Evidence From Ghana Indicates That Childhood Cancer Treatment in Sub- Saharan Africa Is Very Cost Effective: A Report From the Childhood Cancer 2030 Network Purpose No published study to date has examined total cost and cost-effectiveness of maintaining a pediatric oncology treatment center in an African setting, thus limiting childhood cancer advo- cacy and policy efforts. Methods Within the Korle Bu Teaching Hospital in Accra, Ghana, costing data were gathered for all inputs related to operating a pediatric cancer unit. Cost and volume data for relevant clinical services (eg, laboratory, pathology, medications) were obtained retrospectively or prospectively. Salaries were determined and multiplied by proportion of time dedicated toward pediatric patients with cancer. Costs associated with inpatient bed use, outpatient clinic use, administrative fees, and overhead were estimated. Costs were summed for a total annual operating cost. Cost-effectiveness was calculated based on annual patients with newly diagnosed disease, survival rates, and life expectancy. Results The Korle Bu Teaching Hospital pediatric cancer unit treats on average 170 new diagnoses annually. Total operating cost was $1.7 million/y. Personnel salaries and operating room costs were the most expensive inputs, contributing 45% and 21% of total costs. Together, medications, imaging, radiation, and pathology services accounted for 7%. The cost per disability-adjusted life-year averted was $1,034, less than the Ghanaian per capita income, and thus considered very cost effective as per WHO-CHOICE methodology. Lorna Renner Conclusion To our knowledge, this study is the first to examine institution-level costs and cost- Shivani Shah effectiveness of a childhood cancer program in an African setting, demonstrating that operating Nickhill Bhakta such a program in this setting is very cost effective. These results will inform national childhood Avram Denburg cancer strategies in Africa and other low- and middle-income country settings. Sue Horton J Glob Oncol 4. © 2018 by American Society of Clinical Oncology Licensed under the Creative Commons Attribution 4.0 License Sumit Gupta BACKGROUND childhood cancer strategies that conform to local Author affiliations and health system contexts and public resources support information (if Among children diagnosed with cancer in available.5,6 applicable) appear at the high-income countries (HICs), long-term cure end of this article. rates are now > 80%.1 Nearly 90% of the global A major barrier to creating and implementing L.R. and S.S. are co-first pediatric population resides in low- and middle- national childhood cancer strategies is a pau- authors. income countries (LMICs), where childhood city of data on the cost of delivering childhood Corresponding author: Sumit Gupta, MD, PhD, cancer survival rates vary between 10% and cancer treatment. It is commonly believed that 50%.2,3the Hospital for Sick Twinning programs involving financial LMIC health systems are unable to bear the Children, Division of and infrastructural support from HIC institutions costs of pediatric oncology services. Although Haematology/Oncology, have resulted in significant survival improve- recent data suggest this assumption is inaccu- 555 University Ave, 4 7,8 Toronto, Ontario M5G ments in individual LMIC centers. Improving rate, there is scant evidence on the financial 1X8, Canada; Twitter: @ population-based LMIC childhood cancer out- and economic costs of treating childhood cancer Zen_Sumit; e-mail: comes will instead require regional and national in LMICs. Without cost data, policymakers have sumit.gupta@sickkids.ca. 1 jgo.org JGO – Journal of Global Oncology © 2018 by American Society of Clinical Oncology Licensed under the Creative Commons Attribution 4.0 License Downloaded from ascopubs.org by 173.211.107.4 on May 3, 2018 from 173.211.107.004 Copyright © 2018 American Society of Clinical Oncology. All rights reserved. abstract original reports little context-relevant evidence to inform the cre- 2016, approximately 45% of patients were diag- ation or expansion of childhood cancer services. nosed with leukemia or lymphoma, 42% with Indeed, assumptions that childhood cancer solid tumors, and 13% with CNS tumors.10,11 treatment is expensive may prevent policymak- Treatment protocols are based on international ers from even considering pediatric oncology standards but are often modified to account for when setting national health priorities. Studies greater risk of toxicity or lack of resources. For of the cost and