Choroba koronawirusowa 2019 (covid-19) – Rokowania, prognozy i postęp choroby

Choroba koronawirusowa 2019 (covid-19)
Rokowania, prognozy i postęp choroby

Wczesna identyfikacja pacjentów z COVID-19 zagrożonych ciężkim przebiegiem jest kluczowa dla optymalizacji opieki medycznej i alokacji zasobów. W literaturze opisano ponad 600 modeli prognostycznych, z których najczęściej oceniano śmiertelność (w 152 badaniach, 48,4%). Najważniejsze wskaźniki prognostyczne to m.in. liczba limfocytów i płytek krwi, poziom kreatyniny, IL-6, prokalcytoniny, D-dimerów, ferrytyny, LDH, CRP, aminotransferaz (AST, ALT), troponiny T wysokiej czułości, albuminy i kinazy kreatynowej. Czynniki kliniczne o istotnym znaczeniu to wiek (najsilniejszy czynnik ryzyka), choroby współistniejące (POCHP, choroby układu krążenia, nadciśnienie), duszność, zaburzenia świadomości oraz status szczepienia. Modele oparte na uczeniu maszynowym, wykorzystujące m.in. poziom albuminy, LDH, wiek i liczbę neutrofili, osiągają wysoką skuteczność prognostyczną (AUC do 0,905). Wśród zwalidowanych narzędzi prognostycznych wyróżnia się modele Qcovid, PRIEST, Carrs, ISARIC4C, Xie oraz ALDCC, które uwzględniają dane kliniczne i laboratoryjne przy przyjęciu pacjenta.

Choroba koronawirusowa 2019 (COVID-19): Rokowanie (przewidywanie wyników)

Modele prognostyczne w COVID-19 – przegląd aktualnego stanu wiedzy
Kluczowe czynniki prognostyczne w COVID-19
Najlepiej zwalidowane modele prognostyczne
Wykorzystanie uczenia maszynowego w prognozowaniu COVID-19
Prognostyczne wskaźniki laboratoryjne i ich kombinacje
Przewidywanie wyników u pacjentów wymagających ECMO
Prognostyczne znaczenie szczepień przeciwko COVID-19
Ograniczenia obecnych modeli prognostycznych

Prognostyczne implikacje pandemii COVID-19 dla innych schorzeń
Przyszłość modeli prognostycznych w COVID-19

Kolejne rozdziały

Choroba koronawirusowa 2019 (COVID-19): Rokowanie (przewidywanie wyników)

Wczesna identyfikacja pacjentów z COVID-19 zagrożonych ciężkim przebiegiem choroby ma kluczowe znaczenie dla szybkiego wdrożenia odpowiedniej opieki medycznej oraz optymalnej alokacji zasobów ochrony zdrowia. Odpowiednie prognozowanie przebiegu choroby nie tylko umożliwia efektywne kosztowo przydzielanie zasobów opieki zdrowotnej, ale potencjalnie może również zmniejszyć wskaźniki śmiertelności.¹²³

Modele prognostyczne w COVID-19 – przegląd aktualnego stanu wiedzy

Obecne dane na temat modeli prognostycznych COVID-19 są niespójne, a ich kliniczna przydatność pozostaje kontrowersyjna. Pomimo opracowania wielu modeli prognostycznych dla COVID-19, ich zastosowanie w praktyce jest ograniczone ze względu na problemy z uogólnianiem wyników oraz niedoskonałości metodologiczne.¹²

Przeglądy systematyczne zidentyfikowały łącznie setki modeli prognostycznych dla COVID-19. Wśród nich można wymienić:

125 modeli diagnostycznych (w tym 75 opartych na obrazowaniu medycznym)
606 modeli prognostycznych, w tym:
- 13 modeli identyfikujących osoby z grupy ryzyka COVID-19 w populacji ogólnej
- 593 modeli przewidujących różne wyniki u osób z potwierdzonym COVID-19

⁴⁵

Najczęstszym prognozowanym wynikiem była śmiertelność, którą oceniano w 152 (48,4%) badaniach.⁶⁷ Należy jednak zauważyć, że większość opublikowanych badań modeli predykcyjnych była słabo raportowana i obciążona wysokim ryzykiem błędu systematycznego, co prowadzi do prawdopodobnie zawyżonych wyników ich skuteczności predykcyjnej.⁵

Kluczowe czynniki prognostyczne w COVID-19

Badania zidentyfikowały szereg wskaźników laboratoryjnych i klinicznych związanych z niekorzystnym rokowaniem w COVID-19. Do najważniejszych należą:⁸³

