Social media and online resources are increasingly used as a method of health information transmission(Reference Song, Omori and Kim1–Reference Orizio, Schulz and Gasparotti3), with multiple pages and websites dedicated to specific health conditions. Dietitians, nutritionists and general practitioners are often listed as the three most preferred sources of nutrition information and are perceived to be the most trustworthy, credible and effective(Reference Cash, Desbrow and Leveritt4). Specifically for women planning a pregnancy, the ease of access to the internet allows them to search for health-related information both before and after, consulting with health professionals(Reference Lush, Hutchison and Pacella-Ince5). It has been shown that women in the preconception period or in early pregnancy utilise the internet at least once per month to find health-related information; however, women report not discussing this information with their healthcare providers as they perceive the information online to be reliable and useful(Reference Sayakhot and Carolan-Olah6). This has the potential to create an environment where inaccurate information or mistaken beliefs may not be corrected.
A systematic review of online, nutrition-related information found that most content was inaccurate and of low quality. According to a panel of experts, only 17 % of the forty-seven assessed websites and social media posts contained accurate information(Reference Denniss, Lindberg and McNaughton7). Similarly, an analysis of 676 nutrition-related Instagram posts found that only 6 % were of ‘good’ quality, with none rated as ‘excellent’ when using the Principles for Health-Related Information on Social Media tool(Reference Denniss, Lindberg and Marchese8). No study has specifically reviewed online nutrition-related claims in regards to preparing for conception. Concerningly, recent research indicates that a majority of fertility-related social media posts are not authored by health professionals(Reference Peyser, Goldstein and Mullin9), and much of the conception-related health information found online is considered inaccurate by fertility and conception experts(Reference Kedzior, Bianco-Miotto and Breen10). For example, in a 2019 study, Kedzior et al. (Reference Kedzior, Bianco-Miotto and Breen10) reviewed and categorised eighty-nine claims into three themes: ‘conception behaviour and monitoring’, ‘lifestyle and exposures’ and ‘medical’. While 40 % of the reviewed conception-related information was found to be inaccurate, the study did not explicitly identify or assess nutrition-related claims. Given that web searches are a primary source of fertility and conception information for many individuals(Reference Hammarberg, Zosel and Comoy11,Reference Broughton, Schelble and Cipolla12) , understanding the accuracy of information available online is important to guide the development of future public health campaigns aimed at targeting misinformation.
With a clear shift towards the use of online resources and social media by people of reproductive age for health and nutrition information(Reference Orizio, Schulz and Gasparotti3), understanding the quality of online information and how the community perceives online nutrition fertility claims is needed for improved dissemination of scientific knowledge to consumers. This project aims to (1) explore and analyse current online preconception health and nutrition-related claims from Google, YouTube, OpenAI, Instagram and TikTok, (2) assess the identified online preconception health claims against a selection of current guidelines and (3) understand the perceived accuracy of a random sample of identified claims among reproductive-aged men and women living in the community.
Methods
Search strategy and data collection
This study involved a content analysis of online platforms, encompassing websites and social media dedicated to providing guidance on preconception nutrition and lifestyle behaviours for individuals of reproductive age. The intention of the search strategy was to employ lay terms and simple search methodologies. Over a 2-d period in May 2023, through an incognito Google Chrome browser, the terms ‘fertility nutrition’, ‘lifestyle changes when trying to conceive’ and ‘ways to increase fertility’ were searched on Google, the video-sharing platform YouTube and the conversational artificial intelligence (AI) platform OpenAI (see online supplementary material, Supplementary Fig. 1). The search terms were transformed into concise hashtags suitable for use on visually oriented platforms such as Instagram and TikTok with a newly created, generic account. The adapted hashtags included ‘#fertilitydiet’, ‘#TTClifestyle’ and ‘#increasefertility’. Search terms were decided through consultation between study authors, with the aim to simulate a population of individuals searching for fertility-related information. Video media across YouTube, TikTok and Instagram were transcribed verbatim prior to the extraction of claims. Media platforms such as Facebook, where claims were located within a private group, were not included in the study. From each of the three searches per platform, the top ten articles and posts were extracted, as most consumers are unlikely to view search results beyond the first page(Reference Morahan-Martin13).
