verification of cause of death from deathcertificates can be done through select review of medical recordsfor the cause of death, or by a sample survey of the population, or by“verbal autopsy”. Verbal autopsy is an approach used whena significant portion of deaths occurs outside the medical system(e.g., rural areas in LMCs). It involves obtaining the cause of deathby interviewing respondents about the signs and symptomsexperienced prior to the death of someone they knew personally,followed by the assignment of probable cause of death usingphysician review and defined criteria. Verbal autopsy can provideuseful information in conjunction with other methods of dataacquisition.19 WHO has produced a manual of standards for con-ducting and utilizing verbal autopsies to determine cause ofdeath.20 Of course, it is impractical to utilize verbal autopsymethodology for an entire population. It is useful, however, toobtain a sampling of causes of death from a selected subpopulation.

Differences found in cause of death through verificationmethods may suggest a need to improve death reporting methods.The number of deaths from breast cancer within a population alsoprovides an indication of the level of palliative care servicesrequired to provide adequate end-of-life care for breast cancerpatients. While planning for breast cancer treatment facilitiesshould include palliative care services, the overall palliative carerequirements in a given locale should take into account deaths fromall causes since >90% of deaths are not from breast cancer (even inhigh income countries).

Table 1Summary of recommended support system resources for low and middle resource countries.

Resource level Services Record keeping Cancer care facility Breast care center

Basic Diagnostic/PathologyservicesNursing servicesOncology servicesPalliative servicesPsychosocial servicesPrimary care servicesSurgical services

Individual medical records andservice-based patient registration

Health facilityOperating facilityOutpatient care facilityPharmacyExternal consultation pathologylaboratoryHome hospice support

Breast health care accessintegrated into existing healthcare infrastructure

Limited Imaging servicesPeer support servicesRadiation oncology services

Facility-based medical records andcentralized patient registrationHospital level cancer registry

Clinical information systemsHealth system networkImaging facilityInternal pathology laboratoryRadiation therapy

Breast Center with clinician,staff and breast imaging accessBreast prostheses formastectomy patients

Enhanced Cancer follow-upGroup supportScreening programsRehabilitation servicesSurvivorship services

Resource Room(s) for education/outreachFacility-based follow-upRegional cancer registry

Centralized referral cancer center(s)Radiation therapy: low energy linearaccelerator, electrons, brachytherapy, treatmentplanning system

Multidisciplinary breastprogramsOncology nurse specialistsPhysician assistants

Note: The table stratification scheme implies incrementally increasing resource allocation at the basic and limited levels.Definitions: Basic level resources are core resources or fundamental services that are absolutely necessary for any breast health care system to function. Limited level resourcesor services are those that produce major improvements in outcome but that are attainable with limited financial means and modest infrastructure. Source: Adapted fromRef. 15. This material is reproduced with permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.

J.B. Harford et al. / The Breast 20 (2011) S20eS29S22

Critical workforce issues

One of the greatest challenges facing health care delivery inLMCs is the shortage of health care workers d an estimatedunmet need of over 4 million nurses, midwives, physicians, andsupport workers worldwide.21 The problem is most severe in thepoorest of settings, especially sub-Saharan Africa, where theestablishment and sustainability of health care programs havebeen jeopardized by outward migration, lack of positive workenvironments, and a weak knowledge base of medicine andresearch.22

Worldwide, there are at least 60 million health workers, two-thirds providing direct health services (doctors, nurses,midwives, pharmacists, laboratory technicians, etc.) and one-third providing indirect health care services (hospital managers,financial officers, cooks, drivers and cleaners).23 Obstacles toworkforce development in LMCs include lower salaries andpoorer work environments. Quality of life outside the hospital isalso a concern for health care workers. Approximately 20% ofphysicians in many European countries come from LMCs,23

reflecting a phenomenon commonly referred to as “brain flight”or “brain drain”. Brain drain can also occur within a countrywhen health care workers move from rural to urban settings.Moreover, international donors, through their focus on commu-nicable diseases such as HIV/AIDS and tuberculosis, have alsoinfluenced the distribution of the global health workforce,inadvertently depleting the number of medical professionalsworking in cancer and other less-funded diseases, in favor of HIV/AIDS and other communicable diseases.

In 2010, the World Health Assembly adopted a Code of Practiceto try to balance the rights of educated and skilled individuals tomigrate from poor countries to richer ones with the need tosupport the struggling health care systems that they leavebehind.24 Key components include taking measures to educate,retain and sustain in-country health workforce, while ensuring thatmigrant workers’ rights are equal to domestically trained healthworkers. Research has shown that better education, training, andworking environments can help improve critical workforce issuesin LMCs.25

Practical considerations for critical workforce issues

Investments in in-country education and training programs inLMCs are needed to lessen global health disparities and strains onlocal health care systems. In many LMCs there are insufficientnumbers of academically trained medical personnel, such asphysicians, nurses and pharmacists. Without in-country educationand training programs, there is little hope of training adequatenumbers in the near future. As an intermediate solution, theconcept of “task-shifting” for the performance of some health carefunctions may need to be considered. Motivated and highly capableworkers can be trained to conduct focused tasks, relieving physi-cians, nurses and pharmacists to perform more complex tasks,resulting in an expanded program able to treat a larger volume ofpatients. However, task-shifting requires an investment in thesupervision and monitoring of hospital personnel to ensure thattrained workers are delivering services with minimal errors andthat the quality of care is not compromised.