cost-effectiveness of such treat- example, for children with acute lymphoblastic ment in LMIC settings are therefore essential. leukemia, a modified UKALL protocol is used; A preliminary analysis of theoretical cost- induction doses of anthracycline are often omit- effectiveness thresholds suggested that childhood ted to prevent toxicity. cancer treatment may be financially feasible in The pediatric oncology program is primarily LMICs but was limited to two specific malignan- financed by the Ministry of Health. Although a cies and neglected nondrug costs.7,9 A recent National Health Insurance Authority exists in center-level study by our team in El Salvador Ghana, it does not cover all medications and ser- represented the first rigorous and comprehen- vices, meaning that families must absorb these sive LMIC childhood cancer costing analysis, to costs. For example, although common generic our knowledge, finding that a childhood cancer antibiotics are covered, chemotherapy for child- treatment unit represented a very cost-effective hood cancer is not, nor are diagnostic tests such intervention.8 Whether analogous treatment units as computed tomography scans or pathology. in sub-Saharan Africa, with more limited resources Private philanthropic sources of funding exist to and generally lower survival rates, are also cost offset out-of-pocket costs incurred by families; effective is unknown. the most prominent are World Child Cancer, an Our objective was to determine the total cost of international nongovernmental agency, and the maintaining a major pediatric cancer treatment Ghana Parents’ Association for Childhood Can- unit in a sub-Saharan setting. Using administra- cer. Local private and faith-based organizations tive and clinical data from the Korle Bu Teaching also play an important role in the day-to-day Hospital (KBTH) in Accra, Ghana, we also aimed operation and financing of the pediatric oncol- to determine the cost-effectiveness of the program. ogy program, including fundraising and provid- ing financial assistance to low-income families for transportation, meals, and medical services. METHODS Study Setting Data Collection KBTH is the largest hospital in Ghana (2,000 To collect cost data, a detailed abstraction tool beds) and the third largest in sub-Saharan was developed after compartmentalizing costs Africa, serving a catchment area of approxi- into: personnel (both medical and support), mately 19.74 million in the southern half of the room and board for patients and their families country. In addition to Komfo Anoye in Kumasi, (hoteling), outpatient clinic, shared services KBTH is one of two Ghanaian hospitals with (pharmacy, pathology, surgery, radiation, imag- the facilities to treat childhood cancer, diagno- ing, and blood bank), other services (information ses approximately 170 new cases of cancer in technology, training), and other central hospital patients younger than 14 years of age annually, services (utilities, human resources, administra- and contains 30 inpatient beds. An outpatient tive costs). The structure of the abstraction tool clinic sees an average of 77 patients per day. is available in Appendix Table A1. All costs were Forty-one full-time-equivalent medical personnel collected and included regardless of funding were involved in the care of pediatric oncology stream (ie, government v family out of pocket v patients, including two pediatric oncologists and philanthropic). 21 nurses. Patients deemed at high risk of com- plications are kept as inpatients for close moni- Information on the volume and unit cost of items toring, including patients presenting with bulky came from various sources. Medical personnel disease. Because of limited pediatric oncology costs were determined by multiplying salary fig- services in the West African region, KBTH also ures for relevant health care providers by the admits patients from neighboring countries. In self-reported proportion of their time dedicated 2 jgo.org JGO – Journal of Global Oncology Downloaded from ascopubs.org by 173.211.107.4 on May 3, 2018 from 173.211.107.004 Copyright © 2018 American Society of Clinical Oncology. All rights reserved. to pediatric oncology care. Operating room (OR) by the central administration for activities such costs associated with pediatric oncology patients as human resources, legal activities, commu- were determined by obtaining OR records for a nications, and relationships with external orga- 4-week period within the last calendar year and nizations and government. In the absence of determining