PubMed2004
OBJECTIVES: To compare three outpatient methods of endometrial evaluation in terms of performance, patient acceptability and cost-effectiveness. DESIGN: Pragmatic unblinded trial randomised separately within three groups determined by risk of endometrial cancer. SETTING: The gynaecology outpatient clinic of a large city hospital in Edinburgh, Scotland. PARTICIPANTS: Women referred for investigation and management of abnormal bleeding between January 1999 and May 2001. INTERVENTIONS: Investigations were: blind biopsy alone, hysteroscopy with biopsy, ultrasound evaluation including transvaginal ultrasound, and, in the low-risk group, the option of no investigation. Within this design, two devices for obtaining endometrial biopsy were compared, the Pipelle sampler and the Tao brush. MAIN OUTCOME MEASURES: Successful (informative) completion of the investigation, acceptability of the investigation method to women, women's satisfaction with clinic care in the short term and at 10 months and 2 years of follow-up, and cost-effectiveness to the end of investigation. RESULTS: Minor adverse events (e.g. shock, patient distress) did not occur for ultrasound, but occurred in 16% and 10% of women for hysteroscopy and biopsy procedures respectively. Pipelle biopsy provided an acceptable endometrial sample for 79% of moderate-risk women, but only 43% of high-risk women. The Tao brush gave similar performance in moderate-risk women (77%), but was more successful than the Pipelle sampler in postmenopausal (high-risk) women (72%). There were significantly more successful visualizations for ultrasound than for hysteroscopy in both the low-risk and the moderate-risk group, and a similar but non-significant trend in the high-risk group. Ultrasound was significantly better than hysteroscopy at detecting fibroids, but hysteroscopy significantly better for polyps. At the 10-month follow-up, high-risk women who had been investigated by hysteroscopy (with biopsy) had the most positive views of their clinic experience, but this effect had largely disappeared by 24 months. In the moderate-risk group, the subgroup randomised to biopsy alone gave the most negative responses about their clinic experience and health now. Women wishing they had more investigation comprised 22% of moderate-risk women and 38% of low-risk women, but only 14% of postmenopausal women. At follow-up the moderate-risk women (with menstrual bleeding problems), compared with postmenopausal women, had much worse ratings for clinic experience and health now. Resource use tended to be higher in the moderate- and low-risk women. There was minimal difference in cost-effectiveness between investigation options in the high-risk group, with the option involving hysteroscopy being marginally better than ultrasound. The most cost-effective investigation in the moderate-risk group was biopsy alone and in the low-risk group ultrasound. CONCLUSIONS: Decision-making about investigation would be clarified if postmenopausal women were studied separately from premenopausal women with menstrual bleeding problems. For postmenopausal women exclusion of cancer is a main objective, so once investigation has been completed discharge follows, but in the woman with abnormal menstrual bleeding, even if serious pathology is excluded, the original presenting symptoms require management. About 60% of premenopausal women with abnormal bleeding reported that their symptoms were not 'much improved' at 10 months. Research is needed to understand this phenomenon, and to explore ways to integrate patient factors into optimising evaluation and treatment. The significance of benign pathologies in this group also requires clarification. Given the relatively small differences observed in cost-effectiveness, there is justification for allowing other issues (such as clinician preferences and women's perspectives) to influence decisions as to the investigation method. There is scope to make better use of patient factors to inform decisions as to the most efficient and acceptable method of investigation for an individual woman. Additional analyses, using data available as a result of this study, will contribute to this agenda.