Diagnostic Accuracy of an Integrated AI Tool to Estimate Gestational Age From Blind Ultrasound Sweeps

Importance

Accurate assessment of gestational age (GA) is essential to good pregnancy care but often requires ultrasonography, which may not be available in low-resource settings. This study developed a deep learning artificial intelligence (AI) model to estimate GA from blind ultrasonography sweeps and incorporated it into the software of a low-cost, battery-powered device.

Objective

To evaluate GA estimation accuracy of an AI-enabled ultrasonography tool when used by novice users with no prior training in sonography.

Design, Setting, and Participants

This prospective diagnostic accuracy study enrolled 400 individuals with viable, single, nonanomalous, first-trimester pregnancies in Lusaka, Zambia, and Chapel Hill, North Carolina. Credentialed sonographers established the "ground truth" GA via transvaginal crown-rump length measurement. At random follow-up visits throughout gestation, including a primary evaluation window from 14 0/7 weeks' to 27 6/7 weeks' gestation, novice users obtained blind sweeps of the maternal abdomen using the AI-enabled device (index test) and credentialed sonographers performed fetal biometry with a high-specification machine (study standard).

Main Outcomes and Measures

The primary outcome was the mean absolute error (MAE) of the index test and study standard, which was calculated by comparing each method's estimate to the previously established GA and considered equivalent if the difference fell within a prespecified margin of ±2 days.

Results

In the primary evaluation window, the AI-enabled device met criteria for equivalence to the study standard, with an MAE (SE) of 3.2 (0.1) days vs 3.0 (0.1) days (difference, 0.2 days [95% CI, -0.1 to 0.5]). Additionally, the percentage of assessments within 7 days of the ground truth GA was comparable (90.7% for the index test vs 92.5% for the study standard). Performance was consistent in prespecified subgroups, including the Zambia and North Carolina cohorts and those with high body mass index.

Conclusions and Relevance

Between 14 and 27 weeks' gestation, novice users with no prior training in ultrasonography estimated GA as accurately with the low-cost, point-of-care AI tool as credentialed sonographers performing standard biometry on high-specification machines. These findings have immediate implications for obstetrical care in low-resource settings, advancing the World Health Organization goal of ultrasonography estimation of GA for all pregnant people.

Trial Registration

ClinicalTrials.gov Identifier: NCT05433519.

True and false positive HIV point of care test results in a prospective multinational study of at-risk African women: implications for large-scale repeat HIV testing in HIV prevention programs

BACKGROUND

Accurate HIV point of care testing is the cornerstone of prevention and treatment efforts globally, though false (both negative and positive) results are expected to occur.

SETTING

We assessed the spectrum of true and false positive HIV results in a large prospective study of HIV incidence in African women using three contraceptive methods tested longitudinally in Eswatini, Kenya, South Africa, and Zambia.

METHODS

HIV serologic testing was conducted quarterly using two parallel rapid HIV tests. When one or both tests were positive, additional confirmatory testing was conducted, including HIV enzyme immunoassay (EIA) and ribonucleic acid (RNA).

RESULTS

7730 women contributed 48,234 visits: true positive results occurred at 412 visits (0.9%) and false positives at 96 visits (0.2%). Of 412 women with HIV seroconversion, 10 had discordant (i.e., one negative and one positive) rapid tests and 13 had undetectable HIV RNA levels. Of 62 women with false positive rapid HIV results, most had discordant rapid testing but six (9.7%) had dually-positive rapid results and four (6.5%) had false positive or indeterminate EIA results. The positive predictive value of dual positive rapid results was 98.3%.

CONCLUSION

Although the majority of rapid test results were accurate, false positive results were expected and occurred in this population of initially HIV seronegative individuals tested repeatedly and prospectively. When HIV infection occurred, not all cases had textbook laboratory results. Our findings highlight the importance of confirmatory testing, particularly for individuals undergoing repeat testing and in settings where the point prevalence is expected to be low.

Weight Changes and Adverse Pregnancy Outcomes With Dolutegravir- and Tenofovir Alafenamide Fumarate-Containing Antiretroviral Treatment Regimens During Pregnancy and Postpartum

BACKGROUND

We evaluated associations between antepartum weight change and adverse pregnancy outcomes and between antiretroviral therapy (ART) regimens and week 50 postpartum body mass index in IMPAACT 2010.

