Attitudes, knowledge and practice regarding the anti‐müllerian hormone test among general practitioners and reproductive specialists: A cross‐sectional study

To describe clinicians' attitudes, knowledge and practice relating to the anti‐müllerian hormone (AMH) test.

the potential number of retrievable eggs from a controlled ovarian stimulation cycle. 3This helps set expectations and enables adjustment of the gonadotrophin dose, which may reduce the risk of ovarian hyperstimulation syndrome, 4 although large variation in ovarian response remains unexplained. 57][8] Women with low AMH have the same likelihood of conceiving and time to pregnancy as women with normal AMH. 9,10In addition, although AMH levels are associated with menopause timing at a population level, the estimates have wide confidence intervals and cannot predict the extremes.Therefore, it cannot reliably identify a woman at risk of premature menopause or predict age of menopause in individual women. 11,12espite evidence showing that the test is not predictive of fertility, it is increasingly promoted as a way for women to access their reproductive potential or chances of conceiving. 13Up to 36% of Australasian fertility clinic websites were found to include misleading statements, including that the AMH test indicates current and future fertility or is suitable for women who are simply curious about their fertility. 13In addition, online companies in the USA, 14 Australia, the UK and elsewhere have recently started selling the test direct-to-consumers without the involvement of a healthcare provider, promising women detailed insights into their fertility potential. 15Given that the internet is the most common source of fertility-related information, 16 such misleading claims may lead women to seek AMH testing for inappropriate reasons.
Consumers in Australia can either access the test with an out-of-pocket cost through a referral from their general practitioner (GP) or reproductive specialist, or purchase the test online through a direct-to-consumer website.Although accessing the test through a doctor is more likely to enable counselling about the test's limitations before the test being ordered, a small pilot survey about egg freezing with 72 Australian GPs found that 40% of participants incorrectly believed that AMH is a measure of oocyte quality or natural fertility. 17Given increasing concerns about AMH testing for inappropriate reasons, this study aimed to investigate GPs', gynaecologists' and reproductive specialists' attitudes, knowledge and practice relating to the AMH test.

| Design and participants
We conducted a cross-sectional online survey of a convenience sample of GPs, gynaecologists and reproductive specialists currently practising in Australia.Data were collected between May 2021 and April 2022.The study was approved by the University of Sydney Human Research Ethics Committee (2021/039) and all participants gave informed consent.

| Recruitment and procedure
The study was advertised through the mailing lists or websites of relevant professional organisations, including primary healthcare networks, Fertility Society Australia and New Zealand, the Royal Australian College of General Practitioners (RACGP) and the Royal Australian and New Zealand College of Obstetricians and Gynaecologists (RANZCOG).To reduce response bias, the study advertisement did not mention the AMH test (Appendix S1).The study advertisement included a link to the online platform Qualtrics where the survey could be completed anonymously in less than 10 minutes.Participants were given the opportunity to enter a prize draw for the chance to win one of five $100 gift cards upon completion of the survey.

| Measures
8][19] The survey was revised through discussions with the multidisciplinary study team and piloted with five clinicians for comprehension and length.One question used a hypothetical scenario to assess clinicians' decision-making following an inappropriate AMH request.Items assessing attitudes and knowledge included both supported and refuted statements based on current evidence.Sociodemographic and practice characteristics included: specialty, age, gender, years of experience, postcode of primary professional practice, and typical number of reproductive-aged female patients seen per week.All questions were mandatory except for the free-text items.

| Patient and public involvement
The multidisciplinary team included four clinicians with expertise in women's reproductive health.A further five practising clinicians pilot-tested the survey to ensure comprehension, length, quality and relevance.Patients were not involved in the study.