Wskaźniki laboratoryjne:
- Liczba limfocytów i płytek krwi
- Poziom kreatyniny
- Interleukina 6 (IL-6)
- Prokalcytonina (PCT)
- D-dimery
- Ferrytyna
- Dehydrogenaza mleczanowa (LDH)
- Białko C-reaktywne (CRP)
- Aminotransferaza asparaginianowa (AST) i alaninowa (ALT)
- Troponina T wysokiej czułości (hs-TnT)
- Albumina
- Kinaza kreatynowa (CK)
Czynniki kliniczne i demograficzne:
- Wiek (najsilniejszy czynnik ryzyka ciężkiego przebiegu COVID-19)⁹
- Choroby współistniejące, szczególnie POCHP, choroby układu krążenia i nadciśnienie¹⁰
- Duszność (jedyny objaw istotnie związany z ciężkim przebiegiem COVID-19 i przyjęciem na OIT)¹⁰
- Zaburzenia świadomości¹¹
- Status szczepienia (brak szczepienia lub niekompletne szczepienie zwiększa ryzyko ciężkiego przebiegu)⁹

Nowsze badania z wykorzystaniem uczenia maszynowego wskazały, że najsilniejszymi predyktorami ciężkiego przebiegu choroby są: poziom albuminy w surowicy, poziom dehydrogenazy mleczanowej (LDH), wiek oraz liczba neutrofili.¹² Jeden z modeli osiągnął AUC na poziomie 0,905 wykorzystując te cztery cechy, co sugeruje, że dobrze ustrukturyzowane modele predykcyjne mogą być skuteczniejsze niż zwiększanie liczby analizowanych czynników.¹²¹³

Najlepiej zwalidowane modele prognostyczne

W przeglądzie systematycznym zidentyfikowano kilka modeli o niskim ryzyku błędu systematycznego i odpowiedniej wydajności predykcyjnej:⁴¹⁴

Modele Qcovid – przewidują przyjęcie do szpitala i zgon z powodu COVID-19 w populacji ogólnej w Wielkiej Brytanii, oddzielnie dla mężczyzn i kobiet
Skala PRIEST – przewiduje 30-dniową śmiertelność lub konieczność wsparcia narządów u pacjentów z podejrzewanym lub potwierdzonym COVID-19 zgłaszających się na oddział ratunkowy
Model Carrs
Model pogorszenia ISARIC4C – zwalidowany regionalnie w Wielkiej Brytanii
Model Xie

Warto również wspomnieć o modelu ALDCC, który został opracowany w Bostonie (USA) i zewnętrznie zwalidowany w Wuhan (Chiny) na grupie 375 uczestników. Model ten opiera się na informacjach uzyskanych przy przyjęciu, w tym wieku, liczbie limfocytów, poziomie D-dimerów, CRP i kreatyniny.⁷

Wykorzystanie uczenia maszynowego w prognozowaniu COVID-19

Algorytmy uczenia maszynowego oferują obiecujące podejście do prognozowania wyników klinicznych w oparciu o dane, przewyższając tradycyjne modelowanie statystyczne.¹⁵ Przeprowadzono wiele badań wykorzystujących różne techniki uczenia maszynowego do przewidywania wyników COVID-19:¹⁶¹⁷

Sieć neuronowa Elmana i SVM (Support Vector Machine) – skutecznie przewidują trend rozwoju łącznej liczby potwierdzonych przypadków, zgonów i wyleczonych przypadków
LSTM (Long Short-Term Memory) – szczególnie odpowiedni do przewidywania skumulowanej liczby potwierdzonych przypadków
SVM z granulacją rozmytą – skuteczny w przewidywaniu zakresu wzrostu nowych potwierdzonych przypadków, nowych zgonów i nowych wyleczonych przypadków
Sztuczne sieci neuronowe (ANN) – w jednym z badań wykazano, że model ANN osiągnął dokładność 86,25% w przewidywaniu śmiertelności, z czułością 87,50% i swoistością 85,94%¹⁸
Algorytm k-najbliższych sąsiadów (k-NN) – osiągnął dokładność 95,25%, czułość 95,30%, precyzję 92,7%, swoistość 93,30%, wynik F1 93,98% i pole pod krzywą ROC 96,90% w badaniu przeprowadzonym w Etiopii¹⁵

Modele uczenia maszynowego mogą identyfikować zagrożonych pacjentów już w momencie przyjęcia lub podczas hospitalizacji, umożliwiając optymalne wykorzystanie ograniczonych zasobów szpitalnych.¹⁹²⁰

