Self-reported fertility problems and their socio-demographic and health correlates: a cross-sectional online survey of 8,000 Polish adults

This research received no external funding.

Results are reported across two overlapping analytic samples and outcome definitions, summarized in Supplementary Fig. S1 and detailed as follows. The study included 8,000 respondents, the majority of whom were women ( n = 4,990, 62.4%, Table 1 and Supplementary Fig. S1). The overall mean age was 40.3 years (SD 11.9). Women were significantly younger than men (Mann–Whitney U test, P < 0.001; Fig. 1 ). The sample was predominantly urban and highly educated: nearly a third resided in large cities of over 500,000 inhabitants (29.2%) and a further fifth in rural areas (21.0%; Fig. 2A ), while more than half held a master’s degree (56.7%; Fig. 2B ). Most respondents were married (56.8%) and were already biological parents (55.2%). Age in groups of respondents differing by sex and the result of the significance test. Black squares indicate medians, hatched boxes represent the 25–75% interquartile range, and whiskers indicate the minimum and maximum values. Statistical significance was determined using the Mann–Whitney U test ( P < 0.001). Socio-demographic structure of the respondents participating in the National Fertility Test ( n = 8,000). (A) Distribution of place of residence and (B) distribution of education level. Values are in percentages. Overall, 4,412 respondents (55.2%) were biological parents. Based on the fertility-history item, ‘when trying to conceive a child, have you ever encountered any problems with your reproductive health (e.g. trying for a child for over a year/diagnosis of infertility)?’ (see the section titled ‘Methods’), 1,601 respondents (20.0%) of the full sample reported having encountered health problems while trying to conceive, 3,931 (49.1%) reported no such problems, 1,243 (15.5%) had not yet attempted conception, and for 1,225 (15.3%) the item was not applicable. Two complementary operational definitions of self-reported fertility problems were used in this study (see the section titled ‘Methods’): a simple definition based on this item, and a composite definition additionally incorporating a physician-confirmed diagnosis or prolonged attempts to conceive combined with fertility-related help-seeking. Under the composite definition, 1,122 respondents were classified into the fertility-problems group. These fertility-related characteristics are summarized in Table 2 , which reports the distribution within the full survey sample, and Table 3 , which reports the corresponding sex-stratified distribution within the analytical sample of respondents who had attempted to conceive. Descriptive univariable analyses of correlates of self-reported fertility problems, based on the full survey sample ( n = 8,000) using a composite outcome definition (see the section titled ‘Methods’), are summarized below; full results are provided in the Supplementary Material (Tables S1, S2, S3, S4, S5). These analyses retain the broader, composite outcome definition to provide a comprehensive overview of correlates across the full survey sample, irrespective of respondents’ conception status. In this analysis, candidate correlates of self-reported fertility problems were considered, including age, education, lifestyle behaviours, diagnosed health conditions, and sex differences in reported conditions. Regarding age, two categorical age thresholds were examined as secondary, exploratory comparisons (Supplementary Table S1). Although the conventional 35-year cut-off showed no significant association with self-reported fertility problems (OR = 1.01, 95% CI: 0.88; 1.15; P = 0.904), a data-driven 31-year threshold yielded a significant association (OR = 1.53, 95% CI: 1.28; 1.85; P < 0.001). Furthermore, respondents with a master’s degree reported fertility problems more frequently than those with lower educational attainment (OR = 1.47, 95% CI: 1.29; 1.67; P < 0.001; Supplementary Table S2). The univariable analysis revealed also higher odds of self-reported fertility problems among those significantly limiting alcohol consumption (OR = 1.60, 95% CI: 1.40; 1.83; P < 0.001), following a healthy diet (OR = 1.54, 95% CI: 1.35; 1.75; P < 0.001), and not smoking tobacco products (OR = 1.18, 95% CI: 1.03; 1.36; P = 0.021). Regular physical activity was not significantly associated with the outcome (OR = 0.95, 95% CI: 0.83; 1.08; P = 0.359, Supplementary Table S3). Among diagnosed health conditions (Supplementary Table S4), thyroid disease was associated with higher odds (OR = 2.37, 95% CI: 2.07; 2.72; P < 0.001), whereas arterial hypertension was associated with lower odds (OR = 0.72, 95% CI: 0.61; 0.86; P < 0.001). Women and men also differed in the pattern of reported conditions (Supplementary Table S5): compared with men, women had higher odds of long-term health problems (OR = 1.40, 95% CI: 1.27; 1.53), sexually transmitted infections (OR = 1.95, 95% CI: 1.43; 2.66), and thyroid disease (OR = 6.82, 95% CI: 5.78; 8.06), whereas men had higher odds of hypertension (OR = 3.28, 95% CI: 2.93; 3.68) and liver disease (OR = 1.61, 95% CI: 1.32; 1.97). Overall, within the full survey sample ( n = 8,000), women had higher odds of being classified in the fertility-problems group than men (crude OR = 2.52, 95% CI: 2.22; 2.85; P < 0.001). After exclusion of respondents who reported not having attempted conception ( n = 1,243, 15.5%) and those for whom the item was not applicable ( n = 1,225, 15.3%), the analytical sample comprised 5,532 respondents (3,316 women and 2,216 men) with a determinable fertility-problem status ( Table 3 ). Within this analytical sample, 1,601 respondents (28.9%) reported having encountered health-related difficulties while trying to conceive and were accordingly classified into the self-reported fertility problems group ( Table 3 ). The proportion was substantially higher among women (1,215/3,316, 36.6%) than among men (386/2,216, 17.4%; χ 2 = 239, P < 0.001), corresponding to more than twofold higher odds in women (OR = 2.74, 95% CI: 2.41; 3.12). As the sample was self-selected through an online fertility test, this proportion should not be interpreted as a population-level prevalence estimate for Poland (see Limitations). In the age-adjusted models ( Tables 4 and 5 ), correlates of self-reported fertility problems fell into three broad categories: diagnosed medical conditions with established links to impaired fertility, socio-demographic factors, and behaviours and help-seeking practices whose association with