METHODS

Women with human immunodeficiency virus (HIV)-1 in 9 countries were randomized 1:1:1 at 14-28 weeks' gestational age (GA) to start dolutegravir (DTG) + emtricitabine (FTC)/tenofovir alafenamide fumarate (TAF) versus DTG + FTC/tenofovir disoproxil fumarate (TDF) versus efavirenz (EFV)/FTC/TDF. Insufficient antepartum weight gain was defined using Institute of Medicine guidelines. Cox-proportional hazards regression models were used to evaluate the association between antepartum weight change and adverse pregnancy outcomes: stillbirth (≥20 weeks' GA), preterm delivery (<37 weeks' GA), small size for GA (<10th percentile), and a composite of these endpoints.

RESULTS

A total of 643 participants were randomized: 217 to the DTG + FTC/TAF, 215 to the DTG + FTC/TDF, and 211 to the EFV/FTC/TDF arm. Baseline medians were as follows: GA, 21.9 weeks; HIV RNA, 903 copies/mL; and CD4 cell count, 466/μL. Insufficient weight gain was least frequent with DTG + FTC/TAF (15.0%) versus DTG + FTC/TDF (23.6%) and EFV/FTC/TDF (30.4%). Women in the DTG + FTC/TAF arm had the lowest rate of composite adverse pregnancy outcome. Low antepartum weight gain was associated with higher hazard of composite adverse pregnancy outcome (hazard ratio, 1.44 [95% confidence interval, 1.04-2.00]) and small size for GA (1.48 [.99-2.22]). More women in the DTG + FTC/TAF arm had a body mass index ≥25 (calculated as weight in kilograms divided by height in meters squared) at 50 weeks postpartum (54.7%) versus the DTG + FTC/TDF (45.2%) and EFV/FTC/TDF (34.2%) arms.

CONCLUSIONS

Antepartum weight gain on DTG regimens was protective against adverse pregnancy outcomes typically associated with insufficient weight gain, supportive of guidelines recommending DTG-based ART for women starting ART during pregnancy. Interventions to mitigate postpartum weight gain are needed.

Deep learning to estimate gestational age from fly-to cineloop videos: A novel approach to ultrasound quality control

OBJECTIVE

Low-cost devices have made obstetric sonography possible in settings where it was previously unfeasible, but ensuring quality and consistency at scale remains a challenge. In the present study, we sought to create a tool to reduce substandard fetal biometry measurement while minimizing care disruption.

METHODS

We developed a deep learning artificial intelligence (AI) model to estimate gestational age (GA) in the second and third trimester from fly-to cineloops-brief videos acquired during routine ultrasound biometry-and evaluated its performance in comparison to expert sonographer measurement. We then introduced random error into fetal biometry measurements and analyzed the ability of the AI model to flag grossly inaccurate measurements such as those that might be obtained by a novice.

RESULTS

The mean absolute error (MAE) of our model (±standard error) was 3.87 ± 0.07 days, compared to 4.80 ± 0.10 days for expert biometry (difference -0.92 days; 95% CI: -1.10 to -0.76). Based on simulated novice biometry with average absolute error of 7.5%, our model reliably detected cases where novice biometry differed from expert biometry by 10 days or more, with an area under the receiver operating characteristics curve of 0.93 (95% CI: 0.92, 0.95), sensitivity of 81.0% (95% CI: 77.9, 83.8), and specificity of 89.9% (95% CI: 88.1, 91.5). These results held across a range of sensitivity analyses, including where the model was provided suboptimal truncated fly-to cineloops.

CONCLUSIONS

Our AI model estimated GA more accurately than expert biometry. Because fly-to cineloop videos can be obtained without any change to sonographer workflow, the model represents a no-cost guardrail that could be incorporated into both low-cost and commercial ultrasound devices to prevent reporting of most gross GA estimation errors.

High rates of placental inflammation among samples collected by the Multi-Omics for Mothers and Infants consortium

BACKGROUND

The Multi-Omics for Mothers and Infants consortium aims to improve birth outcomes. Preterm birth is a major obstetrical complication globally and causes significant infant and childhood morbidity and mortality.

