Health data have enormous potential to transform healthcare, health service design, research, and individual health management. However, health data collected by institutions tend to remain siloed within those institutions limiting access by other services, individuals or researchers. Further, health data generated outside health services (e.g., from wearable devices) may not be easily accessible or useable by individuals or connected to other parts of the health system. There are ongoing tensions between data protection and the use of data for the public good (e.g., research). Concurrently, there are a number of data platforms that provide ways to disrupt these traditional health data siloes, giving greater control to individuals and communities. Through four case studies, this paper explores platforms providing new ways for health data to be used for personal data sharing, self-health management, research, and clinical care. The case-studies include data platforms: PatientsLikeMe, Open Humans, Health Record Banks, and unforgettable.me. These are explored with regard to what they mean for data access, data control, and data governance. The case studies provide insight into a shift from institutional to individual data stewardship. Looking at emerging data governance models, such as data trusts and data commons, points to collective control over health data as an emerging approach to issues of data control. These shifts pose challenges as to how “traditional” health services make use of data collected on these platforms. Further, it raises broader policy questions regarding how to decide what public good data should be put towards.

Global Health Security Index (GHSI) and Joint External Evaluation (JEE) are two well-known health security and related capability indices. We hypothesised that countries with higher GHSI or JEE scores would have detected their first COVID-19 case earlier, and would experience lower mortality outcome compared to countries with lower scores. We evaluated the effectiveness of GHSI and JEE in predicting countries' COVID-19 detection response times and mortality outcome (deaths/million). We used two different outcomes for the evaluation: (i) detection response time, the duration of time to the first confirmed case detection (from 31st December 2019 to 20th February 2020 when every country's first case was linked to travel from China) and (ii) mortality outcome (deaths/million) until 11th March and 1st July 2020, respectively. We interpreted the detection response time alongside previously published relative risk of the importation of COVID-19 cases from China. We performed multiple linear regression and negative binomial regression analysis to evaluate how these indices predicted the actual outcome. The two indices, GHSI and JEE were strongly correlated (r = 0.82), indicating a good agreement between them. However, both GHSI (r = 0.31) and JEE (r = 0.37) had a poor correlation with countries' COVID-19–related mortality outcome. Higher risk of importation of COVID-19 from China for a given country was negatively correlated with the time taken to detect the first case in that country (adjusted R2 = 0.63–0.66), while the GHSI and JEE had minimal predictive value. In the negative binomial regression model, countries' mortality outcome was strongly predicted by the percentage of the population aged 65 and above (incidence rate ratio (IRR): 1.10 (95% confidence interval (CI): 1.01–1.21) while overall GHSI score (IRR: 1.01 (95% CI: 0.98–1.01)) and JEE (IRR: 0.99 (95% CI: 0.96–1.02)) were not significant predictors. GHSI and JEE had lower predictive value for detection response time and mortality outcome due to COVID-19. We suggest introduction of a population healthiness parameter, to address demographic and comorbidity vulnerabilities, and reappraisal of the ranking system and methods used to obtain the index based on experience gained from this pandemic.

This study aimed to analyse the survival of patients admitted to Brazilian hospitals due to the COVID-19 and estimate prognostic factors. This is a retrospective, multicentre cohort study, based on data from 46 285 hospitalisations for COVID-19 in Brazil. Survival functions were calculated using the Kaplan–Meier's method. The log-rank test compared the survival functions for each variable and from that, hazard ratios (HRs) were calculated, and the proportional hazard model was used in Cox multiple regression. The smallest survival curves were the ones for patients at the age of 68 years or more, black/mixed race, illiterate, living in the countryside, dyspnoea, respiratory distress, influenza-like outbreak, O2 saturation <95%, X-ray change, length of stay in the intensive care unit (ICU), invasive ventilatory support, previous heart disease, pneumopathy, diabetes, Down's syndrome, neurological disease and kidney disease. Better survival was observed in the influenza-like outbreak and in an asthmatic patient. The multiple model for increased risk of death when they were admitted to the ICU HR 1.28, diabetes HR 1.17, neurological disease HR 1.34, kidney disease HR 1.11, heart disease HR 1.14, black or mixed race of HR 1.50, asthma HR 0.71 and pneumopathy HR 1.12. This reinforces the importance of socio-demographic and clinical factors as a prognosis for death.