International collaborations such as the partnership betweenthe Oncology Nursing Society (ONS) and the U.S. Middle EastPartnership for Breast Cancer Awareness and Research are otherstrategies to consider, especially when oncology expertise may bemore readily available outside an LMC. In this example, ONS hasbeen working through the U.S. Department of State Middle EastPartnership Initiative (MEPI) to develop a “train the trainer”program that trains personnel in their own institution as a modelfor sustainability of nursing education in LMCs.26

The 2008 BHGI recommendations for countries at the “basicresource level” include: education of primary care providers in thedetection, diagnosis and treatment of breast cancer; and the educa-tion of nurses regarding cancer patient management and emotionalsupport. BHGI recommendations for countries at the “limitedresource level” include: education of nurses regarding breast cancerdiagnosis, treatment, andpatientmanagement; education of imagingtechnician regarding imaging technique and quality control; and therecruitment of volunteers to address gaps in health care support (seeTable 2). BHGI recommendations for countries at the “enhancedresource level” include: organization of national volunteer networks;specialized nursing oncology training; and onsite cytopathologists.15

Table 2Summary of recommended breast cancer program human resources for low and middle resource countries.

Resource level Patient and family education Human resource capacity building Patient navigation

Basic General education regarding primaryprevention of cancer, early detectionand self-examination

Primary care provider education regardingbreast cancer detection, diagnosisand treatment

Field nurse, midwife or health care provider triagespatients to central facility for diagnosis and treatment

Development of culturally adaptedpatient and family education services

Nursing education regarding cancerpatient management andemotional supportPathology technician education regardingtissue handling and specimen preparationTrained community worker

Limited Group or one-on-one counselinginvolving family and peer supportEducation regarding nutrition andcomplementary therapies

Nursing education regarding breast cancerdiagnosis, treatment and patient managementImaging technician education regardingimaging technique and quality controlVolunteer recruitment corp to support care

Onsite patient navigator (staff member or nurse)facilitates patient triage through diagnosis and treatment

Enhanced Education regarding survivorshipLymphedema educationEducation regarding Home Care

Organization of national volunteer networkSpecialized nursing oncology trainingHome care nursingPhysiotherapist and lymphedema therapistOnsite cytopathologist

Patient navigation team from each discipline supportspatient “handoff” during key transition from specialistto specialist to ensure complete therapy

Note: The table stratification scheme implies incrementally increasing resource allocation at the basic and limited levels.Definitions: Basic level resources are core resources or fundamental services that are absolutely necessary for any breast health care system to function. Limited level resourcesor services are those that produce major improvements in outcome but that are attainable with limited financial means and modest infrastructure. Source: Adapted fromRef. 15. This material is reproduced with permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.

J.B. Harford et al. / The Breast 20 (2011) S20eS29 S23

Global health workforce shortage is one of the greatest imped-iments to health care delivery in LMCs, affecting the quality of careand ultimately the health outcomes of patients. Human resourcesneeded to implement breast cancer control programs includepersonnel to provide patient and family education, social workers,and patient navigators. Lack of resources for training existing andnew medical professionals is also a core concern for capacitybuilding.

Fundamental elements of breast health program design

Multidisciplinary approach and accreditation

It has become increasingly evident that the successfulmanagement of breast cancer requires a multidisciplinaryapproach, with oncologists, radiologists, surgeons, pathologists,nurses and other health care professionals working in concert toefficiently guide patients through a cohesive system of compre-hensive medical services. However, the shortage of specializedhealth care professionals may make it difficult or impossible toadopt a multidisciplinary breast program in LMCs. This is unfor-tunate as the importance of multidisciplinary case managementhas been demonstrated by the associated disparities observed inbreast cancer mortality rates that prevail even within Europe,where the adoption of a multidisciplinary approach and highvolume specialized breast cancer clinics are not evenly distributedamong European countries.27 A recent survey of 338 oncologyspecialists from Arab countries showed that three-fourths regularlyattended multidisciplinary tumor boards and over 90% of thesespecialists thought that tumor boards should be mandatory forcancer treatment planning.28

Practical considerations for multidisciplinary breast healthprograms include the adoption and adaptation of national accred-itation efforts similar to the efforts found in high income countries.For example, national accreditation efforts in the United States havebeen in development since 2006 through the National Accredita-tion Program for Breast Centers (NAPBC)29 and could be adapted forinternational use if standards are adjusted to reflect the limitationsof existing resources in LMCs. Settings with limited resources mayalso benefit from starting with small multidisciplinary teams of twoor three members d a general surgeon and a clinical nurse d andexpand over time as resources become available.15

Setting standards through accreditation can help programs setcommon goals and identify gaps in quality indicators. It can alsohelp accredited organizations to obtain training and clinicalupdates and set them in the right path toward better quality healthcare when more resources become available.

Implementation and assessment of outcomes

All programs for early detection, diagnosis and treatment ofbreast cancer should include a way to measure their impact ondeaths due to breast cancer, stage of disease at diagnosis, andtreatment practices. A situation analysis, as discussed earlier, helpsto measure outcomes by establishing baselines that can be used toevaluate the impact of a given intervention after a specific period oftime. Improving assessment and planning of breast healthprograms is very important and should be considered part of thelessons learned from other health programs.

Themost rigorousdesign,orgold standard, for theevaluationof anintervention is the randomized clinical trial (RCT),whichmaybehardto achieve even in high income countries. LMCsmay need to employmore novel RCT schemes to accommodate for the high-resourcedemandsof traditional RCTs.Anexample is the cluster randomizationemployed in a Shanghai breast self-examination (BSE) trial.30

Staggered introduction of an intervention can also be employed ina randomized fashion, so that the intervention is phased in byrandomly selecting somesegments of thepopulation for interventionbeforeothers, therebycreatingan interventionandcontroldataset forcomparison.31 Irrespective of the method employed, data collectionshould begin with the earliest stage of the program.