OBJECTIVE

We analyzed placental samples (basal plate, placenta or chorionic villi, and the chorionic plate) collected by the 5 Multi-Omics for Mothers and Infants sites, namely The Alliance for Maternal and Newborn Health Improvement Bangladesh, The Alliance for Maternal and Newborn Health Improvement Pakistan, The Alliance for Maternal and Newborn Health Improvement Tanzania, The Global Alliance to Prevent Prematurity and Stillbirth Bangladesh, and The Global Alliance to Prevent Prematurity and Stillbirth Zambia. The goal was to analyze the morphology and gene expression of samples collected from preterm and uncomplicated term births.

STUDY DESIGN

The teams provided biopsies from 166 singleton preterm (<37 weeks' gestation) and 175 term (≥37 weeks' gestation) deliveries. The samples were fixed in formalin and paraffin embedded. Tissue sections from these samples were stained with hematoxylin and eosin and subjected to morphologic analyses. Other placental biopsies (n=35 preterm, 21 term) were flash frozen, which enabled RNA purification for bulk transcriptomics.

RESULTS

The morphologic analyses revealed a surprisingly high rate of inflammation that involved the basal plate, placenta or chorionic villi, and the chorionic plate. The rate of inflammation in chorionic villus samples, likely attributable to chronic villitis, ranged from 25% (Pakistan site) to 60% (Zambia site) of cases. Leukocyte infiltration in this location vs in the basal plate or chorionic plate correlated with preterm birth. Our transcriptomic analyses identified 267 genes that were differentially expressed between placentas from preterm vs those from term births (123 upregulated, 144 downregulated). Mapping the differentially expressed genes onto single-cell RNA sequencing data from human placentas suggested that all the component cell types, either singly or in subsets, contributed to the observed dysregulation. Consistent with the histopathologic findings, gene ontology analyses highlighted the presence of leukocyte infiltration or activation and inflammatory responses in both the fetal and maternal compartments.

CONCLUSION

The relationship between placental inflammation and preterm birth is appreciated in developed countries. In this study, we showed that this link also exists in developing geographies. In addition, among the participating sites, we found geographic- and population-based differences in placental inflammation and preterm birth, suggesting the importance of local factors.

Acceptability of a Randomized Trial of Anti-depressant Medication or Interpersonal Therapy for Treatment of Perinatal Depression in Women with HIV

Postpartum depression (PPD) affects nearly 20% of postpartum women in Sub-Saharan Africa (SSA), where HIV prevalence is high. Depression is associated with worse HIV outcomes in non-pregnant adults and mental health disorders may worsen HIV outcomes for postpartum women and their infants. PPD is effectively treated with psychosocial or pharmacologic interventions; however, few studies have evaluated the acceptability of treatment modalities in SSA. We analyzed interviews with 23 postpartum women with HIV to assess the acceptability of two depression treatments provided in the context of a randomized trial. Most participants expressed acceptability of treatment randomization and study visit procedures. Participants shared perceptions of high treatment efficacy of their assigned intervention. They reported ongoing HIV and mental health stigma in their communities and emphasized the importance of social support from clinic staff. Our findings suggest a full-scale trial of PPD treatment will be acceptable among women with HIV in Zambia.

Leveraging External Validation Data: The Challenges of Transporting Measurement Error Parameters

Approaches to address measurement error frequently rely on validation data to estimate measurement error parameters (e.g., sensitivity and specificity). Acquisition of validation data can be costly, thus secondary use of existing data for validation is attractive. To use these external validation data, however, we may need to address systematic differences between these data and the main study sample. Here, we derive estimators of the risk and the risk difference that leverage external validation data to account for outcome misclassification. If misclassification is differential with respect to covariates that themselves are differentially distributed in the validation and study samples, the misclassification parameters are not immediately transportable. We introduce two ways to account for such covariates: (1) standardize by these covariates or (2) iteratively model the outcome. If conditioning on a covariate for transporting the misclassification parameters induces bias of the causal effect (e.g., M-bias), the former but not the latter approach is biased. We provide proof of identification, describe estimation using parametric models, and assess performance in simulations. We also illustrate implementation to estimate the risk of preterm birth and the effect of maternal HIV infection on preterm birth. Measurement error should not be ignored and it can be addressed using external validation data via transportability methods.

