Predicting date of birth and examining the best time to date a pregnancy

Deep Learning Model Using Continuous Skin Temperature Data Predicts Labor Onset

Background

Changes in body temperature anticipate labor onset in numerous mammals, yet this concept has not been explored in humans.

Methods

We evaluated patterns in continuous skin temperature data in 91 pregnant women using a wearable smart ring. Additionally, we collected daily steroid hormone samples leading up to labor in a subset of 28 pregnancies and analyzed relationships among hormones and body temperature trajectory. Finally, we developed a novel autoencoder long-short-term-memory (AE-LSTM) deep learning model to provide a daily estimation of days until labor onset.

Results

Features of temperature change leading up to labor were associated with urinary hormones and labor type. Spontaneous labors exhibited greater estriol to α-pregnanediol ratio, as well as lower body temperature and more stable circadian rhythms compared to pregnancies that did not undergo spontaneous labor. Skin temperature data from 54 pregnancies that underwent spontaneous labor between 34 and 42 weeks of gestation were included in training the AE-LSTM model, and an additional 40 pregnancies that underwent artificial induction of labor or Cesarean without labor were used for further testing. The model was trained only on aggregate 5-minute skin temperature data starting at a gestational age of 240 until labor onset. During cross-validation AE-LSTM average error (true - predicted) dropped below 2 days at 8 days before labor, independent of gestational age. Labor onset windows were calculated from the AE-LSTM output using a probabilistic distribution of model error. For these windows AE-LSTM correctly predicted labor start for 79% of the spontaneous labors within a 4.6-day window at 7 days before true labor, and 7.4-day window at 10 days before true labor.

Conclusion

Continuous skin temperature reflects progression toward labor and hormonal status during pregnancy. Deep learning using continuous temperature may provide clinically valuable tools for pregnancy care.

Estimated date of delivery with electronic medical records by a hybrid GBDT-GRU model

An accurate estimated date of delivery (EDD) helps pregnant women make adequate preparations before delivery and avoid the panic of parturition. EDD is normally derived from some formulates or estimated by doctors based on last menstruation period and ultrasound examinations. This study attempted to combine antenatal examinations and electronic medical records to develop a hybrid model based on Gradient Boosting Decision Tree and Gated Recurrent Unit (GBDT-GRU). Besides exploring the features that affect the EDD, GBDT-GRU model obtained the results by dynamic prediction of different stages. The mean square error (MSE) and coefficient of determination (R²) were used to compare the performance among the different prediction methods. In addition, we evaluated predictive performances of different prediction models by comparing the proportion of pregnant women under the error of different days. Experimental results showed that the performance indexes of hybrid GBDT-GRU model outperformed other prediction methods because it focuses on analyzing the time-series predictors of pregnancy. The results of this study are helpful for the development of guidelines for clinical delivery treatments, as it can assist clinicians in making correct decisions during obstetric examinations.

AI aided analysis on saliva crystallization of pregnant women for accurate estimation of delivery date and fetal status

Saliva contains similar molecular components to serum. Analysis of saliva can provide important diagnostic information about the body. Here we report an artificial intelligence (AI) aided home-based method that can let pregnant women perform daily monitoring on their pregnant status and accurate prediction on their delivery date by the pattern analysis of their salivary crystals. The method was developed based on the information obtained from our investigation on the saliva samples of 170 pregnant women about the correlation of the salivary crystal pattern with pregnant age and fetal status. It demonstrated that the patterns of salivary crystallization could act as indicators of the pregnant age, fetal state, and some medical conditions of pregnant women. On this basis, with the aid of AI recognition and analysis of the fractal dimension and some characteristic crystals in the salivary crystallization, we performed estimation on the delivery date in both quantitative and qualitative manners. The accuracy of the prediction on 15 pregnant women was satisfactory: 100 % delivering in the predicted week, 93.3 % within the estimated three days, and 86.7 % on the day as the prediction. We also developed a simple smartphone-based AI-aided salivary crystal imaging and analysis device as an auxiliary means to let pregnant women monitor their fetal status daily at home and predict their delivery date with adequate accuracy.

Errors in estimated gestational ages reduce the likelihood of health facility deliveries: results from an observational cohort study in Zanzibar

Background

Most maternal health programs in low- and middle- income countries estimate gestational age to provide appropriate antenatal care at the correct times throughout the pregnancy. Although various gestational dating methods have been validated in research studies, the performance of these methods has not been evaluated on a larger scale, such as within health systems. The objective of this research was to investigate the magnitude and impact of errors in estimated delivery dates on health facility delivery among women enrolled in a maternal health program in Zanzibar.

Methods

This study included 4225 women who were enrolled in the Safer Deliveries program and delivered before May 31, 2017. The exposure of interest was error in estimated delivery date categorized as: severe overestimate, when estimated delivery date (EDD) was 36 days or more after the actual delivery date (ADD); moderate overestimate, when EDD was 15 to 35 days after ADD; accurate, when EDD was 6 days before to 14 days after ADD; and underestimate, when EDD was 7 days or more before ADD. We used Chi-squared tests to identify factors associated with errors in estimated delivery dates. We performed logistic regression to assess the impact of errors in estimated delivery dates on health facility delivery adjusting for age, district of residence, HIV status, and occurrence of past home delivery.

