Identifying relevant asymmetry features of EEG for emotion processing.

The left and right hemispheres of the brain process emotion differently. Neuroscientists have proposed two models to explain this difference. The first model states that the right hemisphere is dominant over the left to process all emotions. In contrast, the second model states that the left hemisphere processes positive emotions, whereas the right hemisphere processes negative emotions. Previous studies have used these asymmetry models to enhance the classification of emotions in machine learning models. However, little research has been conducted to explore how machine learning models can help identify associations between hemisphere asymmetries and emotion processing. To address this gap, we conducted two experiments using a subject-independent approach to explore how the asymmetry of the brain hemispheres is involved in processing happiness, sadness, fear, and neutral emotions. We analyzed electroencephalogram (EEG) signals from 15 subjects collected while they watched video clips evoking these four emotions. We derived asymmetry features from the recorded EEG signals by calculating the log ratio between the relative energy of symmetrical left and right nodes. Using the asymmetry features, we trained four binary logistic regressions, one for each emotion, to identify which features were more relevant to the predictions. The average AUC-ROC across the 15 subjects was 56.2, 54.6, 51.6, and 58.4% for neutral, sad, fear, and happy, respectively. We validated these results with an independent dataset, achieving comparable AUC-ROC values. Our results showed that brain lateralization was observed primarily in the alpha frequency bands, whereas for the other frequency bands, both hemispheres were involved in emotion processing. Furthermore, the logistic regression analysis indicated that the gamma and alpha bands were the most relevant for predicting emotional states, particularly for the lateral frontal, parietal, and temporal EEG pairs, such as FT7-FT8, T7-T8, and TP7-TP8. These findings provide valuable insights into which brain areas and frequency bands need to be considered when developing predictive models for emotion recognition.

Identifying factors that shape whether digital food marketing appeals to children.

OBJECTIVE: Children are frequently exposed to unhealthy food marketing on digital media. This marketing contains features that often appeal to children, such as cartoons or bold colours. Additional factors can also shape whether marketing appeals to children. In this study, in order to assess the most important predictors of child appeal in digital food marketing, we used machine learning to examine how marketing techniques and children's socio-demographic characteristics, weight, height, BMI, frequency of screen use and dietary intake influence whether marketing instances appeal to children. DESIGN: We conducted a pilot study with thirty-nine children. Children were divided into thirteen groups, in which they evaluated whether food marketing instances appealed to them. Children's agreement was measured using Fleiss' kappa and the S score. Text, labels, objects and logos extracted from the ads were combined with children's variables to build four machine-learning models to identify the most important predictors of child appeal. SETTING: Households in Calgary, Alberta, Canada. PARTICIPANTS: 39 children aged 6-12 years. RESULTS: Agreement between children was low. The models indicated that the most important predictors of child appeal were the text and logos embedded in the food marketing instances. Other important predictors included children's consumption of vegetables and soda, sex and weekly hours of television. CONCLUSIONS: Text and logos embedded in the food marketing instances were the most important predictors of child appeal. The low agreement among children shows that the extent to which different marketing strategies appeal to children varies.

Predicting Abnormal Laboratory Blood Test Results in the Intensive Care Unit Using Novel Features Based on Information Theory and Historical Conditional Probability: Observational Study.

