Clinical epidemiology of childhood cancer in Central America and Caribbean countries
M. G. Valsecchi1,
G. Tognoni2,
M. Bonilla3,
N. Moreno3,
F. Baez4,
C. Pacheco4,
A. P. Hernandez5,
F. Antillon-Klussmann6,
S. Machin7,
R. Cabanas8,
M. Navarrete9,
R. Nieves10,
P. De Lorenzo1 and
G. Masera11,*
1 Department of Clinical Medicine, Prevention and Biotechnologies, Section of Medical Statistics, University of Milano-Bicocca, Monza; 2 Consorzio Mario Negri Sud, Santa Maria Imbaro, Italy; 3 Departamento Hematologia y Oncologia Pediatrica de Niños Benjamin Bloom, San Salvador, El Salvador; 4 Pediatric Hemato-Oncology Department Manuel de Jesus Rivera, La Mascota Hospital, Managua, Nicaragua; 5 Servicio de Hemato/Oncologia Pediatrica, Hospital Escuela, Tegucigalpa, Honduras; 6 Unidad National de Oncologia Pediatrica, Guatemala City, Guatemala; 7 Instituto de Hematologia y Immunologia, La Habana; 8 Hospital Pediatrico Juan M. Marquez, La Habana, Cuba; 9 Servicio de Hematologia, Hospital National de Niños, San José, Costa Rica; 10 Servicio de Hematologia, Hospital Infantil Dr Robert Reid Cabral, Santo Domingo, Dominican Republic; 11 Pediatric Clinic, University of Milano-Bicocca, Monza, Italy
Received 28 August 2003; revised 17 December 2003; accepted 2 January 2004
 |
ABSTRACT
|
---|
Background:
Countries with scarce resources have the right to appropriate essential health care but very few reports discuss how this can be achieved. We assessed the survival of a large cohort of pediatric oncological patients to provide hard data on which to base realistic evaluation and planning schemes.
Patients and methods:
This multicenter retrospective survey covered consecutively diagnosed and treated patients admitted to eight national level hospitals in seven countries in Central America and the Caribbean. The research protocol was discussed extensively, so the data to be collected and the criteria for their evaluation were clearly pre-defined. We analysed 2214 patients diagnosed between 1996 and 1999 with various cancers, classified as hemato-oncological disorders (70%) and solid tumors (30%).
Results:
Three-year overall survival was 48.4% [standard error (SE) 1.3]. Detailed analysis of acute lymphoblastic leukemia highlighted the wide intercountry variability: 3-year survival was 62.2% (SE 5.3) in Cuba, 74.2% (SE 3.3) in Costa Rica, 61.7% (SE 4.9) in Nicaragua, and lower in the other four countries.
Conclusions:
The yield of diagnostic-therapeutic protocols depends largely on the context of care in which they are applied. This paper documents the importance of including epidemiological research in interventions for cooperation in complex health areas such as pediatric oncology.
Key words:
clinical epidemiology, countries with limited resources, pediatric oncology, survival analysis
 |
Introduction
|
---|
The overall incidence of childhood cancer in low-income countries and developed countries is reported to be broadly superimposable [1]. Very little is known, however, about the clinical epidemiology and even less about the outcome of pediatric cancer patients, who certainly are not given high public health priority in contexts of care where even essential needs are hardly met. Pediatric cancer patients are therefore at a high risk of severe under-treatment because of lack of resources and organizational facilities [25]. Within the framework of an international collaborative project to favor the transferability of diagnostic and therapeutic protocols, through twinning programs between centers of excellence in Italy and centers in Central and Latin America [6], a proposal for epidemiological assessment of the impact of the quality of care on outcomes was adopted.
