Pamela Matthews is one of our fantastic experts and a senior embryologist now based in Australia. She has been in the IVF industry for almost three decades, working with IVF pioneers around the globe and feels passionate about how IVF success rates are reported to potential patients and here she explains why.
‘We all familiar with the quote made famous by Mark Twain, “There are three types of lies, “Lies, Damned Lies and Statistics”. Even the most well informed health professionals struggle to untangle a clinics ‘success rates’.
I quote one of my colleagues; “We tortured the data until it confessed”, and this is the approach that must be taken in order understand a clinics ‘success rate’.
Reporting ‘success rates’ is an important part of good IVF practice but it is difficult to understand and far from perfect.
It is even more imperfect when used as a marketing tool. Statistics answer very specific questions and knowing the question is essential. Many patients have contributed to that statistic and an individual can fall anywhere within a wide range of data. A single statistic cannot represent the multiple variables influencing the outcome of any treatment cycle. The same treatment protocols in the same patient in consecutive cycles can result in very different outcomes.
IVF outcomes can be expressed in a myriad of ways and it is important to understand exactly what ‘success’ is being reported.
In any IVF treatment cycle there are many hurdles to jump. Firstly a woman must respond to the fertility drugs and have eggs retrieved, eggs must be fertilized and develop into embryos, the length of time the embryos are grown before transfer will influence how many embryos are available for transfer and cryopreservation, how many embryos are transferred each time is also a factor. Then if there is a +ve pregnancy test, there are three stages that can be reported as a pregnancy.
The very first stage is a raised HCG level or chemical pregnancy, if a fetal heart beat is detected on scan it is a clinical pregnancy and finally the most important outcome, a live born baby.
HCG stands for Human Chorionic Hormone, and is produced by the placenta of a developing fetus. A high dose of HCG is also used as the trigger for egg retrieval and stays in the system for some time, which could be detected in an early pregnancy test. A home pregnancy test indicates the presence of HCG but a blood test is more accurate and informative.
Generally, a HCG level below 10 mIU/mL is regarded as a negative pregnancy test, 10-25 mIU/mL is a borderline pregnancy and more than 25 mIU/mL is taken to be a positive pregnancy. Both the day the blood is taken for the HCG test and the concentration are important. Some clinics report anything over 10 as a pregnancy, some anything over 25 and others only report a pregnancy after two readings showing a healthy rise in the level. HCG levels on average double every 48-72 hours. One clinic web site reported that a HCG level of more than 2 mIU/ml, 10-11 days after the trigger as a pregnancy!!! Other clinics can wait until anything up to 20 days after the trigger before testing. Clearly there is significant variation in HCG testing protocols, which will influence a clinics chemical pregnancy rates.
Around 10-20% of chemical pregnancies do not progress. How early the blood test is done and the level of HCG accepted, as a pregnancy will have a large impact on what this rate will be.
A diligent clinic that tests initially on day 16 and requires the level to rise significantly between two tests would be expected to have a lower loss of chemical pregnancies than a clinic that tests on day 12 and accepts 10mIU/ml as a positive test.
A comparison of clinical pregnancy success rates defined as the detection of one or more fetal heart beats, is a more accurate and less pliable parameter than chemical pregnancy success rates. However, it is not necessarily favored by clinics, as it will always be lower.
Finally and sadly the detection of a heartbeat does not always result in the birth of a baby. Generally if a clinical pregnancy does not result in the birth of a baby, a miscarriage has occurred. If a fetal heart has not been detected it is termed a biochemical pregnancy.
A ‘success rate’ is dependent on the parameters used to express it. The lowest and most meaningful to patients; is live born babies/cycle started, and the highest and most appealing marketing tool will be the biochemical pregnancy rate/embryo transfer.
These will be significantly different and it is important to understand exactly what definition of ‘success rate’ the clinic is quoting. Many clinics will only quote the success rate for a particular group of patients. This can be valid but what defines the group must be expressed in unambiguous terms. The mean maternal age of the patient group should always be defined.
Live Baby/Cycle started is the number of babies born for each IVF cycle started.
This should include the number of patients who do not get enough eggs to justify an egg collection, those who do not get fertilization, those that do not get a transfer. Ideally, it should include babies born from frozen embryo transfers, which can make it a very difficult statistic to obtain as frozen embryos can be in storage for many years.
Clinical pregnancy/Embryo transfer is a common “success rate” used and is the number of patients who have an embryo transfer where at least one fetal heart is detected.
The important questions to help determine the true comparative success rate of clinics is the average number of embryos/transfer and the number of patients who do not get a transfer.
Biochemical pregnancy rate/embryo transfer is by far the highest ‘success rate’ and many questions need to be asked.
