A "Prediction - Detection - Judgment" framework for sudden water contamination event detection with online monitoring

The contamination detection technology helps in water quality management and protection in surface water. It is important to detect sudden contamination events timely from dynamic variations due to various interference factors in online water quality monitoring data. In this study, a framework named "Prediction - Detection - Judgment" is proposed with a method framework of "Time series increment - Hierarchical clustering - Bayes' theorem model". Time to detection is used as an evaluation index of contamination detection methods, along with the probability of detection and false alarm rate. The proposed method is tested with available public data and further applied in a monitoring site of a river. Results showed that the method could detect the contamination events with a 100% probability of detection, a 17% false alarm rate and a time to detection close to 4 monitoring intervals. The proposed index time to detection evaluates the timeliness of the method, and timely detection ensures that contamination events can be responded to and dealt with in time. The site application also demonstrates the feasibility and practicability of the framework proposed in this study and its potential for extensive implementation.

Statistically significant differences versus convincing evidence of real treatment effects: an analysis of the false positive risk for single-centre trials in anaesthesia

BACKGROUND

The American Statistical Association has highlighted problems with null hypothesis significance testing and outlined alternative approaches that may 'supplement or even replace P-values'. One alternative is to report the false positive risk (FPR), which quantifies the chance the null hypothesis is true when the result is statistically significant.

METHODS

We reviewed single-centre, randomised trials in 10 anaesthesia journals over 6 yr where differences in a primary binary outcome were statistically significant. We calculated a Bayes factor by two methods (Gunel, Kass). From the Bayes factor we calculated the FPR for different prior beliefs for a real treatment effect. Prior beliefs were quantified by assigning pretest probabilities to the null and alternative hypotheses.

RESULTS

For equal pretest probabilities of 0.5, the median (inter-quartile range [IQR]) FPR was 6% (1-22%) by the Gunel method and 6% (1-19%) by the Kass method. One in five trials had an FPR ≥20%. For trials reporting P-values 0.01-0.05, the median (IQR) FPR was 25% (16-30%) by the Gunel method and 20% (16-25%) by the Kass method. More than 90% of trials reporting P-values 0.01-0.05 required a pretest probability >0.5 to achieve an FPR of 5%. The median (IQR) difference in the FPR calculated by the two methods was 0% (0-2%).

CONCLUSIONS

Our findings suggest that a substantial proportion of single-centre trials in anaesthesia reporting statistically significant differences provide limited evidence of real treatment effects, or, alternatively, required an implausibly high prior belief in a real treatment effect.

CLINICAL TRIAL REGISTRATION

PROSPERO (CRD42023350783).

Is scientific knowledge socially constructed? A Bayesian account of Laboratory Life

In the book Laboratory Life Latour and Woolgar present an account of how scientific "facts" are formed through a process of microsocial interactions among individuals and "inscription devices" in the lab initially described as social construction. The process moves through a series of steps during which the details and nature of the object become more and more certain until all qualifications are dropped, and the "fact" emerges as secure scientific knowledge. An alternative to this account is described based on a Bayesian probabilistic framework which arrives at the same end point. The motive force for the constructivist approach appears to involve social processes of convincing colleagues while the Bayesian approach relies on the consistency of theory and evidence as judged by the participants. The role of social processes is discussed in Bayesian terms, the acquisition and asymmetry of information, and its analogy to puzzle solving. Some parallels between the Bayesian and constructivist accounts are noted especially in relation to information theory.

External validation of Fetal Medicine Foundation competing-risks model for midgestation prediction of small-for-gestational-age neonates in Spanish population

OBJECTIVE

To examine the external validity of the new Fetal Medicine Foundation (FMF) competing-risks model for prediction in midgestation of small-for-gestational-age (SGA) neonates.

METHODS

This was a single-center prospective cohort study of 25 484 women with a singleton pregnancy undergoing routine ultrasound examination at 19 + 0 to 23 + 6 weeks' gestation. The FMF competing-risks model for the prediction of SGA combining maternal factors and midgestation estimated fetal weight by ultrasound scan (EFW) and uterine artery pulsatility index (UtA-PI) was used to calculate risks for different cut-offs of birth-weight percentile and gestational age at delivery. The predictive performance was evaluated in terms of discrimination and calibration.

RESULTS

The validation cohort was significantly different in composition compared with the FMF cohort in which the model was developed. In the validation cohort, at a 10% false-positive rate (FPR), maternal factors, EFW and UtA-PI yielded detection rates of 69.6%, 38.7% and 31.7% for SGA < 10th percentile with delivery at < 32, < 37 and ≥ 37 weeks' gestation, respectively. The respective values for SGA < 3rd percentile were 75.7%, 48.2% and 38.1%. Detection rates in the validation cohort were similar to those reported in the FMF study for SGA with delivery at < 32 weeks but lower for SGA with delivery at < 37 and ≥ 37 weeks. Predictive performance in the validation cohort was similar to that reported in a subgroup of the FMF cohort consisting of nulliparous and Caucasian women. Detection rates in the validation cohort at a 15% FPR were 77.4%, 50.0% and 41.5% for SGA < 10th percentile with delivery at < 32, < 37 and ≥ 37 weeks, respectively, which were similar to the respective values reported in the FMF study at a 10% FPR. The model had satisfactory calibration.

CONCLUSION

The new competing-risks model for midgestation prediction of SGA developed by the FMF performs well in a large independent Spanish population. © 2023 International Society of Ultrasound in Obstetrics and Gynecology.

Multicentre randomised trials in anaesthesia: an analysis using Bayesian metrics

Are the results of randomised trials reliable and are p values and confidence intervals the best way of quantifying efficacy? Low power is common in medical research, which reduces the probability of obtaining a 'significant result' and declaring the intervention had an effect. Metrics derived from Bayesian methods may provide an insight into trial data unavailable from p values and confidence intervals. We did a structured review of multicentre trials in anaesthesia that were published in the New England Journal of Medicine, The Lancet, Journal of the American Medical Association, British Journal of Anaesthesia and Anesthesiology between February 2011 and November 2021. We documented whether trials declared a non-zero effect by an intervention on the primary outcome. We documented the expected and observed effect sizes. We calculated a Bayes factor from the published trial data indicating the probability of the data under the null hypothesis of zero effect relative to the alternative hypothesis of a non-zero effect. We used the Bayes factor to calculate the post-test probability of zero effect for the intervention (having assumed 50% belief in zero effect before the trial). We contacted all authors to estimate the costs of running the trials. The median (IQR [range]) hypothesised and observed absolute effect sizes were 7% (3-13% [0-25%]) vs. 2% (1-7% [0-24%]), respectively. Non-zero effects were declared for 12/56 outcomes (21%). The Bayes factor favouring a zero effect relative to a non-zero effect for these 12 trials was 0.000001-1.9, with post-test zero effect probabilities for the intervention of 0.0001-65%. The other 44 trials did not declare non-zero effects, with Bayes factors favouring zero effect of 1-688, and post-test probabilities of zero effect of 53-99%. The median (IQR [range]) study costs reported by 20 corresponding authors in US$ were $1,425,669 ($514,766-$2,526,807 [$120,758-$24,763,921]). We think that inadequate power and mortality as an outcome are why few trials declared non-zero effects. Bayes factors and post-test probabilities provide a useful insight into trial results, particularly when p values approximate the significance threshold.

Probabilistic co-occurrence assessment for suites of listed species

Section 7 of the Endangered Species Act requires the US Environmental Protection Agency (US EPA) to consult with the Services (US Fish and Wildlife Service and National Marine Fisheries Service) over potential pesticide impacts on federally listed species. Consultation is complicated by the large number of pesticide products and listed species, as well as by lack of consensus on best practices for conducting co-occurrence analyses. Previous work demonstrates that probabilistic estimates of species' ranges and pesticide use patterns improve these analyses. Here we demonstrate that such estimates can be made for suites of sympatric listed species. Focusing on two watersheds, one in Iowa and the other in Mississippi, we obtained distribution records for 13 species of terrestrial and aquatic listed plants and animals occurring therein. We used maximum entropy modeling and bioclimatic, topographic, hydrographic, and land cover variables to predict species' ranges at high spatial resolution. We constructed probabilistic spatial models of use areas for two pesticides based on the US Department of Agriculture Cropland Data Layer and reduced classification errors by incorporating information on the relationships between individual pixels and their neighbors using object-based images analysis. We then combined species distribution and crop footprint models to derive overall probability of co-occurrence of listed species and pesticide use. For aquatic species, we also integrated an estimate of downstream residue transport. We report each separate species-by-use-area co-occurrence estimate and also combine these modeled co-occurrence probabilities across species within watersheds to produce an overall metric of potential pesticide exposure risk for these listed species at the watershed level. We propose that the consultation process between US EPA and the Services be based on such batched estimation of probabilistic co-occurrence for multiple listed species at a regional scale. Integr Environ Assess Manag 2022;18:1088-1100. © 2021 SETAC.

Second-trimester contingent screening for small-for-gestational-age neonate

OBJECTIVES

First, to investigate the additive value of second-trimester placental growth factor (PlGF) for the prediction of a small-for-gestational-age (SGA) neonate. Second, to examine second-trimester contingent screening strategies.