the number of OR hours used by nonmedical personnel cost, we thus used the pediatric oncology patients. OR hours were cate- same ratio of cost of nonmedical to medical gorized as major versus minor on the basis of the personnel (25:75) as for the Pediatric Cancer length of time in surgery (ie, > 1 hour v ≤ 1 hour) Department at the Hospital Nacional de Niños and by the type of surgery (procedures involv- Benjamin Blum, El Salvador, which maintains ing extensive resections, thoracotomies, CNS, separate financial statistics for their pediatric or cardiopulmonary procedures were all consid- cancer unit and thus produced, to our knowl- ered major). Total hours were multiplied by 13 to edge, the first published estimates of the cost of derive an annual figure and then multiplied by running a pediatric cancer unit in an LMIC.8 For the cost of an average hour of OR time in Ghana the cost of central administration, we again used as determined by the WHO (stratified by major v data from Hospital Nacional de Niños Benjamin minor) to provide an estimate of the annual OR Blum, which, by prorating the cost of utilities and budget attributable to children with cancer.12 central administration by the pediatric cancer unit’s share of inpatient admissions, was able to The pediatric cancer unit at KBTH does not determine that such costs came to 11.8% of total maintain financial records separate from those cost of the pediatric cancer unit. Such assump- of the overall hospital. Thus, for a number of tions were necessary to include some estimate items, including diagnostic imaging, radiation, of administrative and nondirect costs and thus and blood products, four 1-week periods within avoid gross underestimates of total cost. All cost- the prior calendar year were randomly chosen ing parameters were summed to determine the and patient charts reviewed to record all of the overall annual cost associated with operating the above services delivered to pediatric oncology KBTH pediatric oncology treatment center. patients. Unit costs were obtained from the appropriate hospital department. Unit costs and average number of items ordered over each Cost-effectiveness Analysis 1-week period were also multiplied by 13 to Cost-effectiveness was calculated using the cost derive estimated annual utilization figures. per new diagnosis combined with the estimated The number and types of laboratory tests and 5-year survival, thus allowing the estimation of medications (supportive and chemotherapeu- the cost per life saved (Table 1 summarizes the tic) ordered for pediatric oncology patients were key parameters). Currently, KBTH is only able to recorded prospectively for 2 weeks; both inpa- track survival for 1 year from diagnosis; at this tients and outpatients were included. Unit costs time period, 58.6% of patients were still alive. To were obtained from appropriate hospital depart- estimate the proportion of patients alive at 5 years ments and, in the case of medications, adjusted from diagnosis, comparable literature was used. based on dosage. Unit costs for diagnostic ser- For example, a study in Chennai, India found that vices incorporated the costs of personnel (eg, 5-year survival in a lower- to middle-income set- laboratory technicians) inherent in providing the ting was 62% of 1-year survival overall for child- service. Unit costs and volumes were multiplied to hood cancers.13 We used this same proportion to determine the total laboratory and medication- convert the 1-year survival at KBTH to a 5-year associated cost of treating children with cancer survival of 35%. Given the uncertainty in this esti- over the 2 weeks and then multiplied by 26 to mate, we conducted sensitivity analyses reducing determine the annual costs. the 5-year survival to 30%. Information on the time devoted by nonmedical The cost per life saved was then converted to personnel (eg, clerical staff) to pediatric oncol- cost per disability-adjusted life-years (DALYs) ogy services was unavailable. Such services averted using Ghana’s life expectancy of 61.5 included registration of patients in the inpatient years,15 with the mean age at diagnosis of 6 and outpatient clinics, data entry into the can- years. Although length of treatment of individual cer registry, and other clerical, technical, and childhood cancers varies, we used 1 year as the administrative tasks. The unit is also supported median duration of therapy, because lymphomas, 3 jgo.org JGO – Journal of Global Oncology Downloaded from ascopubs.org by 173.211.107.4 on May 3, 2018 from 173.211.107.004 