skip to main content

Title: Informative predictors of pregnancy after first IVF cycle using eIVF practice highway electronic health records

The aim of this study is to determine the most informative pre- and in-cycle variables for predicting success for a first autologous oocyte in-vitro fertilization (IVF) cycle. This is a retrospective study using 22,413 first autologous oocyte IVF cycles from 2001 to 2018. Models were developed to predict pregnancy following an IVF cycle with a fresh embryo transfer. The importance of each variable was determined by its coefficient in a logistic regression model and the prediction accuracy based on different variable sets was reported. The area under the receiver operating characteristic curve (AUC) on a validation patient cohort was the metric for prediction accuracy. Three factors were found to be of importance when predicting IVF success: age in three groups (38–40, 41–42, and above 42 years old), number of transferred embryos, and number of cryopreserved embryos. For predicting first-cycle IVF pregnancy using all available variables, the predictive model achieved an AUC of 68% + /− 0.01%. A parsimonious predictive model utilizing age (38–40, 41–42, and above 42 years old), number of transferred embryos, and number of cryopreserved embryos achieved an AUC of 65% + /− 0.01%. The proposed models accurately predict a single IVF cycle pregnancy outcome and identify important predictive variables associated more » with the outcome. These models are limited to predicting pregnancy immediately after the IVF cycle and not live birth. These models do not include indicators of multiple gestation and are not intended for clinical application.

« less
; ; ; ;
Award ID(s):
1914792 1664644 1645681
Publication Date:
Journal Name:
Scientific Reports
Nature Publishing Group
Sponsoring Org:
National Science Foundation
More Like this
  1. Abstract STUDY QUESTION

    Can we derive adequate models to predict the probability of conception among couples actively trying to conceive?


    Leveraging data collected from female participants in a North American preconception cohort study, we developed models to predict pregnancy with performance of ∼70% in the area under the receiver operating characteristic curve (AUC).


    Earlier work has focused primarily on identifying individual risk factors for infertility. Several predictive models have been developed in subfertile populations, with relatively low discrimination (AUC: 59–64%).


    Study participants were female, aged 21–45 years, residents of the USA or Canada, not using fertility treatment, and actively trying to conceive at enrollment (2013–2019). Participants completed a baseline questionnaire at enrollment and follow-up questionnaires every 2 months for up to 12 months or until conception. We used data from 4133 participants with no more than one menstrual cycle of pregnancy attempt at study entry.


    On the baseline questionnaire, participants reported data on sociodemographic factors, lifestyle and behavioral factors, diet quality, medical history and selected male partner characteristics. A total of 163 predictors were considered in this study. We implemented regularized logistic regression, support vector machines, neural networks and gradient boosted decisionmore »trees to derive models predicting the probability of pregnancy: (i) within fewer than 12 menstrual cycles of pregnancy attempt time (Model I), and (ii) within 6 menstrual cycles of pregnancy attempt time (Model II). Cox models were used to predict the probability of pregnancy within each menstrual cycle for up to 12 cycles of follow-up (Model III). We assessed model performance using the AUC and the weighted-F1 score for Models I and II, and the concordance index for Model III.


    Model I and II AUCs were 70% and 66%, respectively, in parsimonious models, and the concordance index for Model III was 63%. The predictors that were positively associated with pregnancy in all models were: having previously breastfed an infant and using multivitamins or folic acid supplements. The predictors that were inversely associated with pregnancy in all models were: female age, female BMI and history of infertility. Among nulligravid women with no history of infertility, the most important predictors were: female age, female BMI, male BMI, use of a fertility app, attempt time at study entry and perceived stress.


    Reliance on self-reported predictor data could have introduced misclassification, which would likely be non-differential with respect to the pregnancy outcome given the prospective design. In addition, we cannot be certain that all relevant predictor variables were considered. Finally, though we validated the models using split-sample replication techniques, we did not conduct an external validation study.