D-dimer level on admission is a promising biomarker to predict mortality in patients with COVID-19. In this study, we reviewed the association between on-admission D-dimer levels and all-cause mortality risk in COVID-19 patients. Peer-reviewed studies and preprints reporting categorised D-dimer levels on admission and all-cause mortality until 24 May 2020 were searched for using the following keywords: ‘COVID-19’, ‘D-dimer’ and ‘Mortality’. A meta-analysis was performed to determine the pooled risk ratio (RR) for all-cause mortality. In total, 2911 COVID-19 patients from nine studies were included in this meta-analysis. Regardless of the different D-dimer cut-off values used, the pooled RR for all-cause mortality in patients with elevated vs. normal on-admission D-dimer level was 4.77 (95% confidence interval (CI) 3.02–7.54). Sensitivity analysis did not significantly affect the overall mortality risk. Analysis restricted to studies with 0.5 μg/ml as the cut-off value resulted in a pooled RR for mortality of 4.60 (95% CI 2.72–7.79). Subgroup analysis showed that the pooled all-cause mortality risk was higher in Chinese vs. non-Chinese studies (RR 5.87; 95% CI 2.67–12.89 and RR 3.35; 95% CI 1.66–6.73; P = 0.29). On-admission D-dimer levels showed a promising prognostic role in predicting all-cause mortality in COVID-19 patients, elevated D-dimer levels were associated with increased risk of mortality.

In chemical process engineering, surrogate models of complex systems are often necessary for tasks of domain exploration, sensitivity analysis of the design parameters, and optimization. A suite of computational fluid dynamics (CFD) simulations geared toward chemical process equipment modeling has been developed and validated with experimental results from the literature. Various regression-based active learning strategies are explored with these CFD simulators in-the-loop under the constraints of a limited function evaluation budget. Specifically, five different sampling strategies and five regression techniques are compared, considering a set of four test cases of industrial significance and varying complexity. Gaussian process regression was observed to have a consistently good performance for these applications. The present quantitative study outlines the pros and cons of the different available techniques and highlights the best practices for their adoption. The test cases and tools are available with an open-source license to ensure reproducibility and engage the wider research community in contributing to both the CFD models and developing and benchmarking new improved algorithms tailored to this field.

We aimed to describe the clinical features in coronavirus disease 2019 (COVID-19) cases. We studied 134 critically ill COVID-19 cases from 30 December 2019 to 20 February 2020 in an intensive care unit (ICU) at Wuhan Jinyintan Hospital. Demographics, underlying diseases, therapy strategies and test results were collected and analysed from patients on admission, admission to the ICU and 48 h before death. The non-survivors were older (65.46 (s.d. 9.74) vs. 46.45 (s.d. 11.09)) and were more likely to have underlying diseases. The blood group distribution of the COVID-19 cases differed from that of the Han population in Wuhan, with type A being 43.85%; type B, 26.92%; type AB, 10% and type O, 19.23%. Non-survivors tend to develop more severe lymphopaenia, with higher C-reactive protein, interleukin-6, procalcitonin, D-dimer levels and gradually increased with time. The clinical manifestations were non-specific. Compared with survivors, non-survivors more likely to have organ function injury, and to receive mechanical ventilation, either invasively or noninvasively. Multiple organ failure and secondary bacterial infection in the later period is worthy of attention.

Personalized PageRank has found many uses in not only the ranking of webpages, but also algorithmic design, due to its ability to capture certain geometric properties of networks. In this paper, we study the diffusion of PageRank: how varying the jumping (or teleportation) constant affects PageRank values. To this end, we prove a gradient estimate for PageRank, akin to the Li–Yau inequality for positive solutions to the heat equation (for manifolds, with later versions adapted to graphs).

