ADDRESSING DISPARITIES IN IVF CYCLE OUTCOME COMPARISONS: Making the case for a fair informed consent for patients and a fair evaluation of clinics
David H Barad, MD, MS, is Director of Clinical IVF and Research and a Senior Scientist at the CHR. He can be reached through the editorial office of the The Reproductive Times (Hello@ReproductiveTimes.com)
Briefing: The author of this article explains the intricacies of fair patient consent and IVF clinic evaluations—both, as the article well demonstrates, closely linked. The article, moreover, is not meant as criticism of the current reporting system of SART (and for that matter also of the CDC’s Congress-mandated system), but as an encouragement for further improvements in the systems. Only once outcome analyses will allow for outcome stratification by detailed patient characteristics and, therefore, for appropriate adjustments, will these registries be able to further enhance IVF outcome information and, therefore, be able to allow the public to compare outcomes between clinics. Until then, one is only left to endorse the statement accompanying both registries—therefore—since their creation that they should not be used to compare IVF cycle outcomes between clinics.
The landscape of assisted reproductive technology (ART) is inherently complex, with success metrics that depend not only on clinical expertise but also significantly on patient demographics and treatment philosophies. This complexity becomes particularly evident when outcomes are scrutinized, such as in quality assurance audits by the Society for Assisted Reproductive Technology (SART), to which a majority of IVF clinics (though by no means all) annually voluntarily report their cycle outcome data. SART, a daughter society of the American Society for Reproductive Medicine (ASRM), in turn, accumulates all data and publishes detailed annual reports. The society then in an additional function statistically evaluates aggregate data and informs individual clinics if they are outliers.
And because the Center for Human Reproduction (CHR)’s patient population is absolutely unique in comparison to almost 500 clinics reporting to SART, the CHR very frequently received notifications for an outlier status. The CHR’s most recent one was triggered by a relatively low reported live birth rate per embryo transfer among women under 38 years for the year 2022 and served not only as the motivation for this article but also demonstrates how such data analyses without proper context can be misleading.
Patient demographics: an uneven playing field
The profile of patients the CHR serves has, of course, a profound impact on reported outcomes. In 2022, 41% of cycles performed for women under 38, demonstrated significantly diminished ovarian reserve (DOR), at that young age defined by an AMH level of less than ~1.5 ng/mL (1). At such a young age, this subset of patients – due to obviously prematurely aging ovaries (POA) - inherently faces lower implantation and live birth rates than women with normal functional ovarian reserve (FOR). Reasons for these lower rates are reduced oocyte quality as well as quantity in association with POA.
Unsurprisingly, many of these patients arrived at the CHR’s shores from all around the world with lengthy histories of unsuccessful treatments at other IVF clinics. Their reproductive backgrounds, moreover, usually go beyond just multiple failed prior IVF cycles; in over half of all cases they, indeed, arrive with advice to pursue donor eggs and refusal by their prior clinics to even consider further treatments with autologous oocytes.
Practices that report exceptionally high live birth rates can – typically - be assumed to selectively treat patients with only more favorable prognoses. These clinics often turn away individuals seeking IVF with their own oocytes if they do not meet certain minimal criteria, whereas at CHR, due to its inclusive patient approach that – after fully transparent informed consent – prioritizes patients’ desires, such individuals usually will still be given access to treatments with autologous oocytes, as long as a pregnancy chance appears still possible, even if remote.
This philosophy has positioned CHR as one of the world’s leading “last resort” IVF centers for patients from all over the world. At the CHR, over 90% of new patients fall into this challenging category. Expectations for pregnancy and live birth expectations, therefore, must be accordingly adjusted. At the same time - rather than discouraging so-affected patients with overly pessimistic estimates – it is important to offer them conservative- realistic expectations about success with their own eggs in comparison to obviously much better chances with young donor oocytes.
Impact of prior failed IVF cycles and non-euploid transfers
During the queried period, many patients we treated at CHR arrived after experiencing multiple failed IVF cycles elsewhere—often three to five attempts—indicating possible underlying factors in their infertility that may still be unrecognized, whether maternal or embryonic in origin. These patients’ medical histories stand in sharp contrast to those undergoing their first IVF attempts, when success rates are naturally expected to be much higher.
