Descriptive analysis of South African women’s experience with hormone-related health issues.

Authored by: Zuha Ajlan – Lead Data Collection and Research Specialist at Femmflo.

Introduction

Historically, women have been excluded from scientific research due to the cyclic variation in their hormones. It had been assumed that results obtained from men applied to women as well (Dresser, 1992; Wainer et al., 2020). This has led to a gross underrepresentation of women in research, delayed symptom recognition, diagnostics and even treatment. In fact, it was decades before we learned that women present different symptoms to men when experiencing a myocardial infarction (heart attack). Several studies across the globe have reported delayed times in women getting a diagnosis compared to their male counterpart. Women have reported to take 2.5 years longer to be diagnosed with cancer to men. It takes 7 years for women to get diagnosed with endometriosis. A cross-sectional study involving 1,385 women, primarily from North America and Europe, and 9% from other regions, found that it took an average of three medical professionals six months to over two years to diagnose polycystic ovary syndrome (PCOS) (Gibson-Helm et al., 2017).

Beyond endocrinological disorders, women experience depression differently during different reproductive health period. Instead, women may experience premenstrual dysphoric disorder (PMDD) before their menstrual period due to elevated ovarian-derived hormones including oestrogen and progesterone. This is because progesterone is anti-inflammatory which can lead to poor sleep and mood. Progesterone levels are high during the luteal phase of the menstrual cycle (after ovulation) and drastically drops during menstruation (or period). This may trigger PMDD in some women (Gordon et al., 2015). Conversely, women over the age of 40 years, begin to experience higher follicle stimulating hormone (FSH) compared to Luteinising Hormone (LH) whereas in younger women, LH levels are higher than FSH to trigger ovulation. The higher FSH compared to LH is indicative of menopause and causes hot flashes, lower oestrogen and progesterone levels mood changes and deteriorating fertility. Recent literature has highlighted the impact of changes in sex hormones on mental health, disease contraction and pathogenesis, physical activity, symptomology and sleep quality (VanWijk and Kolk, 1997; Marroco and McEwen, 2016; Ortona, Pierdominici and Rider, 2019; Barajas-Martínez et al., 2021).

Indeed, literature has demonstrated that oestrogen promotes inflammation, protects against infections, and contributes to good sleep. Conversely, progesterone promotes an anti-inflammatory environment to prepare the body for pregnancy, however, this results in increased susceptibility to infections, low energy levels and poor sleep quality. Furthermore, a recent study in Durban (South Africa) reported elevated sex and thyroid hormones were associated with prediabetes in patients (Sosibo et al., 2024).

Unfortunately, there is a disproportionate delay in diagnosis for women. Endometriosis affects 10% of women worldwide that causes debilitating abdominal cramping, pain during sex, increases pain sensitivity and increases risk of cervical cancer. Yet it takes 7 – 10 years for women to be diagnosed by a healthcare professional (Mecha et al., 2022; Anchan et al., 2023). It was initially assumed that women of African descent had a lower prevalence of endometriosis. However, recent studies have speculated it may be because of the lack of resources, information and diagnosis that led to this assumption (Ohayi, Onyishi and Mbah, 2022). One of the reasons that may lead to a lack of information pertaining to hormonal health is cultural stigma against hormonal health. For instance, menstruation is considered a taboo topic in many cultures and should be kept private. A study reported that young girls in East and South Africa would acquire their information regarding menstruation primarily from their friends, mothers and science teachers (Asumah et al., 2022; Khamisa et al., 2022; Ohayi, Onyishi and Mbah, 2022). There were some girls who were unaware of menstruation and were taken by surprise (Asumah et al., 2022; Khamisa et al., 2022). Unfortunately, this is not an isolated experience as the shame and discomfort surrounding female reproductive health is worldwide.