The first and most important is what the clinic defines as a pregnancy, as rarely do the clinics volunteer such information. It is usually expressed as a pregnancy rate or success rate with scant explanation of their statistics. Questions that need to be asked are not only how many patients don’t get a transfer and how many embryos/transfer, but also how many of these pregnancies result in a live born baby.
If a clinic states that they are unable to follow up on live born babies, then the most important question is what they classify as a chemical pregnancy and ideally what the clinical pregnancy rate is. Those clinics that cannot provide at least a clinical pregnancy rate and the average number of embryos transferred/cycle are not trying very hard.
Finally, cryopreservation is an important part of every treatment cycle and is equally difficult to report.
The pregnancy rate/embryo transfer though a valid statistic is far from definitive in terms of the effectiveness of the cryopreservation program. The number of embryos selected for cryopreservation and the survival rate of the cryopreserved embryos is equally important. If only the very best embryos are selected for cryopreservation and transfer, the pregnancy rate/transfer will be high but the number of babies/cycle started may not be so high.
The cumulative pregnancy rate/cycle started is by far the most comprehensive statistic that can be compiled but also the most difficult as embryos can be in storage for many years and some will never be thawed for transfer.
This is important information for any patient about to start a treatment cycle as a live baby is the end of the journey they are about to commence and all of the hurdles must be jumped before they arrive.
Occasionally a clinic may offer an even more comprehensive cumulative pregnancy rate, the take home baby rate after 1, 2, or 3 treatment cycles, i.e. 40% after one cycle, 60% after two cycles, 70% after 3 cycles of having at least one baby.
This should include every parameter, cancelled cycles, cycles without eggs, with no fertilization, without a transfer, pregnancies from frozen embryos. To enable this, clinics must have excellent data collection and management, which is a vital component of good clinical practice.
These are all statistics and an important part of maintaining good clinical practice.
Protocols should always be developed from evidence-based medicine. Clinics are often very anxious to tick the boxes with the services they provide and often introduce new technologies and treatments with out first establishing that they are beneficial.
However, a patient is not a mean of many treatment cycles, but an individual with unique needs and requirements.
Clinics can have very different patient pre-requisites for treatment. If only good prognosis patients are accepted, success rates will be higher.
Donor egg cycles do better, maternal age and life style is an important indicator of success. Patients undergoing a 4th or 5th treatment cycle have much lower success rates than patients undergoing the 1st treatment cycle.
The patient group that makes up the statistics has an enormous impact on the ‘success rate’ and poor prognosis patients may be disadvantaged by the emphasis on this very imprecise statistic, tempting clinics to refuse treatment to patients that could compromise ‘success rates’.
If only very tightly selected embryos are transferred into very tightly selected patients the pregnancy rate/embryo transfer could be very high but the actual number of babies/cycle started may be quite low.
A clinic with an excellent ‘success rate’ may not have experience in dealing with poor prognosis patients, where as a clinic with a lower ‘success rate’ may do very well with this group.
IVF is a very difficult and demanding process, a patient must feel comfortable with and have complete trust in their provider. There is no doubt that good regulation and oversight by either a government or industry body is an important parameter in providing favorable outcomes.
Oversight by experts in the field not only provides a means of controlling rogue clinics and keeping marketing in check, it also raises the over all standards of clinical practice. This is important to take into account when considering ‘cross border treatment’.
If a clinic is operating in a regulated environment with high levels of accountability the standard of practice will be high even if the ‘success rates’ may not appear to be.
An investigation by the Australian Competition and Consumer Commission found some IVF clinics in Australia made misleading claims about their ‘success rates’ on their websites. As a consequence there is now a more uniform reporting requisite that requires both the clinical pregnancy and live birth rate to be presented. A chemical pregnancy is not reported at all. Most clinics present their results in the form of the table below.
This is adequate and gives a reasonable overview, however this particular clinic has presented their results using many different parameters. What is represented by each graph is clearly articulated.’
It is a great example of how different the same data can look when presented in different ways.
Each graph is part of the story, leading ultimately to the highly informative cumulative pregnancy and live birth rate after two retrieval cycles. Reporting of this complexity is necessary to get an understanding of the quality of care of any clinic.
The maternal age is clearly defined in this data set but the number of patients in each age group is not.
This can be significant as can the patient group. It would be impossible to clearly delineate all the patient groups, i.e. male factor, tubal disease, PCOSS, on a website but it unquestionably impacts on the outcomes.
Compulsory reporting and publishing outcomes is a fine balancing act between providing essential information and pressuring clinics to maintain ‘success rates’ which can lead to a bias against treating poor prognosis patients.
A good clinic is one that puts the patients interest above all else.
Do you have any questions on IVF success rates for Pamela Matthews? She would love to hear from you. Email your question here