Copyright © 2018 American Society of Clinical Oncology. All rights reserved. Table 1. Variables and Sources Used in the Cost-effectiveness Model long-term chronic conditions and premature mor- Variable Value Source tality, we completed a one-way sensitivity analysis. Discount rate 0.03 (0, WHO14 The number of additional years survived after diag- 0.06) nosis was varied by allowing a 15% and a 30% Ghana life expectancy (2015: latest 61.5 World Bank15 reduction of additional years of life expected at age available), years 6 years (ie, base case was survival to normal life Mean age at diagnosis, years 6 Assumed, using HNNBB data expectancy for Ghana of 61.5 years, with variants Duration of disability (length of 1 Assumed being survival to age 53.5 and 45.5). The propor- therapy), years tionate reduction of life expectancy for cancer sur- Disability weight during therapy 0.288 Murray5 vivors was based on data for the United States,17,18 Number of new incident cases per 170 Korle Bu Hospital data10,11 given the lack of comparable data for sub-Saharan year Africa. It is important to note that this is likely a con- Proportion of patients with 1 year 0.586 Korle Bu Hospital data10,11 servative assumption, given that treatment inten- survival sity is far lower in a Ghanaian setting, and most Proportion of patients with 5-year 0.35 Assumed; authors’ best reduction in life expectancy for survivors in HICs survival estimates using 1-year is due to treatment-related effects, such as car- survival data, and ratio of diopulmonary dysfunction and second malignant 5:1-year survival similar to 17 Swaminathan13 neoplasms. Ghana GDP per capita (2016: latest 1,513.5 World Bank15 We used the WHO thresholds for cost-effectiveness.19 available), USD These thresholds suggest that interventions costing Abbreviations: GDP, gross domestic product; HNNBB, Hospital Nacional de Niños Benjamin less than per capita income per DALY averted Blum; USD, US dollars. are very cost effective and those costing less than three times per capita income per DALY retinoblastoma, and Wilms tumors constituted a averted are cost effective. In 2016, the World major portion of cancers treated locally. During Bank15 listed Ghana’s per capita gross national treatment, children suffer from diminished quality product as USD $1,513. of life; we accounted for this using the Global Bur- den of Disease disability weight of 0.288.6 RESULTS As recommended using the WHO–Choosing Inter- ventions That Are Cost-Effective (CHOICE) guide- The annual cost to operate a pediatric cancer lines, discounting was incorporated with a base unit in Accra, Ghana was estimated as $1.7 mil- case of 3%.16 Additional estimates using 0% and lion for the 2016 to 2017 year. On the basis of 6% were also calculated. To estimate the effects of admissions data at KBTH, this equates to $9,781 per pediatric patient newly diagnosed with can- Table 2. Annual Costs of Operating Pediatric Oncology Services by Major Cost cer (Table 2; Appendix Table A1). The largest Category single cost component was personnel (46.2% of Percentage of Total costs), followed by the cost of the operating the- Input Annual Cost (USD) Cost aters (22.7%; Fig 1). Chemotherapy and sup- Medical personnel 574,960 34.6 portive medication accounted for 7.9%; hoteling Nonmedical personnel 193,653 11.5 of patients (room and board) for 5.8%; and cen- Hoteling of patients 377,637 5.8 tral administration costs, including utilities, were estimated as 11.8%. Diagnosis-related costs Medical services (pathology and laboratory costs as well as imag- Pathology and laboratory cost 45,514 2.7 ing) amounted to 4.6%. The balance of costs Pharmacy 131,180 7.9 was attributed to radiation (0.9%) and blood ser- Radiation 15,129 0.9 vices (0.2%). Imaging 31,634 1.9 Using the parameters outlined in the base case Surgery (OR cost) 337,637 22.7 (Table 1), the cost per life saved was $27,946. Blood services 2,691 0.2 The cost per DALY averted in the base case was Subtotal 603,786 36.3 $1,034, less that Ghana’s per capita income Central administration and utilities 196,217 11.8 ($1,513), thus meeting WHO-CHOICE criteria Total 1,662,771 100.0 for being considered very cost-effective. These Abbreviations: OR, operating room; USD, US dollars. results were sustained after adjusting for late 4 jgo.org JGO – Journal of Global Oncology Downloaded from ascopubs.org by 173.211.107.4 on May 3, 2018 from 173.211.107.004 Copyright © 2018 American Society of Clinical Oncology. All rights reserved. DISCUSSION The total cost of the pediatric oncology