    Given a wide range of predictor data, machine learning algorithms can be leveraged to analyze epidemiologic data and predict the probability of conception with discrimination that exceeds earlier work.


    The research was partially supported by the U.S. National Science Foundation (under grants DMS-1664644, CNS-1645681 and IIS-1914792) and the National Institutes for Health (under grants R01 GM135930 and UL54 TR004130). In the last 3 years, L.A.W. has received in-kind donations for primary data collection in PRESTO from,, Sandstone Diagnostics and Swiss Precision Diagnostics. L.A.W. also serves as a fibroid consultant to AbbVie, Inc. The other authors declare no competing interests.



    « less
  2. Ryckman, Kelli K (Ed.)
    Background Technology enables the continuous monitoring of personal health parameter data during pregnancy regardless of the disruption of normal daily life patterns. Our research group has established a project investigating the usefulness of an Internet of Things–based system and smartwatch technology for monitoring women during pregnancy to explore variations in stress, physical activity and sleep. The aim of this study was to examine daily patterns of well-being in pregnant women before and during the national stay-at-home restrictions related to the COVID-19 pandemic in Finland. Methods A longitudinal cohort study design was used to monitor pregnant women in their everyday settings. Two cohorts of pregnant women were recruited. In the first wave in January-December 2019, pregnant women with histories of preterm births (gestational weeks 22–36) or late miscarriages (gestational weeks 12–21); and in the second wave between October 2019 and March 2020, pregnant women with histories of full-term births (gestational weeks 37–42) and no pregnancy losses were recruited. The final sample size for this study was 38 pregnant women. The participants continuously used the Samsung Gear Sport smartwatch and their heart rate variability, and physical activity and sleep data were collected. Subjective stress, activity and sleep reports were collected using amore »smartphone application developed for this study. Data between February 12 to April 8, 2020 were included to cover four-week periods before and during the national stay-at-home restrictions. Hierarchical linear mixed models were exploited to analyze the trends in the outcome variables. Results The pandemic-related restrictions were associated with changes in heart rate variability: the standard deviation of all normal inter-beat intervals (p = 0.034), low-frequency power (p = 0.040) and the low-frequency/high-frequency ratio (p = 0.013) increased compared with the weeks before the restrictions. Women’s subjectively evaluated stress levels also increased significantly. Physical activity decreased when the restrictions were set and as pregnancy proceeded. The total sleep time also decreased as pregnancy proceeded, but pandemic-related restrictions were not associated with sleep. Daily rhythms changed in that the participants overall started to sleep later and woke up later. Conclusions The findings showed that Finnish pregnant women coped well with the pandemic-related restrictions and lockdown environment in terms of stress, physical activity and sleep.« less
  3. ABSTRACT Background

    Inflammation during pregnancy may aggravate iron deficiency (ID) by increasing serum hepcidin and reducing iron absorption. This could restrict iron transfer to the fetus, increasing risk of infant ID and its adverse effects.


    We aimed to assess whether iron bioavailability and/or iron transfer to the fetus is impaired in overweight/obese (OW) pregnant women with adiposity-related inflammation, compared with normal-weight (NW) pregnant women.


    In this prospective study, we followed NW (n = 43) and OW (n = 40) pregnant women who were receiving iron supplements from the 14th week of gestation to term and followed their infants to age 6 mo. We administered 57Fe and 58Fe in test meals mid-second and mid-third trimester, and measured tracer kinetics throughout pregnancy and infancy.