Uniquely since 2014, the CHR’s cycle outcomes also include transfers involving allegedly non-euploid embryos. Questioning the accuracy and validity of routine preimplantation genetic testing for aneuploidy (PGT-A) – then still called preimplantation genetic screening (PGS), the CHR in collaboration with other colleagues started in 2014 to selectively transfer some embryos in testing described “aneuploid.” In 2015 – as the first center in the world – the CHR reported chromosomal-normal births after such transfers (2) and has since steadily transferred allegedly chromosomal-abnormal embryos. After establishing a registry for such cases, we, moreover, regularly have been reporting cycle outcomes in the literature.
Since the CHR – except in rare circumstances – usually does not recommend the use of PGT-A to its patients, here, too, almost all cases involve patients at other centers refused such transfers who moved their allegedly “abnormal” embryos to the CHR. In the year 2022, two women under 38 years old had such transfers, which resulted in one healthy live birth. Like the CHR’s high acceptance rate of patients elsewhere refused treatments with their own eggs, this practice also underscores the CHR’s patient-centric model that respects informed patient decisions, yet by definition places the center’s reported metrics at odds with clinics transferring only genetically screened, euploid embryos.
A more nuanced view of IVF metrics
The SART benchmark of an 18% pregnancy rate, taken as a lower 95% confidence interval for women under 38, fails to reflect the heterogeneity of treatment approaches and patient inclusion criteria. Practices that focus solely on transferring high-quality, euploid blastocysts will naturally report more favorable statistics compared to those that accept challenging cases or – for example - allow transfers of non-euploid embryos.
This difference in clinical practice represents not just varying success rates, but often reflects fundamentally different philosophies of patient care and should be better addressed in the comment section of SART reports because very few patients (and likely even only a minority of physicians) understand this point. Unsurprisingly, despite great improvements in SART reporting, patients are, still, only insufficiently informed about expected outcomes in IVF and colleagues – for whatever reasons – in the CHR’s opinion, still direct many infertile women into third-party egg donation too early.
How then should an infertility center be evaluated?
To present a more accurate assessment of outcomes, stratifying patient populations into subgroups based on patient characteristics such as AMH levels and prior infertility/IVF history is essential. Highlighting these distinctions clarifies whether a clinic prioritizes clinic outcomes or goes beyond that by attempting to optimize cycle success rates of individual patients, based on highly individualized treatment pathways. Additionally, it is crucial to contextualize results with evidence-based literature, showing the challenges faced by patients with diminished ovarian reserve and histories of previous treatment failures.
All of this points toward a need for fair benchmarking by discussing the selection biases inherent in higher success rates featured by many clinics as marketing tools. Patients are not only entitled to choose what kind of clinic they wish to attend; they also should be able to make the judgment with relative ease and that requires significant additional improvements to the current SART and CDC reporting systems.
And how should success be defined?
The disparities in IVF outcome comparisons underscore the importance of a nuanced understanding of reported statistics. The pursuit of equitable, patient-centered reproductive care means embracing all infertility cases, not only the easy ones. More complex cases, of course, will not always enhance reported success rates but will fulfill a vital role in patients’ treatment expectations. The analogy for doing otherwise, would be for a cardiac surgeon to operate only on patients with single vessel disease and then claim better outcome in comparison to a surgeon who even treats patients with multiple small vessel disease.
Success of an IVF clinic is not only defined by the overall success in achieving pregnancies and live births, but in the success rate in giving patients what they (not we doctors) define as success. In other words, directing a woman prematurely into egg donation because egg donation - of course – will improve pregnancy and live birth chances for almost everybody - is not really “success” even if this woman conceives on a first try. If she then for the rest of her life ends up wondering whether she could not have achieved the same goal with her own eggs, we, indeed, would consider such an approach an almost inexcusable failure of patient care.
References
2. Gleicher et al., Fertil Steril 2015;104:e5