cancer Medical personnel center at KBTH was $1.7 million, or $9,781 per Nonmedical personnel newly diagnosed case per year. The cost per Hoteling of patients Pathology DALY averted was $1,034, meeting the WHO- Pharmacy CHOICE criterion of very cost-effective. Radiation Imaging Relatively high shares of personnel costs and Surgery Blood services relatively low allocations to traded consumables Administration have been previously shown in other aspects of health care in sub-Saharan Africa. A compara- tive study of pathology and laboratory medicine20 Medical personnel $574,960 found that a major public teaching hospital in Nonmedical personnel $193,653 Nigeria used < 12% of the annual diagnostic Hoteling of patients $377,637 tests per bed as compared with five other public Pathology $45,514 and private hospitals across a range of countries Pharmacy $131,180 (Kenya, India, Malaysia, and the United States). Radiation $15,129 Presumptive diagnoses without confirmatory Imaging $31,634 tests allow for cost savings on pathology spec- Surgery $337,637 imens and training. However, the number and Blood services $2,691 cost of consequent erroneous treatments are Administration $196,217 unknown. Indeed, administering inappropriate treatment regimens is likely to incur significant Fig 1. Annual costs of op- cost with minimal chance of efficacy. Studies of erating a pediatric oncology effects and early mortality risk at the 0% and service. the cost-effectiveness of improving diagnostic 3% discounting levels (Table 3). When 6% dis- capabilities, both for childhood cancer and other counting was tested, operating the cancer unit conditions, are warranted. remained cost effective. Because of uncertainty regarding the 5-year Similarly, given the high usage of surgery, survival of patients with pediatric cancer treated expenditures on blood services were lower than at KBTH, a sensitivity analysis was conducted expected. Parents must pay out of pocket for using 30% reduction in life expectancy and many medical services, except when support is utilities-adjusted late-effect morbidity with 3% available from private foundations, with conse- discounting. These model parameters represent a quent service underutilization relative to coun- conservative upper bound for cost-effectiveness. tries with more comprehensive health insurance Under these conditions, when 5-year survival coverage. The use of less-intense chemotherapy was adjusted down from 35% to 30%, opera- protocols to avoid toxicities and reduce support- tion of the pediatric cancer unit remained very ive care requirements in LMICs such as Ghana cost-effective ($1,505 per DALY averted). is another major reason for relatively low chemo- therapy costs. Table 3. Cost per Disability-Adjusted Life-Years Averted, Base Case and Sensitivity Analysis Other warranted interventions may include those Discounting targeting abandonment of treatment, which Scenarios of LE* and Late-Effect Morbidity 0% 3% 6% contributes to a substantial portion of treat- 2,21,22 Base case (normal LE, no utility adjustment for $506 $1,034 $1,739 ment failures in childhood cancer in LMICs. late-effect morbidity) Abandonment of treatment can occur for many Normal LE plus utility adjustment for late-effect $535 $1,057 $1,752 reasons, including financial hardship and a morbidity lack of awareness of the disease and necessary 15% reduction in LE plus utility adjustment for $642 $1,146 $1,807 treatment. At KBTH, charitable organizations late-effect morbidity like Ghana Parents’ Association for Childhood 30% reduction in LE plus utility adjustment for $806 $1,290 $1,911 Cancer and World Child Cancer aim to decrease late-effect morbidity abandonment rates by raising financial contribu- Abbreviation: LE, life expectancy. tions to fund medical services when necessary. *Decrements in LE selected based off Yeh et al18 and Armstrong et al.17 In other settings, interventions such as social 5 jgo.org JGO – Journal of Global Oncology Downloaded from ascopubs.org by 173.211.107.4 on May 3, 2018 from 173.211.107.004 Copyright © 2018 American Society of Clinical Oncology. All rights reserved. workers and psychologists, who work alongside decision making about resource allocation toward families providing emotional and psychosocial childhood cancer among competing health pri- support, have dramatically decreased aban- orities on the part of Ghanaian policymakers, as donment rates.21 Although such interventions was recently done in Mexico.26 Of note, childhood would increase the center’s operating costs, cancer programs