FemTech Series
Mobile Apps for Fertility Awareness: A Review of Research
By: Faith Slinger, DO
Director’s Note: The science supporting fertility awareness-based methods (FABMs) and restorative reproductive medicine is continually evolving. At FACTS, we share the best evidence to equip our colleagues to help patients achieve their family planning goals. As the use of fertility-tracking apps and monitoring devices keeps increasing, this month’s FemTech Series will feature some of them. We begin with a summary of research by Symul et al [4] titled, “Assessment of Menstrual Health Status and Evolution through Mobile Apps for Fertility Awareness.” The article was published in Digital Medicine and summarized by FACTS elective participant, Dr. Faith Slinger. Learn more about this topic in Part F of our online CME Course, Fertility Awareness and FemTech. And stay tuned for details about virtual conference on October 18th and 19th where Dr. Thomas Bouchard will address this topic in more depth.
Introduction
Advancements in technology have expanded the options available to detect and track biomarkers with fertility awareness-based methods (FABMs) for family planning. Most evidence-based FABMs have their own applications (apps) that enable users to track cervical mucus, basal body temperature (BBT), cervical changes, and hormones with varying accuracy. The data suggests that by 2016, over 200 million women had started using menstrual tracking applications. [3] With so many apps available and more women taking advantage of the technology, it is important to ensure the application predictions around the fertile window are correct. Two 2016 papers evaluated different digital applications; published by Moglia et al [1] and Duane et al [2] , each study showed that most apps were inaccurate at predicting the fertile window and cycle length.
“With so many apps available and more women taking advantage of the technology, it is important to ensure the application predictions around the fertile window are correct.”
The 2019 study summarized below aimed to understand the typical app user of the sympto-thermal fertility method, which combines BBT and cervical mucus observations, and to compare their tracking and observations with prior data from small studies. A secondary objective was to suggest a statistical framework to estimate ovulation from the self-reported data so the average cycle length and time to ovulation could be compared. The two applications studied were Sympto and Kindara.
Methodology
Sympto and Kindara were selected because they ranked high in a study that compared apps for users trying to avoid pregnancy using FABMs. [2] All data was taken directly from the apps from users who were tracking for at least four cycles. The data was then organized to observe patterns of the menstrual cycle. A Hidden Markov Model (HMM) was used to predict ovulation more accurately and determine the apps’ reliability to predict the fertile window.
Inclusion criteria consisted of those using fertility awareness body signs (BBT, cervical mucus changes, and cervical changes), completed cycles, and standard days of data tracking (8-12 days of body signs documented). Exclusion criteria were first-time cycle tracking or ongoing cycles, incomplete cycles, long tracking gaps, mid-cycle medium/heavy bleeding, and cycles in which only the menstrual period was documented.
Before inclusion/exclusion criteria, there were 39,896 total cycles from Sympto and 719,182 cycles from Kindara. After inclusion/exclusion criteria, 28,453 cycles from Sympto and 80,708 cycles from Kindara had reliable data estimating ovulation.
Results
Users of the Sympto and Kindara apps spanned 150 countries and 5 continents. Users of the Sympto app were European whereas Kindara users were primarily from the US. Users of both applications had an average age of 30 6 and a body mass index (BMI) of 23 5. Most users of the Sympto app were trying to avoid pregnancy; most Kindara app users were trying to achieve pregnancy. Information that could not be assessed included level of education, health conditions, and social or marital status of the population studied.
Tracking behaviors were more consistent in users who were trying to achieve pregnancy. Most users of FABMs tracked an average of 16 days per cycle. When applying the HMM mathematical framework to estimate ovulation, missing temperature recordings altered the estimated day of ovulation more than missing cervical mucus recordings.
Cycle lengths varied, with a higher prevalence of cycles lasting longer than 28 days. Although the median duration of the follicular phase was 16 days, the luteal phase was shorter than normal (less than 11 days) in approximately 20% of cycles. When comparing cycle length to previous studies, the follicular phase seemed to have greater variability than previously stated, while luteal phase length was similar.