the outcome most likely reflects reverse causation. These categories are described separately in the following sections and should not be interpreted as equivalent risk factors. Among women ( Table 4 ), the strongest age-adjusted associations were observed for fallopian tube obstruction (OR = 23.0, 95% CI: 12.6; 41.7), polycystic ovary syndrome (OR = 5.35, 95% CI: 4.20; 6.80), reproductive anatomical abnormalities (OR = 3.95, 95% CI: 2.81; 5.55), endometriosis (OR = 2.82, 95% CI: 2.21; 3.61), and luteal phase defect (OR = 1.93, 95% CI: 1.60; 2.32). A history of sexually transmitted infection was likewise associated with higher odds (OR = 1.89, 95% CI: 1.29; 2.76). Long-term health problems (OR = 1.77, 95% CI: 1.53; 2.05), thyroid disease (OR = 1.76, 95% CI: 1.51; 2.05), ovarian cysts (OR = 1.62, 95% CI: 1.36; 1.92), uterine fibroids (OR = 1.58, 95% CI: 1.28; 1.93), dysmenorrhoea (OR = 1.45, 95% CI: 1.26; 1.67), and dyspareunia (OR = 1.40, 95% CI: 1.17; 1.68) were also significantly associated with the outcome, all remaining significant after Holm–Bonferroni correction. However, among men ( Table 5 ), two clinically plausible correlates remained significant after Holm–Bonferroni correction for multiple comparisons: varicocele or epididymal cysts (OR = 3.08, 95% CI: 1.95; 4.86) and a history of sexually transmitted infection (OR = 4.60, 95% CI: 1.94; 10.9). Three further clinical conditions showed positive unadjusted associations that did not survive this correction and should be regarded as suggestive rather than conclusive: a history of testicular damage due to viral disease (OR = 5.30, 95% CI: 1.77; 15.8; P = 0.003, P adj = 0.075), congenital testicular developmental disorders (OR = 2.99, 95% CI: 1.22; 7.35; P = 0.017, P adj = 0.391), and cancer (OR = 2.19, 95% CI: 1.04; 4.61; P = 0.040, P adj = 0.880). Among women, a history of prior pregnancy was associated with substantially lower odds of reporting fertility problems (OR = 0.26, 95% CI: 0.22; 0.32) and better self-rated mental health was similarly associated with lower odds (OR = 0.76, 95% CI: 0.65; 0.88); higher body weight (≥71 kg: OR = 1.39, 95% CI: 1.20; 1.61) and BMI (≥26.3 kg/m 2 : OR = 1.41, 95% CI: 1.21; 1.64) showed modest positive associations. Educational level and place of residence were not significant after Holm–Bonferroni correction in either sex. Among men, none of the remaining socio-demographic factors examined – body weight, BMI, and self-rated physical or mental health – reached statistical significance. Age, modelled continuously, was inversely associated with the outcome in both sexes (women: OR = 0.96 per year, 95% CI: 0.95; 0.97; men: OR = 0.95 per year, 95% CI: 0.94; 0.96; Tables 4 and 5 ). In both sexes, the use of reproductive supplements was strongly associated with the outcome (women: OR = 2.34, 95% CI: 2.01; 2.72; men: OR = 4.06, 95% CI: 2.94; 5.61). Among women, avoidance of chemical exposure was also significantly associated with higher odds after Holm–Bonferroni correction (OR = 1.31, 95% CI: 1.13; 1.52); the corresponding association among men did not reach significance (OR = 1.27, 95% CI: 0.95; 1.69; P = 0.103). However, lifestyle behaviours (smoking, alcohol, diet, physical activity) were not significant after Holm–Bonferroni correction in either sex ( Tables 4 and 5 ). Beyond the overall association with supplement use, individual supplement components showed markedly heterogeneous associations with self-reported fertility problems ( Tables 6 and 7 ). Age-adjusted associations between the use of specific supplements and conception-related behaviours and self-reported fertility problems among men ( n = 2,216). Each row derives from a separate model adjusted for age. * Each OR derives from a separate binary logistic regression model with the listed variable and age (continuous) as the sole covariate. Both supplement components and conception-related behaviours reflect reported use of these products and practices, and are interpreted as markers of active help-seeking rather than as causal or protective risk factors (see Discussion). OR, odds ratio; CI, confidence interval. Among women, the use of individual supplement components was associated with self-reported fertility problems to varying degrees. The strongest associations were observed for taking inositol (OR = 6.80, 95% CI: 5.19; 8.90)) and carnitine (OR = 3.50, 95% CI: 1.97; 6.23) as supplements, followed by vitamin D, folic acid, vitamin B complex, vitamin E, choline, omega-3 fatty acids, zinc, arginine, taurine, magnesium, and iron, use of which was associated with odds ratios between approximately 2.1 and 3.0 ( Table 6 ). Furthermore, reported use of specific conception-related behaviours showed positive associations with the outcome, including herbal preparations or teas (OR = 5.82, 95% CI: 4.69; 7.22), avoiding hot baths after intercourse (OR = 3.86, 95% CI: 3.27; 4.57), lying with legs raised after intercourse (OR = 3.34, 95% CI: 2.84; 3.93), and stopping contraception well in advance (OR = 3.08, 95% CI: 2.60; 3.65). Additional behaviours showed similarly significant positive associations: special sexual positions (OR = 3.44, 95% CI: 2.89; 4.10), intercourse timed to the evening (OR = 2.66, 95% CI: 2.26; 3.13), female orgasm (OR = 2.40, 95% CI: 2.04; 2.82), and a frequency-based approach to conception (OR = 1.98, 95% CI: 1.68; 2.34). Among men, the pattern of taking supplements was similar in direction to women but generally stronger in magnitude: inositol was associated with markedly higher odds (OR = 13.1, 95% CI: 4.72; 36.5), followed by carnitine (OR = 5.13, 95% CI: 2.87; 9.17), folic acid (OR = 4.62, 95% CI: 3.01; 7.08), and zinc (OR = 4.23, 95% CI: 2.98; 6.00), with the remaining components (choline, arginine, taurine, vitamin E, iron, vitamin D, magnesium, omega-3 fatty acids, and vitamin B complex) showing odds ratios between approximately 2.2 and 3.6 ( Table 7 ). Moreover, a similar pattern of reported help-seeking behaviours was observed among men ( Table 7 ) as among women: stopping contraception in advance (OR = 5.51, 95% CI: 3.86; 7.88), use of herbal preparations or teas (OR = 5.58, 95% CI: 3.90; 7.99), and lying with legs raised after intercourse reported by their partner (OR = 4.66, 95% CI: 3.47; 6.26) were among the behaviours most strongly associated with the outcome. Additional behaviours showed comparably significant associations: intercourse timed to the evening (OR = 4.60, 95% CI: 3.38; 6.25), avoiding hot baths after intercourse (OR = 4.16, 95% CI: 3.09; 5.59), avoiding personal lubricants (OR = 3.85, 95% CI: 2.88; 5.15), special sexual positions (OR = 3.55, 95% CI: 2.64; 4.77), female orgasm (OR = 2.14, 95% CI: 1.60; 2.85), and a frequency-based approach to conception (OR = 1.53, 95% CI: 1.15; 2.04). Given that supplement use is itself a marker of active fertility-related help-seeking (see above), these findings – for both supplement components and conception-related behaviours, in both sexes – most plausibly reflect the intensity of help-seeking behaviour associated with specific choices and practices, rather than a causal or protective role of any individual component.