M-estimation for common epidemiological measures: introduction and applied examples

M-estimation is a statistical procedure that is particularly advantageous for some comon epidemiological analyses, including approaches to estimate an adjusted marginal risk contrast (i.e. inverse probability weighting and g-computation) and data fusion. In such settings, maximum likelihood variance estimates are not consistent. Thus, epidemiologists often resort to bootstrap to estimate the variance. In contrast, M-estimation allows for consistent variance estimates in these settings without requiring the computational complexity of the bootstrap. In this paper, we introduce M-estimation and provide four illustrative examples of implementation along with software code in multiple languages. M-estimation is a flexible and computationally efficient estimation procedure that is a powerful addition to the epidemiologist's toolbox.

Diagnostic performance of deep learning in ultrasound diagnosis of breast cancer: a systematic review

Deep learning (DL) has been widely investigated in breast ultrasound (US) for distinguishing between benign and malignant breast masses. This systematic review of test diagnosis aims to examine the accuracy of DL, compared to human readers, for the diagnosis of breast cancer in the US under clinical settings. Our literature search included records from databases including PubMed, Embase, Scopus, and Cochrane Library. Test accuracy outcomes were synthesized to compare the diagnostic performance of DL and human readers as well as to evaluate the assistive role of DL to human readers. A total of 16 studies involving 9238 female participants were included. There were no prospective studies comparing the test accuracy of DL versus human readers in clinical workflows. Diagnostic test results varied across the included studies. In 14 studies employing standalone DL systems, DL showed significantly lower sensitivities in 5 studies with comparable specificities and outperformed human readers at higher specificities in another 4 studies; in the remaining studies, DL models and human readers showed equivalent test outcomes. In 12 studies that assessed assistive DL systems, no studies proved the assistive role of DL in the overall diagnostic performance of human readers. Current evidence is insufficient to conclude that DL outperforms human readers or enhances the accuracy of diagnostic breast US in a clinical setting. Standardization of study methodologies is required to improve the reproducibility and generalizability of DL research, which will aid in clinical translation and application.

Association of maternal prenatal copper concentration with gestational duration and preterm birth: a multicountry meta-analysis

BACKGROUND

Copper (Cu), an essential trace mineral regulating multiple actions of inflammation and oxidative stress, has been implicated in risk for preterm birth (PTB).

OBJECTIVES

This study aimed to determine the association of maternal Cu concentration during pregnancy with PTB risk and gestational duration in a large multicohort study including diverse populations.

METHODS

Maternal plasma or serum samples of 10,449 singleton live births were obtained from 18 geographically diverse study cohorts. Maternal Cu concentrations were determined using inductively coupled plasma mass spectrometry. The associations of maternal Cu with PTB and gestational duration were analyzed using logistic and linear regressions for each cohort. The estimates were then combined using meta-analysis. Associations between maternal Cu and acute-phase reactants (APRs) and infection status were analyzed in 1239 samples from the Malawi cohort.

RESULTS

The maternal prenatal Cu concentration in our study samples followed normal distribution with mean of 1.92 μg/mL and standard deviation of 0.43 μg/mL, and Cu concentrations increased with gestational age up to 20 wk. The random-effect meta-analysis across 18 cohorts revealed that 1 μg/mL increase in maternal Cu concentration was associated with higher risk of PTB with odds ratio of 1.30 (95% confidence interval [CI]: 1.08, 1.57) and shorter gestational duration of 1.64 d (95% CI: 0.56, 2.73). In the Malawi cohort, higher maternal Cu concentration, concentrations of multiple APRs, and infections (malaria and HIV) were correlated and associated with greater risk of PTB and shorter gestational duration.

CONCLUSIONS

Our study supports robust negative association between maternal Cu and gestational duration and positive association with risk for PTB. Cu concentration was strongly correlated with APRs and infection status suggesting its potential role in inflammation, a pathway implicated in the mechanisms of PTB. Therefore, maternal Cu could be used as potential marker of integrated inflammatory pathways during pregnancy and risk for PTB.