Prognostyczne wskaźniki laboratoryjne i ich kombinacje

Badania wskazują, że kombinacje biomarkerów zapalnych mogą mieć większą wartość prognostyczną niż pojedyncze wskaźniki. W jednym z badań opracowano kombinowane skale oparte na CRP i innych biomarkerach zapalnych (LNR – stosunek limfocytów do neutrofili, dv-NLR – stosunek neutrofili do limfocytów, CLR – stosunek CRP do limfocytów), które wykazały obiecującą wartość predykcyjną dla wsparcia oddechowego za pomocą kaniuli nosowej o wysokim przepływie (HFNC) oraz śmiertelności u pacjentów z ciężkim COVID-19.²¹²²

Dla przykładu, punkty odcięcia dla przewidywania konieczności stosowania HFNC wynosiły:

3,2 dla CRP
0,231 dla LNR
0,90 dla dv-NLR
0,004 dla CLR

²²

Biomarker/Skala	Punkt odcięcia dla przewidywania śmiertelności	Dokładność predykcyjna
CRP	1,11	–
CLR	3,2*10^33	–
C-CRP #3* (skala kombinowana)	2 punkty	0,922

²²

Przewidywanie wyników u pacjentów wymagających ECMO

U pacjentów z ciężkim COVID-19 wymagających pozaustrojowego utlenowania krwi (V-V ECMO) 30-dniowa przeżywalność wynosi około 54%, co zachęca do dalszych badań klinicznych i stosowania tej metody. W tej grupie pacjentów wykorzystanie skal do przewidywania śmiertelności nie jest jednak zalecane przy podejmowaniu decyzji terapeutycznych. V-V ECMO powinno być rozważane, jeśli uznaje się je za korzystne, tak długo jak dostępne są zasoby.²³

Prognostyczne znaczenie szczepień przeciwko COVID-19

Nieszczepienie się lub brak aktualnego statusu szczepień przeciwko COVID-19 zwiększa ryzyko ciężkiego przebiegu choroby. Lekarze powinni uwzględniać wiek pacjenta, obecność chorób współistniejących i innych czynników ryzyka oraz status szczepienia przy określaniu ryzyka ciężkich powikłań związanych z COVID-19 u każdego pacjenta.⁹

Ograniczenia obecnych modeli prognostycznych

Większość opracowanych modeli prognostycznych dla COVID-19 ma istotne ograniczenia:¹⁴²⁴

Niska jakość metodologiczna badań
Małe próby badawcze
Nieprawidłowe postępowanie z brakującymi danymi
Nierozwiązany problem przeuczenia modeli
Definicje COVID-19 oparte na cechach klinicznych, a nie na wynikach laboratoryjnych testów diagnostycznych
Heterogenność ocenianych wyników
Wysokie lub niejasne ryzyko błędu systematycznego w 312 badaniach

Skuteczność tych modeli może się również różnić w czasie i między regionami, co wymaga dalszej walidacji i potencjalnego aktualizowania przed wdrożeniem.¹⁴

Prognostyczne implikacje pandemii COVID-19 dla innych schorzeń

Pandemia COVID-19 miała również wpływ na wyniki leczenia innych chorób. Na przykład, pacjenci z wysiękową postacią zwyrodnienia plamki żółtej związanego z wiekiem (nAMD) zdiagnozowani podczas ograniczeń związanych z COVID-19 mieli gorsze wyniki wzrokowe niż ci zdiagnozowani po tym okresie. Przyczyniło się do tego wiele czynników, w tym między innymi zmniejszona częstotliwość leczenia.²⁵

Przyszłość modeli prognostycznych w COVID-19

Przyszłe duże, wieloośrodkowe i dobrze zaprojektowane badania prospektywne są niezbędne do opracowania modeli predykcyjnych dla COVID-19 o użyteczności klinicznej, które można zastosować w różnych populacjach.¹²⁴

Jakość modeli można dodatkowo poprawić poprzez:²⁴

Właściwe uwzględnienie nieliniowych czynników prognostycznych
Prawidłowe zarządzanie brakującymi wartościami w badaniach
Walidację lub porównanie istniejących modeli w różnych warunkach

Ważnym kierunkiem rozwoju jest integracja algorytmów uczenia maszynowego z kompleksowymi bazami danych szpitalnych, która umożliwia dokładną klasyfikację ryzyka śmiertelności pacjentów z COVID-19.²⁶ Dzięki temu podejściu lekarze mogą opracować lepsze strategie zmniejszania powikłań i poprawy wskaźników przeżywalności pacjentów.¹⁹

Trafne modele predykcyjne mogą poprawić jakość opieki i zwiększyć wskaźniki przeżywalności pacjentów, identyfikując pacjentów wysokiego ryzyka i przyjmując najbardziej skuteczne plany leczenia i opieki. Podejście to może pomóc zmniejszyć niejednoznaczność, oferując klinicystom ilościowe, obiektywne i oparte na dowodach modele stratyfikacji ryzyka i przewidywania.¹⁹

Kolejne rozdziały

Zapraszamy do dalszego czytania naszego leksykonu.

