Key words
assisted reproductive technology (ART) - reproductive medicine - patient information
- register
Schlüsselwörter
assistierte Reproduktion - Patienteninformation - Register - Reproduktionsmedizin
Introduction
Any field of medicine which claims to be scientific is not only subject to the requirement
that it helps the sick, it must also give an account of itself to the general public
concerning the success of its efforts. It is not enough to simply want to help; the
results of medical activities must also be collected, processed, documented, and evaluated.
All fields of scientific medicine must explore whether and to what extent physicians
are actually providing the envisaged help to their patients. All of the above also
applies to reproductive medicine procedures.
The German IVF Register (Deutsches IVF-Register [D.I.R]) has been documenting the
use of IVF and related methods since 1982, that is, since shortly after the first
successful IVF procedures were carried out in Germany. The register analyzes results
and publishes them in the form of annual reports. No other institution can provide
information on procedures and their success rates for all of Germany. This article
looks at the institutional anchoring of the D.I.R in Germany, together with the D.I.Rʼs
documentation of IVF and related methods.
The D.I.R as an Institution
The D.I.R as an Institution
The D.I.R was set up in 1982 as a voluntary association of IVF centers [1] and was approved as an institution of the German Society for Gynecology and Obstetrics
in 1996 [2]. The Medical Association of the federal state of Schleswig-Holstein maintained and
managed the register from the establishment of a national head office in 1996 until
2013 [3], [4]. In 1998, a directive of the German Medical Association on how to carry out assisted
reproduction proposed, for the first time, that the medical associations of the different
German states should come together to establish a documentation center, which would
be known as the German IVF Register (D.I.R) [5] (p. 3169). All centers without exception would be expected to provide data: “every
working group should compile computer-based records in accordance with the questionnaire
of the German IVF Register” [5] (p. 3169). In actual fact, not all centers registered with the D.I.R [6]. In a non-public hearing of the select committee “Law and Ethics in Modern Medicine”
of the German Bundestag held on 26th March 2001, the chair of the D.I.R at the time
demanded “a ‘federal office of reproductive medicine’ following the British model
‘to prevent uncontrolled proliferation’ whose responsibilities would include the licensing
of centers, the documentation of outcomes, and regular, unannounced inspections of
the centers” [6] (p. 54). However, this proposal was not implemented.
In 2009 the D.I.R was incorporated as a registered association based in Berlin. An
association of this type is constitutionally not capable of pushing through the compulsory
membership of all centers of reproductive medicine. From 2013 on, the Medical Association
of Schleswig-Holstein no longer cooperated with the D.I.R [4]. In 2014, the (proposed) guideline was amended in such a way that from then on,
data collection by the D.I.R was no longer recommended; instead, the medical association
of each German federal state would be permitted to decide for itself where the data
should be collected [7]. With this amendment, the D.I.R lost its status as a national data collection center.
It nevertheless remains the only national register in Germany.
At present, there is no obligation for all centers of reproductive medicine in Germany
to pass on their data to the D.I.R [8]. In addition, those centers which are official members of the D.I.R do not necessarily
always supply complete datasets about the course of treatment or the course of pregnancy.
Although the D.I.R now collects the data of one year in September of the following
year, there are still considerable gaps in the datasets. And the next annual report
does not fill these gaps either. In a special evaluation published in 2014, the D.I.R
showed that even for pregnancies reported for 2012, only an average of 84.9 % of pregnancy
outcomes were reported per center [9].
The regulations in the (proposed) guideline on the implementation of assisted reproductive
technologies issued by the German Medical Association (German Medical Association
2006) also specify the type of documentation. According to the (proposed) guideline,
data should only be collected prospectively. The D.I.R had already resolved to do
this in 1997 [1]. The term “prospectively” was defined in such a way that data on the treatment cycle
had to be recorded within eight days after starting hormone treatment. This means
that the requirement is not for a “prospective” collection of data in the strictest
sense of the term. Moreover, it cannot be precluded that all data is only entered
after some delay. The following data should be reported:
“homologous insemination following hormone stimulation, IVF with ET, GIFT, ICSI, heterologous
insemination following hormone stimulation, heterologous IVF/ICSI, PBD, relating to:
patient age, indications for the method, course of stimulation, number and fertilization
rate for inseminated oocytes with IVF/ICSI, number of transferred oocytes following
GIFT, number of transferred embryos with IVF/ICSI, pregnancy rate, birth rate, miscarriages,
ectopic pregnancies, abortions, rate of multiple births, malformations” [10] (p. 1398).
