Page 125 - The diagnostic work-up of women with postmenopausal bleeding

Strengths and weaknesses
This study is the first economic analysis that prospectively compared hysteroscopy with expectant management in women with PMB, a thickened endometrium and a benign result of endometrial sampling alongside a randomised trial. Moreover, this trial was performed in a pragmatic fashion increasing external validity of the trial.
A limitation of this economic evaluation is that important cost components are based on the prices of one university hospital in the Netherlands.We limited the economic evaluation to a hospital perspective in which only direct health care costs were taken into account. It is possible that costs of treatment of endometrial cancer or follow-up strategies differ between hospitals or countries and, therefore, our results cannot unconditionally be generalised to all circumstances. Indirect costs such as the value of lost productivity from time off work were not included in this study. However, especially in women in whom a (pre) cancer was detected these costs may be substantial. Therefore, the costs estimated in this trial are probably an underestimation of the societal costs of performing a direct hysteroscopy as compared to expectant management.
Another limitation is the power of this study.When we started the randomised
trial, we assumed a percentage of recurrent bleeding of 40% that was reduced
to 20% after hysteroscopy.This estimate was based on three available studies.22-24 6 However, the percentage of recurrent bleeding in the untreated group in this study
was only 18%.
The prevalence of 6% (pre) cancer in a preselected group of women with PMB with a benign result of endometrial sampling is higher than anticipated based on previous literature.13-15 An explanation for this could be that in these meta-analyses only a small number of postmenopausal women were included and that blind dilatation and curettage (D&C) was used as a reference standard. D&C is nowadays almost completely replaced by hysteroscopy, because we know that D&C misses 50-85% of all focal intracavitary pathology.25,26 Another explanation could be that the prevalence of endometrial (pre) cancers is different in different populations.2,3,27
In this study we assumed that the prevalence of endometrial cancer was the same in the expectant management group, based on randomisation principles. Moreover, it was assumed that effects in this strategy would be the same as in the hysteroscopy only strategy. Although our study shows that SIS has high specificity and sensitivity (94% and 93%, respectively), SIS cannot be considered to be 100% reliable.Thus, it is possible that in the SIS strategy one or more (pre) cancers would have remained undetected.
Cost-effectiveness
123