MenoScale: A novel digital tool to measure menopause symptoms and subjective quality of life – validation,preliminary insights on the menopausal experience and association with diet quality

Abstract

Background

Menopause symptoms significantly impact women’s quality of life, yet existing assessment tools may not fully capture the menopausal experience. The MenoScale is a new digital tool for reporting the number and impact of menopause-associated symptoms.

Objective

To evaluate the reliability and validity of the MenoScale, and explore its application as a research tool.

Study design

Development of novel MenoScale tool and an online validation study comparing it to the Greene Climacteric Scale (GCS) and RAND 36-item Health Survey 1.0, with 1010 peri- and postmenopausal women aged 37–70 years.

Main outcome measures

Construct validity, internal consistency, test-retest reliability, and associations with quality of life and dietary intake.

Results

The MenoScale showed good agreement between the MenoScale and GCS through Bland-Altman plot analysis. Internal consistency varied across four symptom domains: vasomotor, sexual, psychological and cognitive, and somatic (Cronbach’s α = 0.5–0.8). Structural equation modelling revealed psychological and cognitive symptoms as central to the menopausal experience. Higher Healthy Eating Index scores were associated with slightly lower MenoScale scores (β = −0.132, p < .001). In the first 12 weeks post-launch, 65,181 women from 140 countries completed the online MenoScale.

Conclusion

The MenoScale is a valid and reliable tool for assessing menopause symptoms, offering a comprehensive assessment of the menopausal experience. The association between diet quality and symptom burden warrants further investigation. The rapid uptake of the online MenoScale demonstrates its potential for large-scale data collection and empowering individuals in managing their menopausal experience.

Keywords

Menopause questionnaire validation digital tools perimenopause postmenopause HRT

Introduction

The majority of women spend approximately a third of their life in a peri- and postmenopausal state with symptoms that often significantly impact daily functioning.¹ The menopausal transition, or perimenopause, involves fluctuations in reproductive and hormonal function leading to burdensome symptoms in a significant proportion of women.² Once a woman experiences amenorrhoea for 12 months, caused by significant and permanent declines in ovarian reproductive activity, they are considered to be postmenopausal.¹

Despite affecting approximately half of the global population, menopause is under-researched, with limited understanding of interventions that alleviate symptoms. Currently, the gold standard of care is Hormone Replacement Therapy (HRT), particularly for vasomotor symptoms, for example, hot flushes.³ While evidence is limited compared to HRT, lifestyle factors (diet, exercise, smoking) and other measures (metabolic health, socioeconomic status, obesity) are emerging as potential contributors to menopause onset and symptoms.^4,5 Diet, in particular, is a modifiable lifestyle factor with some evidence for associations with menopause symptoms; for example, the consumption of soy isoflavones may alleviate vasomotor symptoms.⁶ Furthermore, menopausal changes significantly impact health, increasing the likelihood of metabolic syndrome, cardiovascular risk factors, and mood disorders.⁷ Given the recognised influence of nutrition on these health variables, a reliable tool for assessing menopausal symptoms enables research into lifestyle interventions, such as diet, that could provide effective symptom management. Additionally, an open access digital tool to track menopausal symptoms would empower women to autonomously find effective solutions.

Widely used instruments for assessing menopausal symptoms, including the Kupperman Index, Menopause Rating Scale, Menopause Specific Quality of Life and the Greene Climacteric Scale (GCS),⁸ are all paper-based questionnaires. However, these scales were published between 1950 and 1996, sometimes using outdated language, without support from large-scale studies of menopausal experiences and the prevalence of symptoms. The MenoScale is a novel, digital tool for reporting the number and impact of menopause-associated symptoms. It builds on existing tools by combining prevalence-informed symptoms from the ZOE PREDICT 3 cohort (NCT04735835) alongside known symptoms, modern language, and offering open online access to the general public.

This study aimed to evaluate the reliability and validity of the MenoScale in peri- and postmenopausal women from the general UK population in comparison to the most commonly used questionnaire (the GCS) and assess its association with quality of life (QoL) using the 36-item RAND Health Survey 1.0. Secondly, we explored the results of the MenoScale from 65,181 individuals following its public launch. Furthermore, the MenoScale was applied in a research context by exploring the relationship between diet quality and menopause symptoms.

Methods

MenoScale

The MenoScale was developed in collaboration between King’s College London and ZOE Ltd, supported by the British Menopause Society. An early version was used to gather data on over 70,000 women from the PREDICT 3 cohort. Based on the responses given and user research into question wording, the questionnaire was refined to the version that is used in this study (Table 2).

The MenoScale considers 20 menopause symptoms. Participants answer on a 4-point Likert scale (‘not at all’, ‘a little’, ‘quite a bit’, ‘extremely’) to indicate how bothered they are by the symptoms in the previous 7 days. The score ranges from 0 to 100 depending on the number and impact of symptoms and can be divided into four domains: vasomotor, sexual, psychological and cognitive, and somatic.