The requirement that only prospective data should be entered is well founded, because
the temptation to manipulate data when data is collected retrospectively is well known.
In 2004, Stammer et al. pointed out the differences in pregnancy rates following IVF
for the year 1997, with mixed data resulting in a rate of 24.4 % compared to a rate
of 18.5 % for exclusively prospective data [11]. Despite this known distortion, calculation of the baby take-home-rate (BTHR) for
2012 in the 2013 annual report is based on both prospective and retrospective data
[9] (p. 14).
The Definition of “Success” in a National Register
The Definition of “Success” in a National Register
The D.I.R not only publishes outcomes, it also sets standards. As the only national
German register it is also involved in defining what should be considered a successful
outcome in reproductive medicine. What should be decisive: is it the quality of the
services offered in terms of supporting parents in dealing with a disorder [12]? Or is it the birth of a child or the inception of pregnancy? For the patient, inception
of pregnancy is certainly not the crucial outcome. The success of reproductive medicine
depends on the choice of criteria.
The rate of clinical pregnancy per embryo transfer which IVF centers also like to
report on their websites [13] is, naturally, rather higher than the rate of live births. But for patients the
rate of live births is what is important. However, there are also different ways of
calculating rates of live births [14], for example, as a percentage of all initiated treatments or of all embryo transfers
carried out. If only embryo transfers are included (live births divided by embryo
transfers), then all treatments in which embryo transfer was not even attempted will
be missing from the calculation (this corresponds to the calculations in the last
report of the D.I.R [9]).
Cumulative live birth rates are another way of quantifying success. They reflect the
cumulative success for the respective number of cycles carried out. But this figure
depends on the willingness (and the opportunity) to undergo the respective number
of treatment attempts. If a center reports a cumulative live birth rate, information
on the percentage of patients who discontinued treatment at an earlier stage is indispensable.
The analyses done by the D.I.R do not contain any cumulative live birth rates, but
the D.I.R reports on the number of treatments carried out per woman from 1997 to 2013
[9] (p. 29). According to the figures of the D.I.R, more than 40 % of women only underwent
one treatment cycle, 15 % underwent 3 treatment cycles, fewer than 9 % underwent 4
treatment cycles, less than 5 % had 5 cycles, and less than 1 % had 10 cycles.
The D.I.R as a Source of Information for Patients
The D.I.R as a Source of Information for Patients
Only the D.I.R is able to publish annual national data on the use of assisted reproductive
technologies in Germany. This means that the D.I.R could play an important role in
providing information to patients. Admittedly, the information would have to be processed
in such a way that it would be both comprehensible and relevant for patients.
At present, the information does not meet either of these criteria. In 2015, an uncritical
and much too short summary was appended to the main report for the first time [15]. Although the D.I.R does have the results for individual centers, it only publishes
national data, but this data, as previously mentioned, is incomplete and some of the
data was entered retrospectively. No information is given on the success rates for
individual centers. According to the current articles of incorporation of the D.I.R,
the “distribution or publication of data which would permit the identity of a center
to be inferred […] is prohibited” [16] (p. 5). German physicians working in reproductive medicine already demanded the
publication of a national standard a few years ago [17] to help patients choose a center. The publication of standards is quite common in
other countries. In the United Kingdom, for example, the authority which is responsible
for managing the official register also publishes information on the results of individual
centers [18]. For every center, patients can look at the probability of achieving a live birth
when using a specific method for various age groups and compare the figure to the
respective national average. The publications of the D.I.R do not provide this information.