| Statistical analyses
Data were analysed using Stata/BE v17.Participants who did not complete at least 75% of survey questions were excluded from the analysis.Total knowledge scores were calculated by a count of correct responses.Descriptive statistics summarised the sample characteristics and the proportion of clinicians endorsing each item, overall and by specialty (dichotomised into [1] GPs or [2] gynaecologists and other reproductive specialists).A series of simple linear regression models were used to explore the association of clinician specialty and other demographic and practice characteristics (years of practice, gender, age, patient caseload and location) with knowledge.Values of p less than 0.05 were considered statistically significant.
Free-text responses were analysed using inductive content analysis. 20,21Two researchers (TC and LA) independently reviewed the free-text responses and developed an initial list of recurring codes and themes for each freetext question.These lists were integrated to form coding frameworks.Free-text responses were then coded using the frameworks by the two researchers independently, with 30% double-coded to confirm consistency.There was a substantial level of agreement for each question (κ = 0.73-0.85). 22

| Participant characteristics
Of the 464 clinicians who consented and began the survey, 13 were ineligible and 85 did not complete at least 75% of the survey questions, leaving 366 for analysis (79% completion rate).Four GPs indicated 'no' or 'not sure' to having heard of the AMH test, so were excluded from further analysis, resulting in a final sample of 362 participants.The median time taken to complete the survey was 8.7 minutes.
The final sample consisted of 177 GPs, 163 gynaecologists, 13 reproductive endocrinologists and 9 other specialists (e.g.fertility specialist, paediatric gynaecologist; see Table 1).The majority were female (79%), with a spread across years of experience.Patient caseload also varied, with participants seeing a median of 30 reproductive-aged women each week (range 0-120).

| Use of the AMH test in practice
Only 15% of GPs reported ordering the test at least once per month compared with 40% of reproductive specialists, with 15% of specialists ordering it daily (Table 1).Around 20% of GPs and 10% of specialists reported never having ordered an AMH test.
Just over half of GPs (53%) reported that their patients initiated the discussion about testing all the time or more often than them, whereas over two-thirds (69%) of specialists said they raised the idea of testing all the time or more often than their patients.
For GPs, the most common reason for ordering an AMH test was when referring to a fertility specialist, followed by investigation of infertility and to help with reproductive planning (Table 1).For specialists, investigation of infertility was the most common reason, followed by ordering the test to help with reproductive planning and to assess the impact of risk factors for diminished ovarian reserve.
Fewer than half of GPs indicated that they were quite or completely confident interpreting an AMH test result (39%) or explaining (41%) the result to their patients, whereas over half of specialists indicated that they were quite or completely confident interpreting (59%) or explaining (62%) the result to their patients.

| Attitudes about the utility of the AMH test
The proportions of respondents who indicated that the AMH test is moderately or very useful to predict low or high ovarian reserve were 55% and 25%, respectively, and 35% indicated that it is useful to predict response to ovarian stimulation (Figure 1; Table S1).For the reasons to test that are not supported by current evidence, 22% of participants indicated that the test was moderately or very useful in predicting premature menopause, 17% to diagnose polycystic ovarian syndrome, 9% in predicting chance of pregnancy and birth rates after assisted reproductive technology, and 5% in predicting the chance of natural conception and birth.