Study design

This study (clinicaltrials.gov ID: NCT06800560) was conducted online between 13th November 2024 and 2nd December 2024 (favourable ethical opinion received from King’s College London Research Ethics Committee; LRS/DP-24/25-45554). All participants were asked to complete online versions of the MenoScale, GCS and the 36-Item RAND Health Survey 1.0 twice: initial assessment (T0) and follow-up (T1; 7 days later). At T0, participants completed a demographics questionnaire and an optional food frequency questionnaire (FFQ; PREDICT-FFQ-v1).⁹

Participant selection

Participants were required to be assigned female at birth, aged 37 to 70 years, and self-reported to be peri- or postmenopausal. Perimenopausal status was self-determined with agreement to the statement ‘my periods may be irregular, or I may have menopausal symptoms such as: hot flushes, mood changes, sleep disturbances, brain fog’. Postmenopausal status was self-determined with agreement to the statement ‘my periods have completely stopped, at least 1 year ago’. Volunteers were excluded if any of the following applied: currently pregnant and/or lactating; history of hysterectomy or oophorectomy; body mass index (BMI) of less than 18.5; currently taking greater than five prescribed medications; unable to provide informed consent online, understand the participant information sheet, or complete the questionnaires online.

Recruitment was carried out between 13th and 18th November 2024 via electronic advertisement to the ZOE Health Studies mailing list. Interested volunteers were asked to register interest by providing their date of birth and sex assigned at birth via an online questionnaire. Eligible participants were sent a detailed participant information sheet then screened at least 24 hours later against the exclusion criteria via an online questionnaire. At this stage, potential participants were also asked to select their menopause and HRT status.

Additional measures

All questionnaires were administered online via a secure online survey platform (https://www.typeform.com).

Greene climacteric scale

The GCS¹⁰ is a 21-item scale that measures menopause symptoms based on a 4-point Likert scale (‘not at all’, ‘a little’, ‘quite a bit’, ‘extremely’) to indicate how bothered the respondent is by symptoms at the moment. Symptoms are divided into four domains: psychological, somatic, vasomotor and sexual. The score ranges from 0 to 63 depending on prevalence and impact of symptoms.

36-item RAND health survey 1.0

The 36-Item RAND Health Survey 1.0 is derived from the Medical Outcomes Study and measures QoL.¹¹ It is a 36-item questionnaire divided into eight health concepts: physical functioning, bodily pain, role limitations due to physical health problems, role limitations due to personal or emotional problems, emotional well-being, social functioning, energy/fatigue and general health problems.

Dietary intake assessment

At T0, participants were asked to complete an optional FFQ.⁹ The outcomes are run through an internal data warehouse which processes a healthy eating index (HEI) score. The HEI determines how well an individual’s eating habits align with the Dietary Guidelines for Americans 2010¹² developed by the US Department of Agriculture, with a higher score (out of 100) indicating better adherence.¹³ The plant diversity index (PDI) score is calculated by assigning positive scores to plant foods and negative scores to animal foods. Higher scores (theoretical range of 18 – 90¹⁴) indicate greater diversity in plant intake.¹⁵

Statistical analysis

All analyses were carried out using RStudio 2024.09.1 + 394 ‘Cranberry Hibiscus’ Release.¹⁶ In cases of repeated completions of a questionnaire the first completion was retained and the rest deleted. If the first set of results were incomplete then the results from the second completion were retained. If demographic data or responses to the MenoScale were missing the participants were removed from the dataset. Participants with BMI >45 were excluded. Given our prior exclusion for polypharmacy, this BMI threshold served to maintain an unbiased dataset whilst removing extreme outliers. Sample characteristics were expressed as means and standard deviations (SD) for age and BMI, with comparisons made using a T test and ANOVA for normally distributed data and Kruskall-Wallis for non-normal data. Ethnicity and education were expressed as frequency and percentage of the sample with comparisons made using chi-square.

Construct validity of the MenoScale to measure menopause symptoms was assessed using Bland-Altman plots and correlation analyses. An ‘excellent’ limit of agreement was decided a priori as 10 and ‘acceptable’ at 15. Spearman’s rho was used for all correlations involving questionnaires due to the non-normal distribution and presence of 0 values. Unbiased QoL prediction ability of the MenoScale and GCS was compared using random forest machine learning with 500 trees. The R package ‘randomForest’ was used. The algorithm was trained on 70% of the data and 30% for the testing dataset. Spearman’s rho was used to assess the association between the predicted QoL and actual QoL scores. Reported symptom number was compared between the MenoScale and GCS by converting the response to the questions into a binary variable indicating the presence of the symptom regardless of burden. Paired samples T test was used to compare the two questionnaires.

Internal validity of the MenoScale was assessed using confirmatory factor analysis with the R package ‘lavaan’. The model performance was evaluated using root mean square error of approximation and standardised root mean square residual. Internal consistency of the domains was measured using Cronbach’s alpha. A threshold >0.7 was considered acceptable, >0.8 good and >0.9 as excellent. A path diagram was created using the R package ‘semPlot’ where solid lines represent direct effects, dashed lines show fixed paths.

All exploratory analysis was carried out in an age-matched sub-sample of participants (N = 561) with an overlapping age range of 48 to 58 years old. MenoScale score was compared between groups using ANCOVA with age and BMI as covariates. Structural equation modelling (R package: ‘lavaan’) was used to investigate the contribution of symptoms to the menopausal experience. BMI was included as a covariate. Multiple regression was used to explore the relationship between MenoScale score and diet quality (HEI and PDI) with age and BMI as covariates.

Results

Sample characteristics

From 4039 individuals who registered interest in the study, 1819 completed the online screening process, 1473 were eligible and 1100 participants were enrolled into the study. 1010 participants completed the T0 questionnaires (and met BMI inclusion criteria) of which 588 (58.2%) completed the T1 questionnaires (Figure S1). Participant characteristics are shown in Table 1.

Table 1.