Wybierz kolejny rozdział z menu poniżej, aby otworzyć nową podstronę kompedium wiedzy i uzyskać szczegółowe informację o leku, substancji lub chorobie.

Materiały źródłowe

#1 Prognostic models in COVID-19 infection that predict severity: a systematic review
https://pmc.ncbi.nlm.nih.gov/articles/PMC9958330/
Current evidence on COVID-19 prognostic models is inconsistent and clinical applicability remains controversial. […] We performed a systematic review to summarize and critically appraise the available studies that have developed, assessed and/or validated prognostic models of COVID-19 predicting health outcomes. […] While several clinical prognostic models for COVID-19 have been described in the literature, they are limited in generalizability and/or applicability due to deficiencies in addressing fundamental statistical and methodological concerns. […] Future large, multi-centric and well-designed prognostic prospective studies are needed to clarify remaining uncertainties. […] Early identification of COVID-19 patients at risk of critical illness is crucial for early identification of patients requiring urgent medical attention or who would benefit the most from treatment.
#2
https://link.springer.com/article/10.1007/s10654-023-00973-x
Current evidence on COVID-19 prognostic models is inconsistent and clinical applicability remains controversial. […] We performed a systematic review to summarize and critically appraise the available studies that have developed, assessed and/or validated prognostic models of COVID-19 predicting health outcomes. […] While several clinical prognostic models for COVID-19 have been described in the literature, they are limited in generalizability and/or applicability due to deficiencies in addressing fundamental statistical and methodological concerns. […] Future large, multi-centric and well-designed prognostic prospective studies are needed to clarify remaining uncertainties. […] Early identification of COVID-19 patients at risk of critical illness is crucial for early identification of patients requiring urgent medical attention or who would benefit the most from treatment.
#3
https://link.springer.com/article/10.1007/s10654-023-00973-x
In addition, early prediction of the disease course not only enables cost-effective allocation of health care resources, but potentially decreases fatality rates as well. […] Laboratory indicators including, but not limited to, lymphocyte and platelet count, creatinine, interleukin 6 (IL-6), procalcitonin (PCT), d-dimer, ferritin, lactate dehydrogenase (LDH), C-reactive protein (CRP), aspartate aminotransferase (AST), alanine aminotransferase (ALT), high-sensitivity troponin T (hs-TnT), albumin and creatine kinase (CK), have been identified as common predictors of poor outcomes in COVID-19. […] Thus, there is an urgent need to comprehensively and critically assess the available literature and identify the best performing and methodologically rigorous prognostic models for COVID-19 progression.
#4 Prediction models for diagnosis and prognosis of covid-19: systematic review and critical appraisal | The BMJ
https://www.bmj.com/content/369/bmj.m1328
Objective To review and appraise the validity and usefulness of published and preprint reports of prediction models for prognosis of patients with covid-19, and for detecting people in the general population at increased risk of covid-19 infection or being admitted to hospital or dying with the disease. […] Of these 731, 125 were diagnostic models (including 75 based on medical imaging) and the remaining 606 were prognostic models for either identifying those at risk of covid-19 in the general population (13 models) or predicting diverse outcomes in those individuals with confirmed covid-19 (593 models). […] The reported C indexes varied between 0.77 and 0.93 in development studies with low risk of bias, and between 0.56 and 0.78 in external validations with low risk of bias. […] The Qcovid models, the PRIEST score, Carrs model, the ISARIC4C Deterioration model, and the Xie model showed adequate predictive performance in studies at low risk of bias.
#5 Prediction models for diagnosis and prognosis of covid-19: systematic review and critical appraisal | The BMJ
https://www.bmj.com/content/369/bmj.m1328
Most published prediction model studies were poorly reported and at high risk of bias such that their reported predictive performances are probably optimistic. […] Models with low risk of bias should be validated before clinical implementation, preferably through collaborative efforts to also allow an investigation of the heterogeneity in their performance across various populations and settings. […] We identified 13 newly developed models aiming to predict covid-19 related risks in the general population. […] We identified 593 prognostic models for predicting clinical outcomes in patients with covid-19 (368 developments, 225 external validations). […] The studies reported C indexes between 0.49 and 1, with a median of 0.81 (interquartile range 0.75-0.89). […] Seven newly developed prognostic models and 22 external validations of prognostic models were at low risk of bias (n=29, 5%).
#6 Prognostic models in COVID-19 infection that predict severity: a systematic review
https://pmc.ncbi.nlm.nih.gov/articles/PMC9958330/