This study utilized data from the National Fertility Test, a cross-sectional, nationwide socio-educational project conducted in Poland in 2024 via an online survey instrument (Supplementary File (see section on Supplementary materials given at the end of the article)). The test was disseminated online via Medonet, a Polish online health platform, and further distributed through national media and social channels to ensure broad public engagement. Participation was voluntary and anonymous. For the purposes of this study, responses from the initial 8,000 adult respondents were analysed. Recruitment relied on self-selection through online channels, resulting in a convenience sample rather than a probabilistically representative sample of the entire Polish adult population. This sampling strategy may introduce selection biases, potentially limiting the generalizability of findings related to prevalence estimates. The estimated duration of the test completion ranged from 5 to 10 min. Participants did not receive any remuneration for their participation in the study. This study was conducted in accordance with the ethical principles of the Declaration of Helsinki. The study protocol was reviewed by the Bioethics Committee of the Military Chamber of Physicians (Wojskowa Izba Lekarska) in Warsaw, Poland. The Committee determined that formal ethical approval was not required due to the non-interventional nature of the research and the use of fully anonymized data (Decision No. KB 65/2024). All participants were adult (≥18 years) internet users. Prior to commencing the questionnaire, all individuals were fully informed about the study’s objectives, data collection methods, and handling procedures. Participation was strictly voluntary and anonymous, and no compensation was provided. Electronic informed consent was obtained from all participants before they could proceed with the survey. Given that all data were collected and analysed anonymously, specific consent to publish individual data was not applicable, as no identifiable information is included in this report. The online survey used in the study included questions regarding the respondent’s socio-demographic data, as well as questions necessary for assessing the prevalence and correlates of self-reported fertility problems. The first section of the questionnaire focused on gathering comprehensive socio-demographic information, including age (continuous variable), sex (categorical: female/male), current place of residence (categorized by population size: village, city up to 19,000, 20–49,000, 50–99,000, 100–299,000, 300–499,000, >500,000 inhabitants), and highest level of educational attainment (categorized: primary, lower secondary, basic vocational, secondary, post-secondary non-tertiary, bachelor’s/engineer’s degree, master’s degree). Self-reported physical and mental health statuses compared with peers (categorized, e.g. very good, good, average, poor) were collected, along with the presence of any long-term health problems or diagnosed chronic diseases (binary: yes/no), with specific inquiries about conditions, such as hypertension, diabetes, thyroid disorders, kidney and liver diseases, and sexually transmitted infections. Body mass index (BMI) was calculated from self-reported height and weight (kg/m 2 ). Lifestyle habits were assessed through questions encompassing aspects such as dietary patterns (e.g. self-assessment of following a healthy diet: yes/no/sometimes), frequency and intensity of physical activity (e.g. meeting WHO recommendation of ≥150 min/week: yes/no/sometimes), current smoking status (e.g. non-smoker vs current/occasional smoker), and patterns of alcohol consumption (e.g. self-assessment of limiting consumption: yes/no/sometimes). Additional self-care behaviours were assessed using the same yes/no/sometimes format, including fruit and vegetable consumption, adequate hydration, adequate sleep (6–8 h), stress reduction, and avoidance of chemical exposures relevant to reproductive health. Two complementary operational definitions of self-reported fertility problems were used in this study, applied to different analyses. For the primary, age-adjusted regression analyses, fertility problems were defined using a single survey item (the fertility-history item, National Fertility Test 2024 Questionnaire): respondents who answered ‘yes’ to having ever encountered problems with their reproductive health while trying to conceive (e.g. trying for over a year, or a diagnosis of infertility) were classified into the fertility-problems group; those answering ‘no’ were classified as not having fertility problems. This analysis was restricted to respondents who had attempted to conceive ( n = 5,532); those who indicated they had ‘not yet tried to have a child,’ or for whom the item was not applicable, were excluded (see the section titled ‘Results’). For the descriptive univariable analyses, a broader, composite definition was used, based on the full survey sample ( n = 8,000). Respondents were classified into the fertility-problems group if they met either of the following criteria: i) responded ‘yes’ to the fertility-history item (National Fertility Test 2024 Questionnaire) and had received a diagnosis from a medical professional (gynaecologist, andrologist/urologist, or reproductive endocrinologist); or ii) had been trying to conceive for longer than one year and had sought medical or psychological support for fertility-related concerns within the past year (e.g. a reproductive health or infertility treatment clinic, a gynaecologist or andrologist/urologist for fertility reasons, or a psychologist, psychotherapist, or psychiatrist for fertility-related stress). All other respondents were classified into the ‘no fertility problems’ group. Both operational definitions rely on self-report and may be subject to recall bias and potential misclassification. Supplementary Fig. S1 summarizes the derivation of the two analytical samples and the corresponding outcome definitions used in this study. Descriptive