METHODS

This was a prospective observational study in women with singleton pregnancy undergoing routine ultrasound examination at 19-24 weeks' gestation. We used the competing-risks model for prediction of SGA. The parameters for the prior model and the likelihoods for estimated fetal weight (EFW) and uterine artery pulsatility index (UtA-PI) were those presented in previous studies. A folded-plane regression model was fitted in the dataset of this study to describe the likelihood of PlGF. We compared the prediction of screening by maternal risk factors against the prediction provided by a combination of maternal risk factors, EFW, UtA-PI and PlGF. We also examined the additive value of PlGF in a policy that uses maternal risk factors, EFW and UtA-PI.

RESULTS

The study population included 40 241 singleton pregnancies. Overall, the prediction of SGA improved with increasing degree of prematurity, with increasing severity of smallness and in the presence of coexisting pre-eclampsia. The combination of maternal risk factors, EFW, UtA-PI and PlGF improved significantly the prediction provided by maternal risk factors alone for all the examined cut-offs of birth weight and gestational age at delivery. Screening by a combination of maternal risk factors and serum PlGF improved the prediction of SGA when compared to screening by maternal risk factors alone. However, the incremental improvement in prediction was decreased when PlGF was added to screening by a combination of maternal risk factors, EFW and UtA-PI. If first-line screening for a SGA neonate with birth weight < 10^th percentile delivered at < 37 weeks' gestation was by maternal risk factors and EFW, the same detection rate of 90%, at an overall false-positive rate (FPR) of 50%, as that achieved by screening with maternal risk factors, EFW, UtA-PI and PlGF in the whole population can be achieved by reserving measurements of UtA-PI and PlGF for only 80% of the population. Similarly, in screening for a SGA neonate with birth weight < 10^th percentile delivered at < 30 weeks, the same detection rate of 90%, at an overall FPR of 14%, as that achieved by screening with maternal risk factors, EFW, UtA-PI and PlGF in the whole population can be achieved by reserving measurements of UtA-PI and PlGF for only 70% of the population. The additive value of PlGF in reducing the FPR to about 10% with a simultaneous detection rate of 90% for a SGA neonate with birth weight < 3^rd percentile born < 30 weeks, is gained by measuring PlGF in only 50% of the population when first-line screening is by maternal factors, EFW and UtA-PI.

CONCLUSIONS

The combination of maternal risk factors, EFW, UtA-PI and PlGF provides effective second-trimester prediction of SGA. Serum PlGF is useful for predicting a SGA neonate with birth weight < 3^rd percentile born < 30 weeks after an inclusive assessment by maternal risk factors and biophysical markers. Similar detection rates and FPRs can be achieved by application of contingent screening strategies. © 2021 International Society of Ultrasound in Obstetrics and Gynecology.

Competing-risks model for prediction of small-for-gestational-age neonate from estimated fetal weight at 19-24 weeks' gestation

OBJECTIVE

To develop further a new competing-risks model for the prediction of a small-for-gestational-age (SGA) neonate, by including second-trimester ultrasonographic estimated fetal weight (EFW).

METHODS

This was a prospective observational study in 96 678 women with singleton pregnancy undergoing routine ultrasound examination at 19-24 weeks' gestation. All pregnancies had ultrasound biometry assessment, and EFW was calculated according to the Hadlock formula. We refitted in this large dataset a previously described competing-risks model for the joint distribution of gestational age (GA) at delivery and birth-weight Z-score, according to maternal demographic characteristics and medical history, to obtain the prior distribution. The continuous likelihood of the EFW was fitted conditionally to GA at delivery and birth-weight Z-score and modified the prior distribution, according to Bayes' theorem, to obtain individualized distributions for GA at delivery and birth-weight Z-score and therefore patient-specific risks for any cut-offs for GA at delivery and birth-weight Z-score. We assessed the discriminative ability of the model for predicting SGA with, without or independently of pre-eclampsia occurrence. A calibration study was carried out. Performance of screening was evaluated for SGA defined according to the Fetal Medicine Foundation birth-weight charts.

RESULTS

The distribution of EFW, conditional to both GA at delivery and birth-weight Z-score, was best described by a regression model. For earlier gestations, the association between EFW and birth weight was steeper. The prediction of SGA by maternal factors and EFW improved for increasing degree of prematurity and greater severity of smallness but not for coexistence of pre-eclampsia. Screening by maternal factors predicted 31%, 34% and 39% of SGA neonates with birth weight < 10^th percentile delivered at ≥ 37, < 37 and < 30 weeks' gestation, respectively, at a 10% false-positive rate, and, after addition of EFW, these rates increased to 38%, 43% and 59%, respectively; the respective rates for birth weight < 3^rd percentile were 43%, 50% and 64%. The addition of EFW improved the calibration of the model.

CONCLUSION

In the competing-risks model for prediction of SGA, the performance of screening by maternal characteristics and medical history is improved by the addition of second-trimester EFW. © 2021 International Society of Ultrasound in Obstetrics and Gynecology.

Competing-risks model for prediction of small-for-gestational-age neonate from maternal characteristics, serum pregnancy-associated plasma protein-A and placental growth factor at 11-13 weeks' gestation

OBJECTIVES

To expand a new competing-risks model for prediction of a small-for-gestational-age (SGA) neonate, by the addition of pregnancy-associated plasma protein-A (PAPP-A) and placental growth factor (PlGF), and to evaluate and compare PAPP-A and PlGF in predicting SGA.

METHODS

This was a prospective observational study of 60 875 women with singleton pregnancy undergoing routine ultrasound examination at 11 + 0 to 13 + 6 weeks' gestation. We fitted a folded-plane regression model for the PAPP-A and PlGF likelihoods. A previously developed maternal history model and the likelihood models were combined, according to Bayes' theorem, to obtain individualized distributions for gestational age (GA) at delivery and birth-weight Z-score. We assessed the discrimination and calibration of the model. McNemar's test was used to compare the detection rates for SGA with, without or independently of pre-eclampsia (PE) occurrence, of different combinations of maternal history, PAPP-A and PlGF, for a fixed false-positive rate.

RESULTS

The distributions of PAPP-A and PlGF depend on both GA at delivery and birth-weight Z-score, in the same continuous likelihood, according to a folded-plane regression model. The new approach offers the capability for risk computation for any desired birth-weight Z-score and GA at delivery cut-off. PlGF was consistently and significantly better than PAPP-A in predicting SGA delivered before 37 weeks, especially in cases with co-existence of PE. PAPP-A had similar performance to PlGF for the prediction of SGA without PE. At a fixed false-positive rate of 10%, the combination of maternal history, PlGF and PAPP-A predicted 33.8%, 43.8% and 48.4% of all cases of a SGA neonate with birth weight < 10^th percentile delivered at ≥ 37, < 37 and < 32 weeks' gestation, respectively. The respective values for birth weight < 3^rd percentile were 38.6%, 48.7% and 51.0%. The new model performed well in terms of risk calibration.

CONCLUSIONS

The combination of PAPP-A and PlGF values with maternal characteristics, according to Bayes' theorem, improves prediction of SGA. PlGF is a better predictor of SGA than PAPP-A, especially when PE is present. The new competing-risks model for SGA can be tailored to each pregnancy and to the relevant clinical requirements. © 2020 International Society of Ultrasound in Obstetrics and Gynecology.

Competing-risks model for prediction of small-for-gestational-age neonate from biophysical and biochemical markers at 11-13 weeks' gestation

OBJECTIVE

To develop a new competing-risks model for the prediction of a small-for-gestational-age (SGA) neonate, based on maternal factors and biophysical and biochemical markers at 11-13 weeks' gestation.

METHODS

This was a prospective observational study in 60 875 women with singleton pregnancy undergoing routine ultrasound examination at 11 + 0 to 13 + 6 weeks' gestation. All pregnancies had pregnancy-associated plasma protein-A and placental growth factor (PlGF) measurements, 59 001 had uterine artery pulsatility index (UtA-PI) measurements and 58 479 had mean arterial pressure measurements; 57 131 cases had complete data for all biomarkers. We used a previously developed competing-risks model for the joint distribution of gestational age (GA) at delivery and birth-weight Z-score, according to maternal demographic characteristics and medical history. The likelihoods of the biophysical markers were developed by fitting folded-plane regression models, a technique that has already been used in previous studies for the likelihoods of biochemical markers. The next step was to modify the prior distribution by the likelihood, according to Bayes' theorem, to obtain individualized distributions for GA at delivery and birth-weight Z-score. We used the 57 131 cases with complete data to assess the discrimination and calibration of the model for predicting SGA with, without or independently of pre-eclampsia, by different combinations of maternal factors and biomarkers.

RESULTS

The distribution of biomarkers, conditional to both GA at delivery and birth-weight Z-score, was best described by folded-plane regression models. These continuous two-dimensional likelihoods update the joint distribution of birth-weight Z-score and GA at delivery that has resulted from a competing-risks approach; this method allows application of user-defined cut-offs. The best biophysical predictor of preterm SGA was UtA-PI and the best biochemical marker was PlGF. The prediction of SGA was consistently better for increasing degree of prematurity, greater severity of smallness, coexistence of PE and increasing number of biomarkers. The combination of maternal factors with all biomarkers predicted 34.3%, 48.6% and 59.1% of all cases of a SGA neonate with birth weight < 10^th percentile delivered at ≥ 37, < 37 and < 32 weeks' gestation, at a 10% false-positive rate. The respective values for birth weight < 3^rd percentile were 39.9%, 53.2% and 64.4%, and for birth weight < 3^rd percentile with pre-eclampsia they were 46.3%, 66.8% and 80.4%. The new model was well calibrated.