Practical considerations for implementation and assessment ofoutcomes

Before a program can be expanded, it is important to consider itslikely impact on the general population. The impact of theprogram’s expansion can be estimated by multiplying the efficacyof an intervention by the coverage:

impact ¼ efficacy� coverage:

For example, if a pilot program reduced the percentage ofwomen with advanced disease at diagnosis by 50% (the efficacy),but the program only reached 10% of the targeted population (thecoverage), then the overall impact of the program on breast cancerdown-staging will be a 5% reduction in advanced disease(0.50 � 0.10 ¼ 0.05). If the program reached 50% of the population,the overall impact would be a 25% reduction. The WHO has sug-gested that coverage of less than 70% makes it unlikely that a pop-ulation screening program will have a significant impact onmortality.32

The mortality impact from mammography programs in moredeveloped countries has been achieved in no small measure bystriving to maximize participation (coverage) through strategicapproaches, such as reminder/recall systems. More limited capacityand catch-as-you-can approaches, such as having a mobilemammography van drive around a city delivering mammograms tothose who show up and not to those who do not, is unlikely to havea significant impact on the population’s mortality from breastcancer, owing to the limited and rather unpredictable coveragesuch an approach affords. In many LMCs, early detection effortsreach very small portions of a target population. Though well-intended, these programs have limited potential for having animpact on the breast health of the population as awhole because oftheir limited coverage. A cost-benefit analysis of such efforts shouldbe considered.

Implementation science can help shape global breast cancerinitiatives

Implementation science aims to study methods that willimprove and increase the application of research findings to patientcare practice and policy.33 Implementation research can addresscomplex issues at various levels of a health care system, identifyingcritical barriers and suggesting possible solutions, which can thenbe evaluated for efficacy both before and during implementation.This approach results in more effective strategies and greaterbenefit to more individuals in the population. While “what to do” isoften known, “how to do it” may not be, and the answer alwaysdepends on the context. Implementation science shifts the focus ofresearch to address questions such as: Why do proven healthinterventions fail to have their expected result in this particularcountry?; Why do established programs lose effectiveness overtime?; Why do programs validated elsewhere sometimes haveunintended effects in a new setting?; and Howcan interventions beeffectively combined to minimize cost and reduce the splintering ofthe health system into disease-specific programs?34

Implementation research combines tools used in the sciencesand humanities, including medicine, anthropology, sociology,

J.B. Harford et al. / The Breast 20 (2011) S20eS29S24

engineering, business and economics, to suggest ways in whichscientific findings can be applied at the broader population level. Itcombines quantitative research (such as randomized trials), whichallows for comparison of health delivery strategies so as to shapeprograms with goals such as reduced mortality from breastcancer,34 with qualitative and observational studies that considerthe culture and context of a given population and address culturaland social obstacles to care.35 Although it is important to tailorinterventions to a given situation, knowledge generated from onecountry can help inform the implementation of interventions inother countries, especially when looking at countries with similarcapacities and resource levels. Implementation research is an iter-ative process: data from pilot programs and small randomizedstudies can be used to plan new health interventions intended forlarge scale implementation.

Cost-effectiveness analysis (CEA) in LMCs

With ever-increasing health care costs, health care systems areunder more and more pressure to make choices about funding.36

Policymakers often need to make difficult decisions about whichhealth programs and interventions to support, and may need tochoose, for example, whether to fund end-stage palliative care, orfund breast cancer awareness and early detection programs. Cost-effectiveness analysis (CEA) can provide a systematic framework formaking these decisions, by furnishing information about the rela-tive value of different interventions. CEAmay be especially relevantfor countries with both a large burden of disease and relativelysmall health care budgets. For example, most countries in Africaspend less than US$100 per capita on all health care,37,38 makingcost-effectiveness and affordability critical considerations inimplementing health programs.

A standard CEA considers incremental costs and benefits, andexamineswhether a new intervention provides value for themoneyspent compared to the status quo (which may be to do little ornothing). CEA has shown the importance d and economic attrac-tivenessd of diagnosing cancer in earlier stages39 when treatmentis usually more effective, as well as less complex and less costly.

An example of the role of CEA in LMCs is a current CEA of breastcancer control that is ongoing in 10 selected LMCs in an interna-tional collaborative effort between Susan G. Komen for the Cure�,WHO, Radboud University (Nijmegen), and Erasmus University(Rotterdam). This study will estimate the generalized costs andcost-effectiveness of a range of breast cancer interventions for earlydetection, treatment and palliative care. The analysis distinguishestreatments for stage I to IV breast cancers with or without earlydetection programs (such as awareness of early signs and symp-toms, and screening by clinical breast examination (CBE) ormammography). The analysis also evaluates stage IV treatmentversus pain relief and palliative care, which is of particular rele-vance in low income countries where over 70% of cancer cases arediagnosed in stage III and stage IV.

Another example is a study conducted by Okonkwo et al.,40 whoused a Microsimulation Screening Analysis model to estimate thecosts of screening for breast cancer in India; its likely effects onmortality, and its cost-effectiveness, i.e., costs of screening per life-year gained or life saved. The WHO-CHOICE program has alsodevelopedmodels to predict country-specific unit costs for 14WHOsub-regions.41 These estimates can be used when country-specificcost information is not available. Ongoing CEA have the potential toprovide an important knowledge base for assisting countries todevelop policies on breast cancer control and should be encour-aged. However, it is important to keep in mind that a CEA does notanswer questions about how to equitably deliver breast cancertreatments, nor does it allow policymakers to analyze trade-offs

between benefits and harms regarding other valid objectives thatcan or should motivate choices in health care. Therefore, a CEA ofbreast cancer care is only one valuable piece of information thatgoes into the complex process of decision-making.

