Preferences for and intention to use an app for premenstrual mental health symptoms using the Health Behaviour Model (HBM)

The study addresses the need for effective, acceptable digital tools to support premenstrual mental health symptoms. Premenstrual symptoms, including PMS and PMDD, are prevalent and impair wellbeing, functioning, and are linked to suicidality. While mental health apps show efficacy in broader populations and period-tracking apps can aid understanding of cycle–mood links, there are few dedicated apps for PMS/PMDD. The research aimed to identify user preferences for app features and to examine, via the Health Belief Model (HBM), which belief constructs predict intention to use an app for premenstrual mental health symptoms, thereby informing design and implementation strategies to enhance uptake and engagement.

Prior work suggests digital help-seeking is widely accepted in the UK, with high smartphone ownership enabling scalability. Period-tracking apps can improve understanding of the cycle–mood relationship and may reduce depressive symptoms when coupled with health information and lifestyle recommendations. Internet-delivered CBT and online peer support show promise for PMS/PMDD and are candidates for app-based delivery. The HBM has been used to predict uptake of mHealth interventions (e.g., medication adherence, COVID-19 contact tracing apps) and mental health help-seeking, as well as preventive behaviours in women’s health related to menstruation. Meta-analytic evidence indicates perceived barriers often exert the largest influence on health behaviours among core HBM constructs, while perceived susceptibility tends to be a weaker predictor. These insights motivated applying the HBM to understand intention to use a premenstrual mental health app.

Design: Cross-sectional online survey with exploratory analysis and structural equation modelling (SEM) grounded in the Health Belief Model (HBM). Recruitment: January–March 2023 via email, word of mouth, paid ads, and organic social media posts in the UK. Inclusion criteria: age ≥18, UK residence, history of menstruation, not pregnant/breastfeeding, not perimenopausal/menopausal. Incentive: raffle for three £50 vouchers. Sample: 578 began and indicated premenstrual mental health concerns; 530 (91.7%) had complete data and were analysed (mean age 35.85, SD 7.28; predominantly female and White). Measures: - Premenstrual symptoms and impairment: Premenstrual Symptom Screening Tool (PSST), 19 items plus an additional item for romantic/intimate relationships (total 20 items), rated from Not at all to Severe. - App preferences: participants selected top three desired app features from a predefined list (e.g., psychoeducation, symptom monitoring, help with diagnosis/referral). Willingness to pay and preferred payment model were assessed. - Health Belief Model constructs: 25 items co-designed with a psychiatrist and reviewed by people with lived experience to assess perceived benefits (3 items), perceived barriers (5), perceived severity (3), perceived susceptibility (4), self-efficacy (3), cues to action (4), and behavioural intention (3). Items scored on a 6-point Likert scale (1=Strongly disagree to 6=Strongly agree). Procedures: Anonymous Qualtrics survey (15–20 minutes), adaptive branching to show relevant questions. Data cleaning excluded respondents indicating perimenopause/menopause elsewhere in the larger survey. Analyses: Descriptive statistics in Excel. Internal consistency (Cronbach’s alpha) for HBM constructs. Spearman rank-order correlations among HBM constructs and with age, education, income. Group differences (employment vs not in paid employment) via Mann–Whitney U. SEM (Stata 17): confirmatory factor analysis to test HBM latent structure, then structural model predicting behavioural intention from HBM constructs. Non-normality addressed with Satorra–Bentler-scaled chi-square. Model fit assessed with χ², CFI (≥0.90 good), RMSEA (low values indicate good fit), and SRMR (<0.08 good). All exogenous latent variables allowed to correlate.