CONCLUSIONS

This study has presented a single continuous two-dimensional model for prediction of SGA for any desired cut-offs of smallness and GA at delivery, laying the ground for a personalized antenatal plan for predicting and managing SGA, in the milieu of a new inverted pyramid of prenatal care. © 2020 International Society of Ultrasound in Obstetrics and Gynecology.

Bayes' theorem, COVID19, and screening tests

The COVID19 crisis has provided a portal to revisit and understand qualities of screening tests and the importance of Bayes' theorem in understanding how to interpret results and implications of next actions.

Ophthalmic artery Doppler in prediction of pre-eclampsia at 35-37 weeks' gestation

OBJECTIVES

First, to examine the potential value of maternal ophthalmic artery Doppler at 35-37 weeks' gestation in the prediction of subsequent development of pre-eclampsia (PE), and, second, to examine the variability between repeat measurements in the same eye and variability in measurements between the two eyes.

METHODS

This was a prospective observational study in women attending for a routine hospital visit at 35 + 0 to 36 + 6 weeks' gestation. The visit included recording of maternal demographic characteristics and medical history and assessment of flow velocity waveforms from the maternal ophthalmic artery. Waveforms were obtained in sequence from the right eye, left eye and again from the right and then left eye. We recorded the average of the four measurements, two from each eye, for the following four indices: first peak of systolic velocity; second peak of systolic velocity; pulsatility index; and the ratio of the second to first peak of systolic velocity (PSV ratio). The measurements of the four indices were standardized to remove the effects of maternal characteristics and elements from the medical history. The competing-risks model was used to determine the detection rate (DR) of delivery with PE at any time and at < 3 weeks after assessment, at a 10% false-positive rate (FPR), in screening by maternal factors alone and a combination of maternal factors and the adjusted value of each of the four ophthalmic artery indices.

RESULTS

The study population of 2287 pregnancies contained 60 (2.6%) that developed PE, including 19 (0.8%) that delivered with PE at < 3 weeks after assessment. The DR, at 10% FPR, of delivery with PE at any time after assessment by maternal factors was 25.0% (95% CI, 14.7-37.9%), and this increased by 25 percentage points to 50.0% (95% CI, 36.8-63.2%) with the addition of the adjusted PSV ratio (P = 0.005); the respective values for delivery with PE at < 3 weeks after assessment were 31.6% (95% CI, 12.6-56.6%) and 57.9% (95% CI, 33.5-79.8%). The other ophthalmic artery indices did not improve the prediction provided by maternal factors alone. There was good correlation between the first and second measurements of PSV ratio from the same eye (right eye r = 0.823, left eye r = 0.840), but poorer correlation in the first and second measurements between the two eyes (first measurement r = 0.690, second measurement r = 0.682). In screening by maternal factors and PSV ratio for PE with delivery at any stage after assessment, the estimated DR, at 10% FPR, was 50.0% when the average of four measurements was used (two from each eye), 49.1% when the average of one measurement from each eye was used, 47.3% when the average of two measurements from the same eye was used, and 45.8% when only one measurement was used.

CONCLUSIONS

Ophthalmic artery PSV ratio at 35-37 weeks' gestation can predict subsequent delivery with PE, especially if this occurs within 3 weeks after assessment. In the assessment of ophthalmic artery Doppler, it is necessary to use the average of one measurement from each eye to minimize variability of measurements. © 2020 International Society of Ultrasound in Obstetrics and Gynecology.

Competing-risks model for prediction of small-for-gestational-age neonate from maternal characteristics and serum pregnancy-associated plasma protein-A at 11-13 weeks' gestation

OBJECTIVES

To develop a continuous likelihood model for pregnancy-associated plasma protein-A (PAPP-A), in the context of a new competing-risks model for prediction of a small-for-gestational-age (SGA) neonate, and to compare the predictive performance of the new model for SGA to that of previous methods.

METHODS

This was a prospective observational study of 60 875 women with singleton pregnancy undergoing routine ultrasound examination at 11 + 0 to 13 + 6 weeks' gestation. The dataset was divided randomly into a training dataset and a test dataset. The training dataset was used for PAPP-A likelihood model development. We used Bayes' theorem to combine the previously developed prior model for the joint Gaussian distribution of gestational age (GA) at delivery and birth-weight Z-score with the PAPP-A likelihood to obtain a posterior distribution. This patient-specific posterior joint Gaussian distribution of GA at delivery and birth-weight Z-score allows risk calculation for SGA defined in terms of different birth-weight percentiles and GA. The new model was validated internally in the test dataset and we compared its predictive performance to that of the risk-scoring system of the UK National Institute for Health and Care Excellence (NICE) and that of logistic regression models for different SGA definitions.

RESULTS

PAPP-A has a continuous association with both birth-weight Z-score and GA at delivery according to a folded-plane regression. The new model, with the addition of PAPP-A, was equal or superior to several logistic regression models. The new model performed well in terms of risk calibration and consistency across different GAs and birth-weight percentiles. In the test dataset, at a false-positive rate of about 30% using the criteria defined by NICE, the new model predicted 62.7%, 66.5%, 68.1% and 75.3% of cases of a SGA neonate with birth weight < 10^th percentile delivered at < 42, < 37, < 34 and < 30 weeks' gestation, respectively, which were significantly higher than the respective values of 46.7%, 55.0%, 55.9% and 52.8% achieved by application of the NICE guidelines.

CONCLUSIONS

Using Bayes' theorem to combine PAPP-A measurement data with maternal characteristics improves the prediction of SGA and performs better than logistic regression or NICE guidelines, in the context of a new competing-risks model for the joint distribution of birth-weight Z-score and GA at delivery. © 2020 International Society of Ultrasound in Obstetrics and Gynecology.

Competing-risks model for prediction of small-for-gestational-age neonate from maternal characteristics and medical history

BACKGROUND

The established method of identifying a group of women at high risk of delivering a small-for-gestational-age (SGA) neonate, requiring increased surveillance, is use of risk scoring systems based on maternal demographic characteristics and medical history. Although this approach is relatively simple to perform, it does not provide patient-specific risks and has an uncertain performance in predicting SGA. Another approach to predict delivery of a SGA neonate is to use logistic regression models that combine maternal factors with first-trimester biomarkers. These models provide patient-specific risks for different prespecified cut-offs of birth-weight percentile and gestational age (GA) at delivery.

OBJECTIVES

First, to develop a competing-risks model for prediction of SGA based on maternal demographic characteristics and medical history, in which GA at the time of delivery and birth-weight Z-score are treated as continuous variables. Second, to compare the predictive performance of the new model for SGA neonates to that of previous methods.

METHODS

This was a prospective observational study in 124 443 women with singleton pregnancy undergoing routine ultrasound examination at 11 + 0 to 13 + 6 weeks' gestation. The dataset was divided randomly into a training and a test dataset. The training dataset was used to develop a model for the joint distribution of GA at delivery and birth-weight Z-score from variables of maternal characteristics and medical history. This patient-specific joint Gaussian distribution of GA at delivery and birth-weight Z-score allows risk calculation for SGA defined in terms of different birth-weight percentiles and GA. The new model was then validated in the test dataset to assess performance of screening and we compared its predictive performance to that of logistic regression models for different SGA definitions.

RESULTS

In the new model, the joint Gaussian distribution of GA at delivery and birth-weight Z-score is shifted to lower GA at delivery and birth-weight Z-score values, resulting in an increased risk for SGA, by lower maternal weight and height, black, East Asian, South Asian and mixed racial origin, medical history of chronic hypertension, diabetes mellitus and systemic lupus erythematosus and/or antiphospholipid syndrome, conception by in-vitro fertilization and smoking. In parous women, variables from the last pregnancy that increased the risk for SGA were history of pre-eclampsia or stillbirth, decreasing birth-weight Z-score and decreasing GA at delivery of the last pregnancy and interpregnancy interval < 0.5 years. In the test dataset, at a false-positive rate of 10%, the new model predicted 30.1%, 32.1%, 32.2% and 37.8% of cases of a SGA neonate with birth weight < 10^th percentile delivered at < 42, < 37, < 34 and < 30 weeks' gestation, respectively, which were similar or higher than the respective values achieved by a series of logistic regression models. The calibration study demonstrated good agreement between the predicted risks and the observed incidence of SGA in both the training and test datasets.

CONCLUSIONS

A new competing-risks model, based on maternal characteristics and medical history, provides estimation of patient-specific risks for SGA in which GA at delivery and birth-weight Z-score are treated as continuous variables. Such estimation of the a-priori risk for SGA is an essential first step in the use of Bayes' theorem to combine maternal factors with biomarkers for the continuing development of more effective methods of screening for SGA. Copyright © 2020 ISUOG. Published by John Wiley & Sons Ltd.