Content analysis
Thematic analysis was selected as the method for content analysis(Reference Braun and Clarke14). The content of the website or social media post was thoroughly read to identify any health claims, which were then extracted. A piece of text was identified as a claim if it was phrased as a statement or advice regarding behaviours, causes and potential methods used to impact fertility. Once extracted, inductive open coding was utilised to code the health claims using NVivo 12 Pro Plus. Each claim was assigned an alphanumeric identifiable code to link to the online or social media platform. Following the labelling of claims, they were grouped into recurring topics. This created a framework to systematically distribute content into appropriate topics for the remaining articles of information. Any topic unrelated to the aims was excluded; only claims related to nutrition, physical activity and general health topics (i.e. sleep, age and weight) were included. After discussing the topics, the first author developed the initial themes and categories, which were reviewed by senior authors and refined. Themes and categories were agreed upon by all authors. Once assigned to a theme and then category, claims were allotted back under their original platform (Google, YouTube, OpenAI, Instagram, TikTok). Duplicate claims were removed to ensure there were no repeats of a claim within each online or social media type.
Assessing the evidence for the claim
The final list of 191 claims was analysed by a panel of three researchers with expert knowledge in nutrition, physiology and reproductive health. A 2022 systematic review of eleven Australian and International preconception guidelines was used to identify relevant guidelines, of which nine were used to compare against each claim(Reference Dorney, Boyle and Walker15). Two guidelines present in the aforementioned systematic review were excluded from use by the expert panel. These guidelines had a limited scope, offering guidance on preconception care only for people living with human immunodeficient virus and Zika virus. A rubric for assessment of evidence was established by the study authors, whereby a claim present in the set of Australian or International preconception guidelines was considered the highest level of evidence. These included a claim reported in Royal Australian College of General Practitioners Guidelines(16), Royal Australian and New Zealand College of Obstetricians and Gynaecologists(17), South Australian Perinatal Practice Guidelines Preconception Advice(18), Public Health Agency of Canada(Reference Shaw, Barney and Di Meglio19), American College of Obstetricians and Gynecologists(20), Federation of Obstetric and Gynecological Societies of India(21), Centres for Disease Control and Prevention(22), American Academy of Family Physicians (Positions Paper)(23) and Recommendations for Preconception Counselling and Care(Reference Farahi and Zolotor24). The second highest level of evidence was judged against the most recent and comprehensive scoping review of observational studies assessing nutritional intake and female fertility outcomes, published in 2023(Reference Alesi, Habibi and Silva25). The panel assessed each claim as being present in a preconception guideline (yes or no), present in the scoping review of female fertility (yes or no), having limited evidence to suggest an association or no association, having insufficient human evidence to support the claim or having no evidence to support the claim (see online supplementary material, Supplementary Fig. 2). A claim was considered to have limited evidence if it was indeed present in the scoping review of female fertility; however, definitive conclusions could not be made. A claim was considered to have insufficient evidence if, from a formative review of the literature, the expert panel could not identify human research but could identify some in vitro or animal evidence. The experts agreed upon the classification of each claim.
Determining the community perception of claims
Following the content analysis and categorisation of claims, forty online claims were randomly selected (i.e. eight claims from each media source) for assessment by a sample of men and women (‘community’). The claims were randomly ordered to create a survey through REDCap, hosted by the University of Adelaide. The survey was distributed through the university and institute emails and newsletters and shared on social media platforms from October to November 2023. Inclusion criteria were aged 18–49 years and able to read and write in English. The survey was administered anonymously, participants provided informed consent, and participation was voluntary. Prior to undertaking the survey, each participant provided their gender (male, female or other), their date of birth and their occupation category (academic; health professional (fertility related); health professional (not fertility related); or neither an academic nor health professional). For pregnancy planning status, participants were instructed to select a single option that best described their current status from the following mutually exclusive choices: (1) not currently trying to conceive, (2) trying to conceive, (3) currently pregnant or (4) had a prior pregnancy. Each participant could select only one option, and these were treated as self-reported categorical descriptors, not objective reproductive statuses. This variable was not used in the analysis. For each claim, participants rated the likelihood of the claim being true using a 5-point Likert scale (1 = not at all likely, 2 = somewhat likely, 3 = unsure, 4 = quite likely, 5 = highly likely). A claim was considered likely to be true by the community if over 75 % of respondents rated the claim as greater than 4 on the Likert scale. Likewise, a claim was considered unlikely to be true if over 75 % of respondents rated the claim as less than 2 on the Likert scale.
