How common is add-on use and how do patients decide whether to use them? A national survey of IVF patients

Abstract

STUDY QUESTION

What is the prevalence and pattern of IVF add-on use in Australia?

SUMMARY ANSWER

Among women having IVF in the last 3 years, 82% had used one or more IVF add-on, most commonly acupuncture, preimplantation genetic testing for aneuploidy and Chinese herbal medicine.

WHAT IS KNOWN ALREADY

IVF add-ons are procedures, techniques or medicines which may be considered nonessential to IVF, but usually used in attempts to improve the probability of conception and live birth. The use of IVF add-ons is believed to be widespread; however, there is little information about the prevalence and patterns of use in different settings.

STUDY DESIGN, SIZE, DURATION

An online survey was distributed via social media to women in Australia who had undergone IVF since 2017. Women were excluded if they were gestational surrogates, used a surrogate, or underwent ovarian stimulation for oocyte donation or elective oocyte cryopreservation only. The survey was open from 21 June to 14 July 2020.

PARTICIPANTS/MATERIALS, SETTING, METHODS

Survey questions included demographics, IVF and medical history, and use of IVF add-ons including details of the type of add-on, costs and information sources used. Participants were also asked about the relative importance of evidence regarding safety and effectiveness, factors considered in decision-making and decision regret.

MAIN RESULTS AND THE ROLE OF CHANCE

A total of 1590 eligible responses were analysed. Overall, 82% of women had used one or more add-ons and these usually incurred an additional cost (72%). Around half (54%) had learned about add-ons from their fertility specialist, and most reported that the decision to use add-ons was equally shared with the specialist. Women placed a high level of importance on scientific evidence for safety and efficacy, and half (49%) assumed that add-ons were known to be safe. Most women experienced some regret at the decision to use IVF add-ons (66%), and this was more severe among women whose IVF was unsuccessful (83%) and who believed that the specialist had a larger contribution to the decision to use add-ons (75%).

LIMITATIONS, REASONS FOR CAUTION

This retrospective survey relied on patient recall. Some aspects were particularly prone to bias such as contributions to decision-making. This approach to capturing IVF add-on use may yield different results to data collected directly from IVF clinics or from fertility specialists.

WIDER IMPLICATIONS OF THE FINDINGS

There is a very high prevalence of IVF add-on use in Australia which may be generalisable to other settings with similar models of IVF provision. Although women placed high importance on scientific evidence to support add-ons, most add-ons do not have robust evidence of safety and effectiveness. This suggests that IVF patients are not adequately informed about the risks and benefits of IVF add-ons, or are not aware of the paucity of evidence to support their use.

STUDY FUNDING/COMPETING INTEREST(S)

This research was supported by a McKenzie Postdoctoral Fellowship Grant (University of Melbourne), a Department of Obstetrics and Gynaecology Innovation Grant (University of Melbourne) and an NHMRC Investigator Grant (APP1195189). A.P. declares that he provides fertility services at Melbourne IVF (part of Virtus Health). J.W. reports grants from Wellcome Trust, during the conduct of the study, and that publishing benefits his career. The remaining authors report no conflict of interest.

TRIAL REGISTRATION NUMBER

N/A

adjuvant, add-on, adjunct, IVF, timelapse, embryoglue, endometrial scratching, PGT-A, assisted hatching

Introduction

IVF add-ons are often considered to be procedures, techniques or medicines which can be used in addition to standard IVF protocols, usually in attempts to improve success rates. Common add-ons include endometrial scratching, assisted hatching of embryos and complementary therapies such as acupuncture. Most IVF add-ons are not supported by robust evidence that they increase the probability of conception or live birth or are safe to use (Armstrong et al., 2019; Kamath et al., 2019; Lensen et al., 2019). Despite this, their use is thought to be widespread. This has led to extensive debate and discussion, particularly regarding the ethics of offering unproven treatments to a population which may be considered vulnerable, usually at additional cost (Wilkinson et al., 2019; Ben Rafael, 2020).