Alcohol and Parasomnias: The Statistical Evaluation of the Parasomnia Defense in Sexual Assault, Where Alcohol is Involved

Sleep sex may be a defense for alleged sexual assault. The International Classification of Sleep Disorders (ICSD3) states: "Disorders of arousal should not be diagnosed in the presence of alcohol intoxication… The former [alcohol blackouts] are exponentially more prevalent." A panel member of ICSD3, quoting ICSD3 asserts: "alcohol intoxication should rule out a sleep-walking defense". This implies extremely strong support for a prosecution hypothesis (H_p ) over a defense hypothesis (H_d ). I use Bayesian methodology to evaluate the evidential probity of alcohol intoxication. The likelihood ratio, LR, measures the amplification of prior odds of guilt, LR = Posterior odds of guilt after considering alcohol intoxication /Prior odds of guilt before considering alcohol intoxication . By Bayes' theorem, LR = p ( alcohol intoxication, given H p ) / p ( alcohol intoxication, given H d ) . I use data from cross-sectional studies of sexual assault and prevalence of alcohol use, in college students, with data from longitudinal studies, and data from the epidemiology of parasomnias to evaluate LR (alcohol). LR ~1.5 or 5, depending whether alcohol does, or does not, increase the risk of parasomnias. The proposition of extremely strong support for H_p implies a LR ~1,000,000, so the proposition in ICSD3 is not supported by formal analysis. The statistical reasoning in ICSD3 is unclear. There appears to be inversion of the Bayesian conditional (confusing intoxication given assault, and assault given intoxication) and failure to evaluate alcohol intoxication in H_d . Similar statistical errors in R. v Sally Clark are discussed. The American Academy of Sleep Medicine should review the statistical methodology in ICSD3.

How to understand and teach P values: a diagnostic test framework

OBJECTIVES

The aim of the tutorial is to help educators address misconceptions about P values and provide a tool that can be used to teach a more contemporary interpretation.

STUDY DESIGN AND SETTING

A scripted tutorial using problem-based learning and a diagnostic test analogy to deconstruct the misunderstandings about P values and develop a more Bayesian approach to study interpretation.

RESULTS

A diagnostic test analogy is an effective teaching tool. Learners' understanding of Bayes' theorem in diagnostic testing can be used as a bridge to the realization that the prestudy probability of a true difference is crucial for study interpretation. The analogy has several caveats and shortcomings. The limitations of this analogy and the conceptual difficulties with the Bayesian study analyses are addressed.

CONCLUSION

P values do not provide the information many assume they do-they are not equivalent to a probability of a chance finding. This tutorial helps move learners from these incorrect notions to new insights.

Evaluation of immunotherapy and targeted therapy treatment on renal cell carcinoma: A Bayesian network analysis

Clinical trials have previously assessed various therapies for renal cell carcinoma (RCC); however, there is currently a lack of direct comparisons between these therapies. The present study identified published studies on RCC through Web of Science, PubMed, EMBASE, Cochrane Library of Controlled Trials and Clinical trials.gov that were written in the English language and published by February 2019. The data were selected and extracted independently by two reviewers. Standard pair-wise meta-analyses were performed using Stata. Network meta-analyses were subsequently performed using WinBUGS (version 1.4.3). The primary outcome of the present study was progression-free survival (PFS). Secondary outcomes included overall survival (OS), objective response rate (ORR) and adverse events of various targeted therapies. The results were presented as cumulative odds ratio, hazard ratio, corresponding 95% confidence interval and the surface under the cumulative ranking curve, which was used to rank the probabilities and outcome of each treatment in RCC. A total of 31 eligible publications for 18 randomized controlled trials consisting of 11,498 participants were included in the present study. The network meta-analyses revealed that a combination of lenvantinib and everolimus ranked first out of 16 treatments in terms of PFS, OS and ORR (probability of 54.0, 53.4 and 61.0%, respectively).

Flawed statistics and science confirming existing paradigms

BACKGROUND

Part of the scientific community states that implausible methods cannot have a true effect and that epidemiological proof can only lead to false positives.

DISCUSSION

Homeopathy is regarded as an example of an implausible method with false positive evidence. However, epidemiological proof is necessary to falsify the placebo hypothesis. Implausibility is now supposed to rectify selection of a part of all trials, but the applied selection criteria are diverse and not common in conventional medicine. Applying Bayes' theorem only once to demonstrate that a low prior chance does not lead to reasonable probability is flawed application of this theorem.

CONCLUSION

Demanding scientific evidence and then rejecting the same with post-hoc selection of trials and flawed statistics shows unwillingness to falsify the completeness of existing paradigms.

Screening for pre-eclampsia at 35-37 weeks' gestation

OBJECTIVE

To examine the performance of screening for pre-eclampsia (PE) at 35-37 weeks' gestation by maternal factors and combinations of mean arterial pressure (MAP), uterine artery pulsatility index (UtA-PI), serum placental growth factor (PlGF) and serum soluble fms-like tyrosine kinase-1 (sFlt-1).

METHODS

This was a prospective observational study in women with singleton pregnancy attending for an ultrasound scan at 35 + 0 to 36 + 6 weeks as part of routine pregnancy care. Bayes' theorem was used to combine the prior distribution of gestational age at delivery with PE, obtained from maternal characteristics and medical history, with various combinations of biomarker multiples of the median (MoM) values to derive the patient-specific risks of delivery with PE. The performance of such screening was estimated.

RESULTS

The study population of 13 350 pregnancies included 272 (2.0%) that subsequently developed PE. In pregnancies that developed PE, the MoM values of MAP, UtA-PI and sFlt-1 were increased and PlGF MoM was decreased. At a risk cut-off of 1 in 20, the proportion of the population stratified into high risk was about 10% of the total, and the proportion of cases of PE contained within this high-risk group was 28% with screening by maternal factors alone; the detection rate increased to 53% with the addition of MAP, 67% with the addition of MAP and PlGF and 70% with the addition of MAP, PlGF and sFlt-1. The performance of screening was not improved by the addition of UtA-PI. The performance of screening depended on the racial origin of the women; in screening by a combination of maternal factors, MAP, PlGF and sFlt-1 and use of the risk cut-off of 1 in 20, the detection rate and screen-positive rate were 66% and 9.5%, respectively, for Caucasian women and 88% and 18% for those of Afro-Caribbean racial origin.

CONCLUSION

Screening by maternal factors and biomarkers at 35-37 weeks' gestation can identify a high proportion of pregnancies that develop late PE. The performance of screening depends on the racial origin of the women. Copyright © 2018 ISUOG. Published by John Wiley & Sons Ltd.

Screening for pre-eclampsia by maternal factors and biomarkers at 11-13 weeks' gestation

OBJECTIVE

To examine the performance of screening for early, preterm and term pre-eclampsia (PE) at 11-13 weeks' gestation by maternal factors and combinations of mean arterial pressure (MAP), uterine artery (UtA) pulsatility index (PI), serum placental growth factor (PlGF) and serum pregnancy-associated plasma protein-A (PAPP-A).

METHODS

The data for this study were derived from three previously reported prospective non-intervention screening studies at 11 + 0 to 13 + 6 weeks' gestation in a combined total of 61 174 singleton pregnancies, including 1770 (2.9%) that developed PE. Bayes' theorem was used to combine the prior distribution of gestational age at delivery with PE, obtained from maternal characteristics, with various combinations of biomarker multiples of the median (MoM) values to derive patient-specific risks of delivery with PE at < 37 weeks' gestation. The performance of such screening was estimated.

RESULTS

In pregnancies that developed PE, compared to those without PE, the MoM values of UtA-PI and MAP were increased and those of PAPP-A and PlGF were decreased, and the deviation from normal was greater for early than late PE for all four biomarkers. Combined screening by maternal factors, UtA-PI, MAP and PlGF predicted 90% of early PE, 75% of preterm PE and 41% of term PE, at a screen-positive rate of 10%; inclusion of PAPP-A did not improve the performance of screening. The performance of screening depended on the racial origin of the women; on screening by a combination of maternal factors, MAP, UtA-PI and PlGF and using a risk cut-off of 1 in 100 for PE at < 37 weeks in Caucasian women, the screen-positive rate was 10% and detection rates for early, preterm and term PE were 88%, 69% and 40%, respectively. With the same method of screening and risk cut-off in women of Afro-Caribbean racial origin, the screen-positive rate was 34% and detection rates for early, preterm and term PE were 100%, 92% and 75%, respectively.

CONCLUSION

Screening by maternal factors and biomarkers at 11-13 weeks' gestation can identify a high proportion of pregnancies that develop early and preterm PE. © 2018 Crown copyright. Ultrasound in Obstetrics & Gynecology © 2018 ISUOG.

Application of Bayes' Theorem in Valuating Depression Tests Performance

The validity of clinical diagnoses is a fundamental topic in clinical psychology, because now there are some political administrations, as the IOM or the U.K. government, which are focusing on best evidence-based practice in clinical psychology. The most problematic issue in clinical psychology is to avoid wrong diagnoses which can have negative consequences on individual life and on the utility of clinical treatments. In the case of diagnoses based on self-report tests, the diagnostic decision about individual health is based on the comparison between its score and the cutoff, according to the frequentist approach to probability. However, the frequentist approach underestimates the possible risks of incorrect diagnoses based on cutoffs only. The Bayesian approach is a valid alternative to make diagnoses on the basis of the scores from psychological tests. The Bayes' theorem estimates the posterior probability of the presence of a pathology on the basis of the knowledge about the diffusion of this pathology (prior probability) and of the knowledge of sensitivity and specificity values of the test. With all this information, it is possible to estimate the diagnostic accuracy of some self-report tests used for assessing depression. We analyzed the diagnostic accuracy of the most used psychological tests of depression (Zung's Self-Rating Depression Scale, Hamilton Rating Scale for Depression, Center for Epidemiological Studies for Depression and the Beck Depression Inventory), together with a new scale (Teate Depression Inventory) developed with the IRT procedure, by analyzing the published works in which data about sensitivity and specificity of these scales are reported. Except the TDI, none of these scales can reach a satisfactory level of diagnostic accuracy, probably for the absence of an optimal procedure to select test items and subjects with clearly defined pathological symptoms which could allow the reduction of false positives in test scoring.

