ABSTRACT
Contact tracing is gaining its importance in controlling the spread of COVID-19. However, the enormous volume of the frequently sampled tracing data brings major challenges for real-time processing. In this paper, we propose a GPU-based real-time contact tracing system based on spatial proximity queries with temporal constraints using location data. We provide dynamic indexing of moving objects using an adaptive partitioning schema on GPU with extremely low overhead. Our system optimizes the retrieval of contacted pairs to match both the requirements of contact tracing scenarios and GPU centered parallelism. We propose an efficient contacts evaluation mechanism to keep only the spatially and temporally valid contacts. Our experiments demonstrate that the system can achieve sub-second level response for large-scale contact tracing of tens of millions of people, with two magnitudes of performance boost over CPU based approach.
ABSTRACT
The coronavirus disease 2019 (COVID-19) epidemic poses a threat to the everyday life of people worldwide and brings challenges to the global health system. During this outbreak, it is critical to find creative ways to extend the reach of informatics into every person in society. Although there are many websites and mobile applications for this purpose, they are insufficient in reaching vulnerable populations like older adults who are not familiar with using new technologies to access information. In this paper, we propose an AI-enabled chatbot assistant that delivers real-time, useful, context-aware, and personalized information about COVID-19 to users, especially older adults. To use the assistant, a user simply speaks to it through a mobile phone or a smart speaker. This natural and interactive interface does not require the user to have any technical background. The virtual assistant was evaluated in the lab environment through various types of use cases. Preliminary qualitative test results demonstrate a reasonable precision and recall rate.
ABSTRACT
Late-onset severe pneumonia (LOSP) is defined as severe pneumonia developing during the late phase of allogeneic hematopoietic stem cell transplantation (allo-HSCT). Because of the high mortality in patients with LOSP, it is important to identify prognostic factors. In this study, we aimed to develop a risk score system with broad applicability that can help predict the risk of LOSP-associated mortality. We retrospectively analyzed 100 patients with LOSP after allo-HSCT between June 2009 and July 2017. The assessment variables included immune, nutritional, and metabolic parameters at the onset of LOSP. Of these 100 patients, 45 (45%) eventually died, and 55 (55%) were positive for organisms, most commonly viruses. In the multivariate analysis, higher monocyte count (≥0.20 × 109/L versus <0.20 × 109/L; P = .001), higher albumin level (≥30.5 g/L versus <30.5 g/L; P = .044), lower lactic dehydrogenase level (<250 U/L versus ≥250 U/L; P = .008) and lower blood urea nitrogen concentration (<7.2 mmol/L versus ≥7.2 mmol/L; P = .026) at the onset of LOSP were significantly associated with better 60-day survival. A risk score system based on the foregoing results showed that the probability of 60-day survival decreased with increasing risk factors, from 96.3% in the low-risk group to 49.1% in the intermediate-risk group and 12.5% in the high-risk group. Our results indicate that this scoring system using 4 variables can stratify patients with different probabilities of survival after LOSP, which suggests that patients' immune, nutritional, and metabolic status are crucial factors in determining outcome.