Extending the range of symptoms in a Bayesian Network for the Predictive Diagnosis of COVID-19

Butcher, Rachel; Fenton, Norman

doi:10.1101/2020.10.22.20217554

Cited by 13 publications

(8 citation statements)

References 63 publications

(79 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…BNs have been used for a variety of COVID-19 models, including modelling transmission and outbreak response [17,18], decision making [19], risk analysis [20,21], risk assessment and contact tracing [22,23], interpretation of SARs-CoV-2 test results [24], and predictive diagnosis [25]. However, to our knowledge, BNs have not yet been applied for risk-benefit analysis of COVID-19 vaccines.…”

Section: Bayesian Networkmentioning

confidence: 99%

Risk-benefit analysis of the AstraZeneca COVID-19 vaccine in Australia using a Bayesian network modelling framework

Lau

Mayfield

Sinclair

et al. 2021

Preprint

View full text Add to dashboard Cite

Thrombosis and Thromobocytopenia Syndrome (TTS) has been associated with the AstraZencea (AZ) COVID-19 vaccine. Australia has reported low TTS incidence of <3/100,000 after the first dose, with case fatality rate (CFR) of 5-6%. Risk-benefit analysis of vaccination has been challenging because of rapidly evolving data, changing levels of transmission, and age-specific variation in rates of TTS, COVID-19, and CFR. We aim to optimise risk-benefit analysis by developing a model that enables inputs to be updated rapidly as evidence evolves. A Bayesian network was used to integrate local and international data, government reports, published literature and expert opinion. The model estimates probabilities of outcomes under different scenarios of age, sex, low/medium/high transmission (0.05%/0.45%/5.76% of population infected over 6 months), SARS-CoV-2 variant, vaccine doses, and vaccine effectiveness. We used the model to compare estimated deaths from vaccine-associated TTS with i) COVID-19 deaths prevented under different scenarios, and ii) deaths from COVID-19 related atypical severe blood clots (cerebral venous sinus thrombosis & portal vein thrombosis). For a million people aged >70 years where 70% received first dose and 35% received two doses, our model estimated <1 death from TTS, 25 deaths prevented under low transmission, and >3000 deaths prevented under high transmission. Risks versus benefits varied significantly between age groups and transmission levels. Under high transmission, deaths prevented by AZ vaccine far exceed deaths from TTS (by 8 to >4500 times depending on age). Probability of dying from COVID-related atypical severe blood clots was 58-126 times higher (depending on age and sex) than dying from TTS. To our knowledge, this is the first example of the use of Bayesian networks for risk-benefit analysis for a COVID-19 vaccine. The model can be rapidly updated to incorporate new data, adapted for other countries, extended to other outcomes (e.g., severe disease), or used for other vaccines.

show abstract

Section: Bayesian Networkmentioning

confidence: 99%

Risk-benefit analysis of the AstraZeneca COVID-19 vaccine in Australia using a Bayesian network modelling framework

Lau

Mayfield

Sinclair

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…in such a way that no information is wasted in identifying true positive cases. Indeed this is being done using Bayesian networks [8], though dealing with temporally sequenced and sparse information is critical for continuously updating diagnosis in real-time in a large population.…”

Section: Intelligently Integrating Multiple Tests Reduces the Proportion Of False Positivesmentioning

confidence: 99%

Bayesian, Universal COVID Testing

Meyer

Kralj

2021

Preprint

View full text Add to dashboard Cite

During the SARS-COV2 pandemic, there has been a persistent call for universal testing to better inform policy decisions. However, a little considered aspect of this call is the relationship between a test's accuracy and the tested demographic. What are the implications of frequent, universal testing in otherwise asymptomatic demographics? By applying Bayesian statistics, it becomes clear that as the odds of having COVID decreases, there is a non-linear increase in the odds that each positive test is, in fact, a false positive. This phenomenon has precedence in the historical narrative surrounding universal mammogram screening which is no longer recommended due to the unacceptably high rate of false positives. The solution to combat the inflation of false positives is also suggested by Bayesian statistics: intelligently integrating multiple COVID diagnostic tests and symptoms via Bayes' Theorem, an approach conceptually similar to pre-screening for mammograms. This extra information is readily available (e.g olfactory function and fever) and will minimize the economic and emotional costs incurred by false positives while simultaneously improving the information available for policy-makers. In summary, along with the push for universal testing should be an equally rigorous approach to interpreting the test results.

show abstract

“…The revised complete BN model, which includes both the extended background risks and infection possibility described in this report together with the extended set of symptoms described in (Butcher 2020), can be seen in the Appendix.…”

Section: Future Workmentioning

confidence: 99%

“…The model presented here expands substantially on that web of relevant risk factors, bringing in ethnicity, religion, occupation and housing conditions, and refining other factors such as age, obesity and underlying medical conditions. It also incorporates concurrent work on an extended set of symptoms described in (Butcher, 2020).…”

Section: Introductionmentioning

confidence: 99%

Extending the range of COVID-19 risk factors in a Bayesian network model for personalised risk assessment

Fenton

2020

Preprint

Self Cite

View full text Add to dashboard Cite

A need is emerging for individuals to gauge their own risks of coronavirus infection as it becomes apparent that contact tracing to contain the spread of the virus is not working in many societies. This paper presents an extension of an existing Bayesian network model for an application in which people can add their own personal risk factors to calculate their probability of exposure to the virus and likely severity if they do catch the illness. The data need not be shared with any central authority. In this way, people can become more aware of their individual risks and adjust their behaviour accordingly, as many countries prepare for a second wave of infections or a prolonged pandemic. This has the advantage not only of preserving privacy but also of containing the virus more effectively by allowing users to act without the time lag of waiting to be informed that a contact has been tested and confirmed COVID-19 positive. Through a nuanced assessment of individual risk, it could also release many people from isolation who are judged highly vulnerable using cruder measures, helping to boost economic activity and decrease social isolation without unduly increasing transmission risk. Although much has been written and reported about single risk factors, little has been done to bring these factors together in a user-friendly way to give an overall risk rating. The causal probabilistic model presented here shows the power of Bayesian networks to represent the interplay of multiple, dependent variables and to predict outcomes. The network, designed for use in the UK, is built using detailed data from government and health authorities and the latest research, and is capable of dynamic updates as new information becomes available. The focus of the paper is on the extended set of risk factors.

show abstract

Extending the range of symptoms in a Bayesian Network for the Predictive Diagnosis of COVID-19

Cited by 13 publications

References 63 publications

Risk-benefit analysis of the AstraZeneca COVID-19 vaccine in Australia using a Bayesian network modelling framework

Risk-benefit analysis of the AstraZeneca COVID-19 vaccine in Australia using a Bayesian network modelling framework

Bayesian, Universal COVID Testing

Extending the range of COVID-19 risk factors in a Bayesian network model for personalised risk assessment

Contact Info

Product

Resources

About