BACKGROUND: Redundancy in laboratory blood tests is common in intensive care units (ICUs), affecting patients' health and increasing health care expenses. Medical communities have made recommendations to order laboratory tests more judiciously. Wise selection can rely on modern data-driven approaches that have been shown to help identify low-yield laboratory blood tests in ICUs. However, although conditional entropy and conditional probability distribution have shown the potential to measure the uncertainty of yielding an abnormal test, no previous studies have adapted these techniques to include them in machine learning models for predicting abnormal laboratory test results. OBJECTIVE: This study aimed to address the limitations of previous reports by adapting conditional entropy and conditional probability to extract features for predicting abnormal laboratory blood test results. METHODS: We used an ICU data set collected across Alberta, Canada, which included 55,689 ICU admissions from 48,672 patients. We investigated the features of conditional entropy and conditional probability by comparing the performances of 2 machine learning approaches for predicting normal and abnormal results for 18 blood laboratory tests. Approach 1 used patients' vitals, age, sex, and admission diagnosis as features. Approach 2 used the same features plus the new conditional entropy-based and conditional probability-based features. Both approaches used 4 different machine learning models (fuzzy model, logistic regression, random forest, and gradient boosting trees) and 10 metrics (sensitivity, specificity, accuracy, precision, negative predictive value [NPV], F1 score, area under the curve [AUC], precision-recall AUC, mean G, and index balanced accuracy) to assess the performance of the approaches. RESULTS: Approach 1 achieved an average AUC of 0.86 for all 18 laboratory tests across the 4 models (sensitivity 78%, specificity 84%, precision 82%, NPV 75%, F1 score 79%, and mean G 81%), whereas approach 2 achieved an average AUC of 0.89 (sensitivity 84%, specificity 84%, precision 83%, NPV 81%, F1 score 83%, and mean G 84%). We found that the inclusion of the new features resulted in significant differences for most of the metrics in favor of approach 2. Sensitivity significantly improved for 8 and 15 laboratory tests across the different classifiers (minimum P<.001 and maximum P=.04). Mean G and index balanced accuracy, which are balanced performance metrics, also improved significantly across the classifiers for 6 to 10 and 6 to 11 laboratory tests. The most relevant feature was the pretest probability feature, which is the probability that a test result was normal when a certain number of consecutive prior tests was already normal. CONCLUSIONS: The findings suggest that conditional entropy-based features and pretest probability improve the capacity to discriminate between normal and abnormal laboratory test results. Detecting the next laboratory test result is an intermediate step toward developing guidelines for reducing overtesting in the ICU.

Diagnostic approach to Aicardi syndrome: A case report.

Aicardi syndrome is an X-linked-dominant genetic condition that is present almost exclusively in females. To diagnose Aicardi syndrome, the classic triad of agenesis of the corpus callosum, infantile spasms, and chorioretinal lacunae must be present. Here, we described a case of a female newborn baby delivered at 36 weeks of gestation that arrived at the emergency department with stiffening of arms and legs; therefore, an electroencephalogram was performed, showing generalized polypots confirming infantile spasms. Moreover, magnetic resonance was performed, showing complete agenesis of the corpus callosum. The patient was then transferred for an ophthalmoscopic examination, which evidenced multiple hypopigmented chorioretinal lesions corresponding to chorioretinal lacunae. Based on the clinical and radiological findings, the diagnosis of Aicardi syndrome was established, and treatment with anticonvulsive therapy and physiotherapy was initiated. This case report highlights the main characteristics that clinicians should consider to suspect this rare genetic condition, emphasizing the imaging and electroencephalographic findings.

A comparison between the Hilbert-Huang and Discrete Wavelet Transforms to recognize emotions from electroencephalographic signals.

Recent studies have attempted to recognize emotions by extracting features from electroencephalographic (EEG) signals using either linear and stationary, or linear and non-stationary transformations. However, as EEG signals are non-linear and non-stationary, it seems that a non-linear and non-stationary transformation may be more suitable. Despite the attractiveness of this hypothesis, until now, little studies have used such transformation. The current work presents a comparison between an approach to recognize positive and negative emotions using a non-linear and non-stationary transformation (Hilbert-Huang Transformation) with an approach using linear and non-stationary transformation (Discrete Wavelet Transform). The two approaches were compared using 200 EEG signals recorded from 10 subjects. The comparison indicated that an approach using the Hilbert-Huang Transformation statistically significantly classified emotions more accurately than a Wavelet-based approach (P < 0.02). This result implies that Hilbert-Huang Transformation is a promising tool to increase the prediction of emotional states, thereby helping to designing and developing more robust emotion recognition approaches.Clinical relevance- This remarks the potential of the Hilbert-Huang transform to enhance EEG-based emotion recognition systems, which can potentially help to diagnose and treat mental diseases, such as autism and depression.

CNN-Based LCD Transcription of Blood Pressure From a Mobile Phone Camera.