Although IVF add-ons are believed to be widely available, only one study has assessed the patterns and prevalence of their use. A UK-based survey reported that 74% of patients attending fertility clinics used add-ons with an increasing prevalence between 2013 and 2018 (HFEA, 2018a). The most commonly used add-ons in 2017/2018 were endometrial scratching, (used by 23%), EmbryoGlue (used by 23%) and Timelapse imaging of embryos (used by 22%). However, these figures reflect a setting where approximately 35% of IVF is publicly provided (HFEA, 2018a), and where the national regulatory agency publishes patient information about the evidence supporting common add-ons (HFEA, 2018b). The use of IVF add-ons may be greater in settings where IVF delivery is largely privately funded, such as the USA, Switzerland and Australia (Macklon et al., 2019).

In Australia, IVF is predominantly provided by large corporations and receives substantial government subsidies. Funding is conditional only on clinician diagnosis of medical infertility, with no restrictions based on prognostic factors such as the patient age or number of cycles an individual or couple can undertake. These circumstances have led to Australia having one of the highest IVF utilisation rates in the world (Adamson et al., 2018). We have previously demonstrated that IVF add-ons are widely advertised on Australian IVF clinic websites (Lensen et al., 2021). However, this does not indicate the prevalence of IVF add-on use or on the use of add-ons accessed outside the IVF clinic such as alternative or complementary medicines, which are widely used among the Australian population (Xue et al., 2008; Harnett et al., 2019).

The aim of this study is to better understand the use of IVF add-ons in a commercial setting, using Australia as a case-study. Objectives included understanding the prevalence and nature of IVF add-on use, the role of the clinician and patient in the decision-making, the sources of information used to support these decisions, and the importance of scientific evidence, all from the perspective of the IVF patient.

Materials and methods

Eligibility and recruitment

This was an online survey of women who had undertaken IVF (including ICSI) or embryo transfer in Australia since 2017. Survey questions focused on her experience and use of IVF and add-ons, including all IVF and embryo transfer cycles since 2017, to capture recent use and to minimise recall bias. Women who were gestational surrogates or who had used a surrogate and those who underwent IVF procedures for oocyte donation were excluded. Women who had undertaken oocyte cryopreservation were also excluded unless they had returned to use the oocytes since 2017.

The survey invitation and introduction explained that the research aim was to better understand the experience of IVF in Australia. To reduce selection bias, the invitation and introduction did not mention IVF add-ons. The survey was advertised via Facebook, supplemented by additional social media and website posts from relevant organisations, such as the Royal Women’s Hospital, the largest women’s hospital in Australia. The survey was open from 21 June to 14 July 2020.

Survey design

The questionnaire was built in Qualtrics and took between 10 and 20 min to complete (Qualtrics, 2005). It was pilot tested amongst members of the research team and five women with a history of IVF who were recruited via Facebook. In response to feedback, minor modifications to the wording of some questions and response options were implemented.

Screening questions were used to filter out ineligible participants. The survey then captured demographic (age, education level, cause of infertility) and IVF history-related information including number of treatment cycles, costs incurred, and whether the women had conceived and achieved live birth following treatment (Supplementary File). Participants were asked to select which add-ons they had used from a list of 24, developed from a survey of IVF websites in Australia and New Zealand (Lensen et al., 2021). This list included add-ons available at IVF clinics (clinic-based add-ons) and those available outside the IVF clinic, specifically acupuncture and Chinese herbal medicine (alternative add-ons). Women were asked whether they paid extra for these add-ons, which sources of information they had consulted, what factors were considered in decision-making, and the importance of evidence regarding the safety and effectiveness of add-ons. Conditional logic was built into the survey design so that women answered different questions depending on their previous responses. Women who responded that they had paid extra for one or more add-ons were asked to indicate which they had used most recently for the remaining survey questions. Women who had not used any add-on were asked about their decision to not use add-ons. Most questions had multiple-choice response options or rating scales and two questions invited free-text responses. The order of response options within multi-choice questions was randomised to prevent bias from order-effect (Israel and Taylor, 1990). For instance, when participants were selecting which add-ons they used from a list of 24, the order that these appeared on the screen was random and was therefore different for each participant (however, all participants were asked the same questions, in the same order).

