Portal de Periódicos da CAPES

Standardization Process Of Raw Datasus And Consumption Analysis Of Oncology Therapies In The Brazil Public Health Care System: A Comparison Between Raw And Standardized Dataset In Colorectal And Lung Cancer

2015; Elsevier BV; Volume: 18; Issue: 7 Linguagem: Inglês

10.1016/j.jval.2015.09.107

1524-4733

AF Ballalai Ferraz, RP Rosim, P Anaya,

Public Health in Brazil

To compare results of oncology drugs consumption extracted from DataSUS raw database with those obtained after standardizing and cleaning database fields. We used the SUS database available from DataSUS FTP and standardized the oncology treatment fields available in the specific oncology database (DataSUS AQ). Standardization included harmonization of different names used for the same drug name (i.e., cetuximab, cetux, cetukimabe and ketuxim), including generic and brand names. We also converted acronyms use in NCCN and MOC Brazil guidelines to the generic name (i.e.: fluoracil and 5FU; irinotecan and CPT-11). We created a new standardized table with additional fields (regimen name, drugs used, adjuvant therapy and a high/low cost flag). For this analysis we filtered by APAC (High complexity procedures approval) code for colorectal cancer (CRC) and lung cancer (LC) from 2012 to 2014. All blank or not identified regimens were excluded from this analysis. The final sample was composed by 50,729 CRC and 23,525 LC records. It was compared the total number of regimens available at raw data and those standardized. Regarding CRC regimens we found 7,698 different treatments in the raw database and 82 in the standardized dataset, a considerable reduction. In raw data, the most frequent regimen was FOLFOX representing 7.9% of all records, in contrast to standardized dataset where FOLFOX regimen represented 33.6% of all records. We found 262 records written in different ways in the raw database that referred to FOLFOX. Analyzing LC records, the most frequent regimen in raw data was carboplatin+taxo representing 3.1% of LC APACs claims in comparison to the standardized dataset where carboplatin+taxol represented 27.6% of all claims; 278 raw records had different names referring to carboplatin+taxol. DataSUS can be a reliable source on oncology consumption therapies after standardizing data fields that were originally introduced by manual typing.