Participants characteristics. Age represented as mean (SD), body mass index represented as median (IQR). Education and ethnicity represented as frequency (%). Peri-HRT: untreated perimenopause; peri+HRT: perimenopause with current hormone replacement therapy (HRT) use; post-HRT: untreated postmenopause; post+HRT: postmenopause with current HRT use. Significant differences between all groups in age. Significant differences in body mass index between peri-HRT and peri+HRT with post-HRT and post+HRT.

N	Perimenopause	Postmenopause	P	Subgroups				P
	Perimenopause	Postmenopause		Peri-HRT	Peri+HRT	Post-HRT	Post+HRT
	507	503		268	239	262	241
Age (years)	48.72 (4.42)	58.87 (4.64)	<0.001^*	47.47 (4.49)	50.12 (3.90)	59.62 (4.87)	58.06 (4.25)	<0.001^£
Body mass index (kg/m²)	27 (5)	25 (4)	<0.001^$	27 (6)	26 (5)	25 (4)	25 (4)	<0.001^$
Education
Non-university degree	308 (60.7%)	322 (64.0%)	0.452^%	162 (60.4%)	146 (61.1%)	166 (63.4%)	156 (64.7%)	0.762^%
University degree	199 (39.3%)	181 (36.0%)	0.452^%	106 (39.5%)	93 (38.9%)	96 (36.6%)	85 (35.3%)	0.762^%
Ethnicity
White	485 (95.7%)	490 (97.4%)	0.452^%	261 (97.4%)	224 (93.7%)	255 (97.3%)	235 (97.5%)	0.167^%
Black, Black British, Black Welsh, Caribbean or African	1 (0.2%)	1 (0.2%)		1 (0.4%)	0 (0.0%)	0 (0.0%)	1 (0.4%)
Asian, Asian British, Asian Welsh	7 (1.4%)	2 (0.4%)		4 (1.5%)	3 (1.3%)	2 (0.8%)	0 (0.0%)
Mixed or multiple ethnic groups	10 (2.0%)	8 (1.6%)		1 (0.4%)	9 (3.8%)	4 (1.5%)	4 (1.7%)
Prefer to self-describe	4 (0.8%)	2 (0.4%)		1 (0.4%)	3 (1.3%)	1 (0.4%)	1 (0.4%)

^*t test, ^£ANOVA, ^$Kruskal-Wallis, ^%Chi-square test. Significance considered at p < .05.

Internal validity of the MenoScale: Establishing symptom domain structure

Menopause symptoms can be separated into domains based on bodily systems. For the currently used version of the MenoScale, we took a data-driven approach to define a final questionnaire structure of four symptom domains: vasomotor, sexual, psychological and cognitive, and somatic. This structure aligns with accepted knowledge on bodily systems, uses internal validity measures to assess how well each question contributes to its domain, and aligns closely with widely used validated questionnaires.

The MenoScale has an adequate model structure with variable reliability of domains

A four-factor model of the MenoScale was tested using confirmatory factor analysis. The domain structure (Table 2) was decided according to model fit measures, reliability and previous knowledge of bodily systems. This led to an acceptable model fit (CFI = 0.80, TLI = 0.77, RMSEA = 0.08, SRMR = 0.06). Furthermore, Cronbach’s alpha for the model was 0.7 for vasomotor, 0.5 for sexual, 0.8 for psychological and cognitive, and 0.8 for somatic indicating mostly good reliability. Standardised factor loadings and domain structures are summarised in Table 3.

Table 2.

MenoScale. Questionnaire structure and scoring system.

Menopausal symptoms	A little	Quite a bit	Extremely	Symptom domain
Hot flushes	1	3	5	Vasomotor (0–10)
Night sweats	1	3	5	Vasomotor (0–10)
Loss of interest in sex	1	3	5	Sexual (0–10)
Vaginal dryness	1	3	5	Sexual (0–10)
Irritability	1	3	5	Psychological and cognitive (0–30)
Low mood, depression, mood swings	1	3	5
Anxiety	1	3	5
Palpitations (heart beating fast, strongly, or irregularly)	1	3	5
Brain fog	1	3	5
Memory problems	1	3	5
Sleep problems	1	3	5	Somatic (0–50)
Headaches	1	3	5
Tiredness and fatigue	1	3	5
Weight gain	1	3	5
Decreased physical strength/stamina	1	3	5
Digestive symptoms (e.g. bloating, gas, discomfort)	1	3	5
Declining skin quality	1	3	5
Declining hair quality	1	3	5
Joint and muscular discomfort	1	3	5
Bladder problems (e.g. infections, leaks, urgency)	1	3	5

Table 3.

Internal validity and reliability of the 20-item MenoScale. Values are standardised factor loadings for the four-factor model.

	Vasomotor	Sexual	Psychological and cognitive	Somatic
Cronbach’s α	0.7	0.5	0.8	0.8
Hot flushes	0.748
Night sweats	0.712
Loss of interest in sex		0.820
Vaginal dryness		0.371
Irritability			0.646
Low mood, depression and mood swings			0.709
Anxiety			0.653
Palpitations			0.362
Brain fog			0.726
Memory problems			0.668
Sleep problems				0.490
Headaches				0.423
Tiredness and fatigue				0.789
Weight gain				0.492
Decreased physical strength and stamina				0.742
Digestive symptoms				0.461
Declining skin quality				0.463
Declining hair quality				0.357
Joint and muscular discomfort				0.582
Bladder problems				0.269

All factor loadings were statistically significant at p < .001.