Decision regret about use of add-ons was measured using the Decisional Regret Scale, a validated 5-item questionnaire which measures levels of distress or remorse following healthcare decisions (Becerra Pérez et al., 2016). Respondents were asked to reflect on their decision to use, or not to use, IVF add-ons and regret scores were calculated per the instrument’s manual (Supplementary File) (Brehaut et al., 2003). Commonly used cut-offs were applied to the decision regret scores, classifying these as: no regret (score = 0), mild regret (score 5–25) or moderate-strong regret (score >25) (Becerra-Perez et al., 2016).

Statistical analysis

In some cases, free-text responses were provided when women selected ‘Other’ in a multi-choice question. Where possible, these were recoded as belonging to one of the original response options and additional categories were developed to group some responses. Percentages were calculated excluding missing data for each question. Data analysis was restricted to descriptive statistics in the whole survey cohort and among a small number of subgroups, however statistical testing was not performed. Data analysis was undertaken in R (RStudio Team, 2020).

Ethics approval

This project received ethics approval from the University of Melbourne Human Ethics Advisory Group (Reference: 2056802.1).

Results

A total of 1877 responses were received. After removing responses from women who did not complete any survey questions beyond providing demographics on the first survey page (n = 269) or those who were ineligible (no IVF during the study period, n = 16; oocyte donor, n = 1; surrogate, n = 1), the remaining 1590 eligible responses were analysed. Women completing the survey were broadly representative of those undergoing IVF in Australia in terms of age, reason for IVF and use of ICSI for fertilisation (Table I) (Newman et al., 2020). Women had undergone an average of 2.3 stimulated IVF cycles and 2.4 embryo transfers since 2017, and almost one quarter (23%) had also undergone IVF prior to 2017.

Table I

Demographics of survey participants.

Participant characteristics	(n = 1590)
Age (median, IQR)	36	(33–39)
Level of education
Secondary school	156	(9.8)
Certificate	185	(11.6)
Diploma	240	(15.1)
Bachelor degree	601	(37.8)
Postgraduate degree	406	(25.6)
Missing	2
Reason for IVF
Female factor	465	(29.5)
Unexplained	387	(24.6)
Combined male and female	277	(17.6)
Male factor	236	(15.0)
Same-sex couple	86	(5.5)
Single female	70	(4.4)
Genetic	49	(3.1)
Other	4	(0.3)
Missing	16
Use of donor oocytes and embryos
Donor oocytes	59	(3.7)
Donor embryos	10	(0.6)
Missing	3
Stimulated IVF cycles 2017-2020
None	41	(2.6)
1	624	(39.3)
2	369	(23.3)
3	209	(13.2)
4	112	(7.1)
≥5	232	(14.6)
Missing	3
Embryo transfers 2017–2020
None	116	(7.3)
1	485	(30.5)
2	317	(20.0)
3	256	(16.1)
4	135	(8.5)
≥5	278	(17.5)
Missing	3
Insemination type used
IVF only	458	(29.2)
ICSI only	755	(48.2)
IVF and ICSI	331	(21.1)
Not applicable	23	(1.5)
Missing	23
IVF before 2017
Yes	356	(22.5)
No	1224	(77.5)
Missing	10

Numbers are displayed as n (%) unless otherwise stated.

Use of IVF add-ons

In all, 82% of women had used one or more add-ons as part of their IVF treatment (median 2 add-ons, IQR 1–5, range 0–18). The most commonly used add-ons were acupuncture (used by 45% of participants), preimplantation genetic testing for aneuploidy (PGT-A) (28%) and Chinese herbal medicine (26%) (Table II).