statistics were used to characterize the study population, including frequencies and percentages for categorical variables, and measures of central tendency (mean, median) and dispersion (standard deviation (SD), interquartile range (IQR)) for continuous variables. To analyse the prevalence of self-reported fertility problems, the frequency (percentage) was calculated for the entire sample and in subgroups based on key demographic, lifestyle, and health characteristics. For comparative analysis between groups (e.g. men vs women, different age groups), appropriate tests of significance were used: independent samples t -tests or Mann–Whitney U tests for continuous variables (depending on data distribution) and Pearson’s chi-square tests for categorical variables. To examine factors associated with self-reported fertility problems, a series of separate binary logistic regression models was fitted, one for each candidate predictor variable, rather than a single combined multivariable model. Each model included the predictor of interest together with age (continuous, in years) as the sole adjusting covariate. This minimally adjusted approach was chosen to avoid the fallacy, whereby the simultaneous interpretation of multiple coefficients from a single multivariable model yields estimates that do not have a consistent or meaningful interpretation, because the appropriate adjustment set differs across exposure–outcome relationships ( Westreich & Greenland 2013 ). Models were fitted separately for women and men. An additional model, with age as the sole predictor and no other covariate, was fitted to estimate the age-adjusted odds ratio for age itself; this model was not included in the family of tests corrected for multiple comparisons (see below). ‘Currently attempting to conceive’ was similarly excluded from the Holm–Bonferroni correction family. This variable is near-tautologically related to the outcome – respondents not currently attempting conception have limited opportunity to recognize or report a fertility difficulty – and its association is reported as an internal validity check confirming that the questionnaire captures genuine signal, rather than as a substantive clinical or behavioural correlate. Results are presented as age-adjusted odds ratios (ORs) with 95% confidence intervals (CI) and the corresponding P -value for each predictor. General characteristics of 8,000 respondents participating in the National Fertility Test. Data are presented as n (%) for categorical variables and as mean (SD) or median (Q1; Q3) for continuous variables. Fertility-related characteristics of the study sample ( n = 8,000). Values are presented as n (%). Given the exploratory nature of the study and the large number of predictors examined, the increased risk of type I error arising from multiple comparisons was addressed by additionally reporting P -values adjusted using the Holm–Bonferroni method ( P adj ); associations that remained significant after this correction are indicated. The correction family comprised 34 tests among women and 27 tests among men, each corresponding to one candidate predictor variable and excluding the age-only model and the ‘currently attempting to conceive’ internal validity check described above. P -values for individual supplement components and conception-related behaviours are not adjusted for multiple comparisons, as these items are presented as an exploratory elaboration of the single ‘supplement use’ variable, rather than as an independent family of hypothesis tests. Analytical sample restricted to respondents who had attempted to conceive ( n = 5,532), using the single-item outcome classification (the fertility-history item); the OR expresses the odds of reporting fertility problems in women relative to men. χ 2 , chi-square test; OR, odds ratio; CI, confidence interval. The fertility-history item in National Fertility Test 2024 Questionnaire Age-adjusted associations between individual predictor variables and self-reported fertility problems among women ( n = 3,316). * Each odds ratio (OR) is derived from a separate binary logistic regression model, including the listed predictor and age (continuous) as the sole covariate. Bold values remained statistically significant after Holm–Bonferroni correction for multiple comparisons ( P adj <0.05). Excluded from the Holm–Bonferroni correction family: age was modelled to estimate its own effect, and ‘currently trying to conceive’ was treated as a positive control (internal validity check), given its near-tautological relationship to the outcome; a ‘–’ is therefore shown in the P adj column for these rows. OR, odds ratio; CI, confidence interval; BMI, body mass index. The reference category for each binary predictor is its absence. The analytical sample excludes respondents who reported never having tried to conceive. Age-adjusted associations between individual predictor variables and self-reported fertility problems among men ( n = 2,216). * Each odds ratio (OR) is derived from a separate binary logistic regression model, including the listed predictor and age (continuous) as the sole covariate. Bold values remained statistically significant after Holm–Bonferroni correction for multiple comparisons ( P adj <0.05). Excluded from the Holm–Bonferroni correction family: age was modelled to estimate its own effect, and ‘currently trying to conceive’ was treated as a positive control (internal validity check), given its near-tautological relationship to the outcome; a ‘–’ is therefore shown in the P adj column for these rows. OR, odds ratio; CI, confidence interval; BMI, body mass index. The reference category for each binary predictor is its absence. The analytical sample excludes respondents who reported never having tried to conceive. Age-adjusted associations between the use of specific supplements and conception-related behaviours and self-reported fertility problems among women ( n = 3,316). Each row derives from a separate model adjusted for age. * Each OR derives from a separate binary logistic regression model with the listed variable and age (continuous) as the sole covariate. Both supplement components and conception-related behaviours reflect reported use of these products and practices and are interpreted as markers of active help-seeking rather than as causal or protective risk factors (see Discussion). OR, odds ratio; CI, confidence interval. Statistical significance for all analyses was defined as a two-sided P -value < 0.05. The STATISTICA v. 13.3 statistical software package (TIBCO Software Inc., USA) was used for the analysis.