Path analysis of the domain structure

Path analysis was carried out to examine the relationships between the questions and their relevant domains. The results indicate varying strengths of association between the symptoms and their respective domains, most showing moderate to strong relationships. All symptom pathways were statistically significant at p < .001 and standardised factor loadings ranged from 0.8 to 0.2 where anything greater than 0.4 is generally considered a minimum for consideration and 0.7 as ideal.¹⁷ Irritability, loss of interest in sex, hot flushes, and sleep problems were fixed paths (dashed arrows) constrained to 1. Full details can be found in the path diagram (Figure 1). In summary, the MenoScale effectively captures a wide range of menopausal symptoms with varying strengths of association across four symptom domains, supporting its validity as a comprehensive measure of menopausal experiences.

Figure 1.

Path diagram of the final four-factor model of the MenoScale. Solid lines represent direct effects, dashed lines show fixed paths. Values indicate standardised factor loadings, all significant at p < .001. Model fit indices: CFI = 0.80, TLI = 0.77, RMSEA = 0.08, SRMR = 0.06. RMSEA and SRMR suggest acceptable fit; CFI and TLI are below conventional thresholds.

Construct validity of the MenoScale

The MenoScale a valid questionnaire for measuring menopause symptom number and impact

Construct validity, that is, how well the MenoScale measures the concept of menopause symptoms, was assessed using T0 measurements from all participants as a single cohort. First, the strength of the relationship between the MenoScale and GCS was explored. There was a strong, positive correlation between MenoScale and GCS scores (R_s = 0.852, adj. p < .001) (Supplemental figure S2(a)) and the different domains (vasomotor domain (R_s = 0.908), psychological domain (R_s = 0.805) somatic domain (R_s = 0.621) and sexual domain (R_s = 0.283), all Bonferroni adj. p < .001. Next, to assess whether the MenoScale and GCS differ in their ability to capture symptom numbers, we compared the mean number of symptoms reported in each questionnaire. There were statistically significant but modest differences reported in the MenoScale (M = 11, SD = 4) and GCS (M = 9, SD = 4) (t(1009) = 20.01, p < 0.001, 95% CI [1.7, 2.0]) (Supplemental figure S2(b)). Overall and domain scores, and symptom number are shown in Figure 2(a).

Figure 2.

(a) Summary of overall and domain scores from the MenoScale (MS) and greene climacteric scale (GCS). Scores presented as median (IQR). GCS score range: 0 - 63; MenoScale score range: 0 - 100. Peri-HRT: untreated perimenopause; peri+HRT: perimenopause with current hormone replacement therapy (HRT) use; post-HRT: untreated postmenopause; post+HRT: postmenopause with current HRT use. (b) MenoScale construct validity. Bland-Altman plot showing agreement between scales. GCS score has been standardised to a percentage to be comparable to the MenoScale. Mean difference = 1.35 (95% CI 0.90, 1.81). 95% limits of agreement −13.10, 15.81. Proportional bias (β = 0.09, p < .001).

The MenoScale shows good agreement with the greene climacteric scale

To assess the degree of agreement between the two methods, the GCS score was standardised to the MenoScale score by converting to a percentage. Overall, there was fairly large variability between the MenoScale and GCS, indicating that while the two scales generally agree, there can be considerable differences in scores for some individuals. There was no evidence of systematic bias between the two methods (Bland-Altman; Figure 2(b)); mean difference 1.35 out of 100 (95% CI [0.90, 1.81]), indicating that one method does not consistently score higher or lower than the other. The upper and lower 95% limits of agreement were −13.10 and 15.81, respectively. Furthermore, the difference between the MenoScale and GCS increased as the scores increased demonstrated by a small but significant proportional bias using linear regression of the differences against the means (β = 0.09, p < .001).

To account for regression to the mean, that is, the tendency for measurements to become more normal when repeated, and ensure construct validity remained when examining more typical figures, the same analyses were also performed on a sub-sample of 399 participants who completed both the MenoScale and GCS at both timepoints. For each participant, the mean of scores from both timepoints (regressed) was used (original MenoScale score: 22; regressed MenoScale score: 22; original GCS score: 14; regressed GCS score: 13). Results were similar as to when the full cohort was analysed at T0 only. First, there was a strong positive correlation with a negligible difference in coefficient to the full T0-only cohort (R_s = 0.898, adj. p < .001). Furthermore, the Bland-Altman plot (Figure S1) showed the same systematic bias with a mean difference of 1.35 (95% CI 0.45, 1.62) with acceptable upper (−10.68) and lower (12.74) 95% limits of agreement. There was also a similar, significant proportional bias (β = 0.12, p < .001).

Finally, 588 (58.2%) of the total sample of 1010 participants, who completed the MenoScale at endpoint, were used to analyse test-retest reliability. There was a strong positive association between the two time points for both the MenoScale and GCS (MenoScale: R_s = 0.854, p < .001; GCS: R_s = 0.825, p < .001) (Supplemental figure S2(c)-(d)).

Quality of life validation

The MenoScale is highly correlated with and able to predict quality of life

Correlation analyses revealed strong relationships between both menopause questionnaires and physical/emotional QoL. The QoL score (from the 36-Item RAND Health Survey 1.0) was negatively correlated with the MenoScale (R_s = −0.650, p < .001) and GCS scores (R_s = −0.688, p < .001), (Figure 3(a)) as were similar domains; the somatic domain of the menopause questionnaires and the RAND’s physical functioning domain (MenoScale: Rs = −0.478, p < .001; GCS: Rs = −0.390, p < .001); and the psychological domain of the menopause questionnaires and the RAND’s emotional well-being domain (MenoScale: Rs = −0.638, p < .001; GCS: Rs = −0.707, p < .001).