Table II

IVF add-ons used and whether they were associated with an additional cost.

Add-ons (n = 1590)	Used add-on		Paid extra for add-on
Acupuncturea	692	(45.3)	681	(98.4)
Pre-implantation genetic testing for aneuploidy (PGT-A)	422	(27.6)	397	(94.1)
Chinese herbal medicine	397	(26.0)	390	(98.2)
Heparin (clexane)	377	(24.7)	272	(72.1)
Aspirina	366	(24.0)	324	(88.5)
Timelapse imaging of embryos (Embryoscope)	358	(23.4)	98	(27.3)
EmbryoGlue (embryo transfer media, hyaluronan-containing transfer media)	341	(22.3)	203	(59.5)
Melatonina	339	(22.2)	319	(94.1)
Prednisolone (corticosteroids, glucocorticoids)	334	(21.9)	302	(90.4)
Endometrial scratch (endometrial injury, pipelle)	264	(17.3)	192	(72.7)
Androgens (testosterone, DHEA, dehydroepiandrosterone, androderm patch)	204	(13.4)	150	(73.5)
Growth Hormone	180	(11.8)	133	(73.9)
Assisted hatching	120	(7.9)	54	(45.0)
Intralipid infusion	125	(8.2)	116	(92.8)
Physiological intracytoplasmic sperm injection (PICSI)	97	(6.4)	71	(73.2)
Intracytoplasmic morphologically selected sperm injection (IMSI)	88	(5.8)	43	(48.9)
Lipiodol flushing (poppy seed oil, lipiodol bathing)	86	(5.6)	67	(77.9)
Endometrial receptivity array (ERA)	53	(3.5)	47	(88.7)
GM-CSF culture media (Embryogen, Blastgen, Granulocyte-macrophage colony-stimulating factor culture media)	49	(3.2)	23	(46.9)
Artificial Oocyte Activation (Artificial oocyte activation calcium ionophore)	32	(2.1)	21	(65.6)
Viagra (vasodilators)	29	(1.9)	22	(75.9)
Meiotic spindle visualisation (Polscope, Oosight, Polarised light egg visualisation, Egg spindle visualisation)	23	(1.5)	8	(34.8)
Intravenous immunoglobulin (IVIG)	18	(1.2)	14	(77.8)
Platelet Rich Plasma intrauterine infusion or instillation	9	(0.6)	7	(77.8)
No add-ons used	274	(17.9)
Missing	63	(4.0)

a

In the survey, these add-ons were accompanied by the text ‘specifically for IVF purposes’ to distinguish from use of these in other contexts.

Most add-ons (72%) incurred an additional cost to the patient. This varied depending on the add-on used. For example, 98% paid additional costs for acupuncture compared to 27% for Timelapse embryo incubation (Table II). The suggestion to use add-ons most often came from the IVF clinician (54%), followed by friends or family, and internet sources such as forums (Supplementary Table SI). Similarly, clinicians were more likely to raise the suggestion of add-ons more often than patients (71% compared to 18%) (Supplementary Table SII). However, this depended on the type of add-on. For example, alternative add-ons (acupuncture and Chinese herbal medicine) were most often recommended by friends or family (31%) rather than the fertility specialist (16%), whereas clinic-based add-ons were generally recommended by the fertility specialist (73%) (Supplementary Table SI).

Women rated the importance of scientific evidence supporting the use of add-ons very highly. On a scale from 0 to 100, an importance score over 90 was selected by 55% of women for evidence that the add-on improves live birth rates and by 73% of women for evidence that the add-on is safe (Fig. 1). When asked to consider a hypothetical novel add-on which had not been evaluated in a scientific study, most women (72%) thought this should be permitted if it was considered low-risk. Almost half (41%) supported the use of a novel add-on if ‘the IVF patient wanted to use it’ and around one third (37%) supported it if the fertility specialist had experienced previous success with the add-on (Supplementary Table SIII).

Figure 1.