in LMICs have demonstrated resultant improvements in treatment adherence the ability to attract funding from alternative phil- may translate to increased cost-effectiveness. anthropic sources,8 allowing for more children to The cost-effectiveness of specific abandonment potentially access treatment without detracting interventions awaits further study. from other areas of need. Despite the above factors, our main finding was Several limitations merit note. First, we were that the delivery of pediatric cancer services unable to determine the specific costs for non- was very cost effective using WHO-CHOICE medical personnel and for central administration definitions. Even in sensitivity analyses using and were instead forced to rely on estimates from more pessimistic assumptions of 5-year survival El Salvador. Second, many costs were deter- and life expectancy, pediatric cancer treat- mined by extrapolating focal periods of data col- ment remained very cost effective. However, lection to annual figures. This may not account cost-effectiveness is distinct from affordability. for fluctuations in volume or intensity over the The annual cost of treating a patient newly diag- course of a year, adding an additional degree of nosed with cancer is more than six times the uncertainty to our final estimates. Our approach, Ghanaian per capita gross domestic product, however, balances true microcosting with feasi- even with modest expenditures on diagnostics, bility in settings with limited data resources. chemotherapy, and radiation. Finally, we did not include indirect costs borne Historically, health system priorities in many by families. Financial toxicity resulting from out- LMICs have focused on the treatment of com- of-pocket costs is significant among HIC care- municable diseases along with maternal and givers of children with cancer and LMIC adult infant mortality. This has been justifiable, given patients with cancer.27,28 These costs are not well the historically high burden of such diseases and characterized in LMIC pediatric oncology but are highly cost-effective interventions, such as vac- likely significant.29 Incorporating these would cines and HIV prevention strategies.23,24 How- thus raise the overall cost of treating childhood ever, as strides against communicable diseases cancer. Nonetheless, all the above limitations are made and countries undergo demographic are unlikely to change our finding of the KBTH transitions, disease burdens shift from commu- childhood cancer unit being very cost-effective, nicable to noncommunicable ones.3, Many LMIC a finding unchanged in even our most conser- health systems grapple with the changing health vative sensitivity analysis. Indeed, despite these care needs that accompany these epidemiologic limitations, this study nonetheless provides the shifts. Even in countries with universal health most rigorous data to date for a childhood can- insurance, the degree of coverage of noncom- cer unit in sub-Saharan Africa. municable diseases often varies. In Ghana, the National Health Insurance Authority does not We outline the total cost of maintaining a child- currently cover childhood cancer treatment, hood cancer treatment center in Ghana and leaving families with the burden of financing demonstrate that treating childhood cancer is and prone to catastrophic health expenditures. very cost-effective. Similar studies in other LMIC By contrast, progress has been made in Ghana centers of increasing complexity are warranted, in the public financing of specific adult cancers, as are cost-effectiveness analyses of specific with coverage of breast and cervical cancer treat- interventions within such centers. The results of ment. Interestingly, a Ghanaian study in 2012 this study can be used to inform policy decisions found that biennial screening clinical breast to strengthen child cancer outcomes in sub- examinations coupled with treatment were asso- Saharan African and other LMICs. ciated with a cost per DALY averted of $1,299,25 a figure slightly higher than that associated with childhood cancer treatment in this study. DOI: https://doi.org/10.1200/JGO.17.00243 This knowledge can inform context-sensitive Published online on jgo.org on May 3, 2018. 