The purpose of this report is two-fold: (i) to document the degree of feasibility and the yield of a multi-country survey of childhood cancer care; (ii) to provide an initial picture of the extent, the causes and the avoidability of the expected gap between the results achievable with current standards in economically developed and lower-income countries.
 |
Patients and methods
|
---|
A network of centers working in the field of (hemato)-oncology in various countries of Central and Latin America was established in 1996 under the auspices of the Monzas International School of Pediatric HematologyOncology (MISPHO) [7]. At their annual training course in 1999, the participants in the network activated a protocol for retrospective epidemiological evaluation of all incident cases diagnosed with cancer in the period 19961999. An ad hoc form for data collection was designed and tested for feasibility, to provide a register of basic demographic and clinical data: date of birth, sex, date and type of diagnosis, date of last follow-up and vital status, notification and reasons for loss to follow-up. The project was restricted to the area of Central America and the Caribbean, specifically Costa Rica, Cuba, El Salvador, Guatemala, Honduras, Nicaragua and the Dominican Republic, where the hospital(s) that were active members of MISPHO could be considered to represent the whole of the resources available for childhood cancers in their country. Costa Rica only provided data on selected hematological disorders (leukemias, aplastic anemias, myelodysplasias).
The data forms completed by clinicians were submitted to the coordinating and data center in Monza and checked for completeness and consistency; specific queries were verified directly with the individual investigators.
At the MISPHO course in October 2000, the working group reviewed a preliminary report, with specific attention to the diagnostic classification of the 2257 patients who had been entered in the register and grouped under seven different types of hemato-oncological disorders and 23 solid tumors. Main causes of death were classified, whenever possible, in the broad categories of progression, infection, toxicity or other (which could be detailed). Forty-three of the 2257 patients were excluded because of missing or inconsistent data on age at diagnosis (four cases), on dates (35 cases), or because the reported date of diagnosis was not in 19961999 (four cases). The present report covers a total of 2214 children with cancer, followed up to 30 June 2000.
For the purpose of the analysis, the following definitions were adopted: patients who abandoned treatment were identified as drop-outs; patients who lost contact with the clinical center or who were transferred to another treating institution were classified as lost to follow-up. The distinction is important for the use of the KaplanMeier estimator of survival, where lost to follow-up can be dealt with as censored data, as long as the underlying process is non-informative on the death process itself [8]. This estimator was used as follows: for all patients, the date of first observation was the date of diagnosis; for patients who died of any cause, time to death was calculated; patients who abandoned treatment were considered treatment failures at the date of last contact; for patients lost to follow-up, time was censored at the last contact. In the survival curves drop-outs and reported deaths are accounted for as failures. A KaplanMeier analysis, which would consider drop-outs as censored observations, is expected to overestimate survival. Standard error (SE) of the probability of survival was estimated according to Greenwood [8].
 |
Results
|
---|
Study population
The baseline demographic characteristics and the country distribution of the population are summarized in Table 1. The types of diagnosis grouped under the broad definitions of hemato-oncological disorders and solid tumors are listed in Table 2. It is worth recalling that these figures, which differ widely between countries, do not refer to any population denominator, as they are all consecutive subjects seen in the participating hospitals during the study period. Comparison is only legitimate insofar as the focus is on the quality and the outcome of the care available to children with cancer in each country.
View this table:
[in this window]
[in a new window]
|
Table 1. Presenting features at diagnosis, by country and type of tumor, for the 2214 evaluable cases diagnosed with cancer in the period 19961999 in Central America and the Caribbean
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Detailed diagnosis by type of tumor for the 2214 evaluable cases diagnosed with cancer in the period 19961999 in Central America and the Caribbean
|
|
One of the critical components of such care is illustrated in Table 3, which documents the importance and the inter-country variability of the drop-outs. It is reasonable to consider these as failures, because they are most likely to have died of disease progression, due to lack of care. The majority of the 477 drop-outs occurred in the first 3 months after diagnosis (67%) and the remainder, except for 40, within 1 year from diagnosis. The clinical centers were able to retrace the actual date of death only for a few of the drop-outs (65 of 477). A total of 53 patients were lost to follow-up after the end of elective therapy (21 patients) or because they were transferred to another hospital (32 patients).
View this table:
[in this window]
[in a new window]
|
Table 3. Drop-outs and reported deaths (including 65 deaths reported after drop-out, 39 in hemato-oncological disorders and 26 in solid tumors) by country and type of tumor in the 2214 evaluable cases diagnosed with cancer in the period 19961999 in Central America and the Caribbean
|
|
Outcome
Figure 1 summarizes the overall outcome profile in terms of survival curves (death + drop-outs are failures) of children with a hemato-oncological diagnosis or solid tumor. Besides providing a quantitative and qualitative picture of the discouraging survival profile of the population, the curves show the wide intercountry variability.