Importance of scientific evidence that add-ons improve the chance of live birth and are safe to use.

Decision-making and reflection

Women who had paid extra for ≥ 1 add-on were asked to reflect on their decision to use the add-on they had accessed most recently. There was substantial variation in perception of who played the biggest role in the decision to use clinic-based add-ons; 18% believed the clinician had 0% input, and 7% believed the clinician had 100% input. Most responded that the decision was equally shared between the clinician and themselves (51%, IQR 40–85%). Almost half of the respondents believed the selected add-on was risk free (49%), and less than one-quarter (24%) thought there might be some risks (Supplementary Table SIV). Among women who conceived and achieved a live birth following the use of an add-on, 61% believed the add-on ‘probably’ or ‘certainly’ helped them to achieve this outcome and only 8.3% of women believed they would have been successful without using the add-on.

Regarding the decision to use the selected add-on, the median regret score was 15/100 (IQR 0–30). Around one third (30%) of women experienced moderate to severe regret about using add-ons and 34% experienced no regret at all (Table III). The level of regret was associated with the outcome of the IVF cycle in which the add on was used. Regret was lower amongst women who had conceived or achieved a live birth compared to women who had not (Table III). Additionally, women who reported their fertility specialist had more than 50% input into the decision to use the add-on experienced more regret than women who reported their clinician to have a lesser role in the decision (Table III).

Table III

Experience of regret among women reporting to pay extra for one or more IVF add-on.

Regret scores	All respondents (n = 1127)	IVF outcome*		Women’s role in the decision**
		Conceived or had a baby (n = 524)	Did not have a baby (n = 424)	<50% role (n = 195)	>50% role (n = 362)
Median regret score (IQR)	15 (0–30)	5 (0–20)	30 (10–45)	25 (5–40)	10 (0–25)
Regret categories
No regret (score = 0)	363 (33.8)	244 (47.1)	72 (17.2)	47 (24.7)	146 (41.2)
Mild regret (score = 5–25)	392 (36.5)	213 (41.1)	133 (31.8)	65 (34.2)	129 (36.4)
Moderate-strong regret (score >25)	320 (29.8)	61 (11.8)	213 (51.0)	78 (41.1)	79 (22.3)
Missing	52	6	6	5	8

Regret score categories are reported as n(%). This question was only asked of respondents who paid extra for one or more add-on. Numbers missing differ between groups as *respondents were excluded from the analysis if they were unable to be grouped as conceiving/having a baby or not (e.g. answered ‘I don’t know yet’, ‘Unsure/can’t remember’ or data were missing); **only respondents selecting they most recently used a clinic-based add-on were asked to complete this question.

Only one in six (18%) respondents reported no add-on use. The two main reasons were being unaware of add-ons (46%) and being aware but not discussing these with their fertility specialist (38%) (Supplementary Table SV). There were too few women who considered add-ons and decided not to use them to analyse decision regret among this group (n = 11).

Discussion

This survey found that the great majority of Australian women undergoing IVF had used one or more add-ons. There has been substantial speculation and debate in the infertility literature regarding the availability of IVF add-ons, and the factors which may drive their use (Lensen et al., 2019; Macklon et al., 2019; van de Wiel et al., 2020; Wilkinson et al., 2019). It has been suggested that a commercialised IVF industry contributes to add-on use in certain settings. Financial incentives may lead fertility clinics or specialists to recommend IVF add-ons, for example to encourage patients to undergo further IVF, or that IVF patients themselves drive use by requesting add-ons (Ben Rafael, 2020; Blakely et al., 2019; Macklon et al., 2019; Wilkinson et al., 2019). We found that most women had first heard about add-ons from their fertility specialist (54%), and that their specialist generally raised these options during consultations (71%). Relatively few women (18%) reported that they first raised the use of add-ons.