Results for Assisted Reproductive Technologies in Germany
Results for Assisted Reproductive Technologies in Germany
The annual report categorizes the baby take-home-rate according to the method used;
the figures for 2012 were 17.78 % for IVF, 18.02 % for ICSI, 22.09 % for a combination
of IVF/ICSI, and 12.58 % for frozen embryo transfer [9] (p. 14, live births divided by treatments). The mean baby take-home-rate for all
methods was therefore 17.62 %. However, the following should be kept in mind:
-
Despite an “integrated verification that the data are collected prospectively”, which
has been extolled as a “unique selling point of the German IVF register” [1], the calculation is based on both prospective and retrospective data. This does
not comply with the resolution of the D.I.R from 1997 and the recommendation of the
German Medical Association in 2006, according to which all treatment cycles should
be recorded prospectively (defined as eight days after the start of hormone treatment
at the latest).
-
Out of a total of 81 035 reported treatments or 77 368 plausible treatment cycles
for 2012, 2116 outcomes (2.73 % of the plausible treatment cycles) were still unknown
in 2014, even though treatment was started in 2012 and the results must therefore
have been known by September 18, 2014 [9] (p. 13). The percentage of pregnancy outcomes not recorded in 2014 per method ranges
from 11 % (IVF/ICSI) to 17 % (IVF) of pregnancies. In addition, the D.I.R has reported
two other higher baby take-home-rates [9] (p. 14). In one version (A), cycles where the pregnancy outcome is unknown were
not included in the calculations. In this case, the mean baby take-home-rate for all
treatment methods was 18.33 %. In the second version (B), a “probable birth rate”
was assumed [9] (p. 14) (births per pregnancy) for the cycles where the real outcome was unknown
and this figure added to the number of known births. The mean baby take-home-rate
for all treatment methods was then calculated as 20.14 %. Based on the reasonable
assumption that unsuccessful treatment outcomes will be reported less often, both
of the above versions probably result in outcomes that are too high.
-
The reported baby take-home-rates do not refer to all initiated cycles starting from
the time of drug administration but start at a later stage of treatment. This means
that treatment cycles which were unsuccessful or were terminated at an earlier stage
(stimulation for egg collection, follicular puncture, oocyte retrieval, oocyte culture
or preparation for frozen embryo transfer) are omitted [9] (p. 13). What does that mean? According to the D.I.R, the mean baby take-home-rate
for all treatment methods is 17.62 % (this equals the mean of the lowest reported
baby take-home-rates for all methods used, live births for IVF, ICSI, IVF/ICSI) [9]. But if the mean baby take-home-rate is now calculated from the time of stimulation
for oocyte retrieval or the time of preparation for frozen embryo transfer (and not
simply from the time of transfer), the rate for 2012 would be 15.27 % (births divided
by plausible cycles). Moreover, the distortion due to the retrospective entry of data
also needs to be taken into account.
The method chosen by the D.I.R to calculate the baby take-home-rate thus suggests
unrealistically high success rates. This subverts the premise of informed consent.
Patients are already inclined to overestimate the prospects of IVF and related methods
[19], [20]. The publications by the D.I.R reinforce this tendency. It should be noted at this
point that in 1988 the German Medical Association already stated that the then current
success rate of 15 % was in “urgent need of improvement” [21] (p. 36); today, some 27 years later, the success rate is still just 15 %.
Conclusion
The D.I.R is the only register which aims to collect and publish data on the outcomes
after IVF and related methods for all of Germany. These data are indispensable for
a scientific evaluation of procedures, for political decisions on reproductive medicine,
for the general public, and for patients. Without a proper knowledge of the real prospects,
risks and burdens, patients cannot make an informed decision for or against specific
treatment methods.