The menstrual cycles analyzed showed significant variability in the estimated day of ovulation, with 90% of ovulations occurring between days 10 and 24. Ovulation was estimated to occur on day 14 or 15 in only about 24% of cycles.
“The menstrual cycles analyzed showed significant variability in the estimated day of ovulation… Ovulation was estimated to occur on day 14 or 15 in only about 24% of cycles.”
Discussion
The primary goal of this study was to identify the typical users of menstrual tracking apps and to inform health care practitioners of the diverse tracking capabilities. The typical users of the Sympto and Kindara applications are from a western country (Europe or the US) with an average age of 30 and a healthy BMI. The average tracking frequency was higher in users who were trying to achieve pregnancy, with 40% of recorded cycles tracking fertility signs every single day.
“The average tracking frequency was higher in users who were trying to achieve pregnancy, with 40% of recorded cycles tracking fertility signs every single day.”
Given the study population, it is important to consider why there were significantly more cycles lasting longer than 28 days. Information that could not be ascertained included time trying to conceive and the presence of conditions affecting the menstrual cycle, such as polycystic ovary syndrome (PCOS). Having a large population attempting to conceive could have skewed the data toward longer menstrual cycles than previously studied in women who were not trying to conceive, suggesting this population may have subfertility. This is further suggested by the HMM framework showing more variability in the follicular phase compared to previous research. An even split of populations trying to conceive versus avoiding conception would help determine the accuracy of this analysis.
The secondary aim of the study was to propose a mathematical framework that could analyze the menstrual cycle and estimate users’ hormonal status as well as the day of ovulation. This allowed comparison to other studies that analyzed the average menstrual cycle lengths and time to ovulation. With a large number of cycles and 40% tracking daily data, a precise statistical analysis was possible. During analysis, missing BBT data influenced the date of ovulation more than missing cervical mucus data. This suggests the mathematical framework relied on the temperature shift caused by progesterone to estimate ovulation retrospectively.
With technology advancing and FABMs growing in popularity, it is vital to have reliable apps, consistent tracking, and accurate data analysis to interpret the information, particularly for new FABM users and medical professionals without previous FABM training. The HMM framework was able to predict ovulation time more accurately with quantitative data (BBT) than with cervical mucus observations. This suggests the need to develop a mathematical modeling framework that can better interpret cervical mucus changes. Such a model would benefit women who prefer methods that rely on cervical mucus tracking.
“With technology advancing and FABMs growing in popularity, it is vital to have reliable apps, consistent tracking, and accurate data analysis to interpret the information, particularly for new FABM users and medical professionals without previous FABM training.”
Although fertility applications are accessible and user friendly, interpretation by the apps still needs to improve. Having accurate analysis at their fingertips can help women make informed health care decisions with guidance from medical professionals throughout their fertile life.
References
[1] Moglia, M.L., Nguyen, H.V., Chyjek, K., Chen, K.T., and Castano, P.M. Evaluation of smartphone menstrual cycle tracking applications using an adapted applications scoring system. Obstet. Gynecol. 127, 1153-1160 (2016).
[2] Duane, M., Contreras, A., Jensen, E. T. and White, A. The performance of fertility awareness-based method apps marketed to avoid pregnancy. J. Am. Board Fam. Med. 29, 508-511 (2016).
[3] Dreaper, J. Women warned about booming market in period tracker apps – BBC News. BBC (2016). https://www.bbc.com/news/health-37013217
[4] Symul, L., Wac, K., Hillard, P., Salathe, M. Assessment of menstrual health status and evolution through mobile apps for fertility awareness. NPJ. Digital Medicine. 2:64 (2019).
ABOUT THE AUTHOR
Faith Slinger, DO
Dr. Faith Slinger is a recent graduate from Burrell College of Osteopathic Medicine in Las Cruces, NM. She completed her undergraduate education at Juniata College in Huntingdon, PA. Dr. Slinger is a resident in family medicine at UPMC-Altoona PA and is interested in women’s health and osteopathic manipulative treatment. She enrolled in the FACTS elective to learn about natural family planning methods and how to empower her future patients to take charge of their health and reproductive decisions. She hopes to utilize cycle chart reading as a diagnostic tool in her future endeavors.