Accounting for nonmonotone missing data using inverse probability weighting

Inverse probability weighting can be used to correct for missing data. New estimators for the weights in the nonmonotone setting were introduced in 2018. These estimators are the unconstrained maximum likelihood estimator (UMLE) and the constrained Bayesian estimator (CBE), an alternative if UMLE fails to converge. In this work we describe and illustrate these estimators, and examine performance in simulation and in an applied example estimating the effect of anemia on spontaneous preterm birth in the Zambia Preterm Birth Prevention Study. We compare performance with multiple imputation (MI) and focus on the setting of an observational study where inverse probability of treatment weights are used to address confounding. In simulation, weighting was less statistically efficient at the smallest sample size and lowest exposure prevalence examined (n = 1500, 15% respectively) but in other scenarios statistical performance of weighting and MI was similar. Weighting had improved computational efficiency taking, on average, 0.4 and 0.05 times the time for MI in R and SAS, respectively. UMLE was easy to implement in commonly used software and convergence failure occurred just twice in >200 000 simulated cohorts making implementation of CBE unnecessary. In conclusion, weighting is an alternative to MI for nonmonotone missingness, though MI performed as well as or better in terms of bias and statistical efficiency. Weighting's superior computational efficiency may be preferred with large sample sizes or when using resampling algorithms. As validity of weighting and MI rely on correct specification of different models, both approaches could be implemented to check agreement of results.

Inertial Measurement Unit-Assisted Ultrasonic Tracking System for Ultrasound Probe Localization

Ultrasonic tracking is a promising technique in indoor object localization. However, limited success has been reported in dynamic orientational and positional ultrasonic tracking for ultrasound (US) probes due to its instability and relatively low accuracy. This article aims at developing an inertial measurement unit (IMU)-assisted ultrasonic tracking system that enables a high accuracy positional and orientational localization. The system was designed with the acoustic pressure field simulation of the transmitter, receiver configuration, position-variant error simulation, and sensor fusion. The prototype was tested in a tracking volume required in typical obstetric sonography within the typical operation speed ranges (slow mode and fast mode) of US probe movement. The performance in two different speed ranges was evaluated against a commercial optical tracking device. The results show that the proposed IMU-assisted US tracking system achieved centimeter-level positional tracking accuracy with the mean absolute error (MAE) of 12 mm and the MAE of orientational tracking was less than 1°. The results indicate the possibility of implementing the IMU-assisted ultrasonic tracking system in US probe localization.

Reducing Bias in Estimates of Per Protocol Treatment Effects: A Secondary Analysis of a Randomized Clinical Trial

This secondary analysis of a randomized clinical trial evaluates ways of reducing bias in estimates of per protocol treatment effects.

Combination HIV prevention during pregnancy and the post-partum period in Malawi and Zambia: a mathematical modelling analysis

INTRODUCTION

Despite widespread success in reducing vertical HIV transmission, most antenatal care (ANC) programmes in eastern and southern Africa have not emphasized primary prevention of maternal HIV acquisition during pregnancy and lactation/breastfeeding. We hypothesized that combination HIV prevention interventions initiated alongside ANC could substantially reduce maternal HIV incidence.

METHODS

We constructed a multi-state model describing male-to-female HIV transmission in steady heterosexual partnerships during pregnancy and lactation/breastfeeding, with initial conditions based on population distribution estimates for Malawi and Zambia in 2020. We modelled individual and joint increases in three HIV prevention strategies at or soon after ANC initiation: (1) HIV testing of male partners, resulting in HIV diagnosis and less condomless sex among those with previously undiagnosed HIV; (2) initiation (or re-initiation) of suppressive antiretroviral therapy (ART) for male partners with diagnosed but unsuppressed HIV; and (3) adherent pre-exposure prophylaxis (PrEP) for HIV-negative female ANC patients with HIV-diagnosed or unknown-status male partners. We estimated the percentage of within-couple, male-to-female HIV transmissions that could be averted during pregnancy and lactation/breastfeeding with these strategies, relative to base-case conditions in which 45% of undiagnosed male partners become newly HIV diagnosed via testing, 75% of male partners with diagnosed but unsuppressed HIV initiate/re-initiate ART and 0% of female ANC patients start PrEP.

