PC2110: Geographic and Socioeconomic Analysis of Pediatric Cancer Cases in Tanzania

Background: The Tanzanian National Childhood Cancer Network (NCCN) includes eleven hospitals each with individual clinical databases. To determine the geographic variation of presenting patients and potential regional socioeconomic influences, it is important to merge this data into a single National registry. This study represents a comprehensive analysis of regional geographic presentation to pediatric cancer care from participating sites as well as corresponding socioeconomic factors.

Methods: Demographic data, diagnosis, and home region was extracted for all patients presenting to the Tanzanian NCCN in 2019. Regional socioeconomic data was extracted from published sources. Geographic distribution of pediatric cancer was determined and cancer cases per region were normalized to each region’s population of children. Cases were then further normalized on a scale of 0 – 1 for comparison across diagnoses. The scaled data was used to create heat maps of patient home regions for each cancer diagnosis. Normalized cancer cases were then graphed against regional socioeconomic data. Coefficients of determination (R2) were calculated to assess trends and correlations between cancer diagnoses and regional data.

Results: In 2019, a total of 826 patients 0 – 18 Yr presented to the NCCN. 58.7% were male and mean age was 6.45 (std dev 4.45) Yr. The most common diagnoses were Wilms Tumor (n=134), retinoblastoma (n=133), and ALL (n=125). The regions patients most commonly presented from were Dar es Salaam (n=107), Mwanza (n=73), and Mara (n=44), coinciding with the location of the two largest comprehensive childhood cancer treatment hospitals. Regional variations in incidence per 100,000 children were noted, with hotspots for CNS tumor diagnosis in Arusha (0.34, average 0.06), retinoblastoma in Geita (1.0, average 0.42), and Burkitt Lymphoma in Mara (1.15, average 0.20) (Figure 1). Retinoblastoma cases were highest overall in the regions bordering Lake Victoria. The increased rate of Burkitt Lymphoma in Mara has historically been cited due to endemic malaria. Regions with the highest overall rates of cancer positively correlated to the National census reported literacy rate, mobile phone ownership, cement flooring, and percent unemployment (Figure 2).

Conclusion: There is a regional variation among pediatric cancer diagnoses in Tanzania. Regional cancer incidence positively correlated with urban designation, likely due to increased diagnostic testing in larger cities. However, an inverse correlation with incidence was seen with regional percent employment, which will need to be evaluated on an individual basis to confirm this finding. Additionally, the specific cancers that showed these correlations were largely hematologic diagnoses. This study further delineated the regional disparities in pediatric cancer presentation in Tanzania, highlighting the need for research into potential environmental and socioeconomic influences on the observed distribution, and the need for increased knowledge and diagnostic capacity in regions further away from major treatment centers.