The most common outcome was mortality which was evaluated in 152 (48.4%) studies. […] The best reported predictive performance belonged to a model developed in Boston, USA, and externally validated in Wuhan, China, with 375 participants. […] A combined machine learning model (Support Vector Machine (SVM), Gradient Boosted Decision Tree (GBDT), and Neural Network(NN)) based on procalcitonin, [T+B+NK cell] count, IL-6, CRP, IL-2-receptor, T-helper lymphocyte/T-suppressor lymphocyte as predictors of critical illness had the best reported predictive performance, with an AUC of 0.99. […] Overall, the RoB assessment was rated to be at high or unclear RoB in 312 studies. […] This requires a homogeneous definition of COVID-19 and outcomes and appropriate model selection methods to lead user-friendly models that can be externally validated.
#7
https://link.springer.com/article/10.1007/s10654-023-00973-x
The most common outcome was mortality which was evaluated in 152 (48.4%) studies. […] Among studies that specified and reported prediction time for mortality there was variation in reporting. […] The difference between fixed time frame and time-to-event outcomes, lies mainly in the fact that time-to-event reporting takes account of whether an event takes place and the time at which the event occurs, such that both the event and the timing of the event are important. […] The best reported predictive performance belonged to a model developed in Boston, USA, and externally validated in Wuhan, China, with 375 participants. […] The model was based on information acquired on admission, including age, lymphocyte count, d-dimer, CRP and creatinine (ALDCC). […] Prediction time defined as timeframe which the model pre-specified to assess its performance, i.e., from admission to the worsening severity or critical illness was different across studies, and the longest time of follow-up was 30 days.
#8 Prognostic models in COVID-19 infection that predict severity: a systematic review
https://pmc.ncbi.nlm.nih.gov/articles/PMC9958330/
In addition, early prediction of the disease course not only enables cost-effective allocation of health care resources, but potentially decreases fatality rates as well. […] Laboratory indicators including, but not limited to, lymphocyte and platelet count, creatinine, interleukin 6 (IL-6), procalcitonin (PCT), d-dimer, ferritin, lactate dehydrogenase (LDH), C-reactive protein (CRP), aspartate aminotransferase (AST), alanine aminotransferase (ALT), high-sensitivity troponin T (hs-TnT), albumin and creatine kinase (CK), have been identified as common predictors of poor outcomes in COVID-19. […] However, most of the reported models are at high risk of bias due to deficiencies in the methods used, definitions of COVID-19 (e.g., cases defined based on clinical features rather than on the result of laboratory diagnostic test for SARS-CoV-2) and the use of heterogeneous outcomes.
#9 Underlying Conditions and the Higher Risk for Severe COVID-19 | COVID-19 | CDC
https://www.cdc.gov/covid/hcp/clinical-care/underlying-conditions.html
Age is the strongest risk factor for severe COVID-19 outcomes. Patients with one or multiple certain underlying medical conditions are also at higher risk.(1-3) […] Additionally, being unvaccinated or not being up to date on COVID-19 vaccinations also increases the risk of severe COVID-19 outcomes. […] Providers should consider the patient’s age, presence of underlying medical conditions and other risk factors, and vaccination status in determining the risk of severe COVID-19-associated outcomes for any patient. […] Certain underlying medical conditions were associated with an increased risk for severe COVID-19 illness in adults. […] Having multiple conditions was also associated with severe COVID-19 illness. […] Obesity, diabetes with complications, and anxiety and fear-related disorders had the strongest association with death. […] The number of frequent underlying medical conditions (present in 10.0% of patients) increased with age.
#10
https://link.springer.com/article/10.1007/s00038-020-01390-7
COVID-19 has a varied clinical presentation. Elderly patients with comorbidities are more vulnerable to severe disease. This study identifies specific symptoms and comorbidities predicting severe COVID-19 and intensive care unit (ICU) admission. […] Dyspnoea was the only symptom predictive for severe disease (pOR 3.70, 95% CI 1.83-7.46) and ICU admission (pOR 6.55, 95% CI 4.28-10.0). […] COPD was the strongest predictive comorbidity for severe disease (pOR 6.42, 95% CI 2.44-16.9) and ICU admission (pOR 17.8, 95% CI 6.56-48.2), followed by cardiovascular disease and hypertension. […] Dyspnoea was the only symptom predictive for severe COVID-19 and ICU admission. Patients with COPD, cardiovascular disease and hypertension were at higher risk of severe illness and ICU admission. […] The most prevalent symptoms in the severe disease group were cough, fever and fatigue and in the ICU admitted group were cough, fever and dyspnoea. […] Dyspnoea was the only symptom significantly associated with disease severity and ICU admission, alongside various comorbidities (COPD, CVD and hypertension). […] COPD was an extremely strong predictor for both severe disease and ICU admission.