Qualitative research

Qualitative research can be used to study social, psychologicaland system barriers, and facilitators to breast cancer health caredelivery in LMCs. With strong underpinnings in anthropology,psychology and other social sciences, qualitative research can helpinvestigators better understand the opinions and views ofa community from its own cultural context and idiosyncrasies42,43

This is important as certain health beliefs, such as fatalism ordestiny, can result in a delay of care-seeking behavior in womenwith breast cancer, as well as a diminished sense of control oncea women is diagnosed. A study conducted in India showed thatbeliefs about breast cancer influenced help-seeking intentions forthe disease, with rural women respondents in India holding morenegative beliefs about breast cancer than urban women respon-dents, and were more likely to report less positive attitudes towardhelp-seeking.44 Qualitative research helps to build a holistic pictureof an individual, making use of words, descriptions, and percep-tions of community members referred to as informants.45e47

Another example of the role of qualitative research in LMCs isa recent study conducted by Puschel and Thompson48 that madeuse of qualitative research to guide the design of a breast screeningintervention in Chile. In this study, focus groups were used toidentify and better understand predisposing, enabling, and rein-forcing factors associated with mammogram screening amongwomen in Chile. Others qualitative studies have shown thatphysicians may not recommend screening to certain patientsbecause of perceived social or cultural barriers that may or may notexist for that patient, or patients may not seek recommended life-saving measures because of cultural beliefs or social barriers, orthey may instead encounter health system or environmentalbarriers.43,49,50

Qualitative data differs from quantitative data in that it relies onfew cases, but many variables, to obtain a deep, rich understandingof a phenomenon.45,46 Based on complex data from conversations,a qualitative researcher may deduce that many women areashamed of a breast lesion and are prone to hiding it until it is toolate for a cure. Other conversations may disclose that cost isa barrier to obtaining screening or reveal that breast screening maybe prompted by a strong desire to be present for her children andfamily. By understanding what women believe and know aboutbreast cancer, an investigator can generate testable hypothesis toincrease compliance with early detection and treatment. Qualita-tive research can also provide data and stories that crystallize theneed for palliative care services, or test hypotheses aboutthe knowledge and behavior of physicians and nurses who are thefrontline of detection and diagnosis. For example, a recent study inMexico City combined semi-structured interviews of breast cancerpatients and cancer care providers with medical record chartreview to assess factors contributing to delayed breast cancerdiagnosis and treatment. They found that the average interval fromthe first primary care consultation to a confirmed cancer diagnosiswas over six months, demonstrating that protracted referral timefrom primary to specialty care (rather than patient procrastination)accounted for the majority of the delay.51

Patient navigation programs and advocacy

Navigating through any medical system can be confusing andoverwhelming for cancer patients and their families. Patient

J.B. Harford et al. / The Breast 20 (2011) S20eS29 S25

navigation programs have been used successfully in minoritycommunities in the United States, with documented improvementsin patient outcomes.52 A patient navigator oversees the course ofcare and works closely with physicians to ensure patient compli-ance with the details of a treatment and follow-up regimen. InLMCs, patient navigators are generally not health professionals buttrained lay persons. Patient navigators can help patients to antici-pate and resolve barriers that can prevent their receiving care ina timely fashion. These barriers might include lack of insurance ortransportation, or speaking a different language than the healthcare providers. Navigators can also assist patients by answeringquestions, providing emotional support, ensuring that appoint-ments are kept, helping patients find financial resources and fillingout the requisite paperwork. In LMCs, all frontline health workersshould be trained and empowered to be patient navigators.53

Breast cancer advocacy movements are well established inNorth America and Western Europe, and have proven to bea powerful force for the advancement of cancer control. In mostLMCs, advocacy is less developed,54 but is gaining momentum.Patient advocates in LMCs can play an important role and workeffectively with health care professionals and policymakers tostimulate community discussion and develop and implementbreast cancer early detection and outreach programs in LMCs.55e57

The role of breast cancer advocates in most countries generallyevolves along a continuum, from support and education, to devel-oping an expanded sense of social responsibility, and finally tobecoming active change agents in the political and social arena.

LMCs have differing financial needs, resource limitations, socialbarriers and competing illnesses that frame how breast canceradvocacy can and should be implemented.56 However, somebarriers to early detection and down-staging of disease crosscultural and national boundaries, such as lack of awareness ofbreast cancer and stigmatization of cancer patients. These are areasthat can be addressed through global breast cancer advocacyefforts.

The model of breast cancer advocacy practiced in Europe andNorth America may not be appropriate or reproducible in othercountries, especially since cultural, social, political and religiousbarriers may be unique to a given scenario. In fact, it may bedifficult at times for non-natives to address these barriers so localgrassroots efforts are advised.56 International advocacy groups facethe challenge of dealing with different languages, cultures andhealth systems. Despite these differences, creating advocacynetworks among low, middle and high resource countries can beeffective and is recommended.54

International advocacy groups have not only funded outreachprograms, but also evidence-based research. Scientific evidence isan important basis for advocacy strategies, and advocacy effortsbased on emotional appeals can distract from more effective,evidence-based programs. For example, in the United States duringthe 1990s, more than 41,000 women underwent high-dosechemotherapy plus autologous bone marrow transplants (HDC-ABMT) for breast cancer, despite the lack of clinical evidence for itsefficacy and high toxicity of the treatment, in large part because ofthe massive efforts put forth by breast cancer advocates and themedia.58 In LMCs, where resources are more limited, it is criticalthat science be the foundation of advocacy.

Program scale-up, networking and partnership

Once a given intervention has been shown to be effective in thepilot project phase, program expansion, or scaling up, is the nextchallenge. Barriers to successful wide-scale implementation maybe very different from those in the pilot project phase, and fewerresources per targeted person may be available.