This large-scale analysis of the National Fertility Test provides contemporary insights into the epidemiology of self-reported fertility problems in Poland. Our findings indicate a substantial prevalence of such problems within the sample population, most consistently associated with clinically plausible health conditions, such as fallopian tube obstruction, PCOS, and endometriosis in women, and varicocele or epididymal cysts and sexually transmitted infection in men. The strong statistical associations with the use of reproductive supplements and specific conception methods (e.g. resting postures post-intercourse) point to reverse causation, identifying these behaviours as markers of fertility-related help-seeking rather than biological risk factors. These results emphasize the need for continued research, ideally incorporating longitudinal data and clinical verification, to fully elucidate the complex interplay of factors influencing fertility in the Polish context. Ultimately, this study provides valuable data to inform targeted prevention strategies, enhance reproductive health education, and guide future research priorities aimed at mitigating the burden of fertility difficulties in Poland.

This cross-sectional study, leveraging data from the large-scale National Fertility Test survey, aimed to estimate the prevalence of self-reported fertility problems and identify associated correlates within a contemporary Polish adult sample. Within the analytical sample of respondents who had attempted to conceive ( n = 5,532), 28.9% reported having encountered related health difficulties. To the best of our knowledge, no rigorous population-based survey of infertility prevalence has previously been conducted in Poland. Existing estimates of 15–20% are extrapolated by analogy with other developed countries rather than derived from Polish survey data ( Chawłowska et al. 2020 ). Comparable self-report surveys using nationally representative sampling in other countries have reported considerably lower numbers: 12.5% among women and 10.1% among men in the British Natsal-3 survey ( Datta et al. 2016 ); 5.6–6.7% period prevalence of primary infertility in a Taiwanese population-based survey ( Lee et al. 2024 ); and 12.6% period and 17.5% lifetime prevalence in the WHO-commissioned global synthesis ( Cox et al. 2022 ). This higher prevalence in our study most plausibly reflects the self-selected nature of the present sample, recruited through voluntary engagement with an online fertility-testing platform, which is expected to over-represent individuals currently experiencing, or concerned about, difficulties conceiving. The present estimate should, hence, be understood as characterizing the profile of platform users rather than the general Polish population. Several of the individual correlates identified in our analyses warrant particular attention. Age, modelled as a continuous variable, was among the most statistically robust correlates of self-reported fertility problems, identified consistently in both sexes; however, its direction was inverse, which is, at first sight, counter-intuitive given the well-established decline in fecundity with age (women: OR = 0.96, 95% CI: 0.95; 0.97; men: OR = 0.95, 95% CI: 0.94; 0.96). This most plausibly reflects the structure of the analytical sample rather than a biological effect: younger respondents who are actively attempting to conceive represent the group most likely to be currently experiencing, and therefore reporting, difficulties, whereas older respondents more frequently include individuals who have already completed their families. This pattern is consistent with prior cross-sectional surveys recruiting through online and social-panel channels, where those currently attempting to conceive tend to be relatively young ( Bunting et al. 2013 ). The exceptionally strong association for ‘currently attempting to conceive’ is expected, given its near-tautological relationship to the outcome and was treated as a positive control rather than a correlate of interest, consistent with its exclusion from the corrected family alongside age. Beyond this continuous-age model, two categorical age thresholds were also examined as secondary, exploratory analyses (Supplementary Table S1). Notably, the conventional 35-year threshold – a widely used clinical marker of advanced maternal age ( American College of Obstetricians & Gynecologists 2022 ) – showed no significant association with the outcome (OR = 1.01, P = 0.904), whereas the 31-year threshold, derived empirically from the present sample rather than specified a priori , did reach significance (OR = 1.53, P < 0.001; Supplementary Table S1). This contrast illustrates the risk of data-driven dichotomization: empirically optimized cut-points are prone to overfitting and capitalizing on chance, and may not generalize to other samples ( Royston et al. 2006 ). The 31-year threshold should, therefore, be regarded as hypothesis-generating rather than confirmatory, and the continuous age estimates as the primary result. Among clinically plausible correlates, the strongest age-adjusted associations among women were observed for well-established gynaecological and reproductive conditions. Fallopian tube obstruction showed by far the largest effect size (OR = 23.0, 95% CI: 12.6; 41.7; Table 4 ), consistent with its status as one of the most direct, mechanically unambiguous causes of infertility, since tubal patency is a structural prerequisite for natural conception. Furthermore, polycystic ovary syndrome (OR = 5.35), the leading cause of anovulatory infertility and estimated to affect 12–18% of women of reproductive age ( March et al. 2010 ), and endometriosis (OR = 2.82), reported to affect 25–50% of women presenting with infertility ( Bulletti et al. 2010 ), likewise showed strong, biologically coherent associations. Reproductive anatomical abnormalities, luteal phase defect, ovarian cysts, uterine fibroids, dysmenorrhoea, and dyspareunia all remained significant after Holm–Bonferroni correction (OR: 1.40–3.95; Table 4 ), consistent with the broader literature linking structural and hormonal reproductive pathology to impaired fecundity. Long-term health problems also showed a comparably strong association (OR = 1.77), suggesting that overall chronic disease burden, not any single condition, may partly underlie this pattern. Among men, two conditions remained significant after Holm–Bonferroni correction: varicocele or epididymal cysts (OR = 3.08, 95% CI: 1.95; 4.86) and a history of sexually transmitted infection (OR = 4.60, 95% CI: 1.94; 10.9; Table 5 ). Varicocele prevalence is well documented to rise sharply with reproductive impairment, from approximately 15% in the general male population to up to 35% among men with primary infertility ( Alsaikhan et al. 2016 ), consistent with its status as the most common surgically correctable cause of male infertility; our finding is in line with this established gradient. Similarly, the STI association aligns with a meta-analysis reporting comparable effect sizes for genital chlamydial infection (OR = 3.68, 95% CI: 2.24; 6.02; Bragazzi et al. 2026 ), supporting a plausible biological pathway via epididymitis, inflammation, and subsequent obstruction or impaired sperm quality. In contrast to the wide range of gynaecological correlates identified in women, the relative scarcity of significant correlates in men may partly reflect the lower prevalence of diagnosed andrological conditions in this sample, alongside genuine sex differences in the distribution of identifiable causes of infertility. Regarding socio-demographic