Corona virus disease 2019 (COVID-19), caused by the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), was first detected in the city of Wuhan, China in December 2019. Although, the disease appeared in Africa later than other regions, it has now spread to virtually all countries on the continent. We provide early spatio-temporal dynamics of COVID-19 within the first 62 days of the disease's appearance on the African continent. We used a two-parameter hurdle Poisson model to simultaneously analyse the zero counts and the frequency of occurrence. We investigate the effects of important healthcare capacities including hospital beds and number of medical doctors in different countries. The results show that cases of the pandemic vary geographically across Africa with notably high incidence in neighbouring countries particularly in West and North Africa. The burden of the disease (per 100 000) mostly impacted Djibouti, Tunisia, Morocco and Algeria. Temporally, during the first 4 weeks, the burden was highest in Senegal, Egypt and Mauritania, but by mid-April it shifted to Somalia, Chad, Guinea, Tanzania, Gabon, Sudan and Zimbabwe. Currently, Namibia, Angola, South Sudan, Burundi and Uganda have the least burden. These findings could be useful in guiding epidemiological interventions and the allocation of scarce resources based on heterogeneity of the disease patterns.

This study estimates the incubation period of COVID-19 among locally transmitted cases, and its association with age to better inform public health measures in containing COVID-19. Epidemiological data of all PCR-confirmed COVID-19 cases from all restructured hospitals in Singapore were collected between 23 January 2020 and 2 April 2020. Activity mapping and detailed epidemiological investigation were conducted by trained personnel. Positive cases without clear exposure to another positive case were excluded from the analysis. One hundred and sixty-four cases (15.6% of patients) met the inclusion criteria during the defined period. The crude median incubation period was 5 days (range 1–12 days) and median age was 42 years (range 5–79 years). The median incubation period among those 70 years and older was significantly longer than those younger than 70 years (8 vis-à-vis 5 days, P = 0.040). Incubation period was negatively correlated with day of illness in both groups. These findings support current policies of 14-day quarantine periods for close contacts of confirmed cases and 28 days for monitoring infections in known clusters. An elderly person who may have a longer incubation period than a younger counterpart may benefit from earlier and proactive testing, especially after exposure to a positive case.

While most research focuses on the clinical treatment of COVID-19, fewer studies have investigated individuals' responses towards this novel infectious disease. This study aims to report the temporal changes in individuals' psychological wellbeing, perceived discrimination, sociopolitical perceptions and information-seeking behaviours among the general public in Hubei, China. Data were obtained from a two-wave survey of 1902 respondents aged 18–80 in Hubei province during the peak and mitigation stages of the outbreak. The results showed that the prevalence of psychological distress dropped from over 75% to around 15% throughout the study period, but perceived discrimination remained stable. Female, middle-aged, well-educated respondents and those employed in government/public institutions/state-owned enterprises tended to report more distress. While respondents' attention on COVID-19 information kept high and stable, their sources of information diversified across different sociodemographic groups. Over time, people obtained more social support from neighbourhoods than from their friends and relatives or non-government organisations. Over 80% of respondents were satisfied with the performance of the central government, which was notably higher than their ratings on the local government and neighbourhood/village committees. The findings of this research are informative for formulating effective intervention strategies to tackle various psychosocial problems during COVID-19.

We recruited 1591 patients who presented to our fever clinics from 23 January 2020 to 16 February 2020. The different imaging findings between COVID-19 pneumonia and influenza A viruses, influenza B virus pneumonia were also investigated. Most patients were infected by influenza A and B viruses in the flu-season. A laboratory kit is urgently needed to test different viruses simultaneously. Computed tomography can help early screen suspected patients with COVID-19 and differentiate different virus-related pneumonia.