The planning of cancer-related initiatives can be informed bylessons learned from other health care initiatives in LMCs. Forexample, global polio eradication initiatives used existing immu-nization and health system surveillance infrastructures and staff tolead, organize, and supervise the polio eradication effort.59

Program elements that contributed to the success of polio vacci-nation programs that could inform other health programs included:simplifying tasks; utilizing non-medical and less literate or lessskilled staff and volunteers supervised by trained medical staff;using existing advocacy channels to recruit volunteers; providingper diem incentives to cover volunteer and staff out-of-pocketexpenses; using simple management tools and transparentperformance indicators to access outcomes, and providing tech-nical assistance to increase in-country capacity to implementhealth strategies.59e61

Barriers to breast cancer program expansions included lack ofcontinued support from collaborating or funding organizations,high staff turnover and changes in organizational leadership.62

Programs are most successful if they are flexible and easily adapt-able to different targeted populations (i.e., modular approacheswere employed), include community volunteers, directly involvethe target community in planning, and specifically monitoroutcome using evaluation techniques that could help the programadapt to community needs based on participant feedback.62

Breast cancer scale-up projects are currently underway throughthe Global Task Force on Expanded Access to Cancer Care andControl in Developing Countries. For example, in Mexico, oneprogram is focused on increasing breast cancer care options byimproving the range of approved treatments covered by thenational insurance plan.63 In Malawi, Rwanda, and Haiti, a part-nership between poor rural districts with no oncologist and Part-ners in Health (PIH) and the Dana-Faber Cancer Institute, theHarvard Medical School, and the Brigham and Women’s Hospitaloffer pathology processing and interpretation, expert consultationand chemotherapy, all free-of-charge, for the care of cancerpatients.64

The IOM recommends that “international partners should assistin developing and improving cancer centers in LMCs throughtwinning arrangements and other means”.65 Although thisrecommendation is aimed at general cancer centers, the principle isalso applicable to specific breast health services. Many LMCs needhelp in achieving their health care goals, and seek assistance, suchas technical expertise, from partners in higher income countries orother LMCs and/or neighboring countries with more developedhealth systems. The BHGI is an example of an internationalcollaboration between low resource countries and high resourcecountries that provides onsite training and assistance to improvebasic services such as pathology66 and breast surgery, as well asnetworking and educational opportunities.

Corporations can also be part of international partnerships intechnology development and research. Corporate marketingresearch often includes collecting data on access to technology incountries of interest. Companies are beginning to realize that thereis a huge market outside the United States and Europe, if aneffective product could be delivered affordably. Collective bar-gaining efforts, such as those of the PIH, can make alternativetechnology approaches (such as lower-technology equipment) partof corporate considerations, and are one way to bridge the gapbetweenwhat is made available by companies and what is needed.Health care experts joining together through academic and orga-nizational initiatives to network with companies is one way to helpcompanies understand the needs and constraints of LMCs. Anexample of this is the successful breast cancer initiative in Ethiopia,where academics partnered with a pharmaceutical company todevelop a sustainable breast cancer initiative in the capital city.67

J.B. Harford et al. / The Breast 20 (2011) S20eS29S26

Conclusions and discussion

National or regional breast cancer data are important for breastcancer programs to be able to assess the magnitude of the breastcancer problem in a population or subpopulation, but these dataare often not available in LMCs. Alternatives, such as smaller ortargeted registries, may be easier to establish and manage.Mortality data may also not be available; alternative data sourcesinclude hospital records, death certificates and verbal autopsies.Data on health system capacity indicators can help identify systemstrengths and weaknesses, including data on the number ofradiotherapy, ultrasound andmammographymachines (along withtheir distribution) and data on the number of health care profes-sionals and breast cancer specialists, such as surgeons, oncologists,and anesthesiologists (along with their distribution). Investmentsin data collection and research should include quantitative, quali-tative and cost-effectiveness research to help better understand thebreast cancer scenario in a given setting and to expand or scale-upsuccessful pilot programs.

Efforts to identify both system and patient barriers to care, andinitiate patient navigation programs and supporting advocacyefforts are needed in LMCs. It can be challenging to successfullyimplement early detection programs in LMCs, due to a lack ofgeneral education about breast cancer, or major misconceptionsabout the nature or curability of the disease.68 Resource constraintsas well as access barriers to cancer drugs can limit treatmentoptions.69

Breast cancer control efforts should consider not only healthcare delivery (early detection, diagnosis and treatment), but alsopublic participation as a tool to increase awareness and promoteearly detection (see Fig. 1). Breast cancer survivors can help rein-force the importance of early diagnosis and treatment. Their veryexistence provides visible proof to the general population thatfavorable cancer outcomes are possible. The awareness thatsurvival is possible can create an upward spiral, leading to evenmore awareness, more advocacy, and more survivors, who thenthemselves become advocates.

Most of the problems associated with the future of breast cancerdetection, diagnosis and treatment in LMCs appear to be ratherdaunting, especially in light of the anticipated increase in numberof cases in LMCs and the limited resources to address the problem.However, lessons learned in high income countries, as well as othermore successful LMCs, can be applied to less successful LMCs, withthe help of implementation science.

The global disparities in the mortality-to-incidence ratios can bedecreased by more efficacious treatment, and by the introduction

of early detection or screening, resulting in cases being detected atearlier stages with more favorable prognoses. However, althoughscreening will improve outcomes for some breast cancer cases, itwill also identify other cases that, if left undiscovered, may not havebecome clinically manifest or proven fatal, resulting in the problemof over-diagnosis and overtreatment. More benign cases may alsoundergo diagnostic work-up to rule out cancer (false positives),placing added burdens on the health care system.70

LMCs face various challenges in their path to prosperity, oneof which is growing incidence rates and high mortality due tobreast cancer. Much has been learned about breast cancer, butlittle has been applied to benefit the more than one hundredthousand lives of women diagnosed with breast cancer thatcould be saved each year in LMCs if early detection and treatmentwere implemented. Implementation research provides a frame-work for translating knowledge into practical solutions. It is anopportunity that the global health community cannot afford tomiss.

