[ad_1]
The epidemiological data provided by the General Directorate of Health (DGS) are of low quality, concludes a study by the Center for Research in Health Technologies and Services (Cintesis) and the Faculty of Medicine of the University of Porto. “The possible consequences are making suboptimal decisions or even not using data to support decisions,” he writes in the article, initially published on the open access platform medRxiv.
Throughout the country, scientists made themselves available to conduct research using epidemiological data collected by DGS and to be able to actively contribute to the control of the pandemic, since the analysis effort multiplied by several academic teams would produce results that DGS, with the overload of current job, can not produce.
The poor quality of Covid-19 surveillance databases limits their usefulness for making informed decisions or conducting useful research, ”the authors conclude.
A portion of the teams that requested access to the database ended up receiving the first file in April, but the weekly or monthly updates never came. A new update to the database just arrived in August and with several problems. To start, the tables were not the same, the data and variables had a different format which meant that the work already done by the researchers had to be replicated (rather than expanded) in the new database.
Later, and comparing the two files, the team of Cristina Costa Santos, a researcher at Cintesis, found that August’s database, although it had new records, had 4,075 missing records compared to April. Furthermore, in more than half of the cases present in both databases, information on “underlying conditions” (such as a previous illness) had been modified.
There are dead in the first database not listed as dead in the second, or dead in the second who will have died before the first database was sent, where they were still reported as alive. Three men and a 97-year-old woman registered as pregnant. 19 people whose diagnosis date is prior to the first registered case of Covid-19 in Portugal and two patients whose hospital stay is expressed as a negative numerical value.
In the description that accompanies the epidemiological surveillance data of Covid-19, DGS warns that it is data “of a provisional nature”, that “may still be subject to validation” and that “may not coincide with those reported by the bulletin DGS newspaper ”. . The problem, says Cristina Costa Santos’ team, is that there are already at least three scientific articles published with these data.
Pregnant men, ignorance of previous diseases and 90% identifiable deaths. DGS provided incomplete and erroneous data to scientists
The authors first criticize the complexity and slowness of the process, which forces doctors to fill in data on three different platforms, with poorly constructed forms and which favor the accumulation of errors or which allow the user to proceed without filling out a single table (achieving all data blank). Y leave suggestions:
- simplify the data entry process;
- constantly monitor databases;
- increase the perception of health professionals about the importance and have good data;
- provide adequate training for healthcare professionals.
“The availability of accurate data during an epidemic is essential to guide public health measures and policies,” the authors emphasize. The Observer tried to find out from the DGS if these same databases are used to support the decisions of the health authorities or what they are doing to correct the problem, but did not receive a response until the time of publication of this article.
