Do discharge delays explain longer stays at veterans health administration hospitals?

Figures

• 

  Fig. 1. Cohort creation and data linkages. Abbreviations: VHA = veterans health administration affairs medical center; COVID-19 = coronavirus disease 2019. For the definitions of criteria used in this figure, see Additional file 4

• 

  Fig. 2. Relative increase in geometric mean length of discharging stay. Both panels depict the relative increase in geometric mean length of discharging stay for each pandemic year in comparison to the pre-pandemic year. The top panel includes all hospitalizations, while the bottom panel stratifies hospitalizations by discharge destination. (absolute values for geometric mean length of discharging stay are depicted in Additional file 8). The circles represent observed values unadjusted for potential confounders; the sizes of the circles are proportional to the number of annual discharges. The dashed line represents the relative increase in geometric mean length of discharging stay when both avoidable bed days and potential confounders are held to mean values from the pre-pandemic year. In contrast, the solid line represents the relative increase in geometric mean length of discharging stay when potential confounders are held to mean values from the pre-pandemic year while avoidable bed days are set to the mean value for the corresponding year of hospital discharge. The error bars represent 95% confidence intervals. The relative increase for all 3 pandemic years combined (not shown),…

• 

  Fig. 3. Proportion of overall increase in length of discharging stay attributable to avoidable days. The heights of the bars represent the annual increase in geometric mean length of discharging stay in comparison to the pre-pandemic year for all veterans discharged from acute medicine services. Colored bars represent the fractions of annual increases that are attributable to increases in avoidable bed days accrued before discharge, grouped by eventual discharge destination. Clear bars represent the fractions of annual increases that were not attributable to increases in avoidable bed days. For example, 6.4% (95% CI 6.0 to 6.8%) of the increase in length of discharging stay for all 3 pandemic years combined when compared to the pre-pandemic year is attributable to the increase in avoidable bed days accrued from veterans who were discharged home; 4.0% (3.7 to 4.3%) from veterans discharged to at-home post-acute care, and 5.2% (4.9 to 5.6%) from veterans discharged to facility-based post-acute care; the remaining 84.4% (83.4 to 85.3%) of the increase was not attributable to increases in avoidable bed daysci