The most frequently used add-on was acupuncture, which was used by almost half of the women surveyed (45%). Chinese herbal medicine use was also common (26%). These complementary or alternative medicines are usually independent of fertility clinics, and most women reported to hear about these from friends or family. All other add-ons were clinic-based, with seven add-ons used by more than one in five women. As most add-ons were reported to incur a cost to the patient, their use presents a financial burden (in addition to the costs of IVF). The cost associated with some add-ons may be negligible, such as aspirin and heparin, however expenses are likely to be significant for more costly add-ons, such as PGT-A, or for repeated use of add-ons over multiple cycles. Together, these additional costs may represent an opportunity cost to patients who are then unable to afford further IVF.

Women placed a high level of importance on scientific evidence supporting use of add-ons. However, robust evidence for the safety and effectiveness of most commonly used add-ons is lacking. Recent reviews report that none of the most commonly used add-ons are supported by high-quality evidence (Armstrong et al., 2019; Kamath et al., 2019; Lensen et al., 2019, 2021). There are also inconsistencies between the perceived safety of add-ons and existing evidence. Whilst most women believed that the add-ons they were offered were safe and that add-ons should only be permitted if they are low-risk, several commonly used add-ons may present a risk. For example, Chinese herbal medicine may interact with IVF medications, and immune therapies such as corticosteroids have been linked to a possible risk of congenital anomalies, prematurity and low birth weight, as well as various significant side-effects with prolonged use (Robertson et al., 2016).

Clinicians are required to make decisions about whether to recommend or offer add-ons, and patients must decide whether to request or accept them, in the absence of robust evidence of benefit and potential risks. It is likely that both parties assume that the worst-case scenario is that the add-on does not confer benefit, causing them only inconvenience and avoidable cost (Wilkinson et al., 2019). However, in the absence of evidence such an assumption may be flawed. For example, a novel non-invasive PGT-A has been offered by a major Australian IVF clinic despite lack of published evidence to demonstrate its diagnostic accuracy (Cree and Farquhar, 2020). This add-on was subsequently discontinued following a report of an unacceptably high incidence of false positive tests, which may have led to many chromosomally normal embryos being discarded (McArthur, 2020).

An important consideration after any healthcare decision is satisfaction with the decision and the presence of regret. Regret can result from knowledge that better outcomes were possible, or from the belief that a poor decision led to a poor outcome (Connolly and Zeelenberg, 2002). A patient-centred approach to healthcare decision making should aim to reduce regret, particularly when the outcome is subsequently unfavourable. In the IVF setting, regret may compound the significant emotional toll of IVF, especially for those experiencing (repeated) unsuccessful outcomes. Women in this survey largely considered that they and their fertility specialist had contributed equally to the decision to use add-ons. Overall, the median regret score was 15/100, which is similar to levels of regret experienced following health care decisions in oncology and primary care (Becerra Pérez et al., 2016). In this survey, the level of regret was associated with IVF outcome. Among those using one or more add-ons, women achieving pregnancy or live birth experienced less regret (median regret score 5) compared to those with unsuccessful outcomes (median 30). Further, women who achieved live birth often attributed this success to the add-on, with only 8% believing they would have conceived without using the add-on. High levels of regret are common after negative health outcomes (Brehaut et al., 2003; Becerra-Perez et al., 2016), including in the setting of IVF and add-ons (Goldman et al., 2019; Huang et al., 2020; Kaing et al., 2020). However, it has been argued that regret should be minimal even when a healthcare decision leads to a negative health outcome, if patients are informed about the risk and the probability of treatment failure (Sung et al., 2008). Higher regret among women who did not conceive may suggest that they anticipated a better outcome, which goes against the suggestion that patients seek out add-ons to be reassured that ‘they left no stone unturned’ (Wilkinson et al., 2019). Women who reported less input into the decision to use the add-on reported more regret, consistent with previous studies (Brehaut et al., 2003). Due to the necessarily subjective and retrospective nature of assessing regret and decision-making, the direction of this relationship is difficult to establish. Women achieving live birth after use of add-ons may be more likely to recall a greater responsibility for their decisions and those experiencing unsuccessful outcomes may be more likely to view the clinician as responsible. Individuals tend to overclaim responsibility for tasks or decisions, especially when the outcome is successful (Miller and Ross, 1975).

