Machine Learning With The Largest COVID-19 Dataset From The US Is Able To Predict Patient Severity

JAKARTA - After a year of collecting data on COVID-19 cases, the central repository of health records for this virus attack has finally shown results and is being shared in a journal published today, Wednesday, July 14.

The repository is the largest COVID-19 record the United States has to date. This data storage center was built by a group of researchers and data experts to better understand the phenomenon of the COVID-19 pandemic.

The study, published in the journal JAMA Network Open, calculated the risk factors for some of the worst cases of the coronavirus. In addition, it also tracks the development of the virus over time.

To help understand the dataset at hand, researchers designed machine learning to predict the severity of COVID symptoms from patients based on information obtained from day one.

It uses a centralized database, called the National COVID Cohort Collaborative Data Enclave, or N3C. That is, researchers were able to include health records from thousands of patients across America and design analyzes from the collected data.

This study used data from 34 health centers, including more than 1 million adults -174,568 of whom were tested positive for COVID-19 and 1,133,848 were negative. The study also included health records from January 2020 to December 2020.

[see-also]

- https://voi.id/technology/66831/branson-ke-luar-angkasa-pacu-jeff-bezos-dan-elon-musk-segera-mengikutinya

- https://voi.id/technology/66818/hasil-tur-keluar-angkasa-branson-ingin-bangun-hotel-di-bulan

- https://voi.id/technology/66830/tantang-tiktok-youtube-shorts-finally-dipercepat-ke-100-countries

- https://voi.id/technology/66798/hat-jadi-ceo-elon-musk-dituduh-menekan-para-pemegang-saham

- https://voi.id/technology/66780/use-ota-perbaikan-untung-mobil-tak-perlu-dalam-ke-bengkel

[/see_also]

The results of the analysis show how the treatment of COVID-19 continues to change throughout 2020. This change is related to new treatment trials and a deeper understanding of the patient's condition and symptoms of corona virus patients.

It was concluded that patients treated with the anti-malarial drug hydroxychloroquine, promoted by former United States President Donald Trump, fell to 0 percent at the end of May 2020. The use of the steroid drug dexamethasone had increased sharply in June, after a study demonstrated its efficacy in increasing levels of life sustainability.

Not only that, the researchers also confirmed that the survival rate of COVID-19 patients continued to increase throughout 2020. In March and April, 16 percent of the total hospitalized patients died. Meanwhile, in October, the number decreased to only 9 percent.

People who come to the hospital with conditions of higher heart rate, respiratory rate, and temperature are preferred for further treatment such as using ventilation. Patients with these conditions are also the most likely to die during treatment.

Meanwhile, the number of cells, blood acidity, inflammation, and less normal liver function are also directly related to more severe cases. Based on this data, the research team designed a machine learning model that was able to predict which patients would show more severe symptoms.