Contributors statement

The manuscript draft was based on the BHGI 2010 SummitWorking Group consensus meeting, outline review, and Day 3panel discussion and presentations. Working Group memberscontributed sections according to their expertise. JBH and IVOincorporated sections into the main manuscript. All authorsreviewed the draft and provided comments or substantive revi-sions that were incorporated into the final version as per consensusprocess previously described.

Conflict of interest and funding statement

BOA received funding from Breast Health Global Initiative(BHGI); EC received consultancy funding from Bayer ScheringPharma, grants from the Breast Cancer Research Foundation(BCRF) and Poniard Pharmaceuticals, was on the speakers bureausfor Bayer and Bristol Meyers Squibb, and received travel expensesfrom ASCO, UICC, ESMO, and BCRF; DBT and PP have receivedfunding from BHGI for travel; JRG received funding from Genetech,Amgen, Norvartis, and Roche; LMN and SGZ received funding fromBHGI and Susan G. Komen for the Cure�; BT received travelfunding or other support from Fogarty International. All otherauthors and working group members reported no conflict ofinterest.

Awareness

Survivorship

Early Detection

Diagnosis

Treatment

Advocacy

Health Care Delivery Public Participation

Fig. 1. Synergistic relationship between public participation and health care delivery in down-staging breast cancer and improving cancer outcomes (�University of Washingtonwith permission).

J.B. Harford et al. / The Breast 20 (2011) S20eS29 S27

Acknowledgments

On behalf of all of the authors and contributors, BHGI thanksSandra Ripley Distelhorst for her invaluable services andoutstanding contributions as Publication Editor in the preparationof this consensus manuscript.

Appendix. List of members

The BHGI 2010 Summit Problem Solving Consensus workinggroup included Benjamin O. Anderson, Rajendra A. Badwe,Mohammad A. Bushnaq, Eduardo Cazap, William J. Gradishar, JulieR. Gralow, Gabrielle M. Kane, Laurens M. Niëns, Peggy L. Porter,Anne V. Reeler, Paula T. Rieger, Lillie D. Shockney, Lawrence N.Shulman, Tanya Soldak, David B. Thomas, Beti Thompson, ElizabethThompson, David P. Winchester, Joe B. Harford and Isabel V. Otero.

References

1. Ferlay J, Shin H, Bray F, Forman D, Mathers C, Parkin D. GLOBOCAN 2008, cancerincidence and mortality worldwide: IARC cancer base no. 10. Lyon, France:International Agency for Research on Cancer [cited 2010 July]; Available from:http://globocan.iarc.fr; 2008.

2. Porter P. “Westernizing” women’s risks? Breast cancer in lower-incomecountries. N Engl J Med 2008 Jan 17;358(3):213e6.

3. Fregene A, Newman LA. Breast cancer in sub-Saharan Africa: how does it relateto breast cancer in African-American women? Cancer 2005 Apr 15;103(8):1540e50.

4. Kamangar F, Dores GM, Anderson WF. Patterns of cancer incidence, mortality,and prevalence across five continents: defining priorities to reduce cancerdisparities in different geographic regions of the world. J Clin Oncol 2006 May10;24(14):2137e50.

5. Anderson BO, Yip CH, Smith RA, Shyyan R, Sener SF, Eniu A, et al. Guidelineimplementation for breast healthcare in low-income and middle-incomecountries: overview of the Breast Health Global Initiative Global Summit 2007.Cancer 2008 Oct 15;113(8 Suppl.):2221e43.

6. El Saghir NS, Adebamowo CA, Anderson BO, Carlson RW, Bird PA, Corbex M,et al. Survey of utilization of multidisciplinary management tumor boards inArab countries. The Breast; 2011. doi:10.1016/j.breast.2011.01.011.

7. Yip CH, Cazap E, Anderson BO, Bright K, CaleffiM, Cardoso F, et al. Breast cancermanagement in middle-resource countries: consensus statement from theBreast Health Global Initiative. Breast April 2011;20(Suppl. 1). TBD.

8. Shanmugaratnam K. Introduction. In: Jensen O, Parkin D, MacLennan R, Muir C,Skeet R, editors. Cancer registration: principles and methods. IARC scientificpublication No 95. London: Oxford University Press; 1991.

9. Coleman MP, Quaresma M, Berrino F, Lutz JM, De Angelis R, Capocaccia R, et al.Cancer survival in five continents: a worldwide population-based study(CONCORD). Lancet Oncol 2008 Aug;9(8):730e56.

10. IARC. Cancer incidence in five continents, vol. IX. Lyon: International Associationfor Cancer Research (IARC). Available from: http://ci5.iarc.fr/CI5i-ix/ci5i-ix.htm; 2010.

11. International Agency for Research on Cancer (IARC). Section of cancer infor-mation (CIN). 2011 [cited January 2011]; Available from: http://www.iarc.fr/en/research-groups/sec1/index.php.

12. IARC. International Agency for research on cancer (IARC). The GLOBOCANProject 2010: Available from: http://globocan.iarc.fr/.

13. ENCR. European network of cancer registries. France 2011 [cited 2011 January1]; Available from: http://www.encr.com.fr/.

14. NAACCR. North American Association of Central Cancer Registries. France 2011[cited 2011 January 1]; Available from: http://www.naaccr.org.