Figure 3.

MenoScale (MS) and quality of life (QoL). (a) Correlation between overall QoL score and MenoScale (R_s = −0.650) and greene climacteric scale (GCS; R_s = −0.688). (b) Correlation between actual QoL scores and predicted QoL based on MS (R_s = 0.823) and GCS (R_s = 0.873) using random forest machine learning. All correlations p < .001. Purple points mark GCS score and green points mark MS score.

Moreover, the MenoScale and GCS demonstrated comparable and substantial efficacy in predicting quality of life in those experiencing menopause. Random forest machine learning revealed moderate predictive power and accuracy in determining QoL (MenoScale: RMSE = 12.82, MAE = 9.85; GCS: RMSE = 12.24, MAE = 9.58). The MenoScale explains 39.01% of variance in overall QoL, whilst the GCS explains 44.47%. When comparing the correlations between the actual and predicted QoL score both the MenoScale and GCS result in a strong positive association with negligible differences between the two (MenoScale: R_s = 0.823, p < .001; GCS: R_s = 0.873, p < .001) (Figure 3(b)). These findings suggest that both the MenoScale and GCS demonstrate comparable and substantial efficacy in predicting quality of life in those experiencing menopause.

Exploration of the menopausal experience using the MenoScale as a research tool

Cross-sectional analyses were carried out to explore the application of the MenoScale as a research tool in understanding menopausal experiences within the study cohort depending on menopause status and HRT usage.

Menopausal experiences are nuanced by menopause status and hormone replacement therapy

In a sub-sample of 561 participants with an overlapping age range of 48 – 58 years, to account for menopause as an age-related process, the median (IQR) MenoScale score for the groups were as follows: peri-HRT = 23 (21), peri+HRT = 24 (19), post-HRT = 21 (19), post+HRT = 17 (17). After square root transformation of the MenoScale scores, to account for a severe skew in the data, there was a significant effect of group (F(3,556) = 8.443, p < .001) whilst controlling for BMI. Post-hoc analysis with Bonferroni correction revealed that peri-HRT had a marginally higher transformed score compared to post+HRT (mean difference = 0.52, 95% CI [0.05, 0.98], p = .022). Peri+HRT also had a higher transformed score compared to post+HRT (mean difference = 0.72, 95% CI [0.28, 1.15], p < .001). Finally, post-HRT had a higher transformed score than post+HRT (mean difference = 0.52, 95% CI [0.03, 1.01], p = .032) (Figure 4(a)). There was no significant difference between the perimenopausal groups (mean difference = −0.20, 95% CI [−0.64, 0.23], p = 1.000) and -HRT groups (mean difference = −0.01, 95% CI [−0.50, 0.49], p = 1.000). Finally, there was no difference between peri+HRT and post-HRT (mean difference = 0.20, 95% CI [−0.27, 0.67], p = 1.000). To summarise, postmenopausal women using HRT had lower MenoScale scores compared to other groups, while perimenopausal women, regardless of HRT use, and postmenopausal women not using HRT reported similar, higher scores.

Figure 4.

Menopausal experience and dietary associations. (a) MenoScale scores across groups (N = 561, age 48-58 years). ANCOVA showed significant group differences (F(3,556) = 8.433, p < .001), controlling for BMI. Post-hoc: higher scores in peri-HRT, peri+HRT, post-HRT versus post+HRT (all p < .05). (b) Standardised factor loadings of symptoms (CFI = 0.932, TLI = 0.931, RMSEA = 0.068). (c) Multiple regression coefficients for diet quality measures healthy eating index (HEI, β = −0.132, p < .001) and plant diversity index (PDI). Dashed lines: range of actual data.

Psychological and cognitive symptoms are the most central to the menopausal experience

To explore differences in symptom profile depending on menopause status and HRT usage, whilst accounting for the fact that all symptoms are inherently related to each other, structural equation modelling was carried out. ‘Symptoms’ was used as the latent variable incorporating all 20 symptoms in the MenoScale. A nuanced picture of the menopausal experience in women across the menopausal spectrum was revealed (Figure 4(b)). The outcome of the model (CFI = 0.932, TLI = 0.931, RMSEA = 0.068) suggests that psychological and cognitive symptoms are central to the menopausal experience, irrespective of menopausal status or HRT use (factor loadings: λ = 0.614 - 0.909). Conversely, ‘vaginal dryness’ demonstrated weak factor loadings across all four groups (λ = 0.020 - 0.365). However, this was statistically significant only in the peri-HRT (p = .005) and post+HRT (p < .001) groups suggesting that it is not a strong or consistent indicator of the overall latent construct of ‘symptoms’ across groups. All variables and their corresponding factor loadings are summarised in Supplemental Table S1.

Factor loading comparisons across groups revealed notable differences, particularly for hot flushes. This symptom exhibited a significantly stronger factor loading in the post+HRT group (λ = 0.850, p < .001) compared to other groups (peri-HRT: λ = 0.477; peri+HRT: λ = 0.259; post-HRT: λ = 0.457). Interestingly, vasomotor symptoms, including hot flushes (0.259) and night sweats (0.172), showed the lowest centrality among menopause symptoms for perimenopausal women using HRT.