View larger version (18K):
[in this window]
[in a new window]
|
Figure 1. (A) Estimates of survival for 1556 patients with hemato-oncological disorders by country. Three-year estimates (SE, standard error) and percentage of drop-outs (% d.o.) are reported. (B) Estimates of survival for 658 patients with solid tumors by country. Two-year estimates (SE) and percentage of drop-outs (% d.o.) are reported.
|
|
Figures 2 and 3 concentrate on acute lymphoblastic leukemia (ALL), which is by far the most important diagnostic group and the one which provides the best indication of the potential transferability of results of otherwise effective diagnostictherapeutic protocols. Figure 2 is specifically informative, as it compares the actual results with what the survival curve would look like if the drop-outs were considered avoidable: 3-year survival is, respectively, 54.2% (SE 1.9) and 65.6% (SE 2.0) when drop-outs (16% of ALL cases) are considered failures or are censored. The data for Cuba and Costa Rica, where drop-outs are minimal thanks to the quality of the health systems (0.9% and 2.7%, respectively), provide support for this way of presenting outcome (survival 62.2%, SE 5.3 and 74.2%, SE 3.3) with Nicaragua occupying a similar position, presumably because of the recent progress of a national program of pediatric hemato-oncology (drop-outs 7.5%; survival 61.7%, SE 4.9). In countries where registration of patients was possibly more careful (e.g. Guatemala), very early drop-outs (i.e. in the first month) amount to as much as 69% of all drop-outs and mark the first step downward of the survival estimate in hemato-oncological disorders.

View larger version (17K):
[in this window]
[in a new window]
|
Figure 2. Two different estimates of survival for 927 patients with acute lymphoblastic leukemia. The KaplanMeier (KM) estimate is calculated differently with regard to drop-outs. Three-year estimates (standard error) are reported.
|
|

View larger version (13K):
[in this window]
[in a new window]
|
Figure 3. (A) Estimates of survival for 927 patients with acute lymphoblastic leukemia (ALL), by sex. Three-year estimates (standard error) are reported. (B) Estimates of survival for 927 patients with ALL, by age at diagnosis. Three-year estimates (standard error) are reported.
|
|
Figure 3 confirms the overall qualitative comparability of this cohort with those of the literature, showing broadly overlapping profiles of the sex-stratified survival curves and their age dependent variability, with the worst outcomes for children younger than one year or, at the other end, at least 10 years old at diagnosis.
 |
Discussion
|
---|
With respect to the objectives set for this investigation, the extent and the quality of the data contributed by the participating centers are satisfactory. The database provides useful and original insights both on the present state and on the challenges of childhood cancer care in countries with limited economic resources. It is specifically worth underlining that the careful documentation of how many children abandon treatment cannot be regarded as a methodological weakness, provided it is properly handled in the statistical evaluation. It represents indeed important information on what happens in countries where provision and accessibility to care is a structural problem. Two complementary messages summarize the main findings: (i) compared to those available for many years in most countries in the North, the overall survival curves confirm and quantify the gap that exists in terms of rights to life for children with cancer who live in less favored countries; (ii) the differences in survival even between countries in the same geographical and cultural area link the disadvantage not so much to macroeconomic conditions (see Cuba and Nicaragua) as to the existence of an attitude of responsibility for providing care to those in need: early drop-outs account for the largest proportion of avoidable unfavorable outcomes. A long tradition of continuity of care with the concept of health as a basic right produces a survival profile in Cuba overlapping the one in Costa Rica, even in a sector where treatment protocols are expensive, but are considered essential for the life of the small minority constituted by children with ALL. The promising situation in Nicaragua is in part due to the existence of a cooperative project which assures drug availability, but much more to the fact that a highly participatory strategy has been adopted, with parents involved in self-supported activities to ensure the logistics needed for compliance in the various phases of treatment [6].
The limitations of the study are obvious: its retrospective nature exposes it to a selection bias of more favorable cases and to differences in the case registration process in different countries; moreover, only centers that participate in an international program contributed cases. The strengths however are no less important. Besides the unique size, at least in the field of childhood cancer clinical epidemiology, and not only in developing countries, the importance of the cohorts is clear from the completeness of the essential data, which were collected according to pre-specified and carefully applied and controlled criteria. This enabled us to qualify the factors accounting for mortality.
