OBJECTIVE. A major reason for poor survival of childhood acute lymphoblastic leukemia in developing countries is treatment refusal or abandonment. This can be associated with parental socioeconomic status and attitudes of health care providers. Our study examined the influence of 2 socioeconomic status determinants, parental income and education, on treatment in an Indonesian academic hospital.
METHODS. Medical charts of 164 patients who received a diagnosis of acute lymphoblastic leukemia between 1997 and 2002 were abstracted retrospectively. Data on treatment results and parental financial and educational background were collected. Open interviews were conducted with parents and health care providers.
RESULTS. Of all patients, 35% refused or abandoned treatment, 23% experienced treatment-related death, 22% had progressive or relapsed leukemia, and 20% had an overall event-free survival. Treatment results differed significantly between patients with different socioeconomic status; 47% of poor and 2% of prosperous patients refused or abandoned treatment. Although poor and prosperous patients used the same protocol, the provided treatment differed. Poor patients received less individualized attention from oncologists and less structured parental education. Strong social hierarchical structures hindered communication with doctors, resulting in a lack of parental understanding of the necessity to continue treatment. Most poor patients could not afford treatment. Access to donated chemotherapy also was inadequate. Treatment refusal or abandonment frequently resulted. There was no follow-up system to detect and contact dropouts. Health care providers were not fully aware that their own attitude and communication skills were important for ensuring compliance of patients and parents.
CONCLUSIONS. Children's survival of acute lymphoblastic leukemia in developing countries could improve if problems that are associated with parental financial and educational background and medical teams' attitudes to treatment and follow-up could be addressed better.
The survival rate of childhood acute lymphoblastic leukemia (ALL), the most common cancer in children, is as high as 80% in developed countries but frequently is <35% in developing nations. This difference is believed to be attributable primarily to refusal or abandonment of treatment, both of which are almost unknown in developed countries.1–5
Pediatric oncology units in developing countries face specific problems that should be accounted for. The main problems are a consequence of poverty and low educational standards. Lack of parental education and little awareness of health issues delay the seeking of medical help. In addition, limited knowledge and subsequent late recognition by health care providers in rural areas may delay referral to specialized centers. Consequently, children have a more advanced stage of cancer at diagnosis. The number of specialized centers is limited. Facilities often are not adequate, and training of health care providers needs improvement.
Childhood ALL is treated according to protocols. During the first phase of treatment, patients have to achieve complete remission. Subsequently, patients with ALL usually are in relatively good health and receive extensive outpatient treatment during 2 years. As with other chronic diseases, such as asthma, tuberculosis, and diabetes, noncompliance with prolonged treatment schedules is a problem. This noncompliance ranges from sporadic lapses to refusal or total abandonment of treatment. It is noteworthy that adherence failure is not restricted to patients and parents but also can involve physicians.
It is not suitable or possible simply to transfer and implement chemotherapy protocols from developed countries to developing nations. Of course, the aim is to provide adequate treatment that is able to produce equivalent results. However, intensive chemotherapy requires excellent supportive care, which often is not available. Malnutrition and high prevalence of infectious diseases increase the vulnerability of these children and make them less tolerant of chemotherapy. In addition, intensive and possibly dangerous treatment schedules should be avoided as much as possible. Both patients and physicians are more likely to fail to comply with complex and intensive protocols. The latter also are more expensive and demand larger efforts from doctors and nurses. In developing countries, therefore, less intensive, less toxic, and less expensive protocols are required. Chemotherapeutic drugs and antibiotics sometimes are in short supply or not available. Poor parents often are unable to pay for the expensive, prolonged medical care and costly medicines.6–17
Refusal or abandonment of treatment can be associated with parental socioeconomic status (SES). The immense gap between the poor and the more privileged in Indonesia offers possibilities to investigate the influence of several sociocultural and economic variables on treatment outcomes. SES is known to be related to long-term disease-free survival, independent of all other prognostic factors. Two important determinants, income and education, have been identified.6,11
Since 1992, there has been a collaborative project between Dr Sardjito Hospital in Indonesia and Vrije University Medical Center in the Netherlands.18 Pediatric oncologists in both academic hospitals recognize therapy refusal or abandonment as a major problem in treatment of childhood ALL in Indonesia. We conducted this combined retrospective and exploratory descriptive study to gain insight into the influence of parental SES on treatment and its results. If problems that are associated with the financial and educational background of patients can be addressed better, then survival of children with ALL in developing countries may improve.
