Use of the INVENT system for standardized quantification of clinical preferences towards mechanical ventilation C Allerød 1,2 , DS Karbing 1 , P Thorgaard 2 , S Andreassen 1 , S Kjærgaard 2 , SE Rees 1 1 : Center for Model-based Medical Decision Support, Aalborg University, Denmark 2 : Anaesthesia and Intensive Care, Region North Jutland, Aalborg Hospital, Aarhus University, Denmark
19
Embed
Use of the INVENT system for standardized quantification of clinical preferences towards mechanical ventilation C Allerød 1,2, DS Karbing 1, P Thorgaard.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Use of the INVENT system for standardized quantification of clinical preferences towards mechanical ventilation
C Allerød1,2, DS Karbing1, P Thorgaard2, S Andreassen1, S Kjærgaard2, SE Rees1
1: Center for Model-based Medical Decision Support, Aalborg University, Denmark2: Anaesthesia and Intensive Care, Region North Jutland, Aalborg Hospital, Aarhus University, Denmark
Introduction
• Several RCTs indicate that low tidal volumes are beneficial in mechanically ventilated patients
• Large variability in ventilator settings are still seen in the ICU
• But how well do physicians agree when presented to identical patients?
The ARDS Network. N Engl J Med 2000; 342:1301-8.
Young et al. Crit Care Med 2004; 32:1260-5.Esteban et al. Am J respir Crit Care Med 2008; 177:170-7
Using INVENT for quantifying preferences
• INVENT: Model-based decision support system for mechanical ventilation– Advice on FiO2, Vt and f.
J Clin Monit Comput 2006; 20:421-9.J Crit Care 2010; 25:367-74.
Research questions
• What is the variability in physicians’ preferences in the same patient?
• What are physicians’ opinion about other physicians’ advice?
• What are the physicians’ opinion about INVENT’s advice?
Methods• Included 10 senior ICU physicians, representing the 4
Danish University Hospitals
• 10 real patient cases, presented in random order– Reflecting range of respiratory, circulatory and metabolic status in
patients with ALI/ARDS
• Asked to set FiO2, Vt and f– Assuming correct PEEP and I:E ratio, 70 kg body weight and that
model simulations were correct
• Evaluated and ranked other physicians’, and INVENT’s advice (blinded, and in random case order)
Results
• What is the variability in physicians’ preferences in the same patient? Example: Vt
Variability in physicians’ preferences
6 ml/kg
8 ml/kg
Variability in physicians’ preferences
6 ml/kg
8 ml/kg
Results
• What is the variability in physicians’ preferences in the same patient? Example: Vt
• What are physicians’ opinion about other physicians’ advice?
Opinion’s about other physicians’ advice
Classification Average (range)
Good 21.2 % (9.9% – 26.0%)
Acceptable 45.5 % (36.4% – 56.8 %)
Unacceptable 33.3 % (25.4% – 39.9 %)
Opinion’s about other physicians’ advice
Classification Average (range)
Good 21.2 % (9.9% – 26.0%)
Acceptable 45.5 % (36.4% – 56.8 %)
Unacceptable 33.3 % (25.4% – 39.9 %)
Rank Best Worst
Average 5.0 7.1
Range 2-10 2-10
Results
• What is the variability in physicians’ preferences in the same patient? Example: Vt
• What are physicians’ opinion about other physicians’ advice?
• What are the physicians’ opinion about INVENT’s advice?
Opinion about INVENT’s advice
Classification Average (range) INVENT
Good 21.2 % (9.9% – 26.0%) 26.0 %
Acceptable 45.5 % (36.4% – 56.8 %) 41.0 %
Unacceptable 33.3 % (25.4% – 39.9 %) 34.0 %
Opinion about INVENT’s advice
Classification Average (range) INVENT
Good 21.2 % (9.9% – 26.0%) 26.0 %
Acceptable 45.5 % (36.4% – 56.8 %) 41.0 %
Unacceptable 33.3 % (25.4% – 39.9 %) 34.0 %
Rank Best Worst INVENT
Average 5.0 7.1 5.3
Range 2-10 2-10 3-10
Conclusions
• Physiological models may be a beneficial tool for quantifying clinical preferences– Separate variation in preference from variation in patient
pathophysiology
Conclusions
• Physiological models may be a beneficial tool for quantifying clinical preferences– Separate variation in preference from variation in patient
pathophysiology
• Large variability found in physicians’ preferences towards appropriate levels of Vt, FiO2, f, PIP, SaO2 and pH
Conclusions
• Physiological models may be a beneficial tool for quantifying clinical preferences– Separate variation in preference from variation in patient
pathophysiology
• Large variability found in physicians’ preferences towards appropriate levels of Vt, FiO2, f, PIP, SaO2 and pH
• When a physician’s advice was evaluated by other physicians, a third were found unacceptable
Conclusions• Physiological models may be a beneficial tool for quantifying
clinical preferences– Separate variation in preference from variation in patient
pathophysiology
• Large variability found in physicians’ preferences towards appropriate levels of Vt, FiO2, f, PIP, SaO2 and pH
• When a physician’s advice was evaluated by other physicians, a third were found unacceptable
• INVENT ranked third best, evaluated similar to physicians– Prospective testing is necessary