Routine blood pressure (BP) measurement in pregnancy is commonly performed using automated oscillometric devices. Since no wireless oscillometric BP device has been validated in preeclamptic populations, a simple approach for capturing readings from such devices is needed, especially in low-resource settings where transmission of BP data from the field to central locations is an important mechanism for triage. To this end, a total of 8192 BP readings were captured from the Liquid Crystal Display (LCD) screen of a standard Omron M7 self-inflating BP cuff using a cellphone camera. A cohort of 49 lay midwives captured these data from 1697 pregnant women carrying singletons between 6 weeks and 40 weeks gestational age in rural Guatemala during routine screening. Images exhibited a wide variability in their appearance due to variations in orientation and parallax; environmental factors such as lighting, shadows; and image acquisition factors such as motion blur and problems with focus. Images were independently labeled for readability and quality by three annotators (BP range: 34-203 mm Hg) and disagreements were resolved. Methods to preprocess and automatically segment the LCD images into diastolic BP, systolic BP and heart rate using a contour-based technique were developed. A deep convolutional neural network was then trained to convert the LCD images into numerical values using a multi-digit recognition approach. On readable low- and high-quality images, this proposed approach achieved a 91% classification accuracy and mean absolute error of 3.19 mm Hg for systolic BP and 91% accuracy and mean absolute error of 0.94 mm Hg for diastolic BP. These error values are within the FDA guidelines for BP monitoring when poor quality images are excluded. The performance of the proposed approach was shown to be greatly superior to state-of-the-art open-source tools (Tesseract and the Google Vision API). The algorithm was developed such that it could be deployed on a phone and work without connectivity to a network.

A review of fetal cardiac monitoring, with a focus on low- and middle-income countries.

There is limited evidence regarding the utility of fetal monitoring during pregnancy, particularly during labor and delivery. Developed countries rely on consensus 'best practices' of obstetrics and gynecology professional societies to guide their protocols and policies. Protocols are often driven by the desire to be as safe as possible and avoid litigation, regardless of the cost of downstream treatment. In high-resource settings, there may be a justification for this approach. In low-resource settings, in particular, interventions can be costly and lead to adverse outcomes in subsequent pregnancies. Therefore, it is essential to consider the evidence and cost of different fetal monitoring approaches, particularly in the context of treatment and care in low-to-middle income countries. This article reviews the standard methods used for fetal monitoring, with particular emphasis on fetal cardiac assessment, which is a reliable indicator of fetal well-being. An overview of fetal monitoring practices in low-to-middle income counties, including perinatal care access challenges, is also presented. Finally, an overview of how mobile technology may help reduce barriers to perinatal care access in low-resource settings is provided.

Unsupervised hidden semi-Markov model for automatic beat onset detection in 1D Doppler ultrasound.

OBJECTIVE: One dimensional (1D) Doppler ultrasound (DUS) is commonly used for fetal health assessment, during both regular prenatal visits and labor. It is used in preference to ECG and other modalities because of its simplicity and cost. To date, all analysis of such data has been confined to a smoothed, windowed heart rate estimation derived from the 1D DUS signal, reducing the potential of short-term variability information. A first step in improving the assessment of short-term variability of the fetal heart rate (FHR) is through implementing an accurate beat detector for 1D DUS signals. APPROACH: This work presents an unsupervised probabilistic segmentation method enabled by a hidden semi-Markov model (HSMM). The proposed method employs envelope and spectral features for an online segmentation of fetal 1D DUS signal. The beat onsets and fetal cardiac beat-to-beat intervals are then estimated from the segmentations. For this work, two data sets were used, including 1D DUS recordings from five fetuses recorded in Germany, comprising 6521 beats and 45.06 minutes of data (dataset 1). Simultaneous fetal ECG (fECG) was used as the reference for beat timing. Dataset 2, comprising 4044 beats captured from 17 subjects in the UK was hand scored for beat location and was used as an independent held-out test set. Leave-one-out subject cross-validation was used for parameter tuning on dataset 1. No retraining was performed for dataset 2. To assess the performance of the beat onset detection, the root mean square error (RMSE), F1 score, sensitivity, positive predictivity (PPV) and the error in several standard common heart rate variability metrics were used. These metrics were evaluated on three fiducial points: (1) beat onset, (2) beat offset, and (3) middle of beat interval. MAIN RESULTS: In dataset 1, the proposed method provided an RMSE of 20 ms, F1 score of 97.5 %, a Se of 97.6%, and a PPV of 97.3%. In dataset 2, the proposed method achieved an RMSE of 26 ms, an F1 score of 98.5 %, a Se of 98.0 % and a PPV of 98.9 %. It was also determined that the best beat-to-beat interval was derived from the onset of each beat. For the dataset 2, significant correlations were found in all short term heart rate variability metrics tested, both in the time and frequency domain. Only the proportion of successive normal-to-normal interval differences greater than 20 ms (pNN20) exhibited a significant absolute difference. SIGNIFICANCE: This work presents the first-ever description of an algorithm to identify cardiac beats with 1D DUS, closely matching the fetal ECG-derived beats, to enable short-term heart rate variability analysis. The novel algorithm proposed requires no human labeling of data, and could have applicability beyond 1D DUS to other similar highly variable time series.

Estimating birth weight from observed postnatal weights in a Guatemalan highland community.

