Called out on its bullshit, Minas Gerais hides Covid-19 data

</p> <p>All Brazilian states — or all countries, for that matter&nbsp;— have fewer confirmed cases than the real numbers. And while some underreporting is unavoidable, there is a dangerous trend occurring in places such as Minas Gerais. Many patients go to healthcare units presenting Covid-19-like symptoms. These cases are monitored —&nbsp;but most are never tested. In the first few weeks of the pandemic, the Health Minister&#8217;s guidelines <a href="https://brazilian.report/latin-america/2020/03/29/low-testing-casts-doubt-over-covid-19-reality-brazil/">reserved testing only for more severe cases</a>, as well as healthcare workers. Those guidelines have since changed in most states.</p> <p>Among the states which publish testing data, Minas Gerais has the second-lowest rate per million people. Researchers from the Federal University of Minas Gerais have analyzed patterns of hospital admittances for diseases that could be mistakenly diagnosed instead of Covid-19, including acute respiratory distress syndrome notifications, and have estimated that underreporting in Minas Gerais might be occurring at a rate <a href="https://ufmg.br/storage/9/f/3/e/9f3e7dadd7bc59c6d3f9339198dcbcaf_15893195534547_1310211378.pdf">four times higher than the national average</a>.</p> <p>Without testing, a healthcare system simply flies blind —&nbsp;as Nelson Teich, who resigned from the Health Ministry <a href="https://brazilian.report/coronavirus-brazil-live-blog/2020/04/29/health-minister-says-brazil-is-flying-blind-during-pandemic/">admitted</a> yesterday. For those who watched HBO&#8217;s hit show &#8220;Chernobyl,&#8221; it&nbsp; is something like the scene in which comrade Dyatlov considers the <a href="https://www.youtube.com/watch?v=Mg5HOnq7zD0">3.6-roentgen radiation</a> reading “not great, not terrible.” We later learn the real level is 15,000 roentgen or twice the radiation level of the Hiroshima atomic bomb — but the instruments at Chernobyl could only measure as high as 3.6.</p> <h2>&#8220;Myopic vision&#8221;</h2> <p>When questioned about the state’s lack of testing, State Health Secretary Carlos Eduardo Amaral claimed that it is not s a problem large enough to prevent the state from reopening its commerce. &#8220;Focusing on testing is a myopic way to look at the situation,&#8221; he <a href="https://www.em.com.br/app/noticia/gerais/2020/05/14/interna_gerais,1147263/focar-na-testagem-e-uma-visao-miope-diz-secretario-de-saude.shtml">said</a> —&nbsp;going against all World Health Organization recommendations.</p> <p>To avoid any further questions on the matter, the government — led by Romeu Zema, a member of the Novo party (which translates as the &#8220;New party&#8221;), resorted to the oldest trick in the book of untransparent rulers: it simply stopped counting suspected cases. &#8220;Following the definitions of cases put forward by the Health Ministry and the State Health Secretariat, cases that were being logged as &#8216;suspected&#8217; now go into our system as &#8216;not-specific flu syndrome&#8217; as they do not integrally meet all criteria for laboratory investigation,&#8221; declared the state government.</p> <p>Minas Gerais, however, was of the few states which published this piece of data — useful for identifying testing bottlenecks. Its importance could only be attested to in late April when the state began releasing microdata on patients who arrived at hospitals with Covid-19 symptoms.</p> <p>The database had 90,000 lines —&nbsp;one per patient&nbsp;— and brought together demographic data such as age, gender, and city of residence. However, it lacked data on symptoms and pre-existing conditions. In neighboring state Espírito Santo, for instance, 27 variables are measured.</p> <h2>Why data is crucial</h2> <p>A highly-detailed database allows scientists and analysts to use machine learning to identify patterns that are not initially apparent — this could offer precious information on the behavior of the virus. The UK&#8217;s National Health Service has compiled a detailed database of 17 million patients, allowing for a robust study on the <a href="https://opensafely.org/outputs/2020/05/covid-risk-factors/">main factors that lead to a patient&#8217;s death</a>.</p> <figure class="wp-block-embed-twitter wp-block-embed is-type-rich is-provider-twitter"><div class="wp-block-embed__wrapper"> <blockquote class="twitter-tweet" data-width="550" data-dnt="true"><p lang="en" dir="ltr">OUR NEW PAPER.<br><br>Factors associated with COVID-19-related hospital death in the linked electronic health records of 17 MILLION adult NHS patients.<br><br>Largest study of its kind ever, anywhere.<br><br>The power of UK / NHS data, realised.<a href="https://t.co/XeQj4IG8I8">https://t.co/XeQj4IG8I8</a><a href="https://t.co/SpglPKpJSk">https://t.co/SpglPKpJSk</a></p>&mdash; Ben StayAtHome Goldacre (@bengoldacre) <a href="https://twitter.com/bengoldacre/status/1258372975004389379?ref_src=twsrc%5Etfw">May 7, 2020</a></blockquote><script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script> </div></figure> <p>When Minas Gerais began publishing microdata on 90,000-plus patients, it was praised by NGO Open Knowledge Brazil — which produces a weekly state-level transparency ranking on Covid-19 data. The study has created a healthy competition among governors and helped improve data collection in nearly all Brazilian regions.</p> <p>In the next ranking, expect Minas Gerais to fall off the charts.