1 Department of Obstetrics and Gynaecology, St Joseph Hospital, PO Box 7777, 5500 MB Veldhoven and 2 Department of Obstetrics and Gynaecology, Academic Hospital, University of Groningen, Groningen, The Netherlands
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: IVFembryo transfer success rates/prognosis/reproducibility/spontaneous conception chance
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
An alternative for the use of prognostic models could be clinical experience. This clinical experience or `gut-feeling' of clinicians was the only available `tool' before the introduction of prognostic models. In fact, this `gut-feeling' is based on characteristics which are also incorporated in prognostic models in a more formal way. Although identification of prognostic factors in daily clinical practice seems easy, quantification of the relative merit of these factors and their mutual dependence is likely to be more complicated.
The aim of this study was to evaluate the reproducibility of the estimated spontaneous fertility chances and success rates of IVFembryo transfer by clinicians with the help of a set of case histories. A poor reproducibility would indicate a strong need for prognostic models.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The four case histories are presented in Table I. They represent a rather wide spectrum of prognostic profiles. In case 1 the prognostic factors were chosen to be unfavourable, with a long duration of subfertility, a relatively aged female partner and a low sperm quality. In contrast, in case 2 the prognostic factors were supposed to be favourable, with a short duration of subfertility and a relatively young female and optimal sperm quality. Case 3 and 4 represented couples in the centre of the prognostic spectrum.
|
Inter-observer reproducibility was expressed by calculating intra-class correlation coefficients (ICC) and lower 95% confidence limits (95% CL) (Fleiss, 1981). The ICC expresses the degree to which the total variance can be attributed to the true variance: true differences between subjects. It not only assesses the strength of correlation between two measurements but also detects systematic errors. Thus, if a set of estimates made by one observer is systematically lower or higher than the estimates made by another observer, the ICC is correspondingly reduced. Values of the ICC are interpreted as k-statistics: slight (ICC 00.20), fair (ICC 0.210.40), moderate (ICC 0.410.60), substantial (ICC 0.610.80), or almost perfect (ICC 0.811.0). The ICC and 95% CL were calculated for gynaecologists and reproductive endocrinologists separately.
To counsel a subfertile couple in the decision of whether or not to proceed to IVFembryo transfer the added value of IVFembryo transfer as compared to the spontaneous conception chance has to be taken into account. In cases where the pregnancy rate after IVFembryo transfer is equal to or lower than the spontaneous pregnancy rate, IVFembryo transfer should not be offered. In cases where the pregnancy rate after IVFembryo transfer is higher than the spontaneous pregnancy rate, IVFembryo transfer should be considered if the spontaneous pregnancy rate is low, and if the pregnancy rate after IVFembryo transfer is sufficiently high.
To illustrate the impact of the estimations on clinical decision-making, we arbitrarily defined the indications for IVFembryo transfer. We defined IVFembryo transfer to was presumed to be indicated for couples whose estimated 12 month cumulative spontaneous conception rate was <25% and in whom three cycles of IVFembryo transfer were expected to generate a relevant improvement of the conception chances; at least 10% improvement in cases where the spontaneous chance was <10%, at least 15% improvement in cases where the spontaneous chance was 20% and at least 25% improvement in cases where the spontaneous chance was 25%. The area in which IVFembryo transfer is supposed to be offered is shown in Figure 1.
|
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
For case 3, a majority of the gynaecologists and all reproductive endocrinologists expected conception rates after IVFembryo transfer to be higher than spontaneous conception rates. The median difference between the expected IVFembryo transfer and spontaneous conception rates were 10% among gynaecologists, and 15% among reproductive endocrinologists (Figure 2C). The judgements would have indicated IVFembryo transfer in 21 (37%) of the gynaecologists and 16 (50%) of the reproductive endocrinologists.
For case 4, the situation is comparable to case 3, with gynaecologists estimating IVFembryo transfer to be better, whereas reproductive endocrinologists expect comparable conception after IVFembryo transfer and expectant management (Figure 2D). The judgements would have indicated IVFembryo transfer in 25 (44%) of the gynaecologists and 6 (19%) of the reproductive endocrinologists.
The differences in expected conception rates after IVFembryo transfer as well as the difference between expected spontaneous and IVFembryo transfer mediated conception rates were not statistically significant, with P values of 0.30 and 0.31 respectively. However, gynaecologists were significantly more optimistic towards spontaneous conception chances as compared with reproductive endocrinologists (P value 0.01).
As comparison, Table II also shows the conception chances as predicted by several prognostic models. For cases 1, 2 and 3 the models predicting spontaneous conception showed comparable outcomes to those made by the clinicians. For case 4, the models were predicting a lower chance, except for the model of Snick et al. (1997) using six variables. For IVFembryo transfer, the models of Templeton et al. (1997) and Haan et al. (1991) showed comparable results.
