Optimising the use of multi-hospital electronic datasets for childhood cancer care and outcomes in Tanzania: the D2C project

Childhood cancer care in Tanzania has grown rapidly since 2010. The country has gone from having no paediatric oncologists to more than 10, and comprehensive childhood cancer care is now available at four treating hospitals. Yet survival remains far below that in high-income countries, where childhood cancer is highly curable. One barrier is data: records of children with cancer are held in multiple, disconnected hospital databases, many supported by different donors, and Tanzania has no national childhood cancer registry. This makes it difficult to know how many children have cancer, what happens to them after diagnosis, and why their outcomes differ. The D2C (Data2Care) project links existing hospital databases rather than building a new system. It brings together electronic records for children and adolescents aged 0 - 19 years diagnosed with cancer between 1 January 2024 and 31 December 2026. The records come from six treating hospitals, two peripheral hospitals and five regional population-based cancer registries in Tanzania. The databases include CleverCharts, REDCap, CanReg5, SJCare and Excel. To create a reliable harmonised dataset, we apply a five-part data quality framework: Child record checks Comparing each child’s records across databases Record linkage and removal of duplicates Completeness and accuracy of key data items Checks that dates are valid and in order along the care pathway This poster, to be presented at the Geneva Health Forum 2026, describes the study design, the data quality framework and early implementation lessons. These lessons include the lack of a common patient ID, databases that record visits rather than patients, care pathway and survival data held in different systems, and the value of a minimum common dataset and local support. It also outlines ongoing work: data analysis, a feedback loop with hospitals, online dashboards, and patient and public involvement. The harmonised dataset will help describe children accessing cancer services, follow them along the care pathway and identify where care breaks down. In the absence of a national registry, it is an important step towards one, and offers a replicable model for strengthening childhood cancer data in other low-resource settings.