OBJECTIVE: Low birth weight is one of the leading contributors to global perinatal deaths. Detecting this problem close to birth enables the initiation of early intervention, thus reducing the long-term impact on the fetus. However, in low-and middle-income countries, sometimes newborns are weighted days or months after birth, thus challenging the identification of low birth weight. This study aims to estimate birth weight from observed postnatal weights recorded in a Guatemalan highland community. APPROACH: With 918 newborns recorded in postpartum visits at a Guatemalan highland community, we fitted traditional infant weight models (Count's and Reeds models). The model that fitted the observed data best was selected based on typical newborn weight patterns reported in the medical literature and previous longitudinal studies. Then, estimated birth weights were determined using the weight gain percentage derived from the fitted weight curve. MAIN RESULTS: The best model for both genders was the Reeds2 model, with a mean square error of 0.30 kg2 and 0.23 kg2 for male and female newborns, respectively. The fitted weight curves exhibited similar behavior to those reported in the literature, with a maximum weight loss around three to five days after birth, and birth weight recovery, on average, by day ten. Moreover, the estimated birth weight was consistent with the 2015 Guatemalan National Survey, no having a statistically significant difference between the estimated birth weight and the reported survey birth weights (two-sided Wilcoxon rank-sum test; [Formula: see text]). SIGNIFICANCE: By estimating birth weight at an opportune time, several days after birth, it may be possible to identify low birth weight more accurately, thus providing timely treatment when is required.

A Proxy for Detecting IUGR Based on Gestational Age Estimation in a Guatemalan Rural Population.

In-utero progress of fetal development is normally assessed through manual measurements taken from ultrasound images, requiring relatively expensive equipment and well-trained personnel. Such monitoring is therefore unavailable in low- and middle-income countries (LMICs), where most of the perinatal mortality and morbidity exists. The work presented here attempts to identify a proxy for IUGR, which is a significant contributor to perinatal death in LMICs, by determining gestational age (GA) from data derived from simple-to-use, low-cost one-dimensional Doppler ultrasound (1D-DUS) and blood pressure devices. A total of 114 paired 1D-DUS recordings and maternal blood pressure recordings were selected, based on previously described signal quality measures. The average length of 1D-DUS recording was 10.43 ± 1.41 min. The min/median/max systolic and diastolic maternal blood pressures were 79/102/121 and 50.5/63.5/78.5 mmHg, respectively. GA was estimated using features derived from the 1D-DUS and maternal blood pressure using a support vector regression (SVR) approach and GA based on the last menstrual period as a reference target. A total of 50 trials of 5-fold cross-validation were performed for feature selection. The final SVR model was retrained on the training data and then tested on a held-out set comprising 28 normal weight and 25 low birth weight (LBW) newborns. The mean absolute GA error with respect to the last menstrual period was found to be 0.72 and 1.01 months for the normal and LBW newborns, respectively. The mean error in the GA estimate was shown to be negatively correlated with the birth weight. Thus, if the estimated GA is lower than the (remembered) GA calculated from last menstruation, then this could be interpreted as a potential sign of IUGR associated with LBW, and referral and intervention may be necessary. The assessment system may, therefore, have an immediate impact if coupled with suitable intervention, such as nutritional supplementation. However, a prospective clinical trial is required to show the efficacy of such a metric in the detection of IUGR and the impact of the intervention.

An open source autocorrelation-based method for fetal heart rate estimation from one-dimensional Doppler ultrasound.

OBJECTIVE: Open research on fetal heart rate (FHR) estimation is relatively rare, and evidence for the utility of metrics derived from Doppler ultrasound devices has historically remained hidden in the proprietary documentation of commercial entities, thereby inhibiting its assessment and improvement. Nevertheless, recent studies have attempted to improve FHR estimation; however, these methods were developed and tested using datasets composed of few subjects and are therefore unlikely to be generalizable on a population level. The work presented here introduces a reproducible and generalizable autocorrelation (AC)-based method for FHR estimation from one-dimensional Doppler ultrasound (1D-DUS) signals. APPROACH: Simultaneous fetal electrocardiogram (fECG) and 1D-DUS signals generated by a hand-held Doppler transducer in a fixed position were captured by trained healthcare workers in a European hospital. The fECG QRS complexes were identified using a previously published fECG extraction algorithm and were then over-read to ensure accuracy. An AC-based method to estimate FHR was then developed on this data, using a total of 721 1D-DUS segments, each 3.75 s long, and parameters were tuned with Bayesian optimization. The trained FHR estimator was tested on two additional (independent) hand-annotated Doppler-only datasets recorded with the same device but on different populations: one composed of 3938 segments (from 99 fetuses) acquired in rural Guatemala, and another composed of 894 segments (from 17 fetuses) recorded in a hospital in the UK. MAIN RESULTS: The proposed AC-based method was able to estimate FHR within 10% of the reference FHR values 96% of the time, with an accuracy of 97% for manually identified good quality segments in both of the independent test sets. SIGNIFICANCE: This is the first work to publish open source code for FHR estimation from 1D-DUS data. The method was shown to satisfy estimations within 10% of the reference FHR values and it therefore defines a minimum accuracy for the field to match or surpass. Our work establishes a basis from which future methods can be developed to more accurately estimate FHR variability for assessing fetal wellbeing from 1D-DUS signals.