The success rates for IVF and related methods were low from the outset. The temptation
to embellish them was correspondingly great. Already in 1985 Soules demanded greater
honesty with regard to in vitro fertilization pregnancy rates [22]. To obtain meaningful data on the use of assisted reproductive technologies in Germany
it will be necessary, as proposed by the German Medical Association in 2006, to collect
prospective data about the overall course of all treatment cycles [10]. That is the only way to calculate and obtain reliable national data on outcomes
for individual methods and age groups. The fact that not all centers participate in
the D.I.R can rightly be described as a “scandal” [23] (p. 52). Despite the important functions which the D.I.R could perform, its status
remains insecure. Originally, it was developed as a voluntary association between
IVF centers. Following a period in which the D.I.R was recognized and included in
guidelines issued by the German Medical Association, the D.I.R lacks this recognition
today.
The data of the D.I.R is in many ways inadequate: firstly, only data from a selection
of all centers located in Germany are recorded. Secondly, even when treatment cycles
are documented, not all of the data for the cycles (including births) is reported
to the D.I.R. Thirdly, the D.I.R still continues to use both prospective and retrospective
data, even though guidelines and their own resolutions should prevent this. Retrospectively
reported data are easier to manipulate and have been proven to produce unrealistically
high success rates. Fourthly, the D.I.R does not publish outcomes for individual centers.
Thus patients cannot obtain information the way they can in the United Kingdom, where
success rates for individual clinics can be obtained from the website of the HFEA,
the national UK regulator which monitors fertility clinics in the UK [18]. Fifthly: the presentation of results is completely unsuitable for non-specialists
as it is too complicated. The summary for patients, published for the first time in
2015, is too short and, in addition, it is uncritical. The summary for patients does
not give any information on outcomes for individual centers.
The data of the D.I.R creates unrealistic expectations about successful outcomes because
it primarily shows the success rates for different stages in the procedure. The baby
take-home-rate following stimulation is, of course, lower than the successful outcomes
reported for individual stages. But only the lowest rate is relevant for the woman
or couple. The individual stages are important for scientific research, but they are
not suitable as an indicator to inform patients about the expected success rate. To
sum up the issues with the presentation of outcomes: it is not just the quality of
the published figures that is inadequate. If all data were entered prospectively under
careful monitoring and the data were complete, this would represent a scientific achievement,
but the data would still not necessarily be useful for patients. By proceeding in
this way, science is taking itself and not the patient as the reference point of success
[24]. But in doing so, medicine is distancing itself from its real job. The actual success
rate of approximately 15 % per treatment cycle in Germany is sobering, given the figures
used to promote IVF (cf. [13]).
The documentation and communication of outcomes following IVF or related reproductive
medicine treatments was criticized from the outset (cf. [25], [26]). The select committee of the German Bundestag commented in 2002 that not all centers
in Germany passed on their data to the D.I.R [6]. This deficit also has an impact on political decision-making as policymakers do
not have recourse to high quality data. Since the report of the select committee “Law
and Ethics in Modern Medicine”, the D.I.R has had 14 years to supply data of an acceptable
quality. That has not yet happened.
For years, there have been demands that Germany should have a regulator of its own
– similar to the British HFEA – which would also have a supervisory function (e.g.
[27]). The committees of the various federal medical associations are currently working
on new processes for quality management [9], for “[t]he medical associations have decided to improve the existing quality management
procedures for centers of reproductive medicine and favor a nationally coordinated
advancement of quality in reproductive medicine across all of Germany whereby the
new process would be differentiated more clearly from scientific issues and from any
function as a register than was previously the case” [28] (2nd paragraph). It is, however, doubtful whether the new form of quality management
will rectify the known deficiencies of the D.I.R. In an enlightened society it is
unacceptable that no adequate information is available about such an important modern
technology. If the self-governance of the medical profession is not able to provide
this information, then a public authority or an institute under state supervision
will be unavoidable.
Demanding a better documentation of success rates is by no means unrealistic as this
requirement has already been implemented in other countries [29]. There are no convincing arguments why high quality, comprehensive documentation
which would also be comprehensible to non-specialists is not possible in Germany.
Acknowledgements
The authors would like to thank one of the anonymous reviewers for pointing out an
opportunity for data manipulation during the transfer of data to the D.I.R, which
was not previously known to them. They would additionally like to thank Prof. Dr.
Dorothee Kimmich from the German Department of the University of Tübingen.