However, in recent decades, there has been a surge in the use of technology to seek medical advice. Women sought information regarding their symptoms online regarding their health and search for methods to treat and manage their potential disorder. Smartphone applications such as period trackers, Samsung Health and Apple Health are now used to monitor health. Studies have shown that these smartphone apps have improved the management and quality of different disorders including endometriosis, PCOS and menopause (Kristjánsdóttir et al., 2013; Scott et al., 2018; Lee and Lee, 2023; Rohlof et al., 2024).

There is a lack of studies that investigated the prevalence of hormonal disorders in women in South Africa, their perception on the use of smartphone applications and what they look for in using smartphone application. Thus, this cross-sectional study investigated the prevalence of hormonal disorders in 223 South African women, their experience using smart hormonal health applications and what features they look for to develop an AI-driven application that can answer their demand.

Methods

A survey was created on TypeForm by Tal Cook which assessed the demand the use and demand for hormonal and mental health tracking apps in the South African (Supplement A). The survey was specifically targeted at individuals assigned female-sex at birth, aged 20 years old and above and resided in South Africa. This survey comprised both multiple choice and open-ended questions.

According to TypeForm, there was a 70% complete rate of the survey. It was further noted that most respondents skipped the open-ended questions. Thus, a second survey was made by Zuha Ajlan which reduced the changed the open-ended questions to multiple choice questions based on the responses received on TypeForm and had two open-ended questions (Supplement B). The new market survey was distributed to group chats and social media by the Femmflo team, Epicentre and Wits Reproductive Health and HIV institute.

The data was statistically analysed using Microsoft Excel (Microsoft 365) and R-studio with P<0.05 was considered significant.

Results

There were 66 respondents on TypeForm and 159 respondents on the Google Form survey (N = 225). However, two respondents were under the age of 20 years, and were excluded from the analysis. Thus 223 respondents were included in the analysis.

Demographic of participants:

Majority of respondents the respondents were between the 20 and 39 years of age (78%), followed by 19% respondents aged 40 – 49 years and the remaining 3% were over the age of 50 years.

Health tracking method:

Half of the women surveyed (50%) used mobile applications such as Flo, Clue, Samsung Health, and Apple Health to monitor their hormonal and mental health (Figure 2). Medical examinations were the second most used method (26%) for health tracking. Wearable devices and self-tracking methods were the least popular options, with 3% and 1% of respondents, respectively. Notably, 30% of women did not employ any specific method to monitor their health.

Hormonal health disorders in respondents and experiences:

This study further found a Spearman correlation between age group and hormonal disorders – namely: PCOS (P=0.016), endometriosis (P=0.016), oestrogen dominance (P=0.04), low progesterone (P=0.016), thyroid dysfunction (P=0.013), adrenal fatigue (P=0.004), PMS (P=0.004), PMDD (P=0.04), uterine fibroids (P=0.01), other (P=0.004), none (P=0.004) and undiagnosed (P=0.004). The results suggest that the prevalence of these disorders may vary across different age demographics, and that as age increases, the incidence or severity of these hormonal disorders could also change.

The most prevalent hormonal health disorders among our respondents were polycystic ovary syndrome (PCOS) (27.3%), premenstrual syndrome (PMS) (21%), low progesterone (16%) and endometriosis (13%). Moreover, 7% of individuals reported "other" conditions, including HIV, hypertension, chronic migraines and Ehlers-Danlos syndrome. However, 3% women, predominantly between 20 and 29 years old, remained undiagnosed – with two women suspecting PCOS but did could not get a diagnosis due to high costs.

Indeed, 45.9% of participants identified cost as a major barrier to accessing hormonal care, second only to the lack of information and resources (56.8%) (Figure 3). Women also reported lack of personalized healthcare and difficulty in accessing healthcare professionals as challenges for their healthcare. Furthermore, 30% women reported that managing the side effects of treatment were also challenging.

These challenges coupled with their diagnosed and undiagnosed hormonal health disorders resulted in poor sleep quality, low energy levels, impaired mood, feelings of frustration and helplessness, affects their ability to work and even their relationships with their partners as reported by respondents that it affects their sex drive.