mHealth intervention to improve the continuum of maternal and perinatal care in rural Guatemala: a pragmatic, randomized controlled feasibility trial.

BACKGROUND/OBJECTIVE: Guatemala's indigenous Maya population has one of the highest perinatal and maternal mortality rates in Latin America. In this population most births are delivered at home by traditional birth attendants (TBAs), who have limited support and linkages to public hospitals. The goal of this study was to characterize the detection of maternal and perinatal complications and rates of facility-level referral by TBAs, and to evaluate the impact of a mHealth decision support system on these rates. METHODS: A pragmatic one-year feasibility trial of an mHealth decisions support system was conducted in rural Maya communities in collaboration with TBAs. TBAs were individually randomized in an unblinded fashion to either early-access or later-access to the mHealth system. TBAs in the early-access arm used the mHealth system throughout the study. TBAs in the later-access arm provided usual care until crossing over uni-directionally to the mHealth system at the study midpoint. The primary study outcome was the monthly rate of referral to facility-level care, adjusted for birth volume. RESULTS: Forty-four TBAs were randomized, 23 to the early-access arm and 21 to the later-access arm. Outcomes were analyzed for 799 pregnancies (early-access 425, later-access 374). Monthly referral rates to facility-level care were significantly higher among the early-access arm (median 33 referrals per 100 births, IQR 22-58) compared to the later-access arm (median 20 per 100, IQR 0-30) (p = 0.03). At the study midpoint, the later-access arm began using the mHealth platform and its referral rates increased (median 34 referrals per 100 births, IQR 5-50) with no significant difference from the early-access arm (p = 0.58). Rates of complications were similar in both arms, except for hypertensive disorders of pregnancy, which were significantly higher among TBAs in the early-access arm (RR 3.3, 95% CI 1.10-9.86). CONCLUSIONS: Referral rates were higher when TBAs had access to the mHealth platform. The introduction of mHealth supportive technologies for TBAs is feasible and can improve detection of complications and timely referral to facility-care within challenging healthcare delivery contexts. TRIAL REGISTRATION: Clinicaltrials.gov NCT02348840 .

Agile Development of a Smartphone App for Perinatal Monitoring in a Resource-Constrained Setting.

Technology provides the potential to empower frontline healthcare workers with low levels of training and literacy, particularly in low- and middle-income countries. An obvious platform for achieving this aim is the smartphone, a low cost, almost ubiquitous device with good supply chain infrastructure and a general cultural acceptance for its use. In particular, the smartphone offers the opportunity to provide augmented or procedural information through active audiovisual aids to illiterate or untrained users, as described in this article. In this article, the process of refinement and iterative design of a smartphone application prototype to support perinatal surveillance in rural Guatemala for indigenous Maya lay midwives with low levels of literacy and technology exposure is described. Following on from a pilot to investigate the feasibility of this system, a two-year project to develop a robust in-field system was initiated, culminating in a randomized controlled trial of the system, which is ongoing. The development required an agile approach, with the development team working both remotely and in country to identify and solve key technical and cultural issues in close collaboration with the midwife end-users. This article describes this process and intermediate results. The application prototype was refined in two phases, with expanding numbers of end-users. Some of the key weaknesses identified in the system during the development cycles were user error when inserting and assembling cables and interacting with the 1-D ultrasound-recording interface, as well as unexpectedly poor bandwidth for data uploads in the central healthcare facility. Safety nets for these issues were developed and the resultant system was well accepted and highly utilized by the end-users. To evaluate the effectiveness of the system after full field deployment, data quality, and corruption over time, as well as general usage of the system and the volume of application support for end-users required by the in-country team was analyzed. Through iterative review of data quality and consistent use of user feedback, the volume and percentage of high quality recordings was increased monthly. Final analysis of the impact of the system on obstetrical referral volume and maternal and neonatal clinical outcomes is pending conclusion of the ongoing clinical trial.

