Total intravenous versus inhalational anaesthesia for colonoscopy: a prospective study of clinical recovery and psychomotor function

A randomized, prospective study was conducted on 69 patients comparing recovery after two different anaesthetic techniques for ambulatory colonoscopy. Thirty-five patients received an intravenous fentanyl (1 microg/kg), midazolam (0.05 to 0. 075 mg/kg) and propofol (10 to 20 mg boluses as required) combination. 34 patients received sevoflurane in 67% nitrous oxide. Drug administration was titrated to clinical signs. At baseline and 30, 60, 90 and 120 minutes after the procedure patient performance on a comprehensive battery of psychomotor tests was recorded. Emergence times were noted. Depth of sedation was assessed at 5 minute intervals for 30 minutes after the end of the procedure. Emergence times were faster in the fentanyl/midazolam/propofol group by 2.2 minutes. A lower sedation score was detected at 20 minutes in the sevoflurane/nitrous oxide group. Psychomotor impairment was of a greater magnitude and more prolonged by 30 to 90 minutes in the fentanyl/midazolam/propofol group. It is concluded that a sevoflurane/nitrous oxide anaesthetic has a suitable recovery profile for ambulatory colonoscopy and results in faster recovery of cognitive function compared with a fentanyl, midazolam and propofol combination.

UTOPIA: a consultation system for visit-by-visit diabetes management

UTOPIA (UTilities for OPtimizing Insulin Adjustment) is a prototype computer system proposed to support home data analysis and therapy recommendations for the individual patient. The paper describes methods of analysis and their incorporation into an overall system design that matches the iterative practices at the physician-patient consultation from visit to visit. Four modules support home data display and comparison with clinical measurements; extraction of blood glucose trends and daily cycles using time series analysis, learning relationships between insulin adjustments and changes in time series patterns via a parametric, linear systems model; and advice generation by solving the linear equation for candidate insulin adjustments. Concepts and methods are placed in context, with a discussion of comparable and related research.