Diet quality and menopause symptoms

To investigate the relationship between diet quality and menopause symptoms, a multiple linear regression model was conducted to predict the MenoScale score, with the Healthy Eating Index (HEI) and Plant Diversity Index (PDI) as predictors, controlling for age and BMI. The overall model was significant (F(4, 962) = 30.53, R² = 0.109, p < .001). The HEI emerged as a significant predictor of the MenoScale score, albeit with a modest effect size (β = −0.132, 95% CI [-0.234, −0.030], p < .001), indicating that higher HEI scores are associated with slightly lower MenoScale scores (Figure 4(c) and (d)). Conversely, the PDI did not significantly predict the MenoScale score (β = 0.069, 95% CI [−0.062, 0.200], p = .300). The covariates age (β = −0.431, 95% CI [-0.557, −0.304], p < .001) and BMI (β = 0.585, 95% CI [0.405, 0.765], p < .001) both significantly predicted MenoScale score.

Public application of the MenoScale

The MenoScale went live online on the 4th September 2024 at https://zoe.com/menoscale. In the first 12 weeks, a total of 65,181 peri- and postmenopausal women from 140 countries (Figure 5(a)) completed the MenoScale for the first time. Out of 34,939 perimenopausal women 12,762 (36.5%) reported that they were on HRT and 22,177 (63.5%) were untreated. Out of 30,242 postmenopausal women 11,810 (39.1%) reported that they were on HRT and 18,432 (60.9%) were untreated.

Figure 5.

Online MenoScale cohort. Data collected from 65,181 peri- and postmenopausal women in the first 12 weeks from 4th September 2024 of online MenoScale launch (https://zoe.com/menoscale). (a) Global distribution of respondents. (b) Reverse cumulative percentage of symptoms reported, irrespective of severity. (c) Percentage of respondents reporting each individual symptom in groups depending on menopause status and hormone replacement therapy (HRT) usage.

The mean MenoScale score for the peri-HRT respondents was 34.7 (95%CI 34.5, 34.9), for the peri+HRT respondents it was 38.8 (95%CI 38.6, 39.1). Meanwhile, for post-HRT respondents the mean score was 35.3 (95%CI 35.1, 35.6) and for post+HRT it was 35.3 (95%CI 35.0, 35.6). Across all four groups 88 - 93% of respondents reported experiencing at least 10 symptoms (out of 20) and 54 - 62% at least 15 symptoms of menopause in the past week (Figure 5(b)). Full details of the frequencies at which individual symptoms were reported are summarised in Supplemental Table S1 and Figure 5(c).

Discussion

This study in over 1000 women demonstrates the validity of the MenoScale as an instrument to measure the number and impact of menopausal symptoms. Construct validity was confirmed by demonstrating strong associations with the GCS, one of the most widely used menopause questionnaires, in addition to the 36-item RAND Health Survey 1.0 for subjective QoL.

A combination of daily symptom diaries, standardised questionnaires and objective measurements, such as hormonal testing, would be the most comprehensive method for tracking symptoms. However, this would be an unrealistic endeavour with the high user burden of daily symptom diaries, subjective bias in self-reported questionnaires and variability in hormonal markers (e.g. follicle-stimulating hormone)¹⁸ throughout the cycle. Further work investigating clinical and biological markers of symptoms alongside self-reported experiences is needed to establish a true gold standard. However, assessment over time using a validated questionnaire, for example, MenoScale, to assess deviation from the norm may offer a superior method of managing an individual’s menopause experience.

The MenoScale, which is a low user burden option, offers several strengths compared to existing tools. It uses modern and accessible language, avoiding potentially alienating terms, such as ‘excitable’, found in the GCS. Moreover, the MenoScale has consolidated multiple aspects of depression that the GCS separates into several questions. For instance, where the GCS asks about ‘feeling unhappy or depressed’, ‘crying spells’, and ‘loss of interest in most things’ as separate items, the MenoScale captures this in the single question: ‘low mood, depression and mood swings’. This allows the MenoScale to include more questions about somatic symptoms, thereby providing a more thorough assessment of menopausal symptoms whilst maintaining the length of the questionnaire at 20 questions to minimise user burden. Another strength is the MenoScale’s specific timeframe, asking about symptoms ‘in the past week’, which provides a clear reference point for respondents. In contrast, the GCS has a more qualitative phrasing of ‘at the moment’ which may be considered vague and open to interpretation by the respondent. Therefore, the MenoScale enables more accurate long-term and real-time tracking of symptom burden. Future studies should address external validity of the MenoScale by conducting additional validation studies in other, external cohorts across a range of populations.

Vasomotor symptoms, particularly hot flushes, are traditionally considered to be a hallmark of the menopausal experience.¹⁹ However, our findings from both the validation cohort and the online MenoScale cohort present a different picture. We observed that vasomotor symptoms were not the most prevalent symptoms, whilst psychological and cognitive symptoms were reported in a higher proportion of women. These findings are supported by a recent study that clustered over 145,000 menopause symptom logs in pre-, peri- and postmenopausal women, who all reported high rates of fatigue and psychological symptoms.²⁰ Interestingly, vasomotor symptoms and sexual symptoms stood out as domains that had variable impact depending on the menopausal stage, echoing several other cohort studies from the literature.²¹ Although considering specific symptom domains can be helpful, particularly in research contexts investigating biological systems, our data suggests that examining the symptom experience as a whole might be more meaningful and reflective of the lived experience of those experiencing menopause. As such, whilst the menopausal transition is characterised by a complex interplay of psychological, cognitive, vasomotor, somatic and sexual symptoms. These experiences should be considered in conjunction with biological measures such as hormone levels and changes in estrogen reserves over time to provide a more comprehensive understanding of menopause. By moving beyond the traditional focus on vasomotor symptoms, broader effective interventions for symptom management can be investigated.