RESULTS

Increasing uptake of any single strategy by 20 percentage points above base-case levels averted 10%-11% of maternal HIV acquisitions during pregnancy and lactation/breastfeeding in the model. Joint uptake increases of 20 percentage points in two interventions averted an estimated 19%-23% of transmissions, and with a 20-percentage-point increase in uptake of all three interventions, 29% were averted. Strategies achieving 95% male testing, 90% male ART initiation/re-initiation and 40% female PrEP use reduced incident infections by 45%.

CONCLUSIONS

Combination HIV prevention strategies provided alongside ANC and sustained through the post-partum period could substantially reduce maternal HIV incidence during pregnancy and lactation/breastfeeding in eastern and southern Africa.

Multiomic signals associated with maternal epidemiological factors contributing to preterm birth in low- and middle-income countries

Preterm birth (PTB) is the leading cause of death in children under five, yet comprehensive studies are hindered by its multiple complex etiologies. Epidemiological associations between PTB and maternal characteristics have been previously described. This work used multiomic profiling and multivariate modeling to investigate the biological signatures of these characteristics. Maternal covariates were collected during pregnancy from 13,841 pregnant women across five sites. Plasma samples from 231 participants were analyzed to generate proteomic, metabolomic, and lipidomic datasets. Machine learning models showed robust performance for the prediction of PTB (AUROC = 0.70), time-to-delivery (r = 0.65), maternal age (r = 0.59), gravidity (r = 0.56), and BMI (r = 0.81). Time-to-delivery biological correlates included fetal-associated proteins (e.g., ALPP, AFP, and PGF) and immune proteins (e.g., PD-L1, CCL28, and LIFR). Maternal age negatively correlated with collagen COL9A1, gravidity with endothelial NOS and inflammatory chemokine CXCL13, and BMI with leptin and structural protein FABP4. These results provide an integrated view of epidemiological factors associated with PTB and identify biological signatures of clinical covariates affecting this disease.

Maternal mid-upper arm circumference to predict small for gestational age: Findings in a Zambian cohort

OBJECTIVE

To compare the performance of mid upper arm circumference (MUAC) and body mass index (BMI) for prediction of small for gestational age (SGA) in Zambia.

METHODS

This is a secondary analysis of an ongoing clinical cohort that included women with a single gestation and MUAC measured before 24 weeks of pregnancy. We assessed relationships between maternal MUAC and birth weight centile using regression. The performance of MUAC and BMI to predict SGA was compared using receiver operating characteristic curves and the effect of maternal HIV was investigated in sub-group analyses.

RESULTS

Of 1117 participants, 847 (75%) were HIV-negative (HIV-) and 270 (24%) were HIV-positive (HIV+). Seventy-four (7%) delivered severe SGA infants (<3rd centile), of whom 56 (76%) were HIV- and 18 (24%) were HIV+ (odds ratio [OR] 1.01, 95% confidence interval [CI] 0.58-1.75). MUAC was associated with higher birth weight centile (+1.2 centile points, 95% CI 0.7-1.6; P < 0.001); this relationship was stronger among HIV+ women (+1.7 centile points, 95% CI 0.8-2.6; P < 0.001) than HIV- women (+0.9 centile points, 95% CI 0.4-1.4; P = 0.001). The discriminatory power was similar, albeit poor (area under the curve [AUC] < 0.7), between MUAC and BMI for the prediction of SGA. In stratified analysis, MUAC and BMI showed excellent discrimination predicting severe SGA among HIV+ (AUC 0.83 and 0.81, respectively) but not among HIV- women (AUC 0.64 and 0.63, respectively).

CONCLUSION

Maternal HIV infection increased the discrimination of both early pregnancy MUAC and BMI for prediction of severe SGA in Zambia.

CLINICAL TRIAL NUMBER

ClinicalTrials.gov (NCT02738892).

Global research priorities related to the World Health Organization Labour Care Guide: results of a global consultation

BACKGROUND

The World Health Organization (WHO) published the WHO Labour Care Guide (LCG) in 2020 to support the implementation of its 2018 recommendations on intrapartum care. The WHO LCG promotes evidence-based labour monitoring and stimulates shared decision-making between maternity care providers and labouring women. There is a need to identify critical questions that will contribute to defining the research agenda relating to implementation of the WHO LCG.