#11 Journal of Medical Internet Research – Prognostic Modeling of COVID-19 Using Artificial Intelligence in the United Kingdom: Model Development and Validation
https://www.jmir.org/2020/8/e20259/
This analysis demonstrates an adaptive ANN trained on data at a single site, which demonstrates the early utility of deep learning approaches in a rapidly evolving pandemic with no established or validated prognostic scoring systems. […] Our aim was to provide a patient-specific, point-of-admission mortality risk prediction to help inform clinical management decisions at the earliest opportunity. […] Altered mentation, dyspnea, and increasing age were found to be the most salient overall features in predicting mortality. Moderate predictors of mortality included collapse, male gender, new cough, and previous respiratory pathology. […] In this analysis, we provide details of an ANN capable of predicting patient-specific mortality with high sensitivity and specificity.
#12 Predicting coronavirus disease 2019 severity using explainable artificial intelligence techniques | Scientific Reports
https://www.nature.com/articles/s41598-025-85733-5
Predictive models for determining coronavirus disease 2019 (COVID-19) severity have been established; however, the complexity of the interactions among factors limits the use of conventional statistical methods. […] A total of 3,301 patients diagnosed with COVID-19 between February 2020 and October 2022 were included. […] The predictive model achieved an AUC of 0.905 using four features: serum albumin levels, lactate dehydrogenase levels, age, and neutrophil count. […] The highest AUC value was 0.906 (sensitivity, 0.842; specificity, 0.811) in the discovery cohort and 0.861 (sensitivity, 0.804; specificity, 0.675) in the validation cohort. […] Simple and well-structured predictive models were established, which may aid in patient management and the selection of therapeutic interventions.
#13 Predicting coronavirus disease 2019 severity using explainable artificial intelligence techniques | Scientific Reports
https://www.nature.com/articles/s41598-025-85733-5
Age, LDH and Alb levels, and neutrophil count were the most significant factors. […] The best prediction performance of the simple prediction model was the AUC values of 0.906 and 0.861 in discovery and validation cohort, respectively. […] These findings suggest that, in severity prediction, modeling the interactions among factors appropriately may contribute more to improving predictive accuracy than merely increasing the number of factors. […] In this multicenter study, simple and well-structured predictive models for COVID-19 severity in Japanese patients were established using explanatory ML.
#14 Prediction models for diagnosis and prognosis of covid-19: systematic review and critical appraisal | The BMJ
https://www.bmj.com/content/369/bmj.m1328
Most newly developed models and external validations were at unclear (n=32, 5%) or high (n=545, 90%) risk of bias according to assessment with PROBAST, which suggests that the predictive performance when used in practice is probably lower than what is reported. […] Overall, 530 (87%) were at high risk of bias for the analysis domain, and the reporting was insufficiently clear to assess risk of bias in the analysis in 42 (7%). […] We found seven newly developed models at low risk of bias. […] The four Qcovid models predict hospital admission and death with covid-19 in the general population in the UK, separately for men and women. […] The PRIEST score predicts 30 day death or organ support in patients with suspected or confirmed covid-19 presenting at the emergency department. […] The ISARIC 4C Deterioration model was validated regionally within the UK. […] The performance of these models is likely to vary over time and differ between regions, necessitating further validation and potentially updating before implementation.
#15 Machine learning algorithms for predicting COVID-19 mortality in Ethiopia | BMC Public Health | Full Text
https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-024-19196-0
Coronavirus disease 2019 (COVID-19), a global public health crisis, continues to pose challenges despite preventive measures. […] Machine learning algorithms offer a promising approach to data-driven prediction of clinical outcomes, surpassing traditional statistical modeling. […] Therefore, the aim of this study is to develop and validate machine-learning models for accurately predicting mortality in COVID-19 hospitalized patients in Ethiopia. […] The experimental results revealed that the k-nearest neighbor (k-NN) algorithm outperformed than other machine learning algorithms, achieving an accuracy of 95.25%, sensitivity of 95.30%, precision of 92.7%, specificity of 93.30%, F1 score 93.98% and a receiver operating characteristic (ROC) score of 96.90%. […] Our study has developed an innovative model that utilizes hospital data to accurately predict the mortality risk of COVID-19 patients.
#16 Prediction and analysis of Corona Virus Disease 2019 | PLOS One
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0239960