factors, education showed a positive univariable association with self-reported fertility problems, with respondents holding a master’s degree more frequently reporting such problems than those with lower educational attainment (OR = 1.47, 95% CI: 1.29; 1.67; Supplementary Table S2). However, this association did not persist after age adjustment and correction for multiple comparisons in either sex (women: OR = 1.23, P adj = 0.195; men: OR = 1.15, P adj = 1.000; Tables 4 and 5 ). This pattern is consistent with evidence that the relationship between socio-economic status and infertility is age-dependent ( Chen et al. 2023 ). It also aligns with the well-documented tendency for higher-educated individuals to delay childbearing ( Schmidt et al. 2012 ). In addition, place of residence (urban versus rural) did not emerge as a significant independent correlate in the adjusted models for either sex. Furthermore, among women, a history of prior pregnancy was associated with substantially lower odds (OR = 0.26, 95% CI: 0.22; 0.32; Table 4 ). This finding requires cautious interpretation, since the analytic sample necessarily includes women experiencing secondary infertility – difficulty conceiving following a previous successful pregnancy – a condition estimated to affect approximately one in ten women globally ( Mascarenhas et al. 2012 ). Without a clearer distinction in the questionnaire between first and subsequent pregnancy attempts, this association likely reflects the balance of these two processes rather than a straightforward protective effect of parity. Moreover, better self-rated mental health was similarly associated with lower odds of reporting fertility problems (OR = 0.76, 95% CI: 0.65; 0.88); rather than poorer mental health causing fertility problems, this more plausibly reflects the reverse pathway, whereby fertility difficulties themselves constitute a substantial psychosocial burden. It is worth noting that, among men, none of the remaining socio-demographic factors examined reached statistical significance. Several lifestyle factors showed significant univariable associations with self-reported fertility problems, consistent with the broader literature linking lifestyle to reproductive health ( Sharma et al. 2013 , Bala et al. 2021 ). The direction of these associations was counter-intuitive: respondents reporting healthier behaviours – limiting alcohol consumption and following a balanced diet – showed higher odds of reporting fertility problems (Supplementary Table S3). However, none of these associations remained significant after age adjustment and correction for multiple comparisons ( Tables 4 and 5 ). This pattern is more consistent with reverse causation, whereby individuals experiencing difficulty conceiving adopt healthier behaviours in response ( Goossens et al. 2018 ), than with a harmful effect of the reference behaviours. Regular physical activity (≥150 min/week) showed no significant univariable association, in contrast to meta-analytic evidence linking higher physical activity to reduced infertility risk ( Xie et al. 2022 ). This null finding is not unprecedented; a clinic-based study similarly found no association between physical activity and fertility status among women attending an infertility clinic ( Tempest et al. 2022 ), suggesting that this relationship may be less consistent in populations already engaged with reproductive health services or platforms. Among chronic health conditions, diagnosed thyroid disease showed a strong association with self-reported fertility problems in women, both in the univariable analysis on the full sample (OR = 2.37; Supplementary Table S4) and after age adjustment (OR = 1.76, P adj = 0.002; Table 4 ), consistent with established links between thyroid dysfunction and female reproductive health ( Krassas et al. 2010 , Mazzilli et al. 2023 ). Among men, thyroid disease showed a similar direction of association but did not reach significance after adjustment (OR = 1.70, P = 0.071; Table 5 ), potentially reflecting lower statistical power given the lower prevalence of diagnosed thyroid disease among men; this is consistent with evidence linking thyroid function to male infertility ( Mazzilli et al. 2023 ). Diagnosed diabetes showed no significant association with self-reported fertility problems after age adjustment in either sex (women: OR = 1.71, P adj = 0.060; men: OR = 1.38, P adj = 0.760; Tables 4 and 5 ), a borderline null finding that may reflect limited statistical power, given the low prevalence of diagnosed diabetes in this sample, rather than a true absence of association ( Livshits & Seidman 2009 ). Furthermore, diagnosed arterial hypertension was associated with lower odds of self-reported fertility problems in the full survey sample (OR = 0.72; Supplementary Table S4), contrary to established evidence linking maternal hypertension to increased infertility risk ( Xiong et al. 2024 ). This most plausibly reflects the same age-related sample structure discussed above, rather than a protective effect and should not be interpreted clinically. More broadly, the pattern observed across Supplementary Table S5 – with women showing higher prevalence of long-term health problems, sexually transmitted infections, and thyroid disease, and men showing higher prevalence of hypertension and liver disease – is consistent with the well-documented ‘morbidity–mortality paradox’ in sex differences in health, whereby women report a higher burden of chronic, non-fatal conditions and greater healthcare engagement, while men show higher rates of more severe, life-threatening conditions ( Verbrugge 1982 ). This broader pattern of differential health reporting by sex should be borne in mind when interpreting any sex-specific comparison of chronic disease prevalence in this study. The finding that men taking dietary supplements specifically ‘supporting reproductive health’ had higher odds of reporting fertility problems (OR = 4.06, 95% CI: 2.94; 5.61; Table 5 ) is likely an example of another reverse causation or confounding by indication: men experiencing fertility difficulties are precisely those most likely to seek out such supplements, rather than the supplements causing the problem. Similarly, for women, associations with specific conception strategies (e.g. herbal teas, particular positions, lying down post-intercourse) likely reflect behaviours adopted by those already actively trying to conceive, rather than causal risk factors. This is consistent with the wider preconception health literature, in which the substantial majority of individuals actively trying to conceive report adopting such behaviours in response to their reproductive intentions ( Goossens et al. 2018 ), and with evidence that participants in fertility-related surveys who are currently attempting conception differ systematically from those who are not ( Bunting et al. 2013 ). Beyond the overall association with supplement use, individual supplement components showed markedly heterogeneous associations with self-reported fertility problems ( Tables 6 and 7 ). In women, the strongest associations were observed for taking inositol (OR = 6.80, 95% CI: 5.19; 8.90) and carnitine (OR = 3.50, 95% CI: 1.97; 6.23). Inositol is an evidence-based therapeutic agent specifically used in the management of PCOS, for which it has shown mixed but promising effects on reproductive outcomes in clinical trials ( Cowan et al. 2024 ). Carnitine has also been investigated as an adjunct therapy for PCOS-related subfertility, with meta-analytic evidence of benefits for ovulation and pregnancy rates, although it is not yet incorporated into