Data reporting
All data, including the content analysis, expert panel rating and online survey results, are summarised with raw numbers and percentages. Figures were created using BioRender(26).
Results
Content analysis
Two primary themes were generated: nutrition and lifestyle (Table 1). Any claim with mention of food, nutrients, diet or supplements was categorised under the nutrition theme, and any claim related to an individual’s way of living or behaviours was categorised under the lifestyle theme. The nutrition theme contained 159 claims and was divided into three categories (whole foods and their components, dietary patterns and supplements). The lifestyle theme included thirty-two claims and was divided into three categories (sleep and stress, physical activity and personal characteristics). The category whole foods and their components contained the highest number of claims (n 64), while physical activity contained the least (n 5).
All topics of claims included in each category and theme, including the number of claims in each category
The YouTube platform contained the highest number of extracted claims (n 58), closely followed by Google (n 54 claims) (Table 2). Ninety-four percent of Instagram claims and 93 % of TikTok claims were present in the nutrition theme. OpenAI had the highest proportion of lifestyle themes when compared with other platforms, with 29 % (6/21) of claims from the lifestyle theme. All platforms had a high percentage of claims from the category whole foods and their components except for TikTok, with only 15 % (4/26) claims in this category. The social media platforms Instagram and TikTok had no claims in the category physical activity, and TikTok did not have any claims present in the category personal characteristics.
Count and percentage of main themes and categories of claims found through the online searches on the platforms Google, YouTube, OpenAI, Instagram and TikTok
Accuracy of claims
Five percent (10/191) of the identified claims are referred to in current preconception guidelines. These claims included statements relating to folic acid and iron supplementation, excess vitamin A consumption and healthy BMI ranges for conception. Two percent (4/191) of claims, while not present in a preconception guideline, were included in a recent scoping review of female fertility, with potential benefits identified for adherence to a Mediterranean diet and a diet low in trans-SFA. Six percent (11/191) of claims were deemed to have ‘limited evidence to suggest that there was no association’ between the claim and the reported health outcome, while 12 % (21/191) of claims were considered to have ‘limited evidence to suggest an association’ between the claim content and health outcome. Twenty-one percent (40/191) of claims were considered to have ‘insufficient evidence’, while 54 % (103/191) of claims were considered to have ‘no evidence for the health outcome’ (see online supplementary material, Supplementary Fig. 3).
Figure 1 reveals that a substantial percentage of preconception health information found online lacks strong scientific support, with many claims falling into the ‘insufficient evidence’ or ‘no evidence for the health outcome’ categories. TikTok and Instagram were found to have a higher percentage of claims with ‘no evidence for the health outcome’ than the other social media platforms (Fig. 1). The platforms YouTube, Google and Instagram contained claims that could be identified in preconception guidelines (Fig. 1), while only the categories supplementation and personal characteristics contained claims present in guidelines (Fig. 1). These claims were related to folic acid supplementation, iron supplementation, body weight and physical activity. A substantial portion of claims made about whole foods and their components and their impact on fertility had ‘no evidence for the health outcome’ stated in the claim (Fig. 1). No claims found through OpenAI or within the categories sleep and stress or physical activity were supported by high-level evidence, as none of the claims identified were present in preconception guidelines or the scoping review of female fertility.
Heatmap illustrating the level of evidence supporting preconception health information across different online platforms and thematic categories. The rows represent the sources of information (YouTube, Google, Instagram, OpenAI, TikTok) and key preconception health topics (personal characteristics, supplements, food compounds, dietary patterns, exercise, sleep/stress), as well as the number of claims obtained from each source. Colour intensity corresponds to the percentage of claims from each platform or category within each level of evidence
Source: Created in BioRender. Lush, K. (2025) https://BioRender.com/ruztax7.
