For a study, researchers sought to identify if the algorithm-guided withdrawal was related to changes in opioid or benzodiazepine use compared to conventional treatment. A stepped-wedge cluster randomized clinical trial design was used to allocate patients to algorithm and usual care groups randomly. Following the collection of usual care data, sites were randomly assigned to the algorithm. The Respiratory Distress Observation Scale (RDOS) was used to assess patient comfort at baseline when the ventilator was turned off, and every 15 minutes for an hour. From the start of the process until the patient died, parenteral morphine and lorazepam equivalents were utilized. Based on patient discomfort and baseline consciousness, the algorithm recommends either ventilator weaning or one-step discontinuation; medication usage is guided by RDOS scores.

About N=120 for routine care and 48 for algorithm data. Gender and race (White or Black) were distributed equally. All patients in the usual care arm were weaned in one step; 58% of patients in the algorithm arm were weaned in an average of 18+27 minutes. Patients in the algorithm arm exhibited lower RDOS ratings when the ventilator was switched off (t=–2.8, P=.006), after 15 minutes (t=–2.34, P=.021), and at 30 minutes (t=–1.92, P=.057). The control arm received more opioids (t=–2.30, P=.023) and benzodiazepines (t=–2.08, P=.040).

The method was efficient in guaranteeing the respiratory comfort of patients. Surprisingly, more medicine was provided in the usual care arm; however, when distress is objectively evaluated RDOS, as in the algorithm, less medication may be required.