[ad_1]
He presented a report on the data quality issues observed during the handling of COVID-19.
“With good data on the epidemic, we can first know what the current situation is, we can measure the effectiveness of epidemic measures, and most importantly, we can predict the burden on our health care system. The data that I observed were publicly available, published by SAM and the Records Center. “
According to him, the main problem was that the data was not provided correctly and the biggest problem was that only daily data was provided.
“It means that if you didn’t download that data that day, you won’t see it the next day. Formats are constantly changing, more data sources have emerged over time, and that data differs from different sources. Discrepancies between those sources differ in number of deaths, the number of cases.
The data itself, if it is published on a daily basis, does not coincide day by day, cases occur, the evidence disappears, the evidence appears, disappears. On weekends, the data is not always updated, for example, there is a lack of hospitalization data, which has only been published since October, ”said V. Zemlys-Balevičius.
He found it strange that the municipal test data was not published at all, which means that it is not possible to say what is happening in the municipalities.
Unpublished data on outbreaks and essential patient data for epidemic management, that is, what is happening in hospitals, were released only in October and released only together, not from all hospitals. That, in fact, prevents any analysis from being done. “
The Doctor of Mathematics has pointed out that the most important problem for analysis and prediction, which became very clear during the second wave, is that there is a discrepancy in the number of cases.
“For example, when lab data is released every day, how many tests are done, how many positive cases are detected, and there have been eight days out of the last ten days when the number of cases was greater than the number of positive tests is mathematically silly. “
Consequently, he explained, cases of the past are registered and it is not known what the situation is, if it is predicted which cases have been registered or if the epidemic is accelerating.
“It is still possible to work with that data and the main job is to use various statistical methods that compensate for all these shortcomings. But when you use inaccurate data and use various statistical tricks, you lose the opportunity to see with that data the current situation that is happening today.” .
According to the scientist, then you have to wait a week or two to see if something has really changed or not.
“Even in this case, it was possible to see that the situation in Lithuania with the second wave is bad. The exponential growth started in July, continued until September, stopped a bit, and since then we have seen exponential growth again since October. The same thing happened when we looked at the hospital data and we could see exponential growth. “
According to V. Zemlio-Balevičius, the problem with exponential growth is that we see it too late and it was possible to see it with the number of patients.
“At the beginning of the month, we have 200 beds occupied by patients in hospitals, and already 600 per month. If we grow that exponentially, then we will have three times more per month, 1800. So, to handle the exponential growth, we have to make much more decisions. faster than before. They have been adopted so far. “
For the first wave, he said, austerity measures were taken and growth held back, but there was a sharp lag during the second wave.
“The data must be handled correctly to make the correct decisions,” concluded the mathematician.
It is strictly prohibited to use the information published by DELFI on other websites, in the media or elsewhere, or to distribute our material in any way without consent, and if consent has been obtained, it is necessary to indicate DELFI as the source. .
[ad_2]
