Comparison of Livingstone's Estimates of Prevalence of Six Different Rare Blood Cancers with DARWIN EU

OBJECTIVES: Livingstone is an analytical platform that uses real-world data to generate reproducible epidemiology reports. Using Livingstone, we estimated the prevalence of six blood cancers and compared them to the Data Analysis and Real World Interrogation Network’s (DARWIN EU) results for the UK.

METHODS: Livingstone and DARWIN EU used UK data from the Clinical Practice Research Database (CPRD) to describe the lifelong prevalence of acute lymphocytic leukaemia (ALL), acute myeloid leukaemia (AML), chronic lymphocytic leukaemia (CLL), diffuse Large B-Cell Lymphoma (DLBCL), follicular lymphoma (FL) and multiple myeloma (MM). Estimates from DARWIN EU were based on CPRD GOLD whereas Livingstone provided estimates from a) CPRD GOLD/Aurum and b) CPRD GOLD/Aurum and linked Hospital Episode Statistics (HES). DARWIN’s estimates for lifelong point-prevalence (2020) were compared against CPRD GOLD/Aurum (2020) and CPRD GOLD/Aurum/HES (2019). DARWIN’s case-definition excluded cases diagnosed before 2010; this restriction did not apply to Livingstone. A Z-test was conducted to compare the results.

RESULTS: Primary care prevalence estimates per 10,000 persons were significantly lower in DARWIN versus Livingstone (all p-value <0.001): ALL: 1.04 (95% CI: 0.93-1.16) vs. 2.13 (2.06-2.20); AML: 1.31 (1.18-1.44) vs. 1.80 (1.74-1.87); CLL: 6.19 (5.91-6.48) vs. 7.16 (7.03-7.29); DLBCL: 1.36 (1.23-1.49) vs. 1.71 (1.65-1.77); FL: 2.61 (2.43-2.79) vs. 3.08 (3.00-3.17); MM: 3.99 (3.77-4.22) vs. 4.44 (4.34-4.54). When HES data was included, Livingstone’s 2019 estimates were: ALL: 3.06 (2.97-3.15); AML: 2.65 (2.57-2.74); CLL: 8.11 (7.96-8.26); DLBCL: 5.13 (5.01-5.25); FL: 4.66 (4.55-4.78); MM: 5.6 (5.48-5.73).

CONCLUSIONS: For each cancer, there was a significant difference between the prevalence estimates. Increased rates in the primary-care comparison could be due to population differences between Aurum and GOLD and differences in case definition. The difference when HES data was used emphasizes the importance of using secondary care data sources when estimating prevalence of conditions likely to be managed in the hospital-setting.