METHODS

This mixed-methods prioritization exercise, adapted from the Child Health and Nutrition Research Initiative (CHNRI) and James Lind Alliance (JLA) methods, combined a metrics-based design with a qualitative, consensus-building consultation in three phases. The exercise followed the reporting guideline for priority setting of health research (REPRISE). First, 30 stakeholders were invited to submit online ideas or questions (generation of research ideas). Then, 220 stakeholders were invited to score "research avenues" (i.e., broad research ideas that could be answered through a set of research questions) against six independent and equally weighted criteria (scoring of research avenues). Finally, a technical working group (TWG) of 20 purposively selected stakeholders reviewed the scoring, and refined and ranked the research avenues (consensus-building meeting).

RESULTS

Initially, 24 stakeholders submitted 89 research ideas or questions. A list of 10 consolidated research avenues was scored by 75/220 stakeholders. During the virtual consensus-building meeting, research avenues were refined, and the top three priorities agreed upon were: (1) optimize implementation strategies of WHO LCG, (2) improve understanding of the effect of WHO LCG on maternal and perinatal outcomes, and the process and experience of labour and childbirth care, and (3) assess the effect of the WHO LCG in special situations or settings. Research avenues related to the organization of care and resource utilization ranked lowest during both the scoring and consensus-building process.

CONCLUSION

This systematic and transparent process should encourage researchers, program implementers, and funders to support research aligned with the identified priorities related to WHO LCG. An international collaborative platform is recommended to implement prioritized research by using harmonized research tools, establishing a repository of research priorities studies, and scaling-up successful research results.

Different effects for different questions: An illustration using short cervix and the risk of preterm birth

OBJECTIVE

To illustrate the difference between exposure effects and population attributable effects.

METHODS

We examined the effect of mid-pregnancy short cervical length (<25 mm) on preterm birth using data from a prospective cohort of pregnant women in Lusaka, Zambia. Preterm birth was live birth or stillbirth before 37 weeks of pregnancy. For estimation, we used multivariable regression and parametric g-computation.

RESULTS

Among 1409 women included in the analysis, short cervix was rare (2.4%); 13.6% of births were preterm. Exposure effect estimates were large (marginal risk ratio 2.86, 95% confidence interval [CI] 1.80-4.54), indicating that the preterm birth risk was substantially higher among women with a short cervix compared with women without a short cervix. However, the population attributable effect estimates were close to the null (risk ratio 1.06, 95% CI 1.02-1.10), indicating that an intervention to counteract the impact of short cervix on preterm birth would have minimal effect on the population risk of preterm birth.

CONCLUSION

Although authors often refer to "the" effect, there are actually different types of effects, as we have illustrated here. In planning research, it is important to consider which effect to estimate to ensure that the estimate aligns with the research objective.

Missing Outcome Data in Epidemiologic Studies

Missing data are pandemic and a central problem for epidemiology. Missing data reduce precision and can cause notable bias. There remain too few simple published examples detailing types of missing data and illustrating their possible impact on results. Here we take an example randomized trial that was not subject to missing data and induce missing data to illustrate 4 scenarios in which outcomes are 1) missing completely at random, 2) missing at random with positivity, 3) missing at random without positivity, and 4) missing not at random. We demonstrate that accounting for missing data is generally a better strategy than ignoring missing data, which unfortunately remains a standard approach in epidemiology.

Development and external validation of machine learning algorithms for postnatal gestational age estimation using clinical data and metabolomic markers

BACKGROUND

Accurate estimates of gestational age (GA) at birth are important for preterm birth surveillance but can be challenging to obtain in low income countries. Our objective was to develop machine learning models to accurately estimate GA shortly after birth using clinical and metabolomic data.

METHODS

We derived three GA estimation models using ELASTIC NET multivariable linear regression using metabolomic markers from heel-prick blood samples and clinical data from a retrospective cohort of newborns from Ontario, Canada. We conducted internal model validation in an independent cohort of Ontario newborns, and external validation in heel prick and cord blood sample data collected from newborns from prospective birth cohorts in Lusaka, Zambia and Matlab, Bangladesh. Model performance was measured by comparing model-derived estimates of GA to reference estimates from early pregnancy ultrasound.

