Models of care for non-invasive ventilation in the Acute COPD Comparison of three Tertiary hospitals (ACT3) study
ABSTRACT
Background and objective
Non-invasive ventilation (NIV) improves clinical outcomes in hypercapnic acute exacerbations of COPD (AECOPD), but the optimal model of care remains unknown.
Methods
We conducted a prospective observational non-inferiority study comparing three models of NIV care: general ward (Ward) (1:4 nurse to patient ratio, thrice weekly consultant ward round), a high dependency unit (HDU) (1:2 ratio, twice daily ward round) and an intensive care unit (ICU) (1:1 ratio, twice daily ward round) model in three similar teaching tertiary hospitals. Changes in arterial blood gases (ABG) and clinical outcomes were compared and corrected for differences in AECOPD severity (Blood urea > 9 mmol/L, Altered mental status (Glasgow coma scale (GCS) < 14), Pulse > 109 bpm, age > 65 (BAP-65)) and co-morbidities. An economic analysis was also undertaken.
Results
There was no significant difference in age (70 ± 10 years), forced expiratory volume in 1 s (FEV1) (0.84 ± 0.35 L), initial pH (7.29 ± 0.08), partial pressure of CO2 in arterial blood (PaCO2) (72 ± 22 mm Hg) or BAP-65 scores (2.9 ± 1.01) across the three models. The Ward achieved an increase in pH (0.12 ± 0.07) and a decrease in PaCO2 (12 ± 18 mm Hg) that was equivalent to HDU and ICU. However, the Ward treated more patients (38 vs 28 vs 15, P < 0.001), for a longer duration in the first 24 h (12.3 ± 4.8 vs 7.9 ± 4.1 vs 8.4 ± 5.3 h, P < 0.05) and was more cost-effective per treatment day ($AUD 1231 ± 382 vs 1745 ± 2673 vs 2386 ± 1120, P < 0.05) than HDU and ICU. ICU had a longer hospital stay (9 ± 11 vs 7 ± 7 vs 13 ± 28 days, P < 0.002) compared with the Ward and HDU. There was no significant difference in intubation rate or survival.
Conclusion
In acute hypercapnic Chronic obstructive pulmonary disease (COPD) patients, the Ward model of NIV care achieved equivalent clinical outcomes, whilst being more cost-effective than HDU or ICU models.
