OBJECTIVES Estimation of the economic burden related to cancer may be performed through several strategies and helps financial planning. This study aims to review methods used to assess costs related to cancer management through Brazilian observational studies.

METHODS A systematic review was carried out until March/2019 through a structured search on MEDLINE via Pubmed and LILACS databases. Search strategy was built considering a filter for economic studies and descriptors, keywords and MeSH terms were used for each database specifically. Studies with an observational design (cohort, case-control or cross-sectional), enrolling patients with any type of cancer and conducted in Brazil were included. No language or time limits were used.

RESULTS Both databases retrieved 107 references, 18 studies were selected for full text analysis and 12 were included in the review. Lung was the most frequently studied site (n=3, 25%), followed by prostate (n=2, 16,7%). Other cancer sites were breast, skin, larynx, esophagus, cervix, head and neck, multiple myeloma and colon. The majority (n=9, 75%) were retrospective cohorts, two were cross-sectional and one design was not defined by the author. Data were retrieved from medical records in 58.3% of the studies and from secondary databases in 41.7%. Bottom-up approach was the most frequently used to estimate costs (n=9, 75%). Direct costs were the main outcome of the studies and two studies have used these data to assess cost-effectiveness.

CONCLUSIONS Costs related with cancer management in a Brazilian real-world setting are usually assessed through retrospective analysis, based on medical records and a bottom-up approach. The retrospective characteristic is an important limitation in costs estimation due to the dependence on availability of information. Data still shows that only few areas in oncology were studied, which highlights a lack of knowledge about disease burden and the need of further studies with higher methodological quality.