26.05.2020

Incidence, clinical outcomes, and transmission dynamics of severe coronavirus disease 2019 in California and Washington: prospective cohort study

Epidemiology Transversal
Lewnard JA et al
BMJ

Main result

  • 1,840 patients with Covid-19 infection included out of the 9.5 million people in this private American healthcare system. Analyses of length of hospitalization and clinical outcomes included 1,328 patients (534 in Northern California, 711 in Southern California, 83 in Washington).
  • The cumulative incidence for acute hospitalization with Covid-19 was 15.6/100,000 cohort members in Northern California, 23.3/100,000 in Southern California, and 14.7/100,000 in Washington.
  • The median duration of hospitalization for survivors was 9.3 days (95% CI [0.8;32.9]) and 12.7 (95% CI [1.6;37.7]) for non-survivors.
  • The adjusted probability of admission to intensive care was 48.5% (95% CI [41.8;56.3]) for men and 32.0 (95% CI [26.6;38.4%]) for women. Critical patients stayed in the ICU for 10.6 days (95% CI [1.3, 30.8]).
  • The mortality rate was 23.5% (95% CI [19.6;28.2]) in men and 14.9% (95% CI [11.8:18.6]) in women. This rate increased with age.
  • A Re was also estimated in the different regions, ranging from 1.39 to 2.06 in early March and decreasing one month later to 0.78 to 0.90.

Takeaways

The incidence of hospitalization for Covid-19 is highest for the elderly, as is mortality. Admission to the intensive care unit is higher in males.
The Re has decayed with the containment measures.

Strength of evidence Moderate

- Prospective Cohort
- Reliable data from an insurance system, with a consistent database
- Lack of consideration of important confounding factors (co-morbidities)
- Possible information bias during the early stages of the epidemic when diagnostic tests were less common.
- Selection bias also possible because people in this private health system have a higher socio-economic status than those without insurance.

Objectives

Understanding the epidemiology and consequences of severe Covid-19 during the first epidemic wave on the West Coast of the USA

Method

  • Prospective cohort study based on a database of a private healthcare system, the Keiser Permanente healthcare system, in California and Washington State.
  • Collection of all patients from this system admitted to hospital with a diagnosis of Covid-19 infection until April 22, 2020 (or April 9, 2020 for final clinical patient data).
  • Cumulative incidence was calculated among the entire cohort and stratified by 10-year age groups.
  • For duration of hospitalisation, probability of admission to intensive care and mortality, right censoring of patients still hospitalised was taken into account by resampling and unconditional weighting to the general distribution. Weibull distribution parameters were adjusted by maximum likelihood. Generalized linear regression models with a Poisson link function were used to estimate the conditional probability of each hospitalization outcome by duration, age and sex.

bibliovid.org and its content are bibliovid property.

Legal Notice