Possibly one of the best long-term results of this project is the fact that the clinicians who participated in it by revising and up-dating their own diagnostic and therapeutic protocols and practices have acted as epidemiological investigators: from discussion of the protocol, through data collection phase, to assessment of the analyses provided by the coordinating data center. The epidemiology of care with its focus on the difficulties, the variability and the inconsistencies inevitable when formal research protocols must be put into practice, has become a powerful instrument of collective awareness and of methodological training. A prospective epidemiological registry of failures has been set up to investigate how to avoid them, and the idea of comparing and discussing the implications and the lessons of variability in data has been adopted into the practice of an otherwise highly heterogeneous group.
The epidemiology of inequality has become a recurrent theme in the literature [3, 9, 10]. There is no doubt that the frequent emphasis on the clinical importance and the ethical intolerability of the fact that health has become an area where fundamental rights are violated is contributing to a better and wider awareness of the magnitude of the problem. The present report, in a very restricted but highly symbolic area of health, provides another and complementary view: the field experience shows that professionals can take direct responsibility, not just describing but actually tackling problems, with the aim of modifying their outcome.
 |
Acknowledgements
|
---|
We thank the following authors for their contributions: J. Carrillo (Servicio de Hematologia, Hospital National de Ninos, San José, Costa Rica); P. Valverde (Unidad National de Oncologia Pediatrica, Guatemala City, Guatemala); L. Fu Carrasco (Servicio de Hemato/Oncologia Pediatrica, Hospital Escuela, Tegucigalpa, Honduras); A. Sala (Pediatric Clinic, San Gerardo Hospital, Monza, Italy). We thank the following organizations for their support: M. Tettamanti Foundation, Comitato M. L. Verga, Cassa di Risparmio delle Provincie Lombarde Foundation, Associazione Italiana Ricerca Cancro.
 |
FOOTNOTES
|
---|
* Correspondence to: Prof. G. Masera, Pediatric Clinic, University of Milano-Bicocca, San Gerardo Hospital, Via Donizetti, 106, 20052 Monza, Italy. Tel: +39-039-2333513; Fax: +39-039-2301646; E-mail: giuseppe.masera@pediatriamonza.it 
 |
REFERENCES
|
---|
1. Parkin DN, Stiller CA, Draper GJ, Bieber CA. The international incidence of childhood cancer. Int J Cancer 1988; 42: 511.[ISI][Medline]
2. Masera G, Biondi A. Research in low-income countries. Ann Oncol 1999; 10: 137138.[CrossRef][ISI][Medline]
3. Leon DA, Walt G, Gilson L. International perspectives on health inequalities and policy. BMJ 2001; 322: 591594.[Free Full Text]
4. Scopinaro MJ, Casak SJ. Paediatric oncology in Argentina: medical and ethical issues. Lancet Oncol 2002; 3: 111117.[CrossRef][ISI][Medline]
5. Greenberg PL, Gordeuk V, Issaragrisil S et al. Major hematologic diseases in the developing worldnew aspects of diagnosis and management of Thalassemia, malarial anemia and acute leukemia. Hematology 2001; 1: 479498.[CrossRef]
6. Masera G, Baez F, Biondi A et al. NorthSouth twinning in paediatric hemato-oncology: the La Mascota programme, Nicaragua. Lancet 1998; 352: 19231926.[CrossRef][ISI][Medline]
7. Masera G, Lanfranco P, Sala A. The Monzas International School of Paediatric Hematology-Oncology (MISPHO) for countries with limited resources: a three years experience. Med Pediatr Oncol 1998; 31: 219.
8. Marubini E, Valsecchi MG. Analysing Survival Data from Clinical Trials and Observational Studies. Chichester: John Wiley & Sons 1995.
9. Marmot M. The influence of income on health: views of an epidemiologist. Does money really matter? Or is it a marker for something else? Health Aff 2002; 21: 3146.[Abstract/Free Full Text]
10. WHO. Macroeconomics and Health: Investing in Health for Economic Development. Geneva: WHO 2001.