Together, the findings, that women: placed significant importance on scientific evidence and yet elected to use unproven IVF add-ons; viewed the add-ons they used as safe despite a lack of evidence; and experienced more regret when they had less input into the decisions, suggest that IVF patients may not be adequately informed about the benefits and risk of IVF add-ons or may not be aware of the paucity of supportive evidence for safety and effectiveness. This is supported by previous research showing that IVF patients may not always receive accurate information about IVF add-ons or may not be informed about the lack of evidence (Ben Rafael, 2020; Galiano et al., 2020; Lensen et al., 2021). Few IVF clinic websites in Australia and the UK mention the potential negative effects of add-ons or their aspects related to safety or risk (van de Wiel et al., 2020; Lensen et al., 2021). To make informed decisions, patients require accessible and transparent information about the evidence for safety and effectiveness of add-ons. This may reduce later regret about decision making, particularly when IVF is unsuccessful. Whilst some countries offer impartial evidence-based information about add-ons (HFEA, 2018b), no such resource is available in other settings, including in Australia.

Strengths and limitations

This was a large survey with over 1500 responses from women across Australia. We utilised social media to increase survey exposure and attract a representative population of women. The survey respondents were broadly comparable to the IVF population in Australia, in terms of age and cause of infertility (Newman et al., 2020). Compared to national data, we observed similar rates of assisted hatching use (7.9% vs. 8.6%) but higher use of PGT-A (27.6% vs. 12.6%). However, these comparisons are limited by use of different denominators and definitions. We captured the use of add-ons over a 3.5 year period, rather than per IVF cycle or embryo transfer, and, therefore, we are unable to directly compare the frequency of assisted hatching or PGT-A use with national data where it is reported per cycle in which an embryo is either fertilised following stimulated IVF cycle or thawed for transfer (Newman et al., 2020). Further, the reported use of PGT-A is difficult to compare to the national dataset where the PGT data includes both preimplantation genetic diagnosis and screening. We observed high rates of the use of acupuncture and Chinese herbal medicine. This may reflect the higher use of complementary and alternative therapies in Australia than elsewhere. Approximately 9% of Australians use acupuncture and Chinese herbal medicine (Xue et al., 2008; Harnett et al., 2019), which may be higher than acupuncture use in other Western populations (Klein et al., 2015; Cui et al., 2017). The use of acupuncture in this survey (45%) was also higher to the reported use by IVF patients in the UK (22%) (HFEA, 2018a). A further strength of our survey is the omission of information relating to IVF add-ons from the survey recruitment materials and participant information, which aimed to reduce response bias from women based on experience with or attitudes towards IVF add-ons.

The major study limitation is the retrospective design. Women were asked to recall their experiences of IVF and add-ons during the last 3.5 years, which may have led to recall bias and error. For example, women may not have been aware or may not have been easily able to recall which add-ons were used or what they were called, particularly for add-ons which are routinely included in some IVF cycles, such as Embryoglue or Timelapse. We attempted to minimise this by including frequently used terms to describe different add-ons. Additionally, this study only reflects the perspectives of women undergoing IVF, which may differ from those of male partners, fertility specialists or IVF clinics. Patient recall of subjective outcomes, such as contribution to decision making and experience of regret, may be more susceptible to differences in perception or bias, as discussed earlier. Although we used a validated regret scale and asked women to reflect specifically on the decision to use add-ons when completing it, it is possible that this measure captured a broader context of regret, for instance regret in having undertaken IVF. Additionally, it was not possible to summarise regret scores among those not using add-ons, due to the small number of women in this group. We were therefore unable to report on the extent of any regret associated with not using add-ons. The use of additional measures, such as satisfaction with the decision, may have provided a more comprehensive picture of how women view the decision.