| Hypothetical scenario
Regarding the hypothetical scenario of an inappropriate AMH test request (Appendix S2), the majority of GPs (58%) indicated that they would not offer an AMH test, but 10% indicated that they would and 33% stated that they needed further information before making a recommendation.Approximately 50% of reproductive specialists indicated that they would not offer an AMH test, whereas 24% indicated that they would and 27% said they needed further information.
When asked to explain their answer, the top three explanations for not offering the test were that it had no clinical value (60% of 169 explanations, e.g.'complete waste of time and money'), was not indicated or required at this stage (22%, e.g.'she hasn't yet proven that she has a fertility problem') and that it causes harm, such as false reassurance or fear, pressure to conceive or leads to unnecessary fertility treatment (20%, e.g.'Most likely to cause unnecessary anxiety … May have unnecessary fertility treatment').See Appendix S3 for content analysis codes, frequencies and quotes.
The top three explanations for offering the AMH test were to guide the reproductive timeline (39% of 54 explanations e.g.'If her AMH is extremely low she might want to consider starting a family sooner or consider an early referral to a fertility specialist'), patient request (26%, e.g.'only because she raised it'), and only after counselling the patient about the test's limitations (22%, e.g.'I would counsel Lucy about the limitations of using an AMH in this situation, as it may falsely worry or reassure her, and a single reading will not tell us the rate of decline in fertility.However, if she still wished to have the test, I would order it.').
The top three explanations for needing further information were that it depended on the presence of other risk factors (44% of 88 explanations e.g.'Depends on cost, finances, her concerns, family history'), that it was not predictive of fertility or reproductive timeline (26%, e.g.'I would counsel her first that the test is not a guarantee of anything and if normal she may still have fertility issues for other reasons'), and that it depended on whether the result would influence her decisions (25%, e.g.'Depends … if the result would change her decision making-would she want to plan a pregnancy earlier or consider egg freezing').Despite most GPs and reproductive specialists indicating that they would not offer an AMH test or needed further information in the hypothetical scenario, almost 40% reported that they would be comfortable ordering the test if the patient asked for one.
When asked what they would explain about the potential benefits, harms and practical considerations of the AMH test, the top three responses were: that it is not an accurate, reliable or meaningful test (30% of 306 responses, e.g.'Expensive, poor reproducibility, low accuracy if on OCP, disagreement about normal ranges'), that it can be anxiety provoking (20%, e.g.'I know of many women who are worried by the result, feeling that their ovaries are letting them down or getting old who subsequently have naturally conceived, healthy pregnancies'), and that it is not predictive of fertility (20%, e.g.'Not predictive of pregnancy chances'; Appendix S3).
When informed that the hypothetical patient's test result comes back low for her age, the top three recommendations were for her to consider bringing forward plans to conceive (64%), to consider egg freezing (27%) and to consult a fertility specialist (26%, only answered by GPs; Figure 2).

| Knowledge
Correct responses to the items gauging knowledge were summed to calculate a total score out of 7 (mean ± SD, 3.5 ± 1.63; Table 2).Linear regression analyses showed that reproductive specialists had a total knowledge score that on average was 0.58 points higher than that of GPs (p = 0.001, 95% CI 0.24-0.91).Knowledge was also associated with female patient caseload, where for every ten additional patients, knowledge increased on average by 0.10 points (p = 0.007, 95% CI 0.03-0.18).There was no association between knowledge score and clinician age (per 10 years of age; p = 0.11, 95% CI −0.03 to 0.25), gender (p = 0.81, 95% CI −0.37 to 0.47), years of practice (p = 0.11) or state (p = 0.31; Table S2).
The vast majority of clinicians indicated that it was somewhat or very common for patients to be misinformed about the AMH test (Table 1; see Appendix S3 for patient misunderstandings content analysis).
When asked if there were any other important issues or comments participants wanted to raise, the top three topics were that they rarely ordered the test (13% of 273 responses, e.g.'Limited experience with ordering or interpreting this test'), that they would like to know more about the test (12%, e.g.'Expensive, unbiased (i.e.not from pathology or companies) information is needed for GPs and patients') and that it is over-ordered (9%, e.g.'A lot of patients are referred to me for discussion of AMH results by GPs who have ordered the test for them.I feel that there needs to be some education regarding exactly what circumstances the test is useful for'; Appendix S3).

| Main findings
The findings of this study suggest that only around half of clinicians surveyed feel confident interpreting and explaining an AMH result to their patients.Although knowledge scores were higher among reproductive specialists than GPs, only two-thirds of specialists correctly identified that AMH

| Strengths and limitations
Strengths of the study include its novelty and representation from a broad range of clinician specialties involved in women's reproductive health across Australia.A large sample of Australian clinicians was also recruited; however, because of the recruitment methods used, it was not possible to calculate a response rate.Clinicians who responded may be more engaged in women's reproductive health than clinicians from the broader population.The study advertisement deliberately did not mention the AMH test to minimise response bias.4][25] The clinicians who dropped out of the survey and were excluded were more likely to have less experience (i.e.be registrars; p = 0.02); and more likely to be male (p = 0.05; Table S3).We dichotomised knowledge items into correct or incorrect for the inferential analysis, which probably removed some of the nuance in responses.For example, a clinician who is not confident about the use of the test in a certain scenario (and, in practice, might seek further may consequently convey more accurate information to patients than a clinician who confidently holds view unsupported by evidence.For brevity, we also did not assess attitudes on all possible reasons for AMH testing.Finally, no adjustment was made to the familywise error rate for the series of linear regression models given their exploratory nature.

