ABSTRACT
Background Stratification of patients with post-acute sequelae of SARS-CoV-2 infection (PASC, or long COVID) would allow precision clinical management strategies. However, long COVID is incompletely understood and characterised by a wide range of manifestations that are difficult to analyse computationally. Additionally, the generalisability of machine learning classification of COVID-19 clinical outcomes has rarely been tested.Methods We present a method for computationally modelling PASC phenotype data based on electronic healthcare records (EHRs) and for assessing pairwise phenotypic similarity between patients using semantic similarity. Our approach defines a nonlinear similarity function that maps from a feature space of phenotypic abnormalities to a matrix of pairwise patient similarity that can be clustered using unsupervised machine learning.Findings We found six clusters of PASC patients, each with distinct profiles of phenotypic abnormalities, including clusters with distinct pulmonary, neuropsychiatric, and cardiovascular abnormalities, and a cluster associated with broad, severe manifestations and increased mortality. There was significant association of cluster membership with a range of pre-existing conditions and measures of severity during acute COVID-19. We assigned new patients from other healthcare centres to clusters by maximum semantic similarity to the original patients, and showed that the clusters were generalisable across different hospital systems. The increased mortality rate originally identified in one cluster was consistently observed in patients assigned to that cluster in other hospital systems. Interpretation Semantic phenotypic clustering provides a foundation for assigning patients to stratified subgroups for natural history or therapy studies on PASC.
ABSTRACT
Background: Studies have shown that patients with ischemic stroke (IS) and concurrent COVID-19 have increased stroke severity. These analyses were limited by use of prepandemic era controls or by utilization of a sample from the early pandemic period when stroke care delivery was affected by lockdown. Studies on the severity of hemorrhagic stroke (HS) in patients with concurrent COVID-19 are few and limited by small sample sizes. Methods: Using the National Institute of Health (NIH) National COVID Cohort Collaborative (N3C) database, we identified patients diagnosed with stroke between Mar 1, 2020 - Feb 28, 2021. Hospitalized stroke patients with concurrent COVID-19 (stroke within 3 months after or one week prior to positive SARS-COV-2 PCR or AG lab test) were matched to all other hospitalized stroke patients in a 1:3 ratio. Nearest neighbor matching with a caliper of 0.25 was used for most clinical and demographic factors;exact matching for race/ethnicity and site. Within our matched sample, we used Poisson regression to calculate stroke severity incident rate ratio (IRR). Results: Our query identified 10,394 patients hospitalized with IS with available NIHSS scores upon admission (802 with concurrent COVID-19 and 9,592 without) and 2138 patients hospitalized with HS (181 with concurrent COVID-19 and 1957 without). Average NIHSS was greater in concurrent groups with both IS and HS (11.1 vs 7.68, p < 0.001 and 15.7 vs 11.7, p < 0.001 respectively). Propensity matched analysis also demonstrated that stroke patients with concurrent COVID-19 had increased initial NIHSS (IS: IRR = 1.4, 95% CI:1.3-1.5, p-value < 0.001;HS: IRR = 1.3, 95% CI:1.2- 1.5, p < 0.001). Average NIHSS in both IS and HS patients was greater in the Mar-Apr 2020 epoch than in all other 2 month epochs studied in these respective groups. Conclusions: This analysis suggests that the association between increased stroke severity and concurrent COVID-19 that was observed during the early pandemic was present throughout the pandemic as stroke care utilization normalized. Further work will center on the interaction between COVID-19 illness severity and stroke severity.