Use of contraception:

Out of 223 women, 60% reported having used birth control and 40% did not. Of these women who used birth control 90.2% used only one type of birth control while the remaining 9.8% had used at least two different types of birth control. Respondents had reported they were either still using birth control or had discontinued. However, we did not inquire the reason for birth control – regulate their hormonal disorders, to prevent unwanted pregnancy or other reasons – thus weren’t reported. Oral contraceptives (22%) and implants (18%) were the most used form of contraception whereas 2% reported “other” which included the use of rings, patch or tubal ligation.

Women’s perception on the use of AI for hormonal tracking:

A significant majority of women (58.2%) have not utilized AI-powered health monitoring tools to track their health. Furthermore, a substantial proportion (93.8%) have not engaged with hormone health testing via mobile apps. This lack of adoption aligns with the finding that 88.4% of women were unaware of the possibility of hormone testing through app-based platforms.

In the Google Form, participants were asked, "When using health applications or services, what aspects do you find most frustrating?" This question was not included in the TypeForm survey, resulting in responses from 157 participants. Among these respondents, 101 reported that high subscription prices were a significant source of frustration, while 42 expressed dissatisfactions with the lack of personalized care. Additionally, 35 respondents noted that a complicated interface was problematic. Notably, 10 individuals indicated that they had not used health applications before. It is important to highlight that some respondents selected multiple options in their answers.

Discussion

In the age of technology, smartphone applications are used to monitor health. Most smartphone applications track user’s menstrual health, mental health, provide supplement suggestions or only provide hormone testing. None of these applications look at the effect of hormonal imbalance and mental health on users and provide an AI-driven tool that can give suggestions of potential differentials (different possible diagnosis), provide at-home and in-laboratory testing and lifestyle suggestions. Thus, Femmflo is the first smartphone application which provides user with an AI-driven tool that can help accelerate the hormonal imbalance diagnosis, make recommendation for mental health and lifestyle choices and answer users questions.

This cross-sectional study was conducted to evaluate South African women’s experience with hormonal imbalance and mental health. Our survey collated responses from 223 women aged 20 years and above in South Africa. Our study demonstrated that most women between 20 – 39 years of age were diagnosed with a hormonal disorder including polycystic ovary syndrome (PCOS) (27.3%), premenstrual syndrome (PMS) (21%), low progesterone (16%) and endometriosis (13%). Furthermore, some women had disclosed “other” disorders with two women stating they have “cysts on their ovaries” but did not select PCOS, two women disclosed they were HIV positive, some were hypertensive, had pelvic inflammatory disorder and Ehlers-Danlos Syndrome. It should be noted that at the time of launch, Femmflo is geared towards women aged 20 – 45 years old, we have included menopause as a potential differential because menopause is an individual experience, and some may experience menopause symptoms earlier than others. Femmflo will have a tailored package towards menopause for women aged 45 and above in the coming years.

Next, most of undiagnosed women were aged 21 – 29 years of age. It should be noted that while there were 19 undiagnosed women, we cannot distinguish whether the 31 women who selected “none” presented with symptoms, had complaints, medical or family history, or were being treated with any other hormonal disorder . Some respondents further stated that they couldn’t get a diagnosis “because of the high cost/it’s expensive/can’t afford it”.

Furthermore, South African women reported that they experienced irritability, difficulty in losing or gaining weight, low sex drive, and poor sleep which affected their relationships, ability to do work, physical discomfort and pain, and caused fatigue. This aligns with the experiences of women with endometriosis in New Zealand (Ellis, Munro and Wood, 2022), Croatia (Škegro et al., 2021) and Switzerland (Girard et al., 2023), PCOS in the UK who, in addition to the aforementioned symptoms and feelings, felt like they were less ‘feminine’ (Lau et al., 2022).