The outbreak of Corona Virus Disease 2019 (COVID-19) in Wuhan has significantly impacted the economy and society globally. […] A SVM with fuzzy granulation was used to predict the growth range of confirmed new cases, new deaths, and new cured cases. […] The experimental results showed that the Elman neural network and SVM used in this study can predict the development trend of cumulative confirmed cases, deaths, and cured cases, whereas LSTM is more suitable for the prediction of the cumulative confirmed cases. […] The SVM with fuzzy granulation can successfully predict the growth range of confirmed new cases and new cured cases, although the average predicted values are slightly large. […] In the present study, three methods, namely, an Elman neural network, LSTM, and SVM are applied to predict and analyze COVID-19 data from January 23, 2020 to April 6, 2020 in Wuhan, Hubei Province, China, including cumulative confirmed cases, confirmed new cases, cumulative deaths, new deaths, and cumulative cured cases and new cured cases.
#17 Prediction and analysis of Corona Virus Disease 2019 | PLOS One
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0239960
Experimental results showed that the Elman neural network and SVM adopted in this study can accurately predict the development trend of COVID-19, whereas LSTM is more suitable for the prediction of cumulative confirmed cases. […] The SVM with fuzzy granulation is effective for the growth range prediction of confirmed new cases, new deaths, and new cured cases, despite the larger averages. […] The prediction results of the two recurrent neural network models used in this study are unstable because neural networks tend to fall into a local optimum for data with irregular growth. […] For the prediction of the growth range of confirmed new cases, new deaths, and new cured cases, the SVM with fuzzy granulation introduced in this paper was shown to be effective.
#18 Journal of Medical Internet Research – Prognostic Modeling of COVID-19 Using Artificial Intelligence in the United Kingdom: Model Development and Validation
https://www.jmir.org/2020/8/e20259/
Background: The current severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) outbreak is a public health emergency and the case fatality rate in the United Kingdom is significant. Although there appear to be several early predictors of outcome, there are no currently validated prognostic models or scoring systems applicable specifically to patients with confirmed SARS-CoV-2. […] We aim to create a point-of-admission mortality risk scoring system using an artificial neural network (ANN). […] Patient-specific mortality was predicted with 86.25% accuracy, with a sensitivity of 87.50% (95% CI 61.65%-98.45%) and specificity of 85.94% (95% CI 74.98%-93.36%). The positive predictive value was 60.87% (95% CI 45.23%-74.56%), and the negative predictive value was 96.49% (95% CI 88.23%-99.02%). The area under the receiver operating characteristic curve was 90.12%.
#19 Machine learning algorithms for predicting COVID-19 mortality in Ethiopia | BMC Public Health | Full Text
https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-024-19196-0
Valid predictive models can improve the quality of care and increase patient survival rates, by identifying high-risk patients and adopting the most effective assistive and treatment care plans. […] This approach can help reduce ambiguity, by offering clinicians quantitative, objective, and evidence-based models for risk stratification, prediction, and eventually episode of the care plan. […] By adopting this approach, clinicians can devise better strategies to reduce complications and improve patient survival rates. […] Our study aimed to develop a new model for predicting the mortality risk of COVID-19 patients using hospital report data from different countries. […] The main purpose of this model is to prioritize early treatment for high-risk patients and optimize the use of limited healthcare resources during the ongoing pandemic.
#20 Machine learning algorithms for predicting COVID-19 mortality in Ethiopia | BMC Public Health | Full Text
https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-024-19196-0
Our study specifically focused on creating and evaluating machine learning-based prediction models for in-hospital mortality, using 23 key clinical predictors. […] This suggests that our model can effectively predict the mortality risk of hospitalized COVID-19 patients, optimizing the allocation of limited hospital resources. […] Importantly, our model can identify high-risk patients as early as the time of admission or during hospitalization. […] Integrating machine learning algorithms with comprehensive hospital databases allows for accurate classification of COVID-19 patient mortality risk.
#21 Predicting the Outcome of Patients with Severe COVID-19 with Simple Inflammatory Biomarkers: The Utility of Novel Combined ScoresâResults from a European Tertiary/Referral Centre
https://www.mdpi.com/2077-0383/13/4/967
Predicting the Outcome of Patients with Severe COVID-19 with Simple Inflammatory Biomarkers: The Utility of Novel Combined ScoresâResults from a European Tertiary/Referral Centre […] Background: The clinical significance of combinations of inflammatory biomarkers in severe COVID-19 infection is yet to be proved. […] We investigated the prognostic value of combination scores of admission values of inflammatory biomarkers in adults with severe COVID-19. […] The purpose of the present study was to investigate the prognostic value of admission values and combination scores of simple inflammatory biomarkers in predicting the outcome, in terms of (i) escalation of respiratory support with the use of high-flow nasal cannula (HFNC), (ii) admission to Intensive Care Unit (ICU) or (iii) death, in patients with severe COVID-19 infection.