Results

In our data, 28% of the estimated delivery dates were a severe overestimate, 23% moderate overestimate, 41% accurate, and 8% underestimate. Compared to women with an accurate delivery date, women with a moderate or severe overestimate were significantly less likely to deliver in a health facility (OR = 0.71, 95% CI: [0.59, 0.86]; OR = 0.74, 95% CI: [0.61, 0.91]). When adjusting for multiple confounders, women with moderate overestimates were significantly less likely to deliver in a health facility (AOR = 0.76, 95% CI: [0.61, 0.93]); the result moved slightly towards null for women with severe overestimates (AOR = 0.84, 95% CI: [0.69, 1.03]).

Conclusions

The overestimation of women’s EDDs reduces the likelihood of health facility delivery. To address this, maternal health programs should improve estimation of EDD or attempt to curb the effect of these errors within their programs.

Childhood outcomes following preterm prelabor rupture of the membranes (PPROM): a population-based record linkage cohort study

OBJECTIVE

The objective of this study is to determine child health, development and educational outcomes for infants born following preterm prelabor rupture of the membrane (PPROM).

STUDY DESIGN

Population-based record linkage cohort study using data from NSW, Australia, 2001 to 2014.

RESULTS

Of 121 822 births at 20 to 37 weeks, 18 799 (15%) followed PPROM, 56 406 (46%) followed spontaneous labor and 46 617 (38%) were planned. Compared with infants of a similar gestational age born following spontaneous labor or planned delivery, exposure to PPROM did not increase the risk of childhood mortality, childhood hospitalization, developmentally vulnerable at school entry, low reading or numeracy scores. Median latency ranged from 12 days (interquartile range 3 to 37 days) at 25 weeks to 1 day (0 to 2 days) at 36 weeks. Longer latency and more advanced gestational age at birth were associated with better outcomes.

CONCLUSION

Infants born following PPROM are at no greater risk of adverse child health, development and education outcomes than those of similar gestational age born without PPROM.

Long-term outcomes for large for gestational age infants born at term

AIM

Large for gestational age (LGA) babies have increased risks for short-term outcomes such as shoulder dystocia, neonatal hypoglycaemia and longer hospital stay. Little is known of long-term health, development and educational outcomes of LGA babies. The aim of this study was to determine the long-term health, mortality, development and educational outcomes for infants born LGA at term.

METHODS

A population-based record linkage study of live singletons born at term (37-41 weeks of gestation) in New South Wales, Australia, from 2001 to 2006.

RESULTS

This study compared 49 439 LGA (>90th percentile for birthweight, gestational age and sex) and 400 418 appropriate size for gestational age (AGA; 10th-90th percentile) infants. LGA infants had increased risk of birth and neonatal outcomes and hospitalisations, for brachial plexus injury after the neonatal period, and for all causes from 1 to 5 years of age. There were no differences in mortality up to 5 years of age or hospitalisations for type 1 diabetes in childhood. LGA infants had lower rates of developmental vulnerability (in kindergarten) and showed a significant trend (χ² for trend <0.0001) to fewer low scores and more high scores in reading and numeracy (in Year 3) compared with AGA. After adjusting for potential confounders, only the relative risk for higher reading scores was statistically significant.

CONCLUSIONS

LGA infants show positive long-term health, development and educational outcomes. Concerns for LGA infants still remain in the perinatal period as a result of birth trauma; however, these complications usually do not persist in postnatal and early childhood.

Association between interpregnancy interval and the risk of recurrent loss after a midtrimester loss

STUDY QUESTION

After an initial midtrimester loss, is the interval to the next conception associated with the risk of a recurrent loss?

SUMMARY ANSWER

Among women who had a pregnancy loss at 14-19 weeks gestation, conception at least 3 months after this initial loss was associated with a reduced risk of a recurrent loss.

WHAT IS KNOWN ALREADY

A short interpregnancy interval (IPI) has been thought to increase risk but recent studies of pregnancy after a loss have found no effect; however, these studies have been based almost entirely on an initial first trimester (<14 weeks) loss.

STUDY DESIGN, SIZE, DURATION

A retrospective cohort study drawing on over 997 000 linked birth and hospital records from New South Wales, Australia for 2003-2011. Index pregnancies were those of women who had a first recorded pregnancy loss of 14-23 weeks gestation (miscarriage, termination and perinatal death). The study population was 4290 women who conceived again within 2 years.

PARTICIPANTS/MATERIALS, SETTING, METHODS

The index loss was categorized by subgroups: 14-19 weeks gestation versus 20-23 weeks, and by whether spontaneous or a termination. The primary outcome was any loss or perinatal death before 24 weeks in the subsequent pregnancy.

