Prediction of Fertile Window Using a Contact-Free Under-the-Mattress Sensor

The purpose of the presented study was to evaluate the performance of a novel contact-free under-the-mattress sensor in prediction of women's fertile window (FW), at home environment. 30 normal-ovulatory women at the age of 22-40 years took part in the study. Heart rate, heart rate variability, and respiration rate values were measured during a period of 3-6 months. These measurements were recorded at the home environment, during night time, using a contact free sensor positioned under the participant's mattress. Predictions of the FW were obtained using an algorithm based on the sensor measurements and menses dates, as reported by the participants. For each participant, and each cycle, the ground-truth ovulation day was established by daily blood tests to measure hormone levels (LH, estrogen, and progesterone), and transvaginal ultrasound scans to determine follicle size. The accuracy of the sensor FW prediction was calculated as the overlap between the ground-truth and predicted FW (based on sensor's measurements). Comparison were also made to standard methods of FW prediction, which relies on the duration of a woman's previous menstrual cycles. Overall, a total of 216 menstrual cycles were recorded. Predicted FW overlapped 61.3 % ± 4.9 % of ground truth after the first menstrual cycle and 72.2 % ± 5.1 % after three menstrual cycles. Prediction overlap as obtained by the standard method was 53.0 % ± 6.9 % after the first menstrual and reached 63.3 % ± 6.3 % after three menstrual cycles. The results suggest that the use of a contact free sensor enhances the ability to accurately predict the FW as defined by the expert reference. Alternative available methods for fertility prediction, require the user to follow a strict protocol, such as measuring basal body temperature or urine LH levels, on a daily basis at morning time. The new contact free sensor requires minimal compliance from the user and increases the chances to conceive, compared to the standard prediction method based only on the duration of menstrual cycles.