We consider a dynamic network cascade process developed by Duncan Watts applied to a class of random networks, developed independently by Newman and Miller, which allows a specified amount of clustering (short loops). We adapt existing methods for locally tree-like networks to formulate an appropriate two-type branching process to describe the spread of a cascade started with a single active node and obtain a fixed-point equation to implicitly express the extinction probability of such a cascade. In so doing, we also recover a formula that has appeared in various forms in works by Hackett et al. and Miller which provides a threshold condition for certain extinction of the cascade. We find that clustering impedes cascade propagation for networks of low average degree by reducing the connectivity of the network, but for networks with high average degree, the presence of small cycles makes cascades more likely.

A Markov tree is a random vector indexed by the nodes of a tree whose distribution is determined by the distributions of pairs of neighbouring variables and a list of conditional independence relations. Upon an assumption on the tails of the Markov kernels associated to these pairs, the conditional distribution of the self-normalized random vector when the variable at the root of the tree tends to infinity converges weakly to a random vector of coupled random walks called a tail tree. If, in addition, the conditioning variable has a regularly varying tail, the Markov tree satisfies a form of one-component regular variation. Changing the location of the root, that is, changing the conditioning variable, yields a different tail tree. When the tails of the marginal distributions of the conditioning variables are balanced, these tail trees are connected by a formula that generalizes the time change formula for regularly varying stationary time series. The formula is most easily understood when the various one-component regular variation statements are tied up into a single multi-component statement. The theory of multi-component regular variation is worked out for general random vectors, not necessarily Markov trees, with an eye towards other models, graphical or otherwise.

Superbubbles are acyclic induced subgraphs of a digraph with single entrance and exit that naturally arise in the context of genome assembly and the analysis of genome alignments in computational biology. These structures can be computed in linear time and are confined to non-symmetric digraphs. We demonstrate empirically that graph parameters derived from superbubbles provide a convenient means of distinguishing different classes of real-world graphical models, while being largely unrelated to simple, commonly used parameters.

We provide a new proof of the existence of Gibbs point processes with infinite range interactions, based on the compactness of entropy levels. Our main existence theorem holds under two assumptions. The first one is the standard stability assumption, which means that the energy of any finite configuration is superlinear with respect to the number of points. The second assumption is the so-called intensity regularity, which controls the long range of the interaction via the intensity of the process. This assumption is new and introduced here since it is well adapted to the entropy approach. As a corollary of our main result we improve the existence results by Ruelle (1970) for pairwise interactions by relaxing the superstabilty assumption. Note that our setting is not reduced to pairwise interaction and can contain infinite-range multi-body counterparts.

In this paper the behaviour of the failure rate and reversed failure rate of an n-component coherent system is studied, where it is assumed that the lifetimes of the components are independent and have a common cumulative distribution function F. Sufficient conditions are provided under which the system failure rate is increasing and the corresponding reversed failure rate is decreasing. We also study the stochastic and ageing properties of doubly truncated random variables for coherent systems.

We consider a class of multitype Galton–Watson branching processes with a countably infinite type set $\mathcal{X}_d$ whose mean progeny matrices have a block lower Hessenberg form. For these processes, we study the probabilities $\textbf{\textit{q}}(A)$ of extinction in sets of types $A\subseteq \mathcal{X}_d$. We compare $\textbf{\textit{q}}(A)$ with the global extinction probability $\textbf{\textit{q}} = \textbf{\textit{q}}(\mathcal{X}_d)$, that is, the probability that the population eventually becomes empty, and with the partial extinction probability $\tilde{\textbf{\textit{q}}}$, that is, the probability that all types eventually disappear from the population. After deriving partial and global extinction criteria, we develop conditions for $\textbf{\textit{q}} < \textbf{\textit{q}}(A) < \tilde{\textbf{\textit{q}}}$. We then present an iterative method to compute the vector $\textbf{\textit{q}}(A)$ for any set A. Finally, we investigate the location of the vectors $\textbf{\textit{q}}(A)$ in the set of fixed points of the progeny generating vector.