6 jgo.org JGO – Journal of Global Oncology Downloaded from ascopubs.org by 173.211.107.4 on May 3, 2018 from 173.211.107.004 Copyright © 2018 American Society of Clinical Oncology. All rights reserved. AUTHOR CONTRIBUTIONS ASCO's conflict of interest policy, please refer to www. Conception and design: Lorna Renner, Nickhill Bhakta, asco.org/rwc or ascopubs.org/jco/site/ifc. Avram Denburg, Sue Horton, Sumit Gupta Lorna Renner Administrative support: Sumit Gupta No relationship to disclose Provision of study material or patients: Lorna Renner Collection and assembly of data: Lorna Renner, Shivani Shivani Shah Shah, Sue Horton, Sumit Gupta No relationship to disclose Data analysis and interpretation: Nickhill Bhakta, Avram Denburg, Sumit Gupta Nickhill Bhakta Manuscript writing: All authors No relationship to disclose Final approval of manuscript: All authors Avram Denburg Accountable for all aspects of the work: All authors No relationship to disclose Administrative support: Nickhill Bhakta Sue Horton No relationship to disclose AUTHORS' DISCLOSURES OF Sumit Gupta POTENTIAL CONFLICTS OF INTEREST No relationship to disclose The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held ACKNOWLEDGMENT unless noted. I = Immediate Family Member, Inst = My We thank the Garron Family Cancer Centre for providing fi- Institution. Relationships may not relate to the subject nancial support for this study and local Ghanaian partners matter of this manuscript. For more information about and Weeda Zabih for their efforts. Affiliations Lorna Renner, University of Ghana School of Medicine and Dentistry, Accra, Ghana; Shivani Shah, Avram Denburg, Sue Horton, and Sumit Gupta, Hospital for Sick Children, Toronto; Sue Horton, University of Waterloo, Waterloo, Ontario, Canada; and Nickhill Bhakta, St Jude Children’s Research Hospital, Memphis, TN. Support Supported by the Garron Family Cancer Centre at the Hospital for Sick Children. REFERENCES 1. Smith MA, Seibel NL, Altekruse SF, et al: Outcomes for children and adolescents with cancer: challenges for the twenty-first century. J Clin Oncol 28:2625-2634, 2010 2. Rodriguez-Galindo C, Friedrich P, Alcasabas P, et al: Toward the cure of all children with cancer through collaborative efforts: Pediatric oncology as a global challenge. J Clin Oncol 33:3065- 3073, 2015 3. Magrath I, Steliarova-Foucher E, Epelman S, et al: Paediatric cancer in low-income and middle- income countries. Lancet Oncol 14:e104-e116, 2013 4. Barr RD, Antillón Klussmann F, Baez F, et al: Asociación de Hemato-Oncología Pediátrica de Centro América (AHOPCA): A model for sustainable development in pediatric oncology. Pediatr Blood Cancer 61:345-354, 2014 5. Murray CJL, Lopez AD (eds): The Global Burden of Disease: A Comprehensive Assessment of Mortality and Disability From Diseases, Injuries, and Risk Factors in 1990 and Projected to 2020: Summary. Harvard University Press, Cambridge, MA. 1996 6. Salomon JA, Haagsma JA, Davis A, et al: Disability weights for the Global Burden of Disease 2013 study. Lancet Glob Health 3:e712-e723, 2015 7. Bhakta N, Martiniuk AL, Gupta S, et al: The cost effectiveness of treating paediatric cancer in low-income and middle-income countries: A case-study approach using acute lymphocytic leukaemia in Brazil and Burkitt lymphoma in Malawi. Arch Dis Child 98:155-160, 2013 8. Fuentes-Alabi S, Bhakta N, Vasquez RF, et al: The cost and cost-effectiveness of childhood cancer treatment in El Salvador, Central America: A report from the Childhood Cancer 2030 Network. Cancer 124:391-397, 2018 7 jgo.org JGO – Journal of Global Oncology Downloaded from ascopubs.org by 173.211.107.4 on May 3, 2018 from 173.211.107.004 Copyright © 2018 American Society of Clinical Oncology. All rights reserved. 9. Russell HV, Panchal J, Vonville H, et al: Economic evaluation of pediatric cancer treatment: A systematic literature review. Pediatrics 131:e273-e287, 2013 10. World Child Cancer: World Child Cancer–Ghana. 2017. www.worldchildcancer.org/what-we-do/ where-we-work/ghana 11. Korle Bu Teaching Hospital: http://www.kbth.gov.gh/latest-news/2017/02/22/kbth-to-build- hostel-for-cancer-patients 12. World Health Organization: WHO Country-specific unit costs. WHO2017. http://www.who.int/ choice/country/country_specific/en/ 13. Swaminathan R, Rama R, Shanta V: Childhood cancers in Chennai, India, 1990-2001: Incidence and survival. Int J Cancer 122:2607-2611, 2008 14. Tan-Torres Edejer T, Evans D, Lowe J: Some strategies to reduce risk. in World Health Report 2002: Reducing Risks, Promoting Healthy Life. World Health Organization, Geneva, Switzerland, 2002, pp. 99-144. 15. Worldbank: World Data Indicators. 