It should be noted that the delay in diagnosis is not an isolated experience. Women in New Zealand reported feeling relieved, overwhelmed, angry, scared and confused after getting diagnosed (Ellis, Munro and Wood, 2022). They further reported that the negative emotions (angry and scared) were due to their participants fear of endometriosis causing infertility (Ellis, Munro and Wood, 2022). According to Girard and colleagues (2023), the delay in diagnosis was due to the lack of acknowledgement of the menstrual pain experienced by patients – which was considered normal, the lack of information provided by their medical provider about different gynaecological disorders when considering a diagnosis, and the lack of resources, knowledge and information available to the patients.

In addition to the previous factors described by Girard et al., (2023), cultural and societal perceptions should also be considered. For instance, a study in Soweto – a township in Johannesburg, South Africa – reported that African women were unaware that there was a specific name for menopause (Matina, Mendenhall and Cohen, 2024). They further reported seeing menopause as a normal part of aging; a time when menstruation and bearing children ended. However, some women also felt that menopause could lead to a loss of femininity and sexual desire (Matina, Mendenhall and Cohen, 2024). In contrast, adolescent girls (12 – 19 years) feel guilty, ashamed, humiliated and confused when experiencing their first menstrual period (and subsequent menses) which would lead to absenteeism and impact their performance in school (Asumah et al., 2022; Khamisa et al., 2022). Furthermore, cultural norms may tell them that they are “impure” and not allowed to participate in certain activities (Asumah et al., 2022; Khamisa et al., 2022). Thus, creating confusion and resistance from sharing their symptoms, seeking professional help or even acquiring credible resources.

In the current age of technology, there has been a significant integration of technology in healthcare through applications and telemedicine. In fact, studies have shown that the use of applications and smart devices aid in improving clinical symptoms and quality of life of patients (Kristjánsdóttir et al., 2013; Scott et al., 2018; Rohlof et al., 2024). However, only 39% of respondents had reported having used smartphone applications to monitor their hormonal health. 88.4% of women were unaware that smartphone applications could be used to monito their hormonal health. This finding underscores the need to educate women about the use of smartphone applications to monitor their health.

Our respondents further reported that they were dissuaded from these applications because of their high subscription cost, difficult user interface and lack of personalized care. Furthermore, 66% of respondents were willing to pay a monthly subscription fee of R500 or less, 22% between R500 and R2000, and 12% over R2000.

When asked about their ideal application, there was a strong emphasis on comprehensive resources on hormonal disorders, personalized recommendations, and access to healthcare professionals among others (see Table 2 for full list). This aligns with another study that investigated the desired features of a mental health app for pre- and menopausal women, with mental health education resources, symptom tracking and self-help tips being the most sought-after feature (Martin-Key et al., 2024).

Thus, based on these findings, we can conclude that there is a high prevalence of South African women with hormonal disorders. Although 45% were apprehensive to using health monitoring applications, it is likely their reservations stem from their previous experience with high subscription cost, difficult user interface, lack of information pertaining the use of smartphone apps to monitor and test for hormonal health, and lack of personalized care.

Recommendations when developing FEMMflo

When developing the Femmflo artificial intelligence interface, it should be easy to use and provide comprehensive resources that explains their disorders and symptoms in a simple language. This can be done by including infographics, pictures, texts, and external articles. Based on their different differentials and bloodwork data, we can curate personalized recommendations. The AI will need to be trained in pattern recognition of when users are “diagnosed” and provide custom recommendations, update the user of potential flares in symptoms, mood shifts and how they may try and counter this. Furthermore, adding regular symptoms check-in, pain severity and logging in their mood which provides user with a graph to monitor changes would be helpful. Based on their mood and scores, the app can then recommend them a “toolkit” which has crisis numbers, breathing techniques, journalling option etc.

We must ensure that the monthly subscription fee is affordable (R 500 or less) – please note that in Martin-Key et al.’s (2024) study, most women preferred that the mental health app does not have a subscription fee (63.8%), or they preferred to pay a once-off service fee (21.19%) or a subscription-based fee (15.01%).

Finally, a strong marketing campaign that explains the use of AI-driven for differentials which can aid in diagnosis, hormonal testing and features that are offered must be implemented.