Explainer | Menstrual Apps

Explainer: What are menstrual apps and how do they work?

What are menstrual apps?

People have been tracking menstrual cycles for millennia. Some prefer not to track, others use a paper calendar, and most recently menstruators have been able to use apps. Hundreds of different apps for tracking menstrual cycles are available for download or embedded in devices such as fitness trackers and mobile phones. Most apps are free or have free versions, although some have a cost for download and others have subscription fees. The business model of many, if not most, of these apps is to monetise user data, leaving users vulnerable to unwanted sharing of their personal data and information (See this report by Privacy International). Menstrual apps do not have transparent algorithms or code (with the exception of some FABM apps, noted below).

How do menstrual apps work?

The main purpose of menstrual apps is to track the menstrual cycle, and most apps also predict the date of the next menstruation. Some apps also track other symptoms and make predictions about other cycle elements (e.g. ovulation). For these functions, menstrual apps have three main parts: Data input by user, the app algorithm, and app output communicated to the user.

Data input by user: Users are required to input information at signup (e.g. typical cycle length), and are requested to input ongoing information about date of menstruation. Most apps also allow or prompt users to input additional information. Some of this additional information includes measurements that have been established as indicative of menstrual cycle phases (e.g. temperature, cervical fluid, hormone levels). Various other symptoms and activities (e.g. pain, mood, sexual activity, sleep, nutrition, exercise) can also be logged.
App algorithm: Algorithms in most apps calculate their predictions based on previous cycles. These general-purpose menstrual tracking apps appear to employ some user-input data, combined with population-level estimates, to create this output. Specific functions are not transparent, as algorithms are treated as commercially sensitive and not revealed.
App output: Output communicated to the user by nearly all general-purpose menstrual apps includes a predicted date of the next menstruation. Most of these apps also show users a predicted day of ovulation and a range of dates around ovulation often labelled ‘fertile window’. Apps will typically allow users to view past data they have logged, arranged by calendar and/or cycle, as well as predictions of future cycles. Many apps also show users other symptoms they have logged, and some offer users specific reports about their symptoms and association with their cycles.

A very small number of apps are specifically designed for use in preventing pregnancy. Some of these apps are based on established Fertility-Awareness Based Methods of contraception (See this interactive graphic of FABM at the British Medical Journal). Many apps explicitly using an established Fertility Awareness-Based Method have transparent algorithms. Several apps advertise that they have been developed as a contraceptive method; some have published effectiveness and two have received FDA clearance for use as medical devices in the United States (Scherwitzl et al 2017; Jennings et al 2019) but others have not validated their method’s effectiveness (Polis 2018).

In contrast to these apps specifically designed as contraception, other apps are not designed for use in preventing pregnancy and often have a disclaimer to this effect in their terms of service.

How accurate are menstrual apps?

There have been very few independent multi-app studies examining predictions of ovulation and fertile window (Earle et al 2021). Most of these studies were concerned with accuracy for conception (Freis et al 2018, Johnson et al 2018, Setton et al 2016), and one study was primarily concerned with contraception (Duane et al 2016). These studies evaluate accuracy by comparing app predictions of ovulation and fertile window with standardised cycles having an expected ovulation date and fertile window. Two studies compared app predictions of fertile window with the fertile window as calculated by established Fertility-Awareness Based Methods (Duane et al 2016, Freis et al 2018). Standardised cycles may account for some variability in cycle length but do not explicitly or comprehensively test for apps’ ability to accommodate the cycle variability of individual users. There has been to date no multi-app or independent study validating app predictions of ovulation directly with app users’ measurements of ovulation. The existing studies find a wide range of app accuracy and functionality, with the large majority of apps performing inadequately (Duane et al 2016, Freis et al 2018, Johnson et al 2018, Setton et al 2016). These results suggest taking a cautious approach to app predictions of ovulation and fertile window.