The limits of agreement for the MenoScale and GCS was decided a priori to be ‘acceptable’ at 15 and ‘excellent’ at 10, equating to 15% and 10% of the scores. Although still evidence of variability, the MenoScale and GCS fell within this range. While our findings are generally consistent, a more nuanced picture emerges when separated into menopause status and HRT usage. We report that hot flushes and night sweats were the least central to the overall symptom experience for perimenopausal women using HRT, in particular. This suggests that HRT may be especially effective in mitigating these symptoms during the perimenopausal phase. This observation is well supported by previous literature. For example, a systematic review of 24 randomised controlled trials reported that HRT reduced the severity and frequency of hot flushes by 75% compared to placebo.²² However, the stronger loading in postmenopausal women who are using HRT implies that the relationship between HRT and symptom experience may change across menopausal stages. This could be the result of a self-selection bias, wherein postmenopausal women experiencing persistent vasomotor symptoms despite HRT use may be more inclined to participate in menopause symptom studies. Overall, these results should be interpreted cautiously due to potential confounding factors such as type of HRT and individual variability in symptom experiences. Future research could benefit from longitudinal designs to track symptom changes over time and across menopausal transitions, potentially informing more targeted and effective interventions.

Exploratory analysis of dietary data from an optional FFQ at baseline revealed that better diet quality, as measured by the HEI, is associated with lower MenoScale scores. However, the practical significance of this finding is limited due to the small effect size. Specifically, a 10-point increase in HEI corresponded to only a 1.32-point decrease in MenoScale score. However, it has been shown by others in multiple cohorts that a healthy dietary pattern is associated with a later onset and less burdensome menopausal experience.⁶ Therefore, further work is warranted in a long-term cohort with a randomised controlled trial. Future research should focus on longitudinal studies with larger sample sizes and more comprehensive dietary assessments to elucidate the complex relationship between diet and menopausal symptoms.

While our analysis focused on the 65,181 peri- and postmenopausal women who completed the MenoScale for the first time, the total number of responses to the online MenoScale reached 86,260 within 12 weeks of release, when including those who were unsure of their menopause status, chose not to disclose it, or were premenopausal. In fact, 59,038 completions occurred within the first 14 days alone. This demonstrated a substantial demand for a user-friendly, accessible tool to evaluate menopause symptoms. Given the rapid uptake and clear public interest in such a tool, it is crucial that the MenoScale can give comparably accurate information to currently accepted methods. The findings of this study provide evidence for the reliability and validity of the MenoScale as a tool to measure menopause symptoms and subjective impact on QoL.

Since the COVID-19 pandemic, there has been an extraordinary shift into digital health and remote studies.²³ Research of this nature creates opportunities for increased diversity and accessibility. An online tool such as the MenoScale offers multiple benefits in this context. Firstly, it raises awareness by increasing understanding around menopause and its symptoms, which is particularly important considering a recent qualitative study in over 800 women found that half of them did not feel informed about peri/menopause at all.²⁴ Secondly, it empowers those experiencing menopause by allowing them to track and manage their symptoms independently, and therefore understand which interventions are beneficial for them. In fact, symptom monitoring as an intervention has been reported to be beneficial for improving menopause-associated symptoms in several studies.²⁵ Moreover, by offering the MenoScale as an open access tool with no licencing requirements, we aim to reach those who may face difficulties accessing such resources. In addition, it facilitates discussions with healthcare providers with data-backed insights, enabling individuals to have greater autonomy over their own care. Finally, on a wider scale, it presents opportunities for large-scale data collection, contributing to our understanding of menopause experiences across diverse populations.

Supplemental Material

Supplemental Material - MenoScale: A novel digital tool to measure menopause symptoms and subjective quality of life - Validation, preliminary insights on the menopausal experience and association with diet quality

Supplemental Material for MenoScale: A novel digital tool to measure menopause symptoms and subjective quality of life – Validation, preliminary insights on the menopausal experience and association with diet quality by Curie Kim, Lucy Marples, Alexander Platts, Kate M Bermingham, Federica Amati, Haitham Hamoda, George Pounis, Jonathan Wolf, Tim D Spector, Wendy L Hall and Sarah E Berry in Post Reproductive Health.

Footnotes

Acknowledgements

We would like to extend our heartfelt thanks to the British Menopause Society for supporting the development of the MenoScale. We would also like to thank Professor Paul Franks for giving us his time to review the analysis carried out in this study. Last, and most certainly not least, we would like to thank all the participants in this study and express our gratitude for the overwhelming support and interest we received during recruitment.

Contributorship

C.K., L.M., A.P., K.M.B., W.L.H., S.E.B. were involved in protocol development. C.K. gained ethical approval. C.K. conducted participant recruitment and data collection. C.K. conducted the data analysis. C.K. wrote the manuscript. All authors reviewed the manuscript and approved the final version.

Declaration of conflicting interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: T.D.S. and J.W. are co-founders of ZOE Ltd. S.E.B. and W.L.H. are consultants for ZOE Ltd. C.K., L.M., A.P., K.M.B., F.A. are or have been employees of ZOE Ltd. C.K., K.M.B., J.W., T.D.S., S.E.B., F.A., receive options with ZOE Ltd.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by ZOE Ltd.