Community perception of health claims
Eighty participants completed the online survey, and their characteristics are reported in Table 3. Ninety-three percent of the survey participants were female, and the median age of the sample population was 33·9 (IQR 30·6, 37·2) years. Most participants (51·3 %) considered themselves to be neither a researcher nor a health professional, and 50 % of participants were not currently trying to conceive. Of the forty health claims presented, three were considered ‘likely to be true’, of which all three were deemed to have ‘insufficient evidence’ by the expert panel and one was considered to have ‘no evidence for the health outcome’ (Table 4). Three of the forty claims were considered ‘unlikely to be true’ by the community (score <2), of which the expert panel identified one to have ‘insufficient evidence’ to suggest an association and two to have ‘no evidence for the health outcome’ (Table 4). Of the forty claims presented to the community, the expert panel considered 17 claims to have ‘no evidence for the health outcome’; however, only two claims were identified as ‘unlikely to be true’ by the community.
Demographic details of participants who completed the public survey of health claims
Claims rated by the community (n 80) as likely (>4) or unlikely (<2) to be true and the expert panel rating of each claim
* ≥75 % of the community rated the claim as <2 or >4 on the Likert scale when completing the survey, where <2 was considered unlikely to be true and >4 was considered likely to be true. Two claims, which were present in preconception guidelines, did not receive a high enough percentage of respondents rating the claim as >4 to be considered likely to be true by the community.
Discussion
Our analysis highlights a significant gap between publicly shared nutrition and lifestyle claims and established preconception guidelines. While only a small proportion of identified claims aligned with existing recommendations, the majority lacked sufficient supporting evidence for the stated health outcomes. Notably, some claims that were considered likely to be true by the public were not reflected in preconception guidelines, suggesting a disconnect between community perception and evidence-based recommendations. This underscores the need for clearer public health communication and improved integration of emerging evidence into preconception care guidance.
With over six million users between the ages of 18–45, social media platforms can be a valuable way to create behavioural change within the community and to spread information(Reference Evans, Abroms and Broniatowski27). All five platforms included in this study contained claims that were considered to have ‘no evidence for the health outcome’; however, TikTok (75 %) and Instagram (73 %) contained a higher percentage of ‘no evidence’ claims than the other media platforms (Google, YouTube and OpenAI, all <60 %). This is interesting given that TikTok is growing in popularity, with users spending upwards of 1 h/d on the platform(Reference Kemp28). As of 2024, 69 and 62 % of respective TikTok(29) and Instagram users(30) are aged between 18 and 34 years, suggesting that a high proportion of users on TikTok and Instagram are of reproductive age. As such, the information presented through social media platforms has the potential to impact preconception behaviours among this age group. Accessing reliable and consistent information is often perceived to be a barrier to implementing preconception nutrition and lifestyle changes(Reference Khan, Boyle and Lang31). These barriers may be more pronounced in vulnerable populations, such as those experiencing socio-economic disadvantage or having lower levels of education and health literacy(Reference Khan, Boyle and Lang31). However, it is important to distinguish between the ability to access and evaluate reliable information and the likelihood of acting on information encountered online. Research indicates that individuals with lower health literacy may be particularly influenced by the persuasive techniques and relatable narratives found on social media platforms such as Instagram and TikTok(Reference Zhu, Liu and Zhang32). In these cases, behaviour change often occurs not because the information is reliable, but because it is accessible, engaging and presented by trusted content creators. This highlights a critical concern: while these platforms can promote behaviour change, they may also disseminate misleading or lower-quality information. Therefore, ensuring that social media content is accurate, evidence-based and accessible is particularly important for reaching and positively influencing populations with low health literacy.
OpenAI was the only platform where none of the claims presented were in the guidelines or scoping review of female fertility. Emerging research suggests AI will become an integral part of healthcare; however, challenges related to privacy, bias and the need for human expertise in healthcare should be addressed before the broader implementation of AI as a health tool(Reference Alowais, Alghamdi and Alsuhebany33). AI has the advantage of synthesising information from multiple online sources to assist consumers from having to deduce their own conclusions when comparing information online. As a source of information, research suggests that AI is reasonably accurate(Reference Kirk, Van Eijnatten and Camps34,Reference Beilby and Hammarberg35) ; however, it has been shown to become confused when provided with an overwhelming volume of information(Reference Koopman and Zuccon36). OpenAI provided information on which 8/21 (38 %) claims had ‘no evidence for the health outcome’. This suggests that AI needs further training to distinguish and provide accurate preconception information. The operations of AI rely on the information available online. Given there is little accurate information for AI to work with regarding preconception nutrition, vague or unsupported answers may be given to those seeking advice.