Table III shows the ICC and lower 95% CL for the gynaecologists, as well as the reproductive endocrinologists. The ICC of the expected spontaneous conception rates were high, with values of 0.71 and 0.66 respectively. In contrast, the ICC for estimated IVFembryo transfer success rates were low, with values of 0.24 and 0.14 respectively. The ICC for the difference between expected spontaneous conception rates and expected conception rates after IVFembryo transfer were 0.34 for the gynaecologists and 0.46 for the reproductive endocrinologists.
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
We used four case histories, of which the first two contained rather extreme profiles. The assessments of gynaecologists and reproductive endocrinologists were consistent in the sense that the case history with the worst prognostic profile was expected to have low conception rates, whereas the case history with the best prognostic profile was expected to have high conception rates.
The study was performed at the end of 1995. At that time, some prognostic models were already available, and it might be that the answers of some of the participants, especially those of the reproductive endocrinologists, were influenced by these models. However, the case histories were completed during a general session, and none of the participants had any of the prognostic models available.
We asked the participants to appraise expected chances for spontaneous conception and IVFembryo transfer, but we did not ask them what final counselling they would make. It might very well be possible that although participants generated different judgements of the chance of conception, their final treatment decision would not be different. However, the analysis of the differences between estimated spontaneous and IVFembryo transfer success rates according to pre-defined rules showed that there is considerable inconsistency among gynaecologists as well as among reproductive endocrinologists in the counselling of IVFembryo transfer. These differences are of clinical importance, especially in cases 1, 3 and 4, describing patients with the better prognostic profiles.
We found the reproducibility of the estimation of spontaneous conception rates to be substantial among both gynaecologists and reproductive endocrinologists. Since reproducibility is one of the key components of accuracy, this finding might indicate that estimated assessment of spontaneous conception chances is reliable (Kraemer, 1992). This is supported by the predictions of the prognostic models, which seem to be consistent with the estimates in three of the four case histories. However, one should also keep in mind that two of the four case histories had rather extreme profiles. In clinical practice, however, we often deal with patients in the middle of the spectrum, comparable to cases 3 and 4.
Consequently, estimates of the chance of spontaneous conception, which is important in reaching a decision on which couples should be considered for artificial reproductive treatment (i.e. those with a low chance are deemed more suitable), seem to be reliably made by both gynaecologists and reproductive endocrinologists.
In contrast, the reproducibility of assessment of success rates of IVFembryo transfer was only slight to fair, indicating that at present unaided prediction of the results of IVFembryo transfer is likely to be inaccurate. The prognostic models of Haan et al. (1991) and Templeton et al. (1997) also predicted success rates of IVFembryo transfer comparable to the estimates of both gynaecologists and reproductive endocrinologists. However, the validity of these predictive models has not been established. Stolwijk et al. (1996) recently tried to validate other prognostic models for IVFembryo transfer, but found disappointing results. Unfortunately, the present study shows that, especially for the prediction of IVFembryo transfer, validated prognostic models are warranted, since the reproducibility of estimates of pregnancy chances after IVFembryo transfer was found to be only slight to fair.
![]() |
Notes |
---|
3 To whom correspondence should be addressed
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Comhaire, F.H. (1987) Simple model and empirical method for the stimulation of spontaneous pregnancies in couples consulting for infertility. Int. J. Androl., 10, 671680.[ISI][Medline]
Eimers, J.M., Te Velde, E.R., Gerritse, R. et al. (1994) The prediction of the chance to conceive in subfertile couples. Fertil. Steril., 61, 4452.[ISI][Medline]
Fleiss, J.L. (1981) Statistical Methods for Rates and Proportions. John Wiley, New York.
Haan, G., Bernardus, R.E., Hollanders, J.M. et al. (1991) Results of IVF from a prospective multicentre study. Hum. Reprod., 6, 805810.[Abstract]
Kraemer, H.C. (1992) Evaluating Medical Tests. SAGE Publications, Newsbury Park, pp. 517.
Snick, H.K.A., Snick, T.S., Evers, J.L.H. and Collins, J.A. (1997) The spontaneous pregnancy prognosis in untreated subfertile couples: the Walcheren primary care study. Hum. Reprod., 12, 15821588.[Abstract]
Stolwijk, A.M., Zielhuis, G.A. Hamilton, C.J.C.M. et al. (1996) Prognostic models for the probability of achieving an ongoing pregnancy after in-vitro fertilization and the importance of testing their predictive value. Hum. Reprod., 11, 22982303.[Abstract]
Templeton, A., Morris, J.K. and Parslow, W. (1997) Factors that affect outcome of in-vitro fertilisation treatment. Lancet, 348, 14021406.[ISI]
Submitted on June 4, 1998; accepted on November 20, 1998.