formal treatment guidelines and the overall quality of evidence remains low ( Mohd Shukri et al. 2022 , Abu-Zaid et al. 2025 ). Other supplements, including vitamin D, folic acid, vitamin B complex, vitamin E, choline, omega-3 fatty acids, zinc, arginine, taurine, magnesium, and iron, were also associated with fertility difficulties, with odds ratios between approximately 2.1 and 3.0 ( Table 6 ). Among men, the pattern of taking supplements was similar in direction to women but generally stronger in magnitude: inositol was associated with markedly higher odds (OR = 13.1, 95% CI: 4.72; 36.5), followed by carnitine (OR = 5.13, 95% CI: 2.87; 9.17). Inositol is an evidence-based supplement specifically used to treat oligoasthenoteratozoospermia (OAT) in men, with meta-analytic evidence showing genuine improvements in sperm concentration, motility, and morphology ( Ghaemi et al. 2024 ), such that its use likely identifies men with an underlying, diagnosed cause of subfertility. Furthermore, a systematic review of eight randomized trials found that carnitine supplementation in men with idiopathic infertility improved sperm motility and morphology ( Khaw et al. 2020 ). Furthermore, reported use of specific conception-related behaviours showed positive associations with the outcome in both sexes ( Tables 6 and 7 ). The strongest associations were observed for herbal preparations or teas (women: OR = 5.82, 95% CI: 4.69; 7.22; men: OR = 5.58, 95% CI: 3.90; 7.99) and stopping contraception well in advance (women: OR = 3.08, 95% CI: 2.60; 3.65; men: OR = 5.51, 95% CI: 3.86; 7.88), followed by lying with legs raised after intercourse (women: OR = 3.34, 95% CI: 2.84; 3.93; men: OR = 4.66, 95% CI: 3.47; 6.26) and avoiding hot baths after intercourse (women: OR = 3.86, 95% CI: 3.27; 4.57; men: OR = 4.16, 95% CI: 3.09; 5.59). For men specifically, avoidance of wet heat sources more broadly – including hot baths, hot tubs, and saunas – has genuine biological plausibility: Finnish sauna exposure has been shown to induce a significant, though reversible, impairment of spermatogenesis, with adverse effects on sperm concentration, motility, and DNA integrity ( Garolla et al. 2013 ), and heat exposure has similarly been linked to reduced fecundability in a preconception cohort study ( McKinnon et al. 2022 ). This association may, therefore, partly reflect an evidence-based protective behaviour among men, in contrast to the same item reported by women, for whom no analogous physiological mechanism exists. Given that supplement use is itself a marker of active fertility-related help-seeking, these findings – for both supplement components and conception-related behaviours, in both sexes – most plausibly reflect the intensity of help-seeking behaviour associated with specific choices and practices, rather than a causal or protective role of any individual component. This interpretation is consistent with the wider preconception health literature discussed above ( Bunting et al. 2013 , Goossens et al. 2018 ). This study possesses notable strengths, primarily its large sample size drawn from a nationwide online survey and the use of a statistically robust, minimally adjusted modelling strategy that avoids the interpretive pitfalls of a single combined multivariable model ( Westreich & Greenland 2013 ). However, several limitations must be acknowledged. First, the reliance on an online, self-selected convenience sample raises important concerns regarding the generalizability of the prevalence estimate: self-selection through voluntary engagement with an online platform likely inflates the apparent proportion of fertility problems relative to the general population, compounded by a marked demographic skew – 56.7% of respondents held a master’s degree, compared with 45.7% of Polish 25- to 34-year olds (the most highly educated national age cohort) holding any tertiary qualification ( European Commission 2025 ), and respondents were disproportionately drawn from urban areas relative to the 59% urban share of the Polish population nationally; neither the prevalence figure nor the associations reported here can, therefore, be generalized to the wider Polish population without caution. Second, all data were self-reported, making them susceptible to recall bias and social desirability bias, particularly for sensitive topics, such as lifestyle habits and health conditions. Third, the cross-sectional design prevents establishing causality; we can only identify associations, several of which are more plausibly explained by reverse causation than by a causal effect on fertility, as discussed above. Fourth, both operational definitions used to classify fertility problems reflect self-reported engagement with fertility difficulties rather than a verified clinical diagnosis of infertility; the resulting case groups are, therefore, likely heterogeneous, combining individuals with medically confirmed infertility, those undergoing diagnostic work-up, and those experiencing distress related to conception attempts, and may be prone to misclassification. Relatedly, the questionnaire did not distinguish between women attempting a first pregnancy and those with secondary infertility, limiting interpretation of parity-related findings; disentangling these two contributing processes would require more detailed reproductive histories, such as the timing of prior pregnancies relative to current conception attempts, which we identify as a direction for future research. Fifth, P -values for individual supplement components and conception-related behaviours were not corrected for multiple comparisons, and these findings should therefore be interpreted with greater caution than the primary age-adjusted correlates. Finally, while we adjusted for age, residual confounding by unmeasured factors (e.g. detailed occupational exposures, partner’s fertility status, duration of attempts to conceive) is possible. Despite these limitations, our findings provide valuable contemporary insights into the epidemiology of self-reported fertility problems and associated factors in a large Polish sample. Clinically plausible gynaecological and andrological correlates – including fallopian tube obstruction, PCOS, and endometriosis among women, and varicocele and sexually transmitted infection among men – emerged as the most consistent and biologically coherent findings, reinforcing existing clinical priorities for reproductive health screening. By contrast, the markedly attenuated lifestyle associations, the counter-intuitive inverse association with age, and the unexpected findings regarding hypertension in this sample underscore the importance of distinguishing genuine clinical risk factors from artefacts of sample structure and self-report in cross-sectional data of this kind; these findings warrant further, more rigorous research, ideally using longitudinal designs and clinically ascertained samples, to clarify these relationships within the Polish context. To our knowledge, this is among the first studies to provide population-based, rather than clinic-based, data on self-reported fertility problems in Poland and to quantify associations for individual supplement components and conception-related behaviours separately in both sexes – a level of granularity rarely available in the existing literature on fertility-related help-seeking. The results can inform the development of targeted educational materials, support screening initiatives, and guide future research aimed at improving reproductive health outcomes in Poland, complementing recent policy initiatives such as the ‘Yes for In Vitro ’ programme.