RESULTS

Samples were collected from 311 newborns from Zambia and 1176 from Bangladesh. The best-performing model accurately estimated GA within about 6 days of ultrasound estimates in both cohorts when applied to heel prick data (MAE 0.79 weeks (95% CI 0.69, 0.90) for Zambia; 0.81 weeks (0.75, 0.86) for Bangladesh), and within about 7 days when applied to cord blood data (1.02 weeks (0.90, 1.15) for Zambia; 0.95 weeks (0.90, 0.99) for Bangladesh).

CONCLUSIONS

Algorithms developed in Canada provided accurate estimates of GA when applied to external cohorts from Zambia and Bangladesh. Model performance was superior in heel prick data as compared to cord blood data.

Limiting adverse birth outcomes in resource-limited settings (LABOR): protocol of a prospective intrapartum cohort study

Background: Each year, nearly 300,000 women and 5 million fetuses or neonates die during childbirth or shortly thereafter, a burden concentrated disproportionately in low- and middle-income countries. Identifying women and their fetuses at risk for intrapartum-related morbidity and death could facilitate early intervention. Methods: The Limiting Adverse Birth Outcomes in Resource-Limited Settings (LABOR) Study is a multi-country, prospective, observational cohort designed to exhaustively document the course and outcomes of labor, delivery, and the immediate postpartum period in settings where adverse outcomes are frequent. The study is conducted at four hospitals across three countries in Ghana, India, and Zambia. We will enroll approximately 12,000 women at presentation to the hospital for delivery and follow them and their fetuses/newborns throughout their labor and delivery course, postpartum hospitalization, and up to 42 days thereafter. The co-primary outcomes are composites of maternal (death, hemorrhage, hypertensive disorders, infection) and fetal/neonatal adverse events (death, encephalopathy, sepsis) that may be attributed to the intrapartum period. The study collects extensive physiologic data through the use of physiologic sensors and employs medical scribes to document examination findings, diagnoses, medications, and other interventions in real time. Discussion: The goal of this research is to produce a large, sharable dataset that can be used to build statistical algorithms to prospectively stratify parturients according to their risk of adverse outcomes. We anticipate this research will inform the development of new tools to reduce peripartum morbidity and mortality in low-resource settings.

Limiting adverse birth outcomes in resource-limited settings (LABOR): protocol of a prospective intrapartum cohort study

Background: Each year, nearly 300,000 women and 5 million fetuses or neonates die during childbirth or shortly thereafter, a burden concentrated disproportionately in low- and middle-income countries. Identifying women and their fetuses at risk for intrapartum-related morbidity and death could facilitate early intervention. Methods: The Limiting Adverse Birth Outcomes in Resource-Limited Settings (LABOR) Study is a multi-country, prospective, observational cohort designed to exhaustively document the course and outcomes of labor, delivery, and the immediate postpartum period in settings where adverse outcomes are frequent. The study is conducted at four hospitals across three countries in Ghana, India, and Zambia. We will enroll approximately 12,000 women at presentation to the hospital for delivery and follow them and their fetuses/newborns throughout their labor and delivery course, postpartum hospitalization, and up to 42 days thereafter. The co-primary outcomes are composites of maternal (death, hemorrhage, hypertensive disorders, infection) and fetal/neonatal adverse events (death, encephalopathy, sepsis) that may be attributed to the intrapartum period. The study collects extensive physiologic data through the use of physiologic sensors and employs medical scribes to document examination findings, diagnoses, medications, and other interventions in real time. Discussion: The goal of this research is to produce a large, sharable dataset that can be used to build statistical algorithms to prospectively stratify parturients according to their risk of adverse outcomes. We anticipate this research will inform the development of new tools to reduce peripartum morbidity and mortality in low-resource settings.

Effect of weekly 17-hydroxyprogesterone caproate on small for gestational age among pregnant women with HIV in Zambia

The IPOP trial demonstrated a reduced risk of severe small for gestational age among infants born to women with HIV who received weekly intramuscular 17 alpha-hydroxyprogesterone caproate. This secondary analysis examined the 17P treatment effect in subgroups of maternal BMI, parity, timing of antiretroviral therapy (ART) initiation, and ART regimen. We found that 17P was more effective among nulliparous women, women who started ART before pregnancy, and those taking protease inhibitors.