Menu
Menu
Menu
Menu
Background: There exist significant challenges to the receipt of comprehensive oncologic treatment for children diagnosed with cancer in sub-Saharan Africa. To better define those challenges, we investigated treatment outcomes and risk factors for treatment abandonment in a cohort of children diagnosed with cancer at the University Teaching Hospital (UTH), the site of the only pediatric oncology ward in Zambia. Methods: Using an established database, a retrospective cohort study was conducted of children aged 0-15 years admitted to the pediatric oncology ward between July 2008 and June 2010 with suspected cancer. Diagnosis, mode of diagnosis, treatment outcome, and risk factors for abandonment of treatment were abstracted from this database and clinical medical records. Results: Among 162 children treated at the UTH during the study time period that met inclusion criteria, only 8.0% completed a treatment regimen with most of the patients dying during treatment or abandoning care. In multivariable analysis, shorter distance from home to the UTH was associated with a lower risk of treatment abandonment (Adjusted Odds Ratio [aOR] = 0.48 (95% confidence interval [CI] 0.23-0.97). Conversely maternal education less than secondary school was associated with increased risk for abandonment (aOR = 1.65; 95% CI 1.05-2.58). Conclusions: Despite availability of dedicated pediatric oncology treatment, treatment completion rates are poor, due in part to the logistical challenges faced by families, low educational status, and significant distance from the hospital. Alternative treatment delivery strategies are required to bring effective pediatric oncology care to the patients in need, as their ability to come to and remain at a central tertiary care facility for treatment is limited. We suggest that the extensive system now in place in most of sub-Saharan Africa that sustains life-long antiretroviral therapy for children with human immunodeficiency virus (HIV) infection be adapted for pediatric cancer treatment to improve outcome. © 2014 Slone et al.
The study was approved by the Vanderbilt University Institutional Review Board (Nashville, TN, USA) and Ethics Reviews Converge (ERES) (Lusaka, Zambia). Requirements for consent were waived by the IRBs. The Republic of Zambia (Figure 1) is geographically the 39th largest nation in the world, roughly the size of the US state of Texas [6]. Considered a LMIC by the World Bank, 59% of its population of 13 million people live in poverty and have an average life expectancy of 49 years [7]. The UTH is a 2000-bed tertiary care institution in the capital of Lusaka and serves as the country’s principal referral hospital and site of the only medical school in the country (UNZA) at the time of this study. UTH is currently the only government-funded institution in Zambia offering oncology care to both adults and children. In the UTH’s Department of Paediatrics and Child Health, the Hematology-Oncology Unit has a 32 bed-capacity and offers chemotherapy. Radiation therapy became available with the establishment of the Cancer Disease Hospital (CDH) on the grounds of the UTH in 2006. However, there remain inadequate human resources to provide fully adequate services including a typical nurse to patient ratio of 1∶15. There is only one subspecialty trained pediatric hematologist-oncologist in Zambia, and there are limited opportunities for doctors or nurses to obtain further subspecialty training. Basic diagnostic investigations such as complete blood count, liver and kidney function tests are readily available at the UTH at no cost to the patients. The UTH Department of Paediatrics and Child Health instituted routine opt out HIV testing in 2005 for every child admitted [8]. Diagnostic tests such as histopathology are available but the laboratory faces challenges such as frequent interrupted supply of reagents, inadequate staffing, and a demand for specimen processing that exceeds capacity. At the time of this study, immunohistochemistry, cytology, and molecular diagnostics were not available and imaging modalities were limited with one magnetic resonance imaging (MRI) and two computerized tomography (CT) imagers that serve the entire country, negatively impacting initial diagnosis or follow-up strategies. Treatment with chemotherapy, radiation and surgery is offered at the UTH free of charge to citizens of Zambia. Chemotherapy protocols are derived from evidence-based protocols in the literature and are not necessarily adapted for LMIC. However, regardless of the protocol, inconsistent availability of cytotoxic drugs often dictates the regimen delivered to the patients, resulting in a lack of uniformity in treatment of specific cancers. A blood bank is available but demand exceeds available products due to limited resources to safely and efficiently distribute blood products. Many patients travel more than 500 km from their homes to receive care at UTH. Once treatment has commenced, patients and caregivers often must remain on the hospital grounds while awaiting the next cycle of chemotherapy due to lack of local housing and inability to return to home due to cost of travel. A pre-existing database had been established at UTH in partnership with the University of York (UY) to investigate the etiologic linkage between viruses and malignancies, based on prior research by the same investigators in Malawi [9], [10]. All children aged 0–15 years who were admitted to the Paediatric Oncology Ward at the UTH between July 2008 and June 2010 with suspected cancer were enrolled in the UTH-UY database with caregiver consent and were tested for HIV. The database collected medical information like laboratory, pathology and radiology results as well as demographic and family information via an extensive face-to-face interview with the child’s caregiver. We constructed our retrospective cohort from this database then performed a chart review to verify malignant diagnoses and ascertain treatment outcomes (Figure 2). Initial inclusion for the study was limited to patients registered in the database. Patients were subsequently excluded from the study if they did not have a malignant diagnosis confirmed by clinical or histopathological evaluation, were not residents of Zambia, or the clinical record could not be obtained to verify diagnosis and ascertain outcome. Clinical diagnoses were typically established based on history, physical exam, chest x-ray, ultrasound and occasionally CT scan. Based on medical record review, one outcome assignment was determined for each patient: (1) completed treatment/actively undergoing therapy; (2) refused treatment; (3) abandoned treatment; or (4) death from any cause. Each child was classified with the first outcome that was met as determined by the medical record review. Abandonment of treatment was defined as the termination of care by the parent/caregiver and/or not presenting for scheduled treatment for>four weeks from the scheduled date of treatment. If a child returned to UTH after already having met the criteria for abandonment of treatment, his/her classification remained unchanged. Refusal of treatment was defined as no initiation of treatment after the diagnosis of a malignancy [11]–[13]. In compliance with the International Society of Paediatric Oncology (SIOP) Position Statement on abandonment of treatment, we combined refusal and abandonment of treatment for data analysis [14]. Data were collected from the UTH-UY database and the clinical charts by study personnel using a paper collection form. Data were later imported into a password-protected Microsoft Access database. Audits of >30% of the charts were performed to ensure data quality. Continuous variables were expressed as means and standard deviations. Categorical variables were expressed as percentages. Chi square tests or Fisher’s exact tests were used to evaluate statistical significance of associations between categorical variables, as appropriate. Student’s t-tests were used similarly to compare means for continuous variables. Kruskal-Wallis tests were utilized for continuous outcomes with more than two groups. Univariate and multivariable logistic regression were used to determine strength of association between risk factors and outcomes. To investigate abandonment of treatment, all treated patients were compared to the patients who had abandoned or had never initiated treatment. Treated patients included those who died during active treatment, completed treatment or were currently under treatment. Statistical analyses were done using STATA™, version 11 (StataCorp LP, College Station, TX, USA).
N/A
