Type 2 diabetes is a risk factor for hospitalization and death in COVID-19 infection, while the effect of type 1 diabetes is less significant. Machine learning models were used to identify predictors of outcomes among COVID-19-positive individuals with diabetes.
OBJECTIVE Type 2 diabetes is an established risk factor for hospitalization and death in COVID-19 infection, while findings with respect to type 1 diabetes have been diverging. RESEARCH DESIGN AND METHODS Using nationwide health registries, we identified all patients aged >= 18 years with type 1 and type 2 diabetes in Sweden. Odds ratios (ORs) describe the general and age-specific risk of being hospitalized, need for intensive care, or dying, adjusted for age, socioeconomic factors, and coexisting conditions, compared with individuals without diabetes. Machine learning models were used to find predictors of outcomes among individuals with diabetes positive for COVID-19. RESULTS Until 30 June 2021, we identified 365 (0.71%) and 11,684 (2.31%) hospitalizations in 51,402 and 504,337 patients with type 1 and 2 diabetes, respectively, with 67 (0.13%) and 2,848 (0.56%) requiring intensive care unit (ICU) care and 68 (0.13%) and 4,020 (0.80%) dying (vs 7,824,181 individuals without diabetes [41,810 hospitalizations (0.53%), 8,753 (0.11%) needing ICU care, and 10,160 (0.13%) deaths). Although those with type 1 diabetes had moderately raised odds of being hospitalized (multiple-adjusted OR 1.38 [95% CI 1.24-1.53]), there was no independent effect on ICU care or death (OR of 1.21 [95% CI 0.94-1.52] and 1.13 [95% CI 0.88-1.48], respectively). Age and socioeconomic factors were the dominating features for predicting hospitalization and death in both types of diabetes. CONCLUSIONS Type 2 diabetes was associated with increased odds for all outcomes, whereas patients with type 1 diabetes had moderately increased odds of hospitalization but not ICU care and death.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据
分享论文