MAIN RESULTS AND THE ROLE OF CHANCE

After a 14-19 weeks index loss, an IPI of ≤3 months had an increased rate of recurrent loss compared with an IPI of >9-12 months: 21.9% versus 11.3% (adjusted relative risk (aRR) = 2.02, 95% CI 1.44-2.83). For women who had a spontaneous index loss of 20-23 weeks, there was no evidence that a short IPI increased or decreased the risk of recurrent loss. For any gestational age group of index losses, an IPI of >18-24 months increased the risk of a recurrent loss; the risk was highest after a 20-23 weeks index loss (aRR = 2.15, 95% CI 1.18-3.91).

LIMITATIONS, REASONS FOR CAUTION

We do not know how many cycles were required to achieve conception. Pregnancies resulting in early first trimester losses are unlikely to have resulted in hospitalization so would not have been identified.

WIDER IMPLICATIONS OF THE FINDINGS

The risk of recurrent loss after an initial midtrimester loss may differ from the risk after an initial first trimester loss.

STUDY FUNDING/COMPETING INTERESTS

This work was supported by an Australian National Health and Medical Research Council (NHMRC) Centre for Research Excellence Grant (1001066). C.L.R. is supported by an NHMRC Senior Research Fellowship (#APP1021025). J.B.F. is supported by an ARC Future Fellowship (#120100069). The authors have no conflicts of interest to declare.

TRIAL REGISTRATION NUMBER

N/A.

Women's views about the timing of birth

BACKGROUND

Estimated date of birth (EDB) is used to guide the care provided to women during pregnancy and birth, although its imprecision is recognised. Alternatives to the EDB have been suggested for use with women however their attitudes to timing of birth information have not been examined.

AIMS

To explore women's expectations of giving birth on or near their EDB, and their attitudes to alternative estimates for timing of birth.

METHODS

A survey of pregnant women attending four public hospitals in Sydney, Australia, between July and December 2012.

RESULTS

Among 769 surveyed women, 42% expected to birth before their due date, 16% after the due date, 15% within a day or so of the due date, and 27% had no expectations. Nulliparous women were more likely to expect to give birth before their due date. Women in the earlier stages of pregnancy were more likely to have no expectations or to expect to birth before the EDB while women in later pregnancy were more likely to expect birth after their due date. For timing of birth information, only 30% of women preferred an EDB; the remainder favoured other options.

CONCLUSIONS

Most women understood the EDB is imprecise. The majority of women expressed a preference for timing of birth information in a format other than an EDB. In support of woman-centred care, clinicians should consider discussing other options for estimated timing of birth information with the women in their care.

Preterm birth and future risk of maternal cardiovascular disease - is the association independent of smoking during pregnancy?

BACKGROUND

While the association of preterm birth and the risk of maternal cardiovascular disease (CVD) has been well-documented, most studies were limited by the inability to account for smoking during pregnancy - an important risk factor for both preterm birth and CVD. This study aimed to determine whether the increased future risk of maternal cardiovascular disease (CVD) associated with preterm birth is independent of maternal smoking during pregnancy.

METHODS

A population-based record linkage study of 797,056 women who delivered a singleton infant between 1994 and 2011 in New South Wales, Australia was conducted. Birth records were linked to the mothers' subsequent hospitaliation or death from CVD. Preterm births were categorised as late (35-36 weeks), moderate (33-34 weeks), or extreme (≤32 weeks); and as spontaneous or indicated. Cox proportional hazard regression was used to estimate adjusted hazard ratios (AHR) [95 % CI].

RESULTS

During the study period, 59,563 women (7.5 %) had at least one preterm birth. After adjustment for CVD risk factors other than smoking, AHR [95 % CI] of CVD among women who ever had a preterm birth was 1.78 [1.61-1.96]. Associations were greater for extreme (AHR = 1.98 [1.63-2.42]) and moderate (AHR = 2.06 [1.69-2.51]) than late preterm birth (AHR = 1.63 [1.44-1.85]), for indicated (AHR = 2.04 [1.75-2.38]) than spontaneous preterm birth (AHR = 1.65 [1.47-1.86]), and for having ≥ two (AHR = 2.29[1.75-2.99]) than having one preterm birth (AHR = 1.73[1.57-1.92]). A further adjustment for maternal smoking attenuated, but did not eliminate, the associations. Smoking during pregnancy was also independently associated with maternal CVD risks, with associations being stronger for mothers who smoked during last pregnancy (AHR = 2.07 [1.93-2.23]) than mothers who smoked during a prior pregnancy (AHR = 1.64 [1.41-1.90]).

CONCLUSIONS

Associations of preterm birth and maternal CVD risk are independent of maternal smoking during pregnancy. This underscores the importance of smoking cessation in reducing CVD and suggests that a history of preterm delivery (especially if severe, indicated or recurrent) identifies women who could be targeted for CVD screening and preventative therapies.