Indonesia has ∼216 million inhabitants, ∼37% (80 million) of whom are children who are younger than 15 years. A childhood leukemia incidence of 2.5 to 4.0 new cases per 100 000 children leads to an estimated 2000 to 3200 new childhood ALL cases each year. Our study was conducted in Dr Sardjito Hospital of Gadjah Mada University in Yogyakarta on the island of Java. The district of Yogyakarta covers an area with a population of 4 million citizens. Dr Sardjito Hospital is a tertiary care referral hospital, where ∼30 to 40 children annually receive a diagnosis ALL.19 The pediatric department recognizes the following 3 different classes for ALL treatment: third class, second class, and very important person ward. Children are assigned to a specific ward on the basis of the income of their parents. With increasing rank, conditions on the ward improve and number of children per hospital room decreases. Parents could, on request, receive part or all of the chemotherapeutic medication either with a discount or free of charge from a Dutch charity.
This was a combined retrospective and exploratory descriptive study. The inclusion criteria for all children in the retrospective study were newly diagnosed ALL (French-American-British classification L1 or L2), both standard and high-risk patients, and age between 0 and 16 years at diagnosis. Abstracts were made of the medical charts of patients who received a diagnosis between January 1997 and August 2002. It is important to emphasize that all patients who received a diagnosis of ALL were included in our analysis and that patients were not selected.
Children were treated by the Comprehensive Protocol-Acute Lymphoblastic Leukemia-92 (COM-ALL-92) protocol from 1997 until 1999, by the Wijaya Kusuma-Acute Lymphoblastic Leukemia-99 (WK-ALL-99) pilot protocol from 1999 until 2000, and by the WK-ALL-2000 protocol until August 2002.19,20 The pilot protocol was accepted in 2000 as the national protocol for Indonesia. It is endorsed by the International Society of Pediatric Oncology.
Biological features of ALL and treatment results were noted. Treatment failure was defined as refusal or abandonment of treatment, treatment-related death, and progressive or relapsed leukemia. The number of hospital appointments kept and reasons given for abandonment were noted. In Dr Sardjito Hospital, information on parental education and income is obtained routinely during each admission to the clinic. Data on parental financial and educational background also were collected.
Our classification of SES consisted of parental financial status and parental educational status. The parental financial status classified children as coming from either a poor or a more prosperous family and was based mainly on monthly income level of parents. Considering income levels during a period of 6 years, we took Indonesian inflation rates into account, using the Consumer Price Index as determined by the International Monetary Fund.21 The threshold of monthly income for poor versus prosperous families in the period from 1997 to 2002 was set at 300 000 Indonesian Rupiah in 1997, and 450 000, 650 000, 700 000, 750 000, and 800 000 Indonesian Rupiah for the consecutive years thereafter. During this period, the threshold for poor versus prosperous families was a monthly income of ∼$80 US on average. We used the assigned hospital class at diagnosis only when no other unambiguous data were available. Poor patients were assigned to second- or third-class wards and prosperous patients were assigned to the VIP ward at diagnosis. Most Indonesians have no health insurance at all.
Children were classified as coming from families with low, intermediate, or high parental education. The parent with the highest educational level determined the designated level. Families with low parental education consisted of families with parents with no education or with elementary school and junior high school education. Families who were defined as having intermediate parental education had parents with senior high school or vocational training. Families with high parental education consisted of families with a parent who was educated in an academy or a university.
We do acknowledge that there is a correlation between level of parental income and education. However, after the onset of the economic crisis in Asia in August 1997, many well-educated people became unemployed or were forced to work in fields other than those for which they were trained. Assessing parental and financial status separately, we were able to gain insight into the independent role of both determinants on treatment outcome.
On the basis of preliminary results of our retrospective investigation, we performed an exploratory descriptive study. Open interviews were held with parents, members of the parents' organization, pediatric oncologists, residents, nurses, and laboratory personnel. The respondents were interviewed intermittently both inside and outside the clinic during a period of 6 months. Data concerning the status of parental education, communication between doctors and parents, and access to donated chemotherapy were collected.