Sample characteristics: N=530; 95.66% female (n=507); 94.15% White (n=499); mean age 35.85 (SD=7.28). Premenstrual symptoms: 100% endorsed at least one symptom ≥mild. Most frequent symptoms: physical symptoms 96.42% (n=511), anger/irritability 95.85% (n=508), fatigue/lack of energy 93.34%; fatigue was most often severe (36.23%, n=192). Functional impairment: 97.17% (n=515) reported ≥mild impairment in ≥1 domain; most frequent domain impacted: work/studies 83.40% (n=442); most severely impacted: romantic/intimate relationships 15.85% (n=84). App preferences: - Top features: symptom monitoring over time 74.72% (n=396); psychoeducation 57.92% (n=307); help obtaining referral to HCP 48.11%; self-help tips 40.19%; treatment recommendations 40.00%; help obtain a diagnosis 33.02%. - Cost: 52.64% (n=279) unwilling to pay; among those willing (47.36%, n=251), preference for one-off payment 65.74% (n=165) vs subscription 34.26% (n=86). HBM descriptives: Overall alpha=0.76. Means (SD): perceived benefits 4.13 (0.99), barriers 3.08 (0.83), severity 4.05 (1.22), susceptibility 3.84 (0.99), self-efficacy 5.71 (0.48), cues to action 4.83 (0.52), behavioural intention 4.67 (1.02). Bivariate correlations with intention: cues to action r=0.45 (p<0.001), perceived benefits r=0.45 (p<0.001), severity r=0.33 (p<0.001), susceptibility r=0.22 (p<0.001), barriers r=−0.36 (p<0.001); self-efficacy not significant (r=0.83, p=0.057; as reported). SEM model fit and predictors: Good fit: χ²(254)=565.91, p<0.001; CFI=0.939; RMSEA=0.048; SRMR=0.058. HBM constructs explained 58.22% of variance in behavioural intention. Significant predictors: cues to action β=0.49 (p<0.001), perceived barriers β=−0.22 (p<0.001), perceived severity β=0.16 (p=0.012), perceived benefits β=0.10 (p=0.035). Non-significant: self-efficacy β=−0.01 (p=0.667), perceived susceptibility β=0.07 (p=0.332). Sociodemographic associations: - Employment: lower perceived susceptibility (U=5.88, p=0.015) and severity (U=18.76, p<0.001) among those in paid employment vs not. - Age: negatively correlated with perceived severity (r=−0.15, p=0.001). - Education: negatively correlated with perceived benefits (r=−0.11, p=0.013), perceived severity (r=−0.12, p=0.007), and behavioural intention (r=−0.10, p=0.031); positively with self-efficacy (r=0.15, p=0.001). - Income: negatively correlated with perceived benefits (r=−0.11, p=0.014), barriers (r=−0.10, p=0.033), susceptibility (r=−0.19, p=0.001), and behavioural intention (r=−0.09, p=0.046); positively with self-efficacy (r=0.12, p=0.012).

Findings indicate substantial interest in apps for premenstrual mental health, with strong preference for symptom monitoring and psychoeducation. Cost is a major consideration; many prefer free access, and among those willing to pay, one-off payments are favoured. The HBM analysis shows intention to use is primarily driven by cues to action—particularly co-design with and endorsement by healthcare professionals or reputable institutions—followed by minimising perceived barriers (digital discomfort, privacy, app quality), higher perceived severity of symptoms, and highlighting perceived benefits (e.g., better understanding, improved relationships). Self-efficacy and perceived susceptibility did not predict intention in this highly digitally literate sample, suggesting other determinants may be more salient. The results support strategies that: co-design with clinicians, academics, and people with lived experience; promote HCP awareness and recommendation pathways; transparently address privacy/data use; include human-connection features (peer support, data sharing with HCPs); and clearly communicate evidence-based benefits. Sociodemographic patterns (education, income, employment, age) relate to HBM constructs and underscore the need for accessible, low- or no-cost, user-friendly apps to broaden equitable uptake.

There is clear user interest in premenstrual mental health apps, especially those offering symptom monitoring and psychoeducation at no cost. Intention to use is most strongly influenced by cues to action, followed by perceived barriers, severity, and benefits. Developers should prioritise co-design with trusted experts and users, secure HCP endorsement, mitigate barriers through transparent privacy practices and quality evidence, and communicate therapeutic benefits. Future research should: validate HBM items in this population; examine additional behavioural models (e.g., theory of planned behaviour) to explain residual variance; and investigate translation from intention to real-world adoption and sustained engagement.

The sample was predominantly highly educated, higher income, female, and White, recruited online and via social media, potentially limiting generalisability and inflating digital literacy. The focus of recruitment may have attracted individuals with more severe premenstrual symptoms. HBM items were developed for this study, not previously validated, and preselected—some relevant beliefs may be missing or items may not suit all users. Participants may have varied assumptions about what a premenstrual mental health app entails, potentially influencing responses. The study measured intention, not actual app uptake or sustained use. Group differences in ethnicity and gender were not analysed due to small subgroup sizes. Non-normal data required robust estimation, though fit indices indicated good model fit.