15. Harford J, Azavedo E, Fischietto M. Guideline implementation for breasthealthcare in low- and middle-income countries: breast healthcare programresource allocation. Cancer 2008 Oct 15;113(8 Suppl.):2282e96.

16. IOM. Summary. In: Sloan F, Gelband H, editors. Institute of medicine of thenational Academies Committee on cancer control in low- and middle-incomecountries. Washington, DC: The National Academies Press; 2007. p. 13.

17. American Cancer Society. Global cancer controle regional initiatives. Strategic PlanProgress Report 2010 [serial on the Internet] (acspc-026910): Available from:,http://www.cancer.org/acs/groups/content/@nho/documents/document/acspc-026910.pdf; 2010.

18. Passman LA, Ramalho A, Tomazelli J, Abreu DF, Dias MK, Assis M, et al. SIS-MAMA e implementation of an information system for breast cancer earlydetection programs in Brazil. Breast April 2011;20(Suppl 1). TBD.

19. Chandramohan D, Maude GH, Rodrigues LC, Hayes RJ. Verbal autopsies foradult deaths: issues in their development and validation. Int J Epidemiol 1994Apr;23(2):213e22.

20. WHO. Verbal autopsy standards: ascertaining and attributing causes of death[cited 2011 January 1]; Available from: http://www.who.int/whosis/mort/verbalautopsystandards/en/index.html; 2007.

21. Chen L, Evans T, Anand S, et al. Human resources for health: overcoming thecrisis. Lancet 2004 Nov 27eDec 3;364(9449):1984e90.

22. Anyangwe SC, Mtonga C. Inequities in the global health workforce: the greatestimpediment to health in sub-Saharan Africa. Int J Environ Res Public Health 2007Jun;4(2):93e100.

23. WHO.Migration of health workers (Fact sheet N� 301) [cited 2010 July]; Availablefrom:, http://www.who.int/mediacentre/factsheets/fs301/en/index.html; 2010.

24. WHO. WHO global code of practice on the international recruitment of healthpersonnel [cited 2010 July]; Available from: http://www.who.int/hrh/migration/code/practice/en/index.html; 2010.

25. Hongoro C, McPake B. How to bridge the gap in human resources for health.Lancet 2004 Oct 16e22;364(9443):1451e6.

26. Middle East Partnership Initiative (MEPI). U.S. middle east partnership for breastcancer awareness and research [cited 2011 January]. Available from: http://www.bcpartnership.org/partners/united-states.cfm; 2011.

27. Autier P, Boniol M, Lavecchia C, Vatten L, Gavin A, Hery C, et al. Disparities inbreast cancer mortality trends between 30 European countries: retrospectivetrend analysis of WHO mortality database. BMJ 2010;341:c3620.

28. El Saghir NS, El-Asmar N, Hajj C, Eid T, Khatib S, Bounedjar A, et al. Survey ofutilization of multidisciplinary management tumor boards in Arab countries.Breast April 2011;20(Suppl 1). TBD.

29. Winchester DP, Kaufman C, Anderson BO, El-Tamer M, Kurtzman SH, Masood S,et al. The National Accreditation Program for Breast Centers: qualityimprovement through interdisciplinary evaluation and management. Bull AmColl Surg 2008 Oct;93(10):13e7.

30. Thomas DB, Gao DL, Ray RM, Wang WW, Allison CJ, Chen FL, et al. Randomizedtrial of breast self-examination in Shanghai: final results. J Natl Cancer Inst 2002Oct 2;94(19):1445e57.

31. Duflo E, Glennerster R, Kremer M. Using randomization in developmenteconomics research: a toolkit. London: Center for Economic Policy Research2007Contract No.: 6059.

32. WHO. Cancer control: knowledge into action. Geneva [cited 2011 January 1];Available from: http://www.who.int/cancer/modules/en/index.html.

33. Eccles M, Mittman B. Welcome to implementation science. Implement Sci2006;1:1.

34. Madon T, Hofman KJ, Kupfer L, Glass RI. Public health. Implementation science.Science 2007 Dec 14;318(5857):1728e9.

35. Buekens P, Keusch G, Belizan J, Bhutta Z. Evidence-based global health. JAMA2004;291:2639e41.

36. Brown ML, Goldie SI, Draisma G, Harford J, Lipscomb J. Health serviceinterventions for cancer control in developing countries. In: Jamison D,Breman J, Measham A, et al., editors. Disease control priorities in developingcountries. 2nd ed. Washington, DC: Oxford University Press/the World Bank;2006. p. 569e89.

37. WHO. World health statistics 2010. France: World Health Organization,Department of Health Statistics and Informatics of the Information, Evidenceand Research Cluster; 2010.

38. World Bank country economic classification [cited 2010 November 1; the Atlasmethod]. Available from: http://data.worldbank.org/about/country-classifications;2010.

39. Groot MT, Baltussen R, Uyl-de Groot CA, Anderson BO, Hortobagyi GN. Costsand health effects of breast cancer interventions in epidemiologically differentregions of Africa, North America, and Asia. Breast J 2006 Jan-Feb;(12 Suppl.1):S81e90.

40. Okonkwo QL, Draisma G, der Kinderen A, Brown ML, de Koning HJ. Breastcancer screening policies in developing countries: a cost-effectiveness analysisfor India. J Natl Cancer Inst 2008 Sep 17;100(18):1290e300.

41. Torres Edejer TT, Baltussen R, Adam T, Hutubessy R, Acharya A, Evans D, et al.Making choices in health: WHO guide to cost-effectiveness analysis. Geneva,Switzerland: World Health Organization; 2003.

42. Strauss A, Cornin J. Basics of qualitative research: techniques and procedures fordeveloping grounded theory. 3rd ed. ThousandOaks, CA: SAGEPublications; 2008.