Infertility, according to the definition provided by the World Health Organization (WHO), is characterized as the failure to achieve a pregnancy within a 12-month timeframe, despite regular unprotected sexual intercourse. This phenomenon impacts an increasing proportion of couples globally, including within Europe ( Zeng et al. 2025 ). Current estimations suggest that this problem affects approximately one in six people of reproductive age at some point in their lives ( World Health Organization 2023 ). Globally, the prevalence of infertility has become a challenge, originating from a range of determinants, such as alterations in lifestyle, environmental contamination, stress, and the progress of medical technologies that facilitate earlier diagnosis and intervention ( Inhorn & Patrizio 2015 ). While global incidence varies, access to advanced treatments differs ( Njagi et al. 2023 ), and approaches within Europe show heterogeneity ( Agarwal et al. 2015 ); the underlying issue gains particular significance in contexts facing demographic pressures. This is particularly true in Europe due to an ageing population and the escalating number of women delaying motherhood. The rise in the number of women aged 35 and above, alongside deferred decisions regarding family planning, contributes to health issues associated with diminished fertility, underscoring the growing clinical relevance of infertility care within contemporary reproductive medicine ( Tatarczak & Janik 2023 ). In Poland, the problem of infertility is becoming increasingly apparent, affecting an estimated 1.5 million couples, according to reports from the Polish Society of Gynaecologists and Obstetricians. This may be related to numerous factors, including changing social attitudes, the postponement of decisions regarding having children, and a growing awareness of the problem ( Tatarczak & Janik 2023 ). While awareness and utilization of treatments such as IVF have increased, challenges related to access and socio-cultural factors persist ( Serour & Serour 2017 ). Key correlates identified globally and often mirrored in Poland include a woman’s age over 35 and lifestyle factors, such as improper diet, lack of physical activity, smoking, and excessive alcohol consumption, which impact fertility in both sexes ( Sharma et al. 2013 , Bala et al. 2021 , Knap-Wielgus et al. 2023 ). Chronic diseases, such as diabetes, thyroid disorders, hypertension, and obesity, are also increasingly recognized as correlates ( Jeon et al. 2025 ). Socio-economic factors, such as the tendency for rural residents to have children earlier compared with urban populations with higher education who often postpone parenthood, also contribute to the complexity of infertility in Poland ( Kotowska et al. 2008 ). From a demographic perspective, the problem of infertility in Poland has significant long-term consequences. This phenomenon is becoming one of the factors influencing the decline in fertility rates, which are among the lowest in Europe ( Eurostat 2021 ). In 2020, the total fertility rate in Poland was below the replacement level of 2.1 children per woman ( Statistics Poland GUS 2024 ). In the long term, fertility-related problems can lead to deepening difficulties in maintaining the country’s demographic balance, which will consequently affect the labour market, pension systems, and other socio-economic aspects ( Bloom et al. 2010 ). Although numerous studies exist on the impact of lifestyle on fertility, including diet, physical activity, smoking, and alcohol consumption, a critical knowledge gap exists specifically within the Polish context. There is a lack of large-scale, contemporary research in Poland that examines lifestyle factors alongside socio-demographic variables and health status within a single study population. This integrated approach is crucial, particularly for studies that consider factors such as access to health education, social norms, and differences in access to healthcare. These aforementioned factors may indeed differentiate the influence of lifestyle on fertility across various social groups, and understanding their combined effect is vital for developing targeted interventions in Poland. Therefore, this study aimed to estimate the frequency of self-reported fertility problems and to identify socio-demographic, lifestyle, and health factors associated with their occurrence. Utilizing data from the large-scale, online National Fertility Test, we analysed these correlates separately for women and men to provide contemporary insights into the profile of Polish adults actively engaging with digital reproductive health platforms. Understanding these correlates may help identify groups of individuals most vulnerable to fertility problems and inform the development of targeted health interventions to support reproductive health in Poland. Consequently, this study may contribute to the development of recommendations regarding health interventions aimed at improving the reproductive health situation in the context of fertility in Poland.

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

The raw data supporting the conclusions of this article will be made available by the authors on request.

RP conceived and designed the study. SA and MD developed the methodology. MD performed the formal analysis. RP performed the investigation. AZZ provided the resources and contributed to data curation alongside MD. RP and SA were major contributors to writing the original manuscript. BJP, MD, SA, and GM reviewed and edited the manuscript. GM supervised the project.

As all data were collected and analysed anonymously, specific consent to publish individual data was not applicable, as no identifiable information is included in this paper.