ASPRE trial: incidence of preterm pre-eclampsia in patients fulfilling ACOG and NICE criteria according to risk by FMF algorithm

OBJECTIVE

To report the incidence of preterm pre-eclampsia (PE) in women who are screen positive according to the criteria of the National Institute for Health and Care Excellence (NICE) and the American College of Obstetricians and Gynecologists (ACOG), and compare the incidence with that in those who are screen positive or screen negative by The Fetal Medicine Foundation (FMF) algorithm.

METHODS

This was a secondary analysis of data from the ASPRE study. The study population consisted of women with singleton pregnancy who underwent prospective screening for preterm PE by means of the FMF algorithm, which combines maternal factors and biomarkers at 11-13 weeks' gestation. The incidence of preterm PE in women fulfilling the NICE and ACOG criteria was estimated; in these patients the incidence of preterm PE was then calculated in those who were screen negative relative to those who were screen positive by the FMF algorithm.

RESULTS

A total of 34 573 women with singleton pregnancy delivering at ≥ 24 weeks' gestation underwent prospective screening for preterm PE, of which 239 (0.7%) cases developed preterm PE. At least one of the ACOG criteria was fulfilled in 22 287 (64.5%) pregnancies and the incidence of preterm PE was 0.97% (95% CI, 0.85-1.11%); in the subgroup that was screen positive by the FMF algorithm the incidence of preterm PE was 4.80% (95% CI, 4.14-5.55%), and in those that were screen negative it was 0.25% (95% CI, 0.18-0.33%), with a relative incidence in FMF screen negative to FMF screen positive of 0.051 (95% CI, 0.037-0.071). In 1392 (4.0%) pregnancies, at least one of the NICE high-risk criteria was fulfilled, and in this group the incidence of preterm PE was 5.17% (95% CI, 4.13-6.46%); in the subgroups of screen positive and screen negative by the FMF algorithm, the incidence of preterm PE was 8.71% (95% CI, 6.93-10.89%) and 0.65% (95% CI, 0.25-1.67%), respectively, and the relative incidence was 0.075 (95% CI, 0.028-0.205). In 2360 (6.8%) pregnancies fulfilling at least two of the NICE moderate-risk criteria, the incidence of preterm PE was 1.74% (95% CI, 1.28-2.35%); in the subgroups of screen positive and screen negative by the FMF algorithm the incidence was 4.91% (95% CI, 3.54-6.79%) and 0.42% (95% CI, 0.20-0.86%), respectively, and the relative incidence was 0.085 (95% CI, 0.038-0.192).

CONCLUSION

In women who are screen positive for preterm PE by the ACOG or NICE criteria but screen negative by the FMF algorithm, the risk of preterm PE is reduced to within or below background levels. The results provide further evidence to support the personalized risk-based screening method that combines maternal factors and biomarkers. Copyright © 2018 ISUOG. Published by John Wiley & Sons Ltd.

Hypothesis-Testing Demands Trustworthy Data-A Simulation Approach to Inferential Statistics Advocating the Research Program Strategy

In psychology as elsewhere, the main statistical inference strategy to establish empirical effects is null-hypothesis significance testing (NHST). The recent failure to replicate allegedly well-established NHST-results, however, implies that such results lack sufficient statistical power, and thus feature unacceptably high error-rates. Using data-simulation to estimate the error-rates of NHST-results, we advocate the research program strategy (RPS) as a superior methodology. RPS integrates Frequentist with Bayesian inference elements, and leads from a preliminary discovery against a (random) H₀-hypothesis to a statistical H₁-verification. Not only do RPS-results feature significantly lower error-rates than NHST-results, RPS also addresses key-deficits of a “pure” Frequentist and a standard Bayesian approach. In particular, RPS aggregates underpowered results safely. RPS therefore provides a tool to regain the trust the discipline had lost during the ongoing replicability-crisis.

Development and small-scale validation of a novel pigeon-associated mitochondrial DNA source tracking marker for the detection of fecal contamination in harvested rainwater

The current study was aimed at designing and validating (on a small-scale) a novel pigeon mitochondrial DNA (mtDNA) microbial source tracking (MST) marker for the detection of pigeon fecal matter in harvested rainwater. The pigeon mtDNA MST marker was designed to target the mtDNA Cytochrome b gene by employing mismatch amplification mutation assay kinetics. The pigeon marker was validated by screening 69 non-pigeon and 9 pigeon fecal samples. The host-sensitivity of the assay was determined as 1.00 while the host-specificity of the assay was 0.96. Harvested rainwater samples (n=60) were screened for the prevalence of the marker with the mtDNA Cytochrome b marker detected in 78% of the samples. Bayes' theorem was applied to calculate the conditional probability of the marker detecting true pigeon contamination and the marker subsequently displayed a 99% probability of detecting true pigeon contamination in the harvested rainwater samples. In addition, the mtDNA Cytochrome b marker displayed high concurrence frequencies versus heterotrophic bacteria (78.3%), E. coli (73.3%), total coliforms (71.1%) and fecal coliforms (66.7%). This study thus validates that targeting mtDNA for the design of source tracking markers may be a valuable tool to detect avian fecal contamination in environmental waters.

Competing-risks model in screening for pre-eclampsia in twin pregnancy by maternal characteristics and medical history

OBJECTIVE

A survival-time regression model for gestational age at delivery with pre-eclampsia (PE) in singleton pregnancy, using maternal demographic characteristics and medical history, was reported previously. The objective of this study was to extend this model to dichorionic (DC) and monochorionic (MC) twin pregnancy.

METHODS

The study population included 1789 DC and 430 MC twin pregnancies and 93 297 singleton pregnancies. A survival-time model for gestational age at delivery with PE was developed from variables of maternal characteristics and medical history. The risk of PE with delivery < 37 weeks and < 42 weeks in twin pregnancies was determined and compared with that in singleton pregnancies.

RESULTS

In singleton pregnancies comprising women of Caucasian racial origin, mean weight of 69 kg at 12 weeks' gestation, mean height of 164 cm, nulliparous, with spontaneous conception, no family history of PE and no history of diabetes mellitus, systemic lupus erythematosus or antiphospholipid syndrome, the mean of the Gaussian distribution of gestational age at delivery with PE was 55 weeks. In DC twins with PE, mean gestational age at delivery was shifted to the left by 8.2 (95% CI, 7.2-9.1) weeks and in MC twins it was shifted to the left by 10.0 (95% CI, 8.5-11.4) weeks. The risk of delivery with PE occurring at, or before, a specified gestational age is given by the area under the fitted distribution curve. For a reference population with the above characteristics, the estimated risk of PE < 37 weeks' gestation, assuming no other cause of delivery, was 0.6% for singletons, 9.0% for DC twins and 14.2% for MC twins; the respective values for PE < 42 weeks were 3.6%, 27.0% and 36.5%.

CONCLUSIONS

A model based on maternal characteristics and medical history has been developed for estimation of patient-specific risks for PE in DC and MC twin pregnancy. Such estimation of the a-priori risk for PE is an essential first step in the use of Bayes' theorem to combine maternal factors with biomarkers for the continuing development of more effective methods of screening for the disease. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.

A likelihood-based approach to P-value interpretation provided a novel, plausible, and clinically useful research study metric

OBJECTIVES

Interpretation of clinical research findings using the paradigm of null hypothesis significance testing has a number of limitations. These include arbitrary dichotomization of results, lack of incorporation of study power and prior probability, and the confusing use of conditional probability. This study aimed to describe a novel method of P-value interpretation that would address these limitations.

STUDY DESIGN AND SETTING

Published clinical research was reinterpreted using the delta likelihood ratio. The delta likelihood ratio is an application of Bayes' rule incorporating the P-value and study power. Calculation of the delta likelihood ratio allows the determination of the most likely effect size using the maximum likelihood principle.

RESULTS

We showed that the delta likelihood is easily calculated and produces plausible results using the example of several previously published research studies. Empirical evidence of validity was demonstrated by simulation.

CONCLUSION

The delta likelihood ratio and most likely effect size are simple and intuitive metrics to summarize research findings. The delta likelihood ratio incorporates study power and provides a continuous measure of the probability that the research result is a true effect. The most likely effect size is an easily understood metric that should aid the interpretation of research.