2017 16. Bertram MY, Lauer JA, De Joncheere K, et al: Cost-effectiveness thresholds: Pros and cons. Bull World Health Organ 94:925-930, 2016 17. Armstrong GT, Chen Y, Yasui Y, et al: Reduction in late mortality among 5-year survivors of childhood cancer. N Engl J Med 374:833-842, 2016 18. Yeh JM, Nekhlyudov L, Goldie SJ, et al: A model-based estimate of cumulative excess mortality in survivors of childhood cancer. Ann Intern Med 152:409-417, 2010 19. World Health Organization: Macroeconomics and Health: Investing in Health for Economic Development. World Health Organization, Geneva, Switzerland. 2001 20. Horton S, Sullivan R, Flanigan J, et al: Delivering modern, high-quality, affordable pathology and laboratory medicine to low-income and middle-income countries: A call to action. Lancet, 2018 doi: 10.1016/S0140-6736(18)30460-4 21. Arora RS, Eden T, Pizer B: The problem of treatment abandonment in children from developing countries with cancer. Pediatr Blood Cancer 49:941-946, 2007 22. Howard SC, Marinoni M, Castillo L, et al: Improving outcomes for children with cancer in low- income countries in Latin America: A report on the recent meetings of the Monza International School of Pediatric Hematology/Oncology (MISPHO)-Part I. Pediatr Blood Cancer 48:364-369, 2007 23. Nonvignon J, Atherly D, Pecenka C, et al: Cost-effectiveness of rotavirus vaccination in Ghana: Examining impacts from 2012 to 2031. Vaccine 10.1016/j.vaccine.2017.11.080 [epub ahead of print on December 6, 2017] 24. Galárraga O, Colchero MA, Wamai RG, et al: HIV prevention cost-effectiveness: A systematic review. BMC Public Health 9:S5, 2009 (suppl 1) 25. Zelle SG, Nyarko KM, Bosu WK, et al: Costs, effects and cost-effectiveness of breast cancer control in Ghana. Trop Med Int Health 17:1031-1043, 2012 26. Gupta S, Rivera-Luna R, Ribeiro RC, et al: Pediatric oncology as the next global child health priority: The need for national childhood cancer strategies in low- and middle-income countries. PLoS Med 11:e1001656, 2014 27. Tsimicalis A, Stevens B, Ungar WJ, et al: A prospective study to determine the costs incurred by families of children newly diagnosed with cancer in Ontario. Psychooncology 21:1113-1123, 2012 28. Knaul F, Horton S, Yerramilli P, et al: Financing cancer care in low-resource settings, in Gelband H, Jha P and Sankaranaryanan R, (eds): Disease Control Priorities. World Bank, Washington, DC, 2015 pp. 281-294. 29. Mostert S, Arora RS, Arreola M, et al: Abandonment of treatment for childhood cancer: Position statement of a SIOP PODC Working Group. Lancet Oncol 12:719-720, 2011 8 jgo.org JGO – Journal of Global Oncology Downloaded from ascopubs.org by 173.211.107.4 on May 3, 2018 from 173.211.107.004 Copyright © 2018 American Society of Clinical Oncology. All rights reserved. Appendix Table A1. Details of Annual Costs Input Quantity, if Relevant Unit Cost (USD) if Relevant Total Annual Cost (USD) Medical personnel* 40.7 FTE Varies by qualifications 574,960 Nonmedical personnel† n/a Varies by qualifications 191,653 Hoteling (food and bed costs, 30 patients/d $5.80 (bed plus food, local charge), 96,096 patients) Hoteling (sanitation and $5.80 per hospital admission 3,828 documentation fee) Outpatient clinic‡ 77 patients/d n/a — Pathology (anatomic)§ 15,009 Laboratory tests|| 30,504 Pharmacy|| 131,180 Radiation§ 15,129 Imaging§ 31,634 Surgery (use of OR)§ 377,637 Blood services‖ 2,691 Subtotal, pediatric service costs 1,443,534 Central hospital costs (utilities, 196,217 central administration)¶ Total 1,662,711 Abbreviations: FTE, full-time equivalent; n/a, not applicable; OR, operating room; USD, US dollars. *Two pediatric oncologists, two general pediatricians (fellows), 21 nurses, four house officers; five junior residents; 2.6 pharmacists; one FTE general surgeon; one FTE orthopedic surgeon; 2.1 FTE neurosurgeons. †Support staff (clerical, administrative, working with patients): no data. Assume same cost ratio of support staff to medical as in El Salvador (ie, 33.3%). ‡Outpatient diagnostic tests and treatment and medical staff costs are combined with those for inpatients. No data were available on space costs for outpatient clinic. §On the basis of number of tests run and charge per test for 4 randomly chosen weeks in past year. ‖On the basis of number of tests run and charge per test for 2 weeks prospectively chosen. ¶No data available: assumed similar ratio to that of Hospital Nacional de Niños Benjamin Blum, El Salvador (ie, an additional cost of 13.38% of subtotal for pediatric services; ie, 11.8% of total cost of pediatric service). 9 jgo.org JGO – Journal of Global Oncology Downloaded from ascopubs.org by 173.211.107.4 on May 3, 2018 from 173.211.107.004 Copyright © 2018 American Society of Clinical Oncology. All rights reserved.