Template-based Quality Assessment of the Doppler Ultrasound Signal for Fetal Monitoring.

One dimensional Doppler Ultrasound (DUS) is a low cost method for fetal auscultation. However, accuracy of any metrics derived from the DUS signals depends on their quality, which relies heavily on operator skills. In low resource settings, where skill levels are sparse, it is important for the device to provide real time signal quality feedback to allow the re-recording of data. Retrospectively, signal quality assessment can help remove low quality recordings when processing large amounts of data. To this end, we proposed a novel template-based method, to assess DUS signal quality. Data used in this study were collected from 17 pregnant women using a low-cost transducer connected to a smart phone. Recordings were split into 1990 segments of 3.75 s duration, and hand labeled for quality by three independent annotators. The proposed template-based method uses Empirical Mode Decomposition (EMD) to allow detection of the fetal heart beats and segmentation into short, time-aligned temporal windows. Templates were derived for each 15 s window of the recordings. The DUS signal quality index (SQI) was calculated by correlating the segments in each window with the corresponding running template using four different pre-processing steps: (i) no additional preprocessing, (ii) linear resampling of each beat, (iii) dynamic time warping (DTW) of each beat and (iv) weighted DTW of each beat. The template-based SQIs were combined with additional features based on sample entropy and power spectral density. To assess the performance of the method, the dataset was split into training and test subsets. The training set was used to obtain the best combination of features for predicting the DUS quality using cross validation, and the test set was used to estimate the classification accuracy using bootstrap resampling. A median out of sample classification accuracy on the test set of 85.8% was found using three features; template-based SQI, sample entropy and the relative power in the 160 to 660 Hz range. The results suggest that the new automated method can reliably assess the DUS quality, thereby helping users to consistently record DUS signals with acceptable quality for fetal monitoring.

Attacks against health care in Syria, 2015-16: results from a real-time reporting tool.

BACKGROUND: Collecting credible data on violence against health services, health workers, and patients in war zones is a massive challenge, but crucial to understanding the extent to which international humanitarian law is being breached. We describe a new system used mainly in areas of Syria with a substantial presence of armed opposition groups since November, 2015, to detect and verify attacks on health-care services and describe their effect. METHODS: All Turkey health cluster organisations with a physical presence in Syria, either through deployed and locally employed staff, were asked to participate in the Monitoring Violence against Health Care (MVH) alert network. The Turkey hub of the health cluster, a UN-activated humanitarian health coordination body, received alerts from health cluster partners via WhatsApp and an anonymised online data-entry tool. Field staff were asked to seek further information by interviewing victims and other witnesses when possible. The MVH data team triangulated alerts to identify individual events and distributed a preliminary flash update of key information (location, type of service, modality of attack, deaths, and casualties) to partners, WHO, United Nations Office for the Coordination of Humanitarian Affairs, and donors. The team also received and entered alerts from several large non-health cluster organisations (known as external partners, who do their own information-gathering and verification processes before sharing their information). Each incident was then assessed in a stringent process of information-matching. Attacks were deemed to be verified if they were reported by a minimum of one health cluster partner and one external partner, and the majority of the key datapoints matched. Alerts that did not meet this standard were deemed to be unverified. Results were tabulated to describe attack occurrence and impact, disaggregated where possible by age, sex, and location. FINDINGS: Between early November, 2015, and Dec 31 2016, 938 people were directly harmed in 402 incidents of violence against health care: 677 (72%) were wounded and 261 (28%) were killed. Most of the dead were adult males (68%), but the highest case fatality (39%) was seen in children aged younger than 5 years. 24% of attack victims were health workers. Around 44% of hospitals and 5% of all primary care clinics in mainly areas with a substantial presence of armed opposition groups experienced attacks. Aerial bombardment was the main form of attack. A third of health-care services were hit more than once. Services providing trauma care were attacked more than other services. INTERPRETATION: The data system used in this study addressed double-counting, reduced the effect of potentially biased self-reports, and produced credible data from anonymous information. The MVH tool could be feasibly deployed in many conflict areas. Reliable data are essential to show how far warring parties have strayed from international law protecting health care in conflict and to effectively harness legal mechanisms to discourage future perpetrators. FUNDING: None.