Protocol for the prospective validation study: 'Screening programme for pre-eclampsia' (SPREE)

Pre-eclampsia (PE), which affects about 2% of pregnancies, is a major cause of maternal and perinatal morbidity and mortality. Early detection of PE can improve pregnancy outcome by providing timely intervention and closer monitoring. The current guideline from the UK National Institute for Health and Care Excellence (NICE) recommends that, at the booking visit, women identified with one major risk factor or more than one moderate risk factor for PE should be advised to take low-dose aspirin daily from 12 weeks until delivery. However, performance of the current method of screening is poor and identifies only about 35% of PE. Extensive studies in the last decade have established that the best performance for early prediction of PE can be achieved by using a novel Bayes' theorem-based method that combines maternal characteristics and medical history together with measurements of mean arterial pressure (MAP), uterine artery pulsatility index (UtA-PI), serum placental growth factor (PlGF) and pregnancy-associated plasma protein-A (PAPP-A) at 11-13 weeks' gestation. This forms the 'combined test', which could be simplified to the 'mini combined test' when only maternal factors, MAP and PAPP-A are taken into consideration. We present the protocol (version 3.1, 14 November 2016) for the 'Screening programme for pre-eclampsia' (SPREE) study, a prospective multicenter cohort study that will be carried out in seven National Health Service maternity hospitals in England. Eligible pregnant women attending their routine scan at 11-13 weeks' gestation will be invited to participate in this study. Maternal characteristics and history and measurements of MAP, UtA-PI, serum PAPP-A and PlGF will be recorded according to standardized protocols. The patient-specific risk for PE will be calculated and data on pregnancy outcomes collected. We hypothesize that the first-trimester mini combined test and combined test for PE screening, using the Bayes' theorem-based method, are likely to be superior to the current method recommended by NICE that is based on maternal demographics and history alone. Enrollment for the study commenced in April 2016. The study is registered on the International Standard Randomised Controlled Trial Number (ISRCTN) registry. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.

Accuracy of competing-risks model in screening for pre-eclampsia by maternal factors and biomarkers at 11-13 weeks' gestation

OBJECTIVE

To examine the diagnostic accuracy of a previously developed model for prediction of pre-eclampsia (PE) by a combination of maternal factors and biomarkers at 11-13 weeks' gestation.

METHODS

This was a prospective first-trimester multicenter study of screening for PE in 8775 singleton pregnancies. A previously published algorithm was used for the calculation of patient-specific risk of PE in each individual. The detection rates (DRs) and false-positive rates (FPRs) for delivery with PE < 32, < 37 and ≥ 37 weeks were estimated and compared with those for the dataset used for development of the algorithm.

RESULTS

In the study population, 239 (2.7%) cases developed PE, of which 17 (0.2%), 59 (0.7%) and 180 (2.1%) developed PE < 32, < 37 and ≥ 37 weeks, respectively. With combined screening by maternal factors, mean arterial pressure, uterine artery pulsatility index and serum placental growth factor, the DR was 100% (95% CI, 80-100%) for PE < 32 weeks, 75% (95% CI, 62-85%) for PE < 37 weeks and 43% (95% CI, 35-50%) for PE ≥ 37 weeks, at a 10% FPR. These DRs were similar to the estimated rates for the dataset used for development of the model: 89% (95% CI, 79-96%) for PE < 32 weeks, 75% (95% CI, 70-80%) for PE < 37 weeks and 47% (95% CI, 44-51%) for PE ≥ 37 weeks.

CONCLUSION

Assessment of a combination of maternal factors and biomarkers at 11-13 weeks provides effective first-trimester screening for preterm PE. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.

Multicenter screening for pre-eclampsia by maternal factors and biomarkers at 11-13 weeks' gestation: comparison with NICE guidelines and ACOG recommendations

OBJECTIVE

To compare the performance of screening for pre-eclampsia (PE) based on risk factors from medical history, as recommended by NICE and ACOG, with the method proposed by The Fetal Medicine Foundation (FMF), which uses Bayes' theorem to combine the a-priori risk from maternal factors, derived by a multivariable logistic model, with the results of various combinations of biophysical and biochemical measurements.

METHODS

This was a prospective multicenter study of screening for PE in 8775 singleton pregnancies at 11-13 weeks' gestation. A previously published FMF algorithm was used for the calculation of patient-specific risk of PE in each individual. The detection rates (DRs) and false-positive rates (FPRs) for delivery with PE < 32, < 37 and ≥ 37 weeks were estimated and compared with those derived from application of NICE guidelines and ACOG recommendations. According to NICE, all high-risk pregnancies should be offered low-dose aspirin. According to ACOG, use of aspirin should be reserved for women with a history of PE in at least two previous pregnancies or PE requiring delivery < 34 weeks' gestation.

RESULTS

In the study population, 239 (2.7%) cases developed PE, of which 17 (0.2%), 59 (0.7%) and 180 (2.1%) developed PE < 32, < 37 and ≥ 37 weeks, respectively. Screening with use of the FMF algorithm based on a combination of maternal factors, mean arterial pressure (MAP), uterine artery pulsatility index (UtA-PI) and serum placental growth factor (PlGF) detected 100% (95% CI, 80-100%) of PE < 32 weeks, 75% (95% CI, 62-85%) of PE < 37 weeks and 43% (95% CI, 35-50%) of PE ≥ 37 weeks, at a 10.0% FPR. Screening with use of NICE guidelines detected 41% (95% CI, 18-67%) of PE < 32 weeks, 39% (95% CI, 27-53%) of PE < 37 weeks and 34% (95% CI, 27-41%) of PE ≥ 37 weeks, at 10.2% FPR. Screening with use of ACOG recommendations detected 94% (95% CI, 71-100%) of PE < 32 weeks, 90% (95% CI, 79-96%) of PE < 37 weeks and 89% (95% CI, 84-94%) of PE ≥ 37 weeks, at 64.2% FPR. Screening based on the ACOG recommendations for use of aspirin detected 6% (95% CI, 1-27%) of PE < 32 weeks, 5% (95% CI, 2-14%) of PE < 37 weeks and 2% (95% CI, 0.3-5%) of PE ≥ 37 weeks, at 0.2% FPR.

CONCLUSION

Performance of screening for PE at 11-13 weeks' gestation by the FMF algorithm using a combination of maternal factors, MAP, UtA-PI and PlGF, is by far superior to the methods recommended by NICE and ACOG. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.

Competing-risks model in screening for pre-eclampsia by maternal factors and biomarkers at 35-37 weeks' gestation

OBJECTIVE

To develop a model for prediction of term pre-eclampsia (PE) based on a combination of maternal factors and late third-trimester biomarkers.

METHODS

Data were derived from prospective screening for adverse obstetric outcomes in women attending their routine hospital visit at 35-37 weeks' gestation in two maternity hospitals in the UK. Uterine artery pulsatility index (UtA-PI) was measured in 5362 pregnancies, mean arterial pressure (MAP) in 5386 and serum placental growth factor (PlGF) and serum soluble fms-like tyrosine kinase-1 (sFlt-1) in 3920. Bayes' theorem was used to combine the a-priori risk of PE from maternal factors with various combinations of biomarkers, expressed as multiples of the median (MoM). Five-fold cross-validation was used to estimate the performance of screening for PE, requiring delivery at some stage after assessment. The empirical performance of screening was compared to model predictions.

RESULTS

In pregnancies that developed PE, the values of MAP, UtA-PI and sFlt-1 were increased and PlGF was decreased compared to unaffected pregnancies. For all biomarkers evaluated, the deviation from normal was inversely related to the gestational age at which delivery became necessary for maternal or fetal indications. Screening by maternal factors and by a combination of maternal factors with all biomarkers predicted 35% and 84% of PE, respectively, at a 10% false-positive rate.

CONCLUSION

A combination of maternal factors and biomarkers at 35-37 weeks' gestation can provide effective screening for term PE. Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

Maternal serum PAPP-A and free β-hCG at 12, 22 and 32 weeks' gestation in screening for pre-eclampsia

OBJECTIVE

To examine the distribution of maternal serum pregnancy-associated plasma protein-A (PAPP-A) and free β-human chorionic gonadotropin (β-hCG) at 12, 22 and 32 weeks' gestation in singleton pregnancies which develop pre-eclampsia (PE) and examine the performance of these biomarkers in screening for PE.

METHODS

Serum PAPP-A and free β-hCG were measured in 94 989 cases at 11-13 weeks, 7597 at 19-24 weeks and 8088 at 30-34 weeks' gestation. Bayes' theorem was used to combine the a-priori risk from maternal characteristics and medical history with PAPP-A and free β-hCG. The empirical and model-based performance of screening for preterm PE requiring delivery < 37 weeks' gestation and term PE with delivery ≥ 37 weeks was estimated.

RESULTS

Combined screening with maternal factors and serum PAPP-A at 11-13 and 30-34 weeks and with maternal factors and serum free β-hCG at 19-24 and 30-34 weeks improved the prediction provided by maternal factors alone for preterm PE. The detection rate, at a 10% false-positive rate, for preterm PE by screening with maternal factors was about 45% which improved to 51% and 53% by combined screening with PAPP-A at 11-13 weeks and 30-34 weeks, respectively, and 55% and 54% by combined screening with free β-hCG at 19-24 weeks and 30-34 weeks, respectively. Measurement of serum PAPP-A and free β-hCG was not useful in the prediction of term PE.

CONCLUSIONS

Measurement of serum PAPP-A and free β-hCG could improve the prediction of preterm PE provided by maternal characteristics and medical history alone. Copyright © 2016 ISUOG. Published by John Wiley & Sons Ltd.

Contingent screening for preterm pre-eclampsia

OBJECTIVE

Effective screening for pre-eclampsia resulting in delivery < 37 weeks' gestation (preterm PE) is provided by assessment of a combination of maternal factors, mean arterial pressure (MAP), uterine artery pulsatility index (UtA-PI) and serum placental growth factor (PlGF) at 11-13 or 19-24 weeks' gestation. This study explores the possibility of carrying out routine screening for preterm PE by maternal factors and MAP in all pregnancies and reserving measurements of UtA-PI and PlGF for a subgroup of the population, selected on the basis of the risk derived from screening by maternal factors and MAP alone.