The relationship between treatment results and parental financial or educational status was evaluated by the χ2 test. The probability of event-free survival was estimated by the method of Kaplan and Meier; estimates were compared using the log-rank test. Event-free survival was measured from the date when the patient received a diagnosis of leukemia to the first treatment failure (refusal or abandonment of treatment, treatment-related death, or progressive or relapsed leukemia) or the date of last follow-up.
From January 1997 to August 2002, 164 patients received a diagnosis of ALL in Dr Sardjito Hospital and met the study's inclusion criteria. These 164 patients consisted of 94 (57%) boys and 70 (43%) girls. Ages ranged between 0 and 15 years. Similar to other studies, an incidence peak was observed between the ages of 2 and 4 years.
Of all 164 patients, 21 did not start therapy (18 refused treatment, and 3 died before treatment could begin). In total, 143 patients began therapy; 36 (25%) followed the COM-ALL-92 protocol, 26 (18%) followed the WK-ALL-99 pilot protocol, and 81 (57%) followed the WK-ALL-2000 protocol.
Refusal or abandonment of therapy was the most common cause of treatment failure (Fig 1). Of 164 patients, 57 (35%) refused or abandoned treatment. The distribution of refusal or abandonment per phase of treatment is as follows: 18 (32%) before start of treatment, 25 (44%) during induction, 8 (14%) during consolidation, 2 (4%) during reinduction, and 4 (7%) during maintenance treatment. Reasons given for refusal or abandonment were in 53 (93%) cases financial difficulties, in 1 (2%) case adverse effects, in 1 (2%) case fear of treatment and preference for traditional medicine, in 1 (2%) case a family problem, and in 1 (2%) case unknown. It is noteworthy that our investigation of reasons for refusal or abandonment was not based on interviews with the families concerned; the information was obtained from medical charts and team members.
The second most common cause of treatment failure was treatment-related death. In total, 38 (23%) patients died of a treatment-related cause: infection (25), hemorrhage (6), and other causes (7). The least common cause of treatment failure was progressive or relapsed leukemia, which occurred in 36 (22%) children: progressive disease (5) and relapse (31). In summary, we found that of all patients, 35% refused or abandoned treatment, 23% had treatment-related death, 22% had progressive or relapsed disease, and 20% had an event-free survival (Fig 1).
Of 164 patients with ALL, 120 (73%) came from a poor family, and 44 (27%) were from a more prosperous one. The most likely treatment outcome in poor patients (n = 120) was refusal or abandonment of therapy (n = 56; 47%), whereas in prosperous patients (n = 44), it was event-free survival (n = 20; 45%). Figure 2 shows the outcome of treatment in poor and prosperous patients. The differences in treatment outcome (refusal or abandonment, treatment-related death, progressive or relapsed leukemia, and event-free survival) between poor and prosperous patients were significant (P = .000). Figure 3 shows that the event-free survival estimate was significantly higher in prosperous patients than in poor patients (P = .000). Note that Fig 2 shows actual percentages, whereas the Kaplan-Meier estimates in Fig 3 show time-dependent probability estimates.
Of 164 patients with ALL, 61 (37%) had parents with a low educational level, 66 (40%) had parents with an intermediate educational level, and 31 (19%) had parents with a high educational level. No data were available for 6 (4%) children. Figure 4 illustrates treatment results in patients with low, intermediate, and high parental education. The differences in treatment outcome between patients with different parental educational levels were significant (P = .002). Figure 5 shows that the event-free survival estimates differed significantly between patients with high, intermediate, or low parental education (P = .001). It is noteworthy that 97% of patients with low parental education, 73% of patients with intermediate parental education, and 26% of patients with high parental education were poor.
