Regardless of the modality used to assess the size of the specimen, preoperative estimates agreed with actual weight in 79% of cases
In a poster presented at the 2015 ACOG Annual Clinical Meeting in San Francisco, Neal Marc Lonky, MD, and colleagues from the Southern California Permanente Group assessed the clinical acumen of physicians in estimating uterine size prior to elective hysterectomy for benign indications. They found that the correlation between estimates and actual uterine weight was 0.79 (P<.001), with a very low conversion rate for the surgery.1
Lonky and colleagues collected preoperative uterine estimates and actual specimen weights prospectively for 1,079 cases of benign hysterectomy. The surgeries were performed by 186 primary surgeons and assistant surgeons at 10 Kaiser Permanente Southern California medical centers. Surgeons based the route of hysterectomy on estimates of uterine size, which were calculated using bimanual examination, ultrasonography, or both. Linear regression was used to measure and compare the relationship between estimated uterine size and the pelvic specimen weight.
Uterine size estimates ranged from 4 cm to 40 cm, and specimen weights ranged from 2 g to 4,607 g. The mean (SD) estimate of uterine size was 11.7 (4.43) cm, and the mean actual specimen weight was 334.6 (401.42) g.
The mean age of women in the sample was 47.2 (8.35) years. Overall, 379 women (35.1%) were Hispanic, 325 (30.1%) were non-Hispanic white, 281 (26.0%) were non-Hispanic black, and 81 (7.5%) were Asian/Pacific Islander. The mean body mass index (BMI) was 30.0 (6.37) kg/m2, with a range of 16.8 to 67.9 kg/m2.
“This is real world research,” said Dr. Lonky. “It’s called comparative effectiveness research. Basically, all patients who are undergoing the procedure are entered in the registry, and the clinical acumen of the physician—either using or not using ultrasound—is assessed.”
“We looked at whether or not we had a bias toward one patient age group, race/ethnicity, BMI, or estimated uterine size. But there were no clusters, so this was truly a random distribution,” said Dr. Lonky.
“These findings may be population-specific to my group of doctors,” he added. “They should be replicated in other settings. It may be that residents are not going to be as linear.”