METHODS

Study data were derived from prospective screening for adverse obstetric outcomes in women attending their routine hospital visit at 11-13 and/or 19-24 weeks' gestation. Bayes' theorem was used to derive the a-priori risk for preterm PE from maternal factors and MAP. The posterior risk was obtained by the addition of UtA-PI and PlGF. We estimated the detection rate (DR) of preterm PE, at an overall false-positive rate (FPR) of 10%, from a policy in which first-stage screening by a combination of maternal factors and MAP defines screen-positive, screen-negative and intermediate-risk groups, with the latter undergoing second-stage screening by UtA-PI and PlGF.

RESULTS

At 11-13 weeks' gestation, the model-based DR of preterm PE, at a 10% FPR, when screening the whole population by maternal factors, MAP, UtA-PI and PlGF was 74%. A similar DR was achieved by two-stage screening, with screening by maternal factors and MAP in the first stage and reserving measurement of UtA-PI and PlGF for the second stage and for only 50% of the population. If second-stage screening was offered to 30% of the population, there would be only a small reduction in DR from 74% to 71%. At 19-24 weeks, the model-based DR of preterm PE, at a 10% FPR, when screening the whole population by maternal factors, MAP, UtA-PI and PlGF was 84%. A similar DR was achieved by two-stage screening with measurements of UtA-PI and PlGF in only 70% of the population; if second-stage screening was offered to 40% of the population, the DR would be reduced from 84% to 81%.

CONCLUSIONS

High DR of preterm PE can be achieved by two-stage screening in the first and second trimesters with maternal factors and MAP in the whole population and measurements of UtA-PI and PlGF in only some of the pregnancies. Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

Mean arterial pressure at 12, 22, 32 and 36 weeks' gestation in screening for pre-eclampsia

OBJECTIVE

To examine the distribution of mean arterial pressure (MAP) at 12, 22, 32 and 36 weeks' gestation in singleton pregnancies which develop pre-eclampsia (PE) and examine the performance of this biomarker in screening for PE.

METHODS

MAP was measured in 77 343 cases at 11-13 weeks, in 31 120 cases at 19-24 weeks, in 29 802 at 30-34 weeks and 5543 at 35-37 weeks. Bayes' theorem was used to combine the a-priori risk from maternal characteristics and medical history with MAP. The performance of screening for PE requiring delivery < 32, at 32 + 0 to 36 + 6 and ≥ 37 weeks' gestation was estimated.

RESULTS

In pregnancies that developed PE, MAP was increased and the separation in multiples of the median (MoM) values from normal was greater with an earlier, compared to later, gestational age at which delivery for PE became necessary. Additionally, the slope of the regression lines of MAP MoM with gestational age at delivery in pregnancies that developed PE increased with advancing gestational age at screening. The detection rate (DR), at a false-positive rate of 10%, for PE delivering < 32 weeks was 66% and 72% with screening at 12 and 22 weeks, respectively. The DR for PE delivering at 32 + 0 to 36 + 6 weeks was 54%, 56% and 81% with screening at 12, 22 and 32 weeks. The DR for PE delivering ≥ 37 weeks was 45%, 43%, 49% and 59% with screening at 12, 22, 32 and 36 weeks, respectively.

CONCLUSIONS

The performance of combined screening with maternal factors and MAP is superior in screening for early, compared to late, PE and, to a certain extent, improves with advancing gestational age at screening. Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

Maternal serum placental growth factor at 12, 22, 32 and 36 weeks' gestation in screening for pre-eclampsia

OBJECTIVE

To examine the distribution of maternal serum placental growth factor (PlGF) at 12, 22, 32 and 36 weeks' gestation in singleton pregnancies which develop pre-eclampsia (PE) and examine the performance of this biomarker in screening for PE.

METHODS

Serum PlGF was measured in 40 212 cases at 11-13 weeks, in 10 282 cases at 19-24 weeks, in 10 400 at 30-34 weeks and 4043 at 35-37 weeks. Bayes' theorem was used to combine the a-priori risk from maternal characteristics and medical history with serum PlGF. The performance of screening for PE requiring delivery < 32, at 32 + 0 to 36 + 6 and ≥ 37 weeks' gestation was estimated.

RESULTS

In pregnancies that developed PE, serum PlGF was decreased and the separation in multiples of the median (MoM) values from normal was greater with earlier, compared to later, gestational age at which delivery for PE became necessary. Additionally, the slope of the regression lines of PlGF MoM with gestational age at delivery in pregnancies that developed PE increased with advancing gestational age at screening. The detection rates (DRs), at a false-positive rate (FPR) of 10%, for PE delivering < 32 weeks were 79% and 97% with screening at 12 and 22 weeks, respectively. The DRs for PE delivering at 32 + 0 to 36 + 6 weeks were 57%, 65% and 90% with screening at 12, 22 and 32 weeks. The DRs for PE delivering ≥ 37 weeks were 40%, 37%, 54% and 64% with screening at 12, 22, 32 and 36 weeks, respectively.

CONCLUSIONS

The performance of combined screening with maternal factors, medical history and PlGF is superior in screening for early, compared to late, PE and improves with advancing gestational age at screening. Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

Maternal serum soluble fms-like tyrosine kinase-1 at 12, 22, 32 and 36 weeks' gestation in screening for pre-eclampsia

OBJECTIVE

To examine the distribution of maternal serum soluble fms-like tyrosine kinase-1 (sFlt-1) at 12, 22, 32 and 36 weeks' gestation in singleton pregnancies that develop pre-eclampsia (PE) and examine the performance of this biomarker in screening for PE.

METHODS

Serum sFlt-1 was measured in 7066 cases at 11-13 weeks, 8079 cases at 19-24 weeks, 8472 at 30-34 weeks and 4043 at 35-37 weeks. Bayes' theorem was used to combine the a-priori risk from maternal characteristics and medical history with serum levels of sFlt-1. The performance of screening for PE in women requiring delivery < 32, between 32 + 0 and 36 + 6 and ≥ 37 weeks' gestation was estimated.

RESULTS

In pregnancies that developed PE, serum sFlt-1 was increased and the separation in multiples of the median (MoM) values from normal was greater with earlier, compared to later, gestational age at which delivery for PE became necessary. In pregnancies that developed PE, the slope of the regression lines of sFlt-1 MoM with gestational age at delivery increased with advancing gestational age at screening. Measurement of sFlt-1 at 11-13 weeks did not improve the prediction of PE achieved by maternal factors alone, sFlt-1 at 19-24 weeks improved the prediction of PE delivering < 37 weeks but not for PE delivering ≥ 37 weeks, sFlt-1 at 30-34 weeks improved the prediction of PE delivering < 37 and PE delivering ≥ 37 weeks and sFlt-1 at 35-37 weeks improved the prediction of PE delivering ≥ 37 weeks. The detection rates (DRs), at a false-positive rate (FPR) of 10%, of PE delivering < 32 weeks were 52% and 65% with screening at 12 and 22 weeks, respectively. The DRs for PE delivering between 32 + 0 and 36 + 6 weeks were 44%, 44% and 93% with screening at 12, 22 and 32 weeks. The DR for PE delivering ≥ 37 weeks were 37%, 37%, 52% and 69% with screening at 12, 22, 32 and 36 weeks, respectively.

CONCLUSIONS

The performance of combined screening with maternal factors, medical history and serum sFlt-1 is superior for detection of early, compared to late, PE and improves with advancing gestational age at screening. Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

Maternal serum alpha-fetoprotein at 12, 22 and 32 weeks' gestation in screening for pre-eclampsia

OBJECTIVE

To examine the distribution of maternal serum alpha-fetoprotein (AFP) at 12, 22 and 32 weeks' gestation in singleton pregnancies which develop pre-eclampsia (PE) and examine the performance of this biomarker in screening for PE.

METHODS

Serum AFP was measured in 17 071 cases at 11-13 weeks, in 8583 cases at 19-24 weeks and 8609 cases at 30-34 weeks' gestation. Bayes' theorem was used to combine the a-priori risk from maternal characteristics and medical history with AFP. The performance of screening for PE requiring delivery < 32, at 32 + 0 to 36 + 6, < 37 and ≥ 37 weeks' gestation was estimated.

RESULTS

In pregnancies that developed PE, serum AFP multiples of the median (MoM) was increased at 11-13 and 19-24 weeks' gestation, but not at 30-34 weeks, and the values were inversely related to gestational age at delivery. Combined screening with maternal factors and serum AFP improved the prediction provided by maternal factors alone for PE delivering < 37 weeks, but not for PE delivering ≥ 37 weeks. The performance of screening for preterm PE was better at 19-24 weeks than at 11-13 weeks and the detection rate (DR) for a given false-positive rate (FPR) was higher for PE delivering < 32 weeks than for PE delivering at 32 + 0 to 36 + 6 weeks. The DRs, at 10% FPR, of combined screening at 11-13 weeks for PE delivering < 32 and at 32 + 0 to 36 + 6 weeks were 54% and 45%, respectively, and these improved to 72% and 53% with screening at 19-24 weeks.

