Sinclair DR, et. al, Anesth 1999; 91:109-118

Reviewed by: R. Prasad, MD

For the complete MathPad-ready version of their model (slightly modified), click here. Just copy it, paste into MemoPad, sync, then copy/past into MathPad.

Comments:
-Cumbersome formula (not bad with a Palm, though)
-Would like to see it compared with other predictive formulas, in other patient populations.
-Could be helpful in designing cost-effective PONV mgmt scheme (e.g., risk <10% get x med, 10-30 get y med, etc.) that could be implemented preoperatively, perhaps based on questionaire?

Study: Prospectively studied 17, 638 outpatients, looking at a variety of patient- and surgery- related factors. Population split into 2 groups; one to develop their model, another to test it. Did multiple regression analysis to develop a predictive model. Highlights/problems:
-90% pts had GA (10110) or MAC (6301)
-Certain procedures (e.g., breast augmentation, strabismus, lap BTL, varicose vein stripping, dental, shoulder orthopedic), as opposed to type of surgery (ortho, ophtho, ...) had high risk of PONV. The model may not reflect these well.
-Only 29.8% responded for 24hour followup, with variability in groups responding. For example, response rates were higher than average in: urologic, general, ENT, dental, orthopedic, ophtho procedures; lower than average for gyn, chronic pain block, local or regional anesthesia.
-Opioid use was found to be higher in pts with PONV, but was not accounted for in model. Also, history of motion sickness.
-Pts received morphine and dimenhydrinate prn in PACU. Our ambulatory patients usually receive neither one ... any implications?

The model (slightly modified):
PercentPONV= 100/ (1+ e(5.97+ 0.014*age+ 1.03*sex+ 0.42*smoker- 1.14*PONV- 0.92*duration- 2.36*GA- 1.48*ENT- 1.77*ophth- 1.90*plastic- 1.20*gyn- 1.04*knee- 1.78*shoulder- 0.94*other))
age in years
sex: 1=male
1=yes for: smoker, PONV, GA, ENG, ophth, plastic, gyn, knee, shoulder, other (ortho)
duration in hours