Mortality, healthcare use, and costs of weaning center survivors and matched prolonged ICU stay controls: a database study

Background — Quantification of long-term survival, healthcare utilization, and costs of prolonged ventilator dependence informs patient/family decision-making, healthcare policy, and understanding of specialized weaning centers (SWCs) as alternate care models. Our objective was to compare survival trajectory, healthcare utilization, and costs of SWC survivors with a matched cohort of ≥ 21-d-stay ICU patients.

Methods — This was a retrospective longitudinal (12 y) case-control study linking to health administrative databases with matching on age, sex, Charlson comorbidity index, income quintiles, and days in ICU and hospital in preceding 12 months.

Results — We matched 201 SWC subjects to 201 prolonged ICU survivors (402-subject cohort); 42% had a Charlson score of > 4. Risk of death at 12 months was lower in SWC subjects (hazard ratio [HR] 0.70 [95% CI 0.54-0.91]) adjusting for length of hospital admission (HR 1.02 [95% CI 1.00-1.04]) and number of care location transfers (HR 0.84 [95% CI 0.75-0.93]). By follow-up end, more SWC subjects died, 149 (73%) versus 127 (62%). We found no difference in discharge to home. At 12 months, acute healthcare utilization was comparable for the entire cohort, except hospital readmission rates (median interquartile range [IQR] 2 [1-3) vs 1 [1-2] d). Median (IQR) cost 12 months after unit discharge was CAD $68,165 ($19,894-$153,475). 12-month costs were higher in the SWC survivors (CAD $82,874 [$29,942-$224,965] vs CAD $55,574 [$6,572-$128,962], P < .001). SWC survivors had higher community healthcare utilization. Regression modeling demonstrated cost was associated with stay and care transfers but not SWC admission. Over 12-y follow-up, healthcare utilization and costs were higher in SWC survivors.

Conclusions — SWC admission may confer some medium-term survival advantage; however, this may be influenced by selection bias associated with admission criteria.