43. Fowler BA. An outsider’s experiences in conducting field research in an African-American community. J Natl Black Nurses Assoc 2002 Jul;13(1):31e7.

44. Grunfeld EA, Kohli N. Beliefs about breast cancer and help-seeking intentionsfor the disease among women in India. Women Health 2010 Jun;50(4):327e41.

45. Creswell JW. Qualitative inquiry & research design: choosing among five tradi-tions. Thousand Oaks, CA: SAGE Publications; 1998.

46. Denzin NK, Lincoln YS, editors. Handbook of qualitative research. 2nd ed. Lon-don: SAGE Publications; 2000.

47. Schwandt TA. Dictionary of qualitative inquiry. 2nd ed. Thousand Oaks, CA:SAGE Publication; 2001.

48. Puschel K, Thompson B. Mammogram screening in Chile: Using mixedmethods to implement health policy planning at the primary care level. TheBreast; 2011. doi:10.1016/j.breast.2011.02.002.

49. Berry DA, Cronin KA, Plevritis SK, et al. Effect of screening and adjuvant therapyon mortality from breast cancer. N Engl J Med 2005 Oct 27;353(17):1784e92.

50. Morgan DL. Focus groups as qualitative research. 2nd ed. London: SAGE Publi-cations; 1997.

51. Bright K, BarghashM,DonachM,Gutierrezde BarreraM, Schneider RF, Formenti SC.The role of health system factors in delaying final diagnosis and treatment of breastcancer in Mexico City, Mexico. Breast April 2011;20(Suppl. 1). TBD.

52. Freeman HP. Patient navigation: a community based strategy to reduce cancerdisparities. J Urban Health 2006 Mar;83(2):139e41.

J.B. Harford et al. / The Breast 20 (2011) S20eS29S28

53. Reeler A, Qiao Y, Dare L, Li J, Zhang AL, Saba J. Women’s cancers in developingcountries: from research to an integrated health systems approach. Asian Pac JCancer Prev 2009 Jul-Sep;10(3):519e26.

54. Sloan F, Gelband H, editors. Cancer control opportunities in low- and middle-income countries. Committee on Cancer Control in low- and middle-incomecountries. Washington, D.C: National Academies Press; 2007.

55. Koon KP, Soldak T, Gralow JR. Breast cancer advocacy: changing perceptions.Salud Publica Mex 2009;(51 Suppl. 2):s323e8.

56. Errico KM, Rowden D. Experiences of breast cancer survivor-advocates andadvocates in countries with limited resources: a shared journey in breastcancer advocacy. Breast J 2006 JaneFeb;(12 Suppl. 1):S111e6.

57. Agarwal G, Ramakant P, Forgach ER, Rendon JC, Chaparro JM, Basurto CS, et al.Breast cancer care in developing countries. World J Surg 2009 Oct;33(10):2069e76.

58. Mello MM, Brennan TA. The controversy over high-dose chemotherapy withautologous bone marrow transplant for breast cancer. Health Aff (Millwood)2001 SepeOct;20(5):101e17.

59. Aylward RB, Linkins J. Polio eradication: mobilizing and managing the humanresources. Bull World Health Organ 2005 Apr;83(4):268e73.

60. Narasimhan V, Brown H, Pablos-Mendez A, Adams O, Dussault G, Elzinga G,et al. Responding to the global human resources crisis. Lancet 2004 May 1;363(9419):1469e72.

61. Barrett S. Eradication versus control: the economics of global infectious diseasepolicies. Bull World Health Organ 2004 Sep;82(9):683e8.

62. Norton W, Mittman B. Scaling-up health promotion/disease preventionprograms in community settings: barriers, facilitators, and initial

recommendations. Report Submitted to Patrick and Catherine Weldon Dona-ghue Medical Research Foundation; 2010.

63. Knaul FM, NigendaG, LozanoR, Arreola-OrnelasH, LangerA, Frenk J. Breast cancerin Mexico: a pressing priority. Reprod Health Matters 2008 Nov;16(32):113e23.

64. Shulman L. Delivering cancer care in resource-limited settings: possibilities ofa partnership. Chicago, IL: Breast Health Global Initiative Summit; 2010.

65. IOM. Expanding the role of the global community in cancer control. In: Sloan F,Gelband H, editors. Institute of Medicine of the National Academies Committee oncancer control in low- and middle-income countries. Washington, DC: TheNational Academies Press; 2007. p. 285e304.

66. Stalsberg H, Awuah B, Ibarra JA, Nsiah-Asare A. Re-establishing a surgicalpathology service in Kumasi, Ghana: case report and discussion of barriers andkey elements of a successful collaboration between low- and high-resourcecountries. Cancer 2008 Oct 15;113(8 Suppl.):2338e46.

67. Reeler AV, Sikora K, Solomon B. Overcoming challenges of cancer treatmentprogrammes in developing countries: a sustainable breast cancer initiative inEthiopia. Clin Oncol (R Coll Radiol) 2008 Mar;20(2):191e8.

68. Yip CH, Smith RA, Anderson BO, Miller AB, Thomas DB, Ang ES, et al. Guidelineimplementation for breast healthcare in low- and middle-income countries:early detection resource allocation. Cancer 2008 Oct 15;113(8Suppl.):2244e56.

69. Yip CH, Smith RA, Anderson BO, Miller AB, Thomas DB, Ang ES, et al. Guidelineimplementation for breast healthcare in low- and middle-income countries:treatment resource allocation. Cancer 2008 Oct 15;113(8 Suppl.):2269e81.

70. Harford JB. Dowhat you can versus one-size-fits-all: Perspectives onbreast cancerearly detection in low- and middle-income countries. Lancet Oncol, in press.

J.B. Harford et al. / The Breast 20 (2011) S20eS29 S29