Two claims present in the online survey were readily available in preconception guidelines; however, neither was rated as ‘likely to be true’ by the community. These claims included ‘achieving a healthy weight and BMI is beneficial while trying to conceive’ and ‘engaging in moderate exercise is beneficial before pregnancy and can lead to healthy pregnancy outcomes’. All Australian and International preconception guidelines have stated that aiming for a healthy body weight prior to conception is beneficial(16–Reference Farahi and Zolotor24). However, no studies to date have examined the use of preconception guidelines by the community; thus, it is unknown if the community is aware of the recommendations made in such guidelines. There is potential that the community did not rate body weight-related claims as ‘likely to be true’ due to the recent emergence of body weight messaging across social media and online, where weight neutrality and body positivity movements gain traction(Reference Clark, Lee and Jingree37). Conversely, the participants surveyed seemingly did not consider the weight of significance when trying to conceive. This is notable as weight stigma, the widespread, stereotypical and harmful belief that being of a higher weight is unhealthy, has been shown to adversely affect fertility outcomes(Reference Hill and Incollingo Rodriguez38). Efforts to promote healthy nutrition and lifestyle behaviours in a manner that is not stigmatising to those of a higher body weight may increase the uptake of preconception healthcare and positively impact the uptake of healthcare advice(Reference Hill and Incollingo Rodriguez38). This will contribute to improved pregnancy outcomes and positively influence long-term health for both the mother and child.
No claims within the category sleep and stress were identified in preconception guidelines, despite this category representing nearly half of the claims within the lifestyle theme. Interestingly, five Australian and International preconception guidelines mention physical activity (16–Reference Shaw, Barney and Di Meglio19,21) for weight loss or weight management rather than a tool for improving health. In contrast, the most recent Australian pregnancy guidelines, which were updated in 2023, advocate for the inclusion of physical activity throughout pregnancy to lower the risk of complications and promote a healthy pregnancy(39). These guidelines provide details regarding duration, intensity and type of physical activity to participate in during pregnancy, in addition to suggestions for activities to avoid. The discord between the absence of lifestyle behaviours in preconception guidelines and the high number of lifestyle-related claims identified in this search highlights a need for more high-quality research to be conducted to establish the impacts of pre-pregnancy lifestyle factors on conception and pregnancy, as well as for the expansion of guidelines to include a more holistic approach to preconception health as relevant high-level evidence emerges.
Three claims that were rated as ‘likely to be true’ by the community were deemed by the expert panel to have ‘limited’ or ‘insufficient’ evidence. When assessing the quality of claims, a common practice of juxtaposing a factually correct statement with a health outcome that lacked significant evidence was observed. For example, ‘avoid alcohol before trying to conceive. Alcohol impacts ovulation and causes genetic disorders in the oocytes’. While the National Health and Medical Research Council(40) and preconception guidelines recommend reducing or abstaining from alcohol prior to pregnancy, ovulation and disruption of genetic material in oocytes are not considered the reasons for the avoidance. Instead, foetal alcohol spectrum disorders are evidenced for this recommendation in preconception guidelines(16–Reference Farahi and Zolotor24). If a claim was structured in this way, it was considered to have ‘no’ or ‘limited evidence’ by the expert panel, despite the former fragment of the statement being correct.
Correcting potential harm caused by online misinformation is a key goal of public health campaigns. This is demonstrated through the preconception guidelines(16–Reference Farahi and Zolotor24), the National Health and Medical Research Council in Australia(40) and the WHO(41), which all recommend, for example, abstinence from alcohol prior to conception in an effort to reduce the risk of foetal alcohol spectrum disorder, cognitive disabilities and birth defects(Reference Dejong, Olyaei and Lo42). Based on evidence, the public health campaign in Western Australia, ‘One Drink’, which promotes abstinence from alcohol prior to and during pregnancy(Reference Pettigrew, Booth and McCausland43), was well received and considered acceptable by target audiences(Reference Pettigrew, Booth and McCausland44). Exposure to the public health campaign created favourable intentions to change alcohol related behaviours(Reference Pettigrew, Booth and McCausland43). However, the view of alcohol avoidance when trying to conceive has not been echoed across media platforms, with an Instagram claim stating that ‘four standard drinks per week is safe while trying to conceive’. One study suggested that social media is a vital part of successful public health campaigns(Reference Golechha45). The present study demonstrates a clear lack of high-level evidence utilised across social media platforms, despite professional organisations publishing clear recommendations regarding preconception alcohol use. As social media platforms heavily impact consumer knowledge base(Reference Ghahramani, De Courten and Prokofieva46), the incorporation of preconception guidelines into targeted social media public health campaigns would assist in disseminating high-level evidence to women and men accessing this information online. Engaging with consumers to better understand their internet use behaviours when accessing fertility-related information would facilitate effective strategies to target misinformation online.