The use of social media for recruitment may pose a limitation as not all eligible women may be active social media users. The survey was only available in English, which may have restricted participation to Australian residents who are fluent in English. It is possible that a women’s experience of IVF may have influenced their decision to participate; for example, women who experienced unsuccessful IVF treatment may have been more or less likely to take part. Further, the survey did not capture information about race or ethnicity, so we were unable to explore cultural differences in the use of add-ons.

A further limitation is that there is no standardised definition of an IVF add-on. In this survey, we defined add-ons as procedures, techniques or medicines which can be used in addition to standard IVF protocols with the aim of improving the probability of conception and live birth. The list of add-ons was based on those commonly referred to in the literature, and a review of IVF clinic websites in Australia (Lensen et al., 2021). However, the selection of add-ons included in this survey may be viewed as a limitation as a different research team may have compiled a different list. Some interventions such as Timelapse and EmbryoGlue are routinely used by some IVF clinics for all patients, and may not necessarily be considered add-ons in these cases. However, such interventions are known to be offered in additional to standard IVF in many settings, and we therefore elected to include them in this survey (van de Wiel et al., 2020; Lensen et al., 2021). We also included alternative therapies such as acupuncture and Chinese herbal medicine as add-ons, which others may not refer to as add-ons. However, we excluded other add-ons that may be in common use such as nutritional supplements aimed at enhancing fertility or IVF success. Additionally, we excluded purely diagnostic add-ons such as DNA sperm fragmentation testing. Finally, we observed some illogical or unlikely response combinations. For example, the Endometrial Receptivity Array was used by 53 women of whom 47 reported that this incurred additional costs. However, it would be anticipated this would always cost patients extra, unless they were participating in a research study.

The optimal approach to evaluating the extent of add-on use would be to collect data directly from IVF clinics. However, in Australia only data on assisted hatching and PGT are collected at a national level (Newman et al., 2020). National data collection relating to add-ons would improve understanding of add-on use and permit tracking for possible safety signals. While this survey reports on some of the drivers and motivations underlying add-on use from the patient perspective, better understanding of the decision-making process may come from qualitative interviews with patients and IVF clinic staff.

Conclusion

In summary, the use of IVF add-ons is widespread in Australia and the use of some add-ons is likely to be associated with a significant financial burden. Of the five most commonly used add-ons, three were clinic based (PGT-A, heparin, aspirin) and two were alternative therapies (acupuncture and Chinese herbal medicine). Women placed great importance on scientific evidence of safety and effectiveness of IVF add-ons, and very few had contemplated potential risks. However, IVF add-ons are not supported by high-quality evidence of safety and effectiveness, and may potentially pose risk to patients. Regret was higher among women who had unsuccessful IVF and who viewed the clinician as having driven the decision to use the add on. Together, this suggests that patients may not be adequately informed about the lack of evidence to support the effectiveness and safety of add-ons.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Acknowledgements

The authors wish to thank the women who contributed to the testing and piloting of the survey: Erin McKenna, Nicole Drummond, Nicola Norton, Katy Lindemann.

Authors’ roles

S.L. conceived the idea for the survey, designed and analysed the survey, and drafted the manuscript. K.H., A.P., J.W., S.W., M.P. and M.H. contributed to the survey design and question structure, interpretation of the data and preparation of the manuscript.

Funding

This research was supported by a McKenzie Postdoctoral Fellowship Grant (University of Melbourne), a Department of Obstetrics and Gynaecology Innovation Grant (University of Melbourne) and an NHMRC Investigator Grant (APP1195189). J.W. is supported by a Wellcome Institutional Strategic Support Fund award (204796/Z/16/Z). M.H. is supported by an NHMRC Practitioner Fellowship (APP1058935).

Conflict of interest

A.P. declares that he provides fertility services at Melbourne IVF (part of Virtus Health). J.W. reports grants from Wellcome Trust, during the conduct of the study, and that publishing benefits his career. The remaining authors report no conflict of interest.