CONCLUSIONS

Measurement of serum AFP at 11-13 and 19-24 weeks' gestation improves the prediction of preterm PE provided by maternal factors alone. Copyright © 2015 ISUOG. Published by John Wiley & Sons Ltd.

Rational Irrationality: Modeling Climate Change Belief Polarization Using Bayesian Networks

Belief polarization is said to occur when two people respond to the same evidence by updating their beliefs in opposite directions. This response is considered to be "irrational" because it involves contrary updating, a form of belief updating that appears to violate normatively optimal responding, as for example dictated by Bayes' theorem. In light of much evidence that people are capable of normatively optimal behavior, belief polarization presents a puzzling exception. We show that Bayesian networks, or Bayes nets, can simulate rational belief updating. When fit to experimental data, Bayes nets can help identify the factors that contribute to polarization. We present a study into belief updating concerning the reality of climate change in response to information about the scientific consensus on anthropogenic global warming (AGW). The study used representative samples of Australian and U.S.

PARTICIPANTS

Among Australians, consensus information partially neutralized the influence of worldview, with free-market supporters showing a greater increase in acceptance of human-caused global warming relative to free-market opponents. In contrast, while consensus information overall had a positive effect on perceived consensus among U.S. participants, there was a reduction in perceived consensus and acceptance of human-caused global warming for strong supporters of unregulated free markets. Fitting a Bayes net model to the data indicated that under a Bayesian framework, free-market support is a significant driver of beliefs about climate change and trust in climate scientists. Further, active distrust of climate scientists among a small number of U.S. conservatives drives contrary updating in response to consensus information among this particular group.

A unified approach to risk assessment for fetal aneuploidies

OBJECTIVE

To examine the potential impact of combining measures from cell-free DNA (cfDNA) testing with maternal age and first-trimester biomarkers in screening for fetal trisomies.

METHODS

This was a theoretical study using Bayes' theorem to combine the a priori risk for fetal trisomy 21 derived from maternal age with likelihoods from nuchal translucency thickness, serum pregnancy-associated plasma protein-A, serum free β-human chorionic gonadotropin and plasma cfDNA. We adopted a binomial counting model for the cfDNA likelihoods and developed a model to account for errors in estimating fetal fraction.

RESULTS

When Bayes' theorem was used to combine the a priori risk for trisomy 21 derived from the first-trimester combined test with likelihoods from the cfDNA test, and when the true fetal fraction was known, the detection rate increased from 62% at a fetal fraction of 4% to 100% at a fetal fraction of ≥ 9%; the positive likelihood ratio (trisomic/euploid) increased from 620 to 1000 and the negative likelihood ratio (euploid/trisomic) increased from 3 to > 10 000. When the fetal fraction is < 4%, the cfDNA test has traditionally been considered to be a failure, but the cfDNA results can be used to improve the performance of screening by the combined test.

CONCLUSIONS

In contingent policies that use the first-trimester combined test for first-line screening to select the subgroup for cfDNA testing, the data from the latter should be used to update the risk from the former. Individual patient results from cfDNA testing depend crucially on the fetal fraction and the precision of its measurement.

Prospective evaluation of a rapid nanoparticle-based lateral flow immunoassay (STic Expert(®) HIT) for the diagnosis of heparin-induced thrombocytopenia

A rapid lateral flow immunoassay (LFIA) (STic Expert(®) HIT), recently developed for the diagnosis of heparin-induced thrombocytopenia (HIT), was evaluated in a prospective multicentre cohort of 334 consecutive patients. The risk of HIT was estimated by the 4Ts score as low, intermediate and high in 28·7%, 61·7% and 9·6% of patients, respectively. Definite HIT was diagnosed in 40 patients (12·0%) with positive results on both enzyme-linked immunosorbent assay (Asserachrom(®) HPIA IgG) and serotonin release assay. The inter-reader reproducibility of results obtained was excellent (kappa ratio > 0·9). The negative predictive value of LFIA with plasma samples was 99·6% with a negative likelihood ratio (LR) of 0·03, and was comparable to those of the particle gel immunoassay (H/PF4-PaGIA(®) ) performed in 124 cases. Positive predictive value and positive LR were 44·4% and 5·87, respectively, and the results were similar for serum samples. The probability of HIT in intermediate risk patients decreased from 11·2% to 0·4% when the LFIA result was negative and increased to 42·5% when it was positive. In conclusion, the STic Expert(®) HIT combined with the 4Ts score is a reliable tool to rule out the diagnosis of HIT.

Epidemiology visualized: the prosecutor's fallacy

The "prosecutor's fallacy" (the assumption that Pr [probability] (A|B) = Pr (B|A)) arises often in epidemiology but is often unrecognized as such, in part because investigators do not have strong intuitions about what the fallacy means. Here, we help inform such intuitions and remind investigators of this fallacy by using visualizations. In figures, we demonstrate the prosecutor's fallacy, as well as show conditions under which Pr (A|B) can be assumed to be equal to Pr (B|A). Visualizations can help build intuition around statistical concepts such as the prosecutor's fallacy and should be more widely considered as teaching tools.

How biological vision succeeds in the physical world

Biological visual systems cannot measure the properties that define the physical world. Nonetheless, visually guided behaviors of humans and other animals are routinely successful. The purpose of this article is to consider how this feat is accomplished. Most concepts of vision propose, explicitly or implicitly, that visual behavior depends on recovering the sources of stimulus features either directly or by a process of statistical inference. Here we argue that, given the inability of the visual system to access the properties of the world, these conceptual frameworks cannot account for the behavioral success of biological vision. The alternative we present is that the visual system links the frequency of occurrence of biologically determined stimuli to useful perceptual and behavioral responses without recovering real-world properties. The evidence for this interpretation of vision is that the frequency of occurrence of stimulus patterns predicts many basic aspects of what we actually see. This strategy provides a different way of conceiving the relationship between objective reality and subjective experience, and offers a way to understand the operating principles of visual circuitry without invoking feature detection, representation, or probabilistic inference.

Use of LC-MS/MS and Bayes' theorem to identify protein kinases that phosphorylate aquaporin-2 at Ser256

In the renal collecting duct, binding of AVP to the V2 receptor triggers signaling changes that regulate osmotic water transport. Short-term regulation of water transport is dependent on vasopressin-induced phosphorylation of aquaporin-2 (AQP2) at Ser256. The protein kinase that phosphorylates this site is not known. We use Bayes' theorem to rank all 521 rat protein kinases with regard to the likelihood of a role in Ser256 phosphorylation on the basis of prior data and new experimental data. First, prior probabilities were estimated from previous transcriptomic and proteomic profiling data, kinase substrate specificity data, and evidence for kinase regulation by vasopressin. This ranking was updated using new experimental data describing the effects of several small-molecule kinase inhibitors with known inhibitory spectra (H-89, KN-62, KN-93, and GSK-650394) on AQP2 phosphorylation at Ser256 in inner medullary collecting duct suspensions. The top-ranked kinase was Ca2+/calmodulin-dependent protein kinase II (CAMK2), followed by protein kinase A (PKA) and protein kinase B (AKT). Liquid chromatography-tandem mass spectrometry (LC-MS/MS)-based in vitro phosphorylation studies compared the ability of three highly ranked kinases to phosphorylate AQP2 and other inner medullary collecting duct proteins, PKA, CAMK2, and serum/glucocorticoid-regulated kinase (SGK). All three proved capable of phosphorylating AQP2 at Ser256, although CAMK2 and PKA were more potent than SGK. The in vitro phosphorylation experiments also identified candidate protein kinases for several additional phosphoproteins with likely roles in collecting duct regulation, including Nedd4-2, Map4k4, and 3-phosphoinositide-dependent protein kinase 1. We conclude that Bayes' theorem is an effective means of integrating data from multiple data sets in physiology.

Individual risk prediction model for incident cardiovascular disease: a Bayesian clinical reasoning approach

BACKGROUND

A Bayesian clinical reasoning model was developed to predict an individual risk for cardiovascular disease (CVD) for desk-top reference.

METHODS

Three Bayesian models were constructed to estimate the CVD risk by sequentially incorporating demographic features (basic), six metabolic syndrome components (metabolic score) and conventional risk factors (enhanced model). By considering clinical weights (regression coefficients) of each model as normal distribution, individual risk can be predicted making allowance for uncertainty of clinical weights. A community-based cohort that enrolled 64,489 participants free of CVD at baseline and followed up over five years to ascertain newly diagnosed CVD cases during the period through 2000 to 2004 was used for the illustration of the three proposed models (full empirical data are available from website http://homepage.ntu.edu.tw/~chenlin/CVD_prediction_data.rar).

RESULTS

The proposed models can be applied to predicting the CVD risk with any combination of risk factors. For a 47-year-old man, the five-year risk for CVD with the basic model was 11.2% (95% CI: 7.8%-15.6%). His metabolic syndrome score, leading to 1.488 of likelihood ratio, enhanced the risk for CVD up to 15.8% (95% CI: 11.0%-21.5%) and put him in highest deciles. As with the habit of smoking over 2 packs per-day and family history of CVD, yielding the likelihood ratios of 1.62 and 1.47, respectively, the risk was further raised to 30.9% (95% CI: 20.7%-39.8%).

CONCLUSIONS

We demonstrate how to make individual risk prediction for CVD by incorporating routine information with a sequential Bayesian clinical reasoning approach.