| Interpretation
Although clinicians acknowledged the limitations of the AMH test in predicting natural conception and would not offer the AMH test in the hypothetical scenario, many reported having used the test in their clinical practice for reasons not supported by the evidence (e.g.reproductive planning).Further, many were willing to provide the test to the hypothetical patient if they specifically requested it and, if the result came back low, would advise bringing forward plans to conceive, consider egg freezing or consult a fertility specialist.These findings appear contradictory.Potential explanations include that clinicians may think the test has at least some predictive ability or that they are willing to provide the test upon patient request, 26 particularly if they feel the patient is well informed of the limitations.][31] Another possible explanation for these incongruous findings could be that some clinicians may use the test as an educational tool to supplement general counselling around the age-related decline in fertility to encourage patients to maximise their chance of conception.
In addition to mixed attitudes and gaps in clinicians' knowledge, patient misperceptions were also reported to be common by clinicians.Free-text responses indicated that clinicians would like more education and evidence-based resources about when the test is useful, including information for patients.The American College of Obstetricians and Gynecologists strongly discourages testing in those without infertility, 32 but there are currently no committee recommendations or guidelines from the Australian colleges (e.g.RANZCOG, RACGP) regarding the AMH test.Concerningly, a recent population-level cross-sectional survey examining AMH test usage found that 30% of women aged 18-55 years in Australia had had an AMH test because they were curious about their fertility or wanted to understand their chances of conceiving. 33As the use of AMH testing is likely to increase with increasing interest in elective egg freezing and increasing marketing and availability of the test, 17 it is vital that clinicians are supported and adequately prepared to provide accurate information and interpretations of AMH to avoid unwarranted distress or false reassurance.

| CONCLUSION
Clinicians play a key educational role in women's reproductive health.These findings demonstrate gaps in knowledge and non-evidence-based use of the AMH test in clinical practice.These results indicate a need for clear, evidencebased guidance and resources about when testing is indicated to better support clinicians to adequately respond to patient requests and counsel about the test's limitations.

AU T HOR C ON T R I BU T ION S
TC conceived the study.TC, RT, JD, KH, MP, SL, DL, BWM and KM were involved in designing the study and developing the methods.TC and EC conducted the quantitative analyses.TC and LA conducted the content analyses.TC drafted the manuscript.All authors critically revised the manuscript and approved the final version.

AC K NO W L E D GE M E N T S
The authors would like to thank the clinicians who participated in this study.Open access publishing facilitated by The University of Sydney, as part of the Wiley -The University of Sydney agreement via the Council of Australian University Librarians.

F U N DI NG I N FOR M AT ION
This project was supported by a National Health and Medical Research Council (NHMRC) Centre for Research Excellence grant (1104136), Program grant (1113532), and Emerging Leader Research Fellowship (2009419).These awarded grants included external peer review for scientific quality.The funder had no role in the design and conduct of the study; collection, management, analysis and interpretation of the data; preparation, review or approval of the manuscript; and decision to submit the manuscript for publication.

C ON F L IC T OF I N T E R E S T S TAT E M E N T
BWM reports consultancy for ObsEva and Merck and travel support from Merck.DL is the Medical Director of City Fertility NSW and reports consultancy for Organon, Ferring, Besins and Merck.KH reports consultancy and travel support from Merck and Organon.All other authors: none declared.

DATA AVA I L A BI L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author.The data are not publicly available due to privacy or ethical restrictions.

E T H IC S A PPROVA L
The study was approved by the University of Sydney Human Research Ethics Committee (2021/039).

T A B L E 1
Participant demographic characteristics and use of the anti-müllerian hormone test in practice by specialty.
Opinions regarding the utility of the anti-müllerian hormone test by specialty (%).GP, general practitioner; PCOS, polycystic ovarian syndrome.
Selected recommendation (%) by specialty for hypothetical patient who receives a low anti-müllerian hormone result for her age.Participants could select more than one option, so categories are not mutually exclusive.GP, general practitioner.Knowledge items (N = 349 # ).
a Answer categorised as correct.