Strengths and limitations
This study had several key strengths. The chosen methodology, including searching three separate terms across five platforms, allowed us to simulate the consumer experience when exploring preconception information-seeking behaviours. This approach identified a wide breadth of claims encompassing nutrition and lifestyle behaviours and created a search environment that could be viewed by health professionals, family members or those trying to conceive. Another strength includes the use of video platforms that were searched in this study, with YouTube, TikTok and Instagram reels available for assessment. This allowed us to capture the most popular media platforms frequently utilised by a reproductive-aged population. In addition to research institutions, social media platforms were used to distribute the community survey. This facilitated an understanding of the community perceptions of health claims among a sample of reproductive-aged people living in Australia.
There are also limitations to consider. The community survey did not employ a formal sample size and therefore represents a convenience sample. This limits the generalisability of the findings and may mean the results do not capture the full diversity of the broader population. Nevertheless, our sample is larger than those used in previous Australian-based studies assessing product content claims, which included twenty-six(Reference Thompson, McMahon and Watson47) and thirty-six(Reference Chan, Patch and Williams48) men and women, respectively. Despite this limitation, our sample offers valuable insight into how the community perceives fertility-related nutrition and lifestyle claims. To reduce participant burden, we presented participants with a smaller sample of claims, which limits our understanding of how the broader community may understand and interpret all categories of claims. Furthermore, not all social media platforms were screened, including media platforms with closed groups. As such, the balance of claims available through online platforms may be different from that observed here. Additionally, our search was conducted on newly created social media profiles using web browsers in incognito mode. This approach differs from the experience of individuals routinely seeking preconception health information, as engagement with content can influence the algorithm to recommend additional profiles and posts. As a result, the likelihood of repeated exposure to unevidenced claims may be higher for regular users. We searched English platforms alone, so the information may not be directly translatable to non-English speaking individuals. We acknowledge the use of binary terms when recording gender identity and how this may have influenced responses. Sex at birth was not required for the purpose of this study as it explores social and cultural concepts, so the term gender was used to allow respondents to self-identify for the survey. Allowing further options for participants to indicate non-binary gender identities or gender diverse experiences would have provided a more inclusive understanding of perspectives on nutrition-related fertility claims and better captured the diversity of lived experiences they may hold.
Conclusions
There is a wealth of inaccurate information across social media platforms regarding preconception nutrition and lifestyle behaviours, with many claims lacking support from high-level evidence. Preconception fertility-related health claims that were routinely present in Australian and International preconception guidelines were not perceived by the community as likely to be true. High-level evidence for preconception health needs to be established and disseminated, and targeted public health campaigns via social media would assist in improving the awareness of preconception guidelines and their content. Despite minimal representation in preconception guidelines, many claims authored online centre on lifestyle behaviours, suggesting there is an interest in improving overall preconception health. Further research assessing preconception health, nutrition and lifestyle behaviours is needed to contribute and strengthen the evidence base for peri-conception recommendations.
Supplementary material
For supplementary material accompanying this paper, visit https://doi.org/10.1017/S1368980025100876.
Acknowledgements
None.
Financial support
None.
Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Authorship
J.G. conceptualised the research. K.L. conducted the initial investigation, data curation and wrote the initial draft. All authors contributed to the methodology and formal analysis. J.G. and A.H. reviewed and edited the manuscript and provided supervision.
Ethics of human subject participation
This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving research study participants were approved by the University of Adelaide Human Research and Ethics Committee (H-2023-187). Written informed consent was obtained from all subjects/patients.
Open access