Code availability

The scripts for the statistical analysis are freely available upon request to ZOE Ltd. Application is via data.papers@joinzoe.com. Code will be made available within 2 months of the request.

Guarantor

S.E.B.

Ethical approval

Favourable ethical opinion was received from the research ethics committee at King’s College London (reference number: LRS/DP-24/25-45554). The study was registered on clinicaltrials.gov (INSERT ID).

Informed consent

Digital informed consent was collected from all participants via an online consent questionnaire.

ORCID iDs

Curie Kim

Haitham Hamoda

Data Availability Statement

The study data can be released to bona fide researchers submitting a research proposal approved by a subpanel of our scientific advisory board. We have meetings once per month with independent members to assess proposals. The data will be anonymised and conform to UK General Data Protection Regulation standards. Access request proposals should be sent to data.papers@joinzoe.com. The MenoScale and associated terms terms of use are available at . Please refer to the FAQs for full information.

Supplemental material

Supplemental material for this article is available online.

References

Gatenby

Simpson

. Menopause: physiology, definitions, and symptoms. Best Pract Res Clin Endocrinol Metabol 2024; 38: 101855.

Santoro

. Perimenopause: from research to practice. J Womens Health (Larchmt) 2016; 25: 332–339.

Cameron

Cohen

Sewell

, et al. The art of hormone replacement therapy (HRT) in menopause management. J Pharm Pract 2024; 37: 736–740.

Schoenaker

DAJM

Jackson

Rowlands

, et al. Socioeconomic position, lifestyle factors and age at natural menopause: a systematic review and meta-analyses of studies across six continents. Int J Epidemiol 2014; 43: 1542–1562.

Bermingham

Linenberg

Hall

, et al. Menopause is associated with postprandial metabolism, metabolic health and lifestyle: the ZOE PREDICT study. EBioMedicine 2022; 85: 104303.

Yelland

Steenson

Creedon

, et al. The role of diet in managing menopausal symptoms: a narrative review. Nutr Bull 2023; 48: 43–65.

Davis

Pinkerton

Santoro

, et al. Menopause-Biology, consequences, supportive care, and therapeutic options. Cell 2023; 186: 4038–4058.

Sourouni

Zangger

Honermann

, et al. Assessment of the climacteric syndrome: a narrative review. Arch Gynecol Obstet 2021; 304: 855–862.

Bermingham

Linenberg

Francis

, et al. The development and validation of a novel digital food-frequency questionnaire to assess habitual diet: the PREDICT food frequency questionnaire. UNDER REVIEW.

10.

Greene

. Constructing a standard climacteric scale. Maturitas 1998; 29: 25–31.

11.

Riesenberg

Glass

. The medical outcomes study: an application of methods for monitoring the results of medical care. JAMA 1989; 262: 943.

12.

McGuire

. U.S. department of agriculture and U.S. department of health and human services, dietary guidelines for Americans. 7th ed. Washington, DC: U.S. Government Printing Office, 2010, pp. 293–294. January 2011. Adv Nutr 2011; 2.

13.

Guenther

Casavale

Reedy

, et al. Update of the healthy eating index: HEI-2010. J Acad Nutr Diet 2013; 113: 569–580.

14.

Satija

Bhupathiraju

Rimm

, et al. Plant-based dietary patterns and incidence of type 2 diabetes in US men and women: results from three prospective cohort studies. PLoS Med 2016; 13: e1002039.

15.

Satija

Bhupathiraju

Spiegelman

, et al. Healthful and unhealthful plant-based diets and the risk of coronary heart disease in U.S. adults. J Am Coll Cardiol 2017; 70: 411–422.

16.

RStudio Team . RStudio: integrated development environment for R. https://www.rstudio.com/ 2020.

17.

Cheung

Cooper-Thomas

Lau

, et al. Reporting reliability, convergent and discriminant validity with structural equation modeling: a review and best-practice recommendations. Asia Pac J Manag 2024; 41: 745–783.

18.

Burger

Dudley

Robertson

, et al. Hormonal changes in the menopause transition. Recent Prog Horm Res 2002; 57: 257–275.

19.

Thurston

Joffe

. Vasomotor symptoms and menopause: findings from the study of women’s health across the nation. Obstet Gynecol Clin N Am 2011; 38: 489–501.

20.

Aras

Grant

Konhilas

. Clustering of > 145,000 symptom logs reveals distinct pre, peri, and menopausal phenotypes. Sci Rep 2025; 15: 1–10.

21.

Woods

Mitchell

. Symptoms during the perimenopause: prevalence, severity, trajectory, and significance in women’s lives. Am J Med 2005; 118(Suppl 12B): 14–24.

22.

Maclennan

Broadbent

Lester

, et al. Oral estrogen and combined estrogen/progestogen therapy versus placebo for hot flushes. Cochrane Database Syst Rev 2004; 2009: CD002978.

23.

Whitelaw

Mamas

Topol

, et al. Applications of digital technology in COVID-19 pandemic planning and response. Lancet Digit Health 2020; 2: e435–e440.

24.

Aljumah

Phillips

Harper

. An online survey of postmenopausal women to determine their attitudes and knowledge of the menopause. Post Reprod Health 2023; 29: 67–84.

25.

Andrews

Hale

John

, et al. Evaluating the effects of symptom monitoring on menopausal health outcomes: a systematic review and meta-analysis. Front Glob Womens Health 2021; 2: 757706.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.10 MB