To gain more insight into the reasons for the disproportionate amount of refusal or abandonment of therapy in poor and less well-educated patients, we conducted interviews with 21 health care providers and 17 parents. We found that despite that both poor and prosperous patients were treated according to the same protocol, the actual implementation of the treatment differed markedly. The pediatric oncologists faced large numbers of patients, and their time was limited. There was a disparity in time and attention provided by the oncologists for poor as opposed to prosperous patients. Only prosperous patients had the same personal oncologist during the whole course of treatment. This oncologist took personal care of prosperous patients during hospitalization, as well as during their visits to the special VIP policlinic. Most patients, however, were treated in the second- or third-class wards and the general policlinic, where they were treated by various residents. Poor patients therefore were deprived of the enhanced and individualized attention that prosperous patients and parents received from oncologists. Structured parental education about leukemia and its treatment and information about the availability of external funding was lacking for poor patients and their parents. Residents provided the parental education. However, residents lacked both experience in the field of pediatric oncology and knowledge about funds; consequently, the information frequently was not complete. Strong social hierarchical structures hindered communication between doctors and poor and less well-educated parents. As a result, many parents did not understand why it was very important for children to take their medication daily during the 2 years of treatment. Their children seemed healthy and experienced no immediate consequences when doses of drug were missed. Most parents could not afford the prolonged treatment costs (hospitalization costs, policlinic visits, transportation, loss of daily wages, bone marrow punctures, lumbar punctures, radiographs, blood transfusions, antibiotics, and chemotherapy). If needed, parents could receive part or all of their chemotherapeutic medication from charitable sources. Oncologists, however, had to inform parents about this possibility. In practice, not many parents were aware of this possibility. The procedure to request the donation was complex and not at all transparent. There were no clear rules about who could receive which medicines and at which discount. The access to donated chemotherapy consequently was inadequate. Not all health care providers believed it possible to cure poor patients with a complex disease, such as leukemia, which requires an expensive, prolonged treatment. Lack of finances and insufficient insight in the necessity to continue treatment frequently resulted in refusal or abandonment of treatment in these poor families. The hospital had no follow-up system to check whether patients attended their clinic appointments. When patients abandoned treatment, no action was taken to contact the dropouts and, consequently, the patients were lost.
Our study showed that refusal or abandonment of childhood ALL therapy was the main reason for treatment failure in Yogyakarta, Indonesia. The differences in dropout rates between poor and prosperous patients and between children from parents with low, intermediate, or high education were sizable and statistically significant. Reasons for refusal or abandonment of therapy need to be explored further in future studies. Outcome of childhood ALL treatment could be improved substantially by interventions that help to prevent refusal or abandonment of therapy. Because most dropouts occur during the diagnostic process and the first weeks of treatment, any effort to reduce their magnitude should be implemented when patients first enter the clinic. This refusal and early abandonment of treatment can be expected to have a disproportionately adverse effect on survival, in comparison with abandonment in the later part of maintenance chemotherapy.
Kaplan-Meier estimates of event-free survival differed significantly between poor and prosperous patients and significantly among patients with low, intermediate, or high parental education. In our study, income seemed a more important determinant than education.
Regarding the interviews with health care providers and parents, we realize that the number of interviewees was limited and that one has to be careful not to extrapolate from qualitative research. However, we also believe that these interviews provided us with valuable insights. The interviews showed that despite that the same protocol is being used, there is a striking disparity in the way treatment is provided to poor, as opposed to prosperous, patients. Only prosperous patients and their parents had 1 oncologist as a consistent caregiver throughout the whole treatment. Consequently, prosperous families received more time and attention from oncologists. Previous reports of successful strategies to reduce rates of therapy abandonment in children with cancer in resource-poor countries did not mention the necessity of poor families' also having individualized attention from oncologists.1,5 In developing countries, health care providers are faced with poor patients who die of relatively simple infectious diseases that require a relatively inexpensive, short therapy. Therefore, it is understandable that not all health care providers believed in the possibility of curing poor patients with a complicated disease, such as leukemia, which requires an expensive and prolonged treatment. The influence of the physician's own beliefs on the curability of leukemia in general and in the case of poor patients in particular on the treatment that actually is provided needs to be analyzed further. In underprivileged countries, doctors have limited resources and must decide, on a daily basis, how to allocate them with presumed maximum benefit. Some health care providers may believe that cancer is not curable and treatment therefore not worthwhile.22 Poor survival rates of childhood ALL in developing countries may be attributed partly to poor communication between physicians and parents. Structured and complete parental education by physicians was lacking, particularly among poor patients and their parents. Health care providers did not seem fully aware that their own attitude, their communication skills, their interest in the patient, knowledge, willingness to provide information, and their belief in the efficacy of treatment all were important and potentially are able to increase compliance of patients and parents. The access to donated chemotherapy was insufficient. In developing countries, it is important to verify whether donations reach those in need, by questioning not only the supervisors but also the recipients. Most poor families could not afford the extended and prolonged medical treatment. Lack of finances and insight into the necessity to complete the 2 years of treatment often resulted in refusal or abandonment of treatment. There was no follow-up system in place to ensure that patients with ALL attended their hospital appointments and to contact families in case of dropout.