#22 Predicting the Outcome of Patients with Severe COVID-19 with Simple Inflammatory Biomarkers: The Utility of Novel Combined ScoresâResults from a European Tertiary/Referral Centre
https://www.mdpi.com/2077-0383/13/4/967
Results: One hundred and fifteen patients (60% males, mean age 57.7 years) were included. […] As far as HFNC is concerned, the cut-off point was 3.2 for CRP, 0.231 for LNR, 0.90 for dv-NLR and 0.004 for CLR. […] Two points of C-CRP #1 and 2 points of C-CRP #3 predicted HFNC with a probability as high as 0.625 (p = 0.005) and 0.561 (p < 0.001), respectively. [...] For death, the optimal cut-off point for CRP was 1.11 and for CLR 3.2*1033. [...] Two points of C-CRP #3* with an accuracy of 0.922 predicted mortality (p = 0.0038) in severe COVID-19. [...] Conclusions: The combination scores of CRP and inflammatory biomarkers, based on admission values, are promising predictors for respiratory support using HFNC and for mortality in patients suffering from severe COVID-19 infection. [...] However, a comprehensive evaluation in large, multicenter studies seems mandatory in order to establish the clinical utility of these biomarkers and to draw generalizable conclusions.
#23 Outcome Prediction in Patients with Severe COVID-19 Requiring Extracorporeal Membrane OxygenationâA Retrospective International Multicenter Study
https://www.mdpi.com/2077-0375/11/3/170
Thirty-day-survival in COVID-19 patients treated with V-V ECMO and evaluated in our registry is 54 percent, encouraging further clinical research and application. […] The use of scores for the prediction of mortality cannot be recommended for treatment decisions in severe COVID-19 ARDS and V-V ECMO should be considered if deemed beneficial as long as resources are available. Scoring results below or above a specific cut-off value may be considered as an additional tool in the evaluation of prognosis but cannot be used for triaging this patient population.
#24
https://link.springer.com/article/10.1007/s10654-023-00973-x
Overall, the RoB assessment was rated to be at high or unclear RoB in 312 studies. […] This could be explained by shortcomings such as poor methodological quality, small sample size, poor handling of missing data, failure to deal with overfitting, definitions of COVID-19 based on clinical features rather than on the result of laboratory diagnostic test for SARS-CoV-2 and its severity with studies using heterogeneous outcomes. […] Future studies should therefore consider validating or comparing the existing models in different settings. […] Quality can be further improved by properly addressing the non-linear prognostic factors and missing values in studies. […] Future large, multi-center and well-designed prospective studies are needed for the development of predictive models for COVID-19 with clinical utility that can be applied to diverse populations.
#25 Comparing visual outcomes of nAMD treatment during and after the COVID-19 restrictions period | PLOS One
https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0323253
Compare treatment outcomes of newly diagnosed neovascular age-related macular degeneration (nAMD) during and after the COVID-19 restrictions. […] Patients diagnosed during the COVID-19 restrictions period had worse visual outcomes than those diagnosed thereafter. Multiple factors, including, but not limited to reduced treatment frequency, likely contributed to worse visual outcomes. […] The pandemic was associated with poorer visual outcomes for patients diagnosed during the restricted period (Group 1) compared to those diagnosed after this period (Group 2). […] Patients in Group 1 had worse visual outcomes at 12 months and last follow-up compared to Group 2, with a lower proportion achieving a BCVA20/40 and a higher proportion with BCVA20/70. […] Our findings emphasize the importance of developing adaptable strategies to maintain effective patient care in the face of external challenges posed by a global pandemic.
#26 Machine learning algorithms for predicting COVID-19 mortality in Ethiopia | BMC Public Health | Full Text
https://bmcpublichealth.biomedcentral.com/articles/10.1186/s12889-024-19196-0
By integrating machine learning with comprehensive hospital databases, our model effectively classifies patients’ mortality risk, enabling targeted medical interventions and improved resource management. […] The top five predictors are gender (female), intensive care unit (ICU) admission, alcohol drinking, smoking, and symptoms of headache and chills. […] This advancement holds great promise in enhancing healthcare outcomes and decision-making during the pandemic. […] By providing services and prioritizing patients based on the identified predictors, healthcare facilities and providers can improve the chances of survival for individuals. […] The accurate prognosis of COVID-19 clinical outcome is challenging due to the wide range of illness severity, which makes appropriate triage and resource allocation crucial for enhancing patient care within health-care systems.