    In total, 38 NW and 36 OW women completed the study to pregnancy week 36, whereas 30 NW and 27 OW mother–infant pairs completed the study to 6 mo postpartum. Both groups had comparable iron status, hemoglobin, and serum hepcidin throughout pregnancy. Compared with the NW, the OW pregnant women had 1) 43% lower fractional iron absorption (FIA) in the third trimester (P = 0.033) with median [IQR] FIA of 23.9% [11.4%–35.7%] and 13.5% [10.8%–19.5%], respectively; and 2) 17% lower maternal–fetal iron transfermore »from the first tracer (P = 0.051) with median [IQR] maternal–fetal iron transfer of 4.8% [4.2%–5.4%] and 4.0% [3.6%–4.6%], respectively. Compared with the infants born to NW women, infants born to OW women had lower body iron stores (BIS) with median [IQR] 7.7 [6.3–8.8] and 6.6 [4.6–9.2] mg/kg body weight at age 6 mo, respectively (P = 0.024). Prepregnancy BMI was a negative predictor of maternal–fetal iron transfer (β = −0.339, SE = 0.144, P = 0.025) and infant BIS (β = −0.237, SE = 0.026, P = 0.001).


    Compared with NW, OW pregnant women failed to upregulate iron absorption in late pregnancy, transferred less iron to their fetus, and their infants had lower BIS. These impairments were associated with inflammation independently of serum hepcidin.

    This trial was registered at as NCT02747316.

    « less
  4. Abstract Background: The U.N. health and well-being goals for 2030 focus on maternal and child health outcomes, among others. Challenges to meeting those goals vary widely throughout Nepal owing to the range of sociocultural factors, infrastructural limitations, physical geography and altitudes. This article explores sociocultural and biological influences on fertility and child survival among ethnically Tibetan women in Nepal. Methods: This is a cross sectional study of 430 women, age 46-86 years old, citizens of Nepal and native residents above 3500m in Mustang District, who provided interview and physiological data. Univariate Poisson regression analyses selected significant variables to include in multivariate Poisson regressions investigating the number of pregnancies, livebirths, child survival and death outcomes. Results: Earlier age at first pregnancy, later age at last pregnancy, and miscarriages associated with more pregnancies. Miscarriages and stillbirths associated with fewer livebirths. Higher maternal BMI and FEV6 associated with fewer children dying before age 15. Marital characteristics (status, type, continuity), contraceptive use, relative wealth, and education influenced these covariates. Conclusions: Low maternal pulmonary function and nutritional status predict poorer child survival in Upper Mustang. Addressing poor lung function and nutrition may improve reproductive outcomes among ethnically Tibetan women living at high altitude. Keywords: Childmore »survival; fertility; high-altitude; Nepal; women..« less
  5. Consistent with national trends, only about ½ of students who intend to major in STEM disciplines at Maryville College (MC) complete bachelor’s degrees in these fields. The Scots Science Scholars (S3) program was funded through the National Science Foundation’s STEM Talent Extension Program to increase the number of students graduating with STEM degrees from MC. The S3 program enrolls college freshmen who have an interest in STEM majors and math ACT scores between 21 and 27, with emphasis on students from groups underrepresented in STEM and first-generation college students. The program consists of a summer bridge, a living-learning community, early engagement in STEM research, a seminar series that exposes students to STEM careers and research fields, academic support through a first-year seminar class, peer tutoring, and time-management counseling. The program has enrolled 6 cohorts of students (n = 97) since 2013, (54% female, 22% underrepresented minorities and 35% first-generation college students). From 2013-2017, S3 compared favorably to the general college population: 96% of all S3 completed the first year of college, 69% declared STEM majors, and 85% returned to the college for a second year (compared to 71%, p < 0.001). Overall, S 3 students persist at the college longermore »than non-S3 students (P<0.01). Compared to a matched control group, S 3 had significantly higher STEM major declaration rates (68% vs. 38%), higher rates of STEM retention through the junior year (41% vs. 20%), and improved overall college persistence (P< 0.01). Students report high levels of satisfaction with the summer program. At the end of the summer program, students report gains in skills and attitudes that are important for success in STEM. They also perform significantly better on math and chemistry assessments after completing the program. College-wide, the number of students enrolled in STEM majors at Maryville has increased by 52% since the inception of S3 , and STEM undergraduate research productivity has increased markedly. Our data suggest the S3 program is an important component of institutional changes that are increasing the STEM population and building a robust and productive STEM culture at a liberal arts college.« less