On the basis of our retrospective and exploratory descriptive study, we recommend the following: (1) pediatric oncology centers in developing countries should provide not only prosperous but also poor patients with 1 oncologist as a consistent caregiver during the course of treatment; (2) it is crucial to provide complete parental education at diagnosis about leukemia, the necessity to complete treatment, and the availability of funds to prosperous, as well as poor, patients and parents; presenting an educational film and providing written information is recommended, because it gives complete information and saves time for the health care providers; (3) hospitals need a follow-up system for detecting and contacting promptly families who miss hospital appointments. These insights probably also are valuable for treatment of other diseases in developing countries for which compliance is known to be a problem, such as asthma, tuberculosis, and diabetes. These measures most likely will have a positive influence on patient compliance and on outcome of treatment.
The Ponte di Legno Statement principle emphasizes that all children with leukemia, also in developing countries, have the right to receive treatment.23,24 In line with this principle, we urge continuous, close monitoring of access to chemotherapy for poor patients in resource-poor nations. This well may enable more children worldwide to be cured of leukemia.
Our project is supported by grants from the Dutch Cancer Society Koningin Wilhelmina Fonds (research), Estella Foundation (donation of medicines), and Frank ter Linden Foundation (education of nurses).
We thank Purwanto for secretarial assistance.
- Accepted June 28, 2006.
- Address correspondence to Saskia Mostert, MD, Department of Pediatrics, Vrije University Medical Center, Amsterdam, Netherlands. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Cella DF, Orav EJ, Kornblith AB, et al. Socioeconomic status and cancer survival. J Clin Oncol.1991;9 :1500– 1509
- Magrath I, Shad A, Epelman S, et al. Pediatric oncology in countries with limited resources. In: Pizzo PA, Poplack DG, eds. Principles and Practice of Pediatric Oncology. Philadelphia, PA: Lippincott-Raven Publishers; 1997:1395–1419
- Lilleyman JS, Lennard L. Non-compliance with oral chemotherapy in childhood leukaemia. An overlooked and costly cause of late relapse. BMJ.1996;313 :1219– 1220
- ↵Hicsonmez G, Ozsoylu S, Yetgin S, Zamani V, Gurgey A. Poor prognosis of childhood acute lymphoblastic leukaemia. BMJ.1983;286 :1437
- Viana MB, Fernandes RAF, Oliveira BM, Murao M, de Andrade Paes C, Duarte AA. Nutritional and socio-economic status in the prognosis of childhood acute lymphoblastic leukemia. Haematologica.2001;86 :113– 120
- Rajajee S, Desikulu MV, Pushpa V. Survival of childhood acute lymphoblastic leukemia: experience Chennai. J Trop Pediatr.1999;45 :367– 370
- Viana MB, Murao M, Ramos G, et al. Malnutrition as a prognostic factor in lymphoblastic leukaemia: a multivariate analysis. Arch Dis Child.1994;71 :304– 310
- ↵Sutaryo, Sumadiono, Suhadi, et al. The Pilot Protocol of Wijaya Kusuma Acute Lymphoblastic Leukemia of Childhood 1999. Indonesian Multicentre Study. Yogyakarta, Indonesia: Gadjah Mada University Press; 1999
- ↵Sutaryo, Sumadiono, Suhadi, et al. The Protocol of Wijaya Kusuma Acute Lymphoblastic Leukemia of Childhood 2000. Indonesian Multicentre Study. Yogyakarta, Indonesia: Gadjah Mada University Press; 2000
- ↵Ramakrishnan U, Vamvakidis A. Forecasting Inflation in Indonesia. Available at: http://ssrn.com/abstract=879845. Accessed September 27, 2006
- ↵Agarwal B, Dalvi R. Treatment of childhood leukemias in underprivileged countries. In: Pui CH, ed. Treatment of Acute Leukemias: New Directions for Clinical Research. Totowa, NJ: Humana Press; 2003:321–329
- Copyright © 2006 by the American Academy of Pediatrics