True Events Research Articles

Abstract 10532: Physician Interpretation and Characterization of Intracardiac Electrograms at the Time of Non-Cardiac Death

Introduction: Electrograms (EGMs) generated by implantable cardiac devices at the time of non-cardiac death may have artifact from non-arrhythmic sources, such as coagulated blood and air bubbles. While few studies have described the nature of terminal EGMs in non-cardiac death, it is clinically and medicolegally important to characterize these EGMs to elucidate events that may otherwise be misinterpreted as device failure. Here we investigated terminal EGM features that may aid clinicians in discerning cardiac from non-cardiac death. Methods: Terminal EGMs were analyzed by independent physician reviewers. Deaths were adjudicated as non-cardiac based on direct clinical data at the time of death using a modified Cardiac Arrhythmia Suppression Trial approach. In total, 16 non-cardiac terminal EGMs were identified and compared to a control cohort of true arrhythmic events. Using a subset of identified EGMs, an 8-question survey was administered to electrophysiologists blinded to cause of death. Several variables were identified as important for the correct classification of EGMs. The statistical performance of each variable for detecting non-cardiac death and true arrhythmias in our sample was evaluated. These results were shared, and the survey was readministered. Results: In our sample, absence of organized atrial activity had a sensitivity (SN) of 78.5%, specificity (SP) of 100%, and positive predictive value (PPV) of 100% for non-cardiac terminal EGMs. Cycle length <130 ms and absence of discernible deflections both had SPs and PPVs of 100% for non-cardiac death but were less sensitive (60% and 20%, respectively). An abrupt change following therapy and response to therapy both had SNs, SPs, and PPVs of 100% for true arrhythmias. Survey response rate was 68.8% and 62.5% for the pre- and post-survey, respectively. Pre-survey scores ranged from 50 – 87.5% correct with a mean of 72.7%. Post-survey scores ranged from 75 – 100% with a mean of 86.3% (p< 0.01). Conclusions: There is variability and inaccuracy among expert physicians' interpretation of EGMs during non-cardiac death. The variables identified in this study will aid in the development of a clinical risk score used to discern non-cardiac death and true device malfunction based on terminal EGMs.

Abstract 12728: The Addition of Social Determinants of Health Improves the Predictive Accuracy of the Pooled Cohort Equations for 10-year Ascvd Events in African Americans

Introduction: The gold standard risk prediction tool for atherosclerotic cardiovascular disease (ASCVD), the Pooled Cohort Equations (PCE) underestimates risk among the socially disadvantaged, yet social determinants of health are absent from the PCE. This may be contributing to racial inequities in CVD since due to systemic and structural racism, socioeconomic disadvantage is disproportionately concentrated in African American (AA) communities. This analysis aimed to determine whether SDOH variables were important for the prediction of 10-year ASCVD risk in a large cohort of AA and to test whether the addition of these variables improved PCE model performance. Methods: Retrospective observational study of 3,470 Jackson Heart Study (JHS) participants aged 40-79 years old without prior ASCVD at baseline study visit. 10-year ASCVD event was defined as stroke, MI, or death from MI. First, we used competing risk random survival forests to determine variable importance of risk factors, then we used cox-proportional hazards to identify which variables among those found to be important, were independently associated with our outcome. We used time dependent AUC to compare the performance of PCE alone to models with the PCE plus SDOH to assess concordance with the true event rates for prediction of 10-year ASCVD. Results: Our sample was 35% male, 100% AA and 46% of participants had an income of <$35K/year. Of 30 SDOH factors tested, the following 6: BMI, depression, weekly stress, insurance status, family income, and neighborhood violence were determined most important for prediction and were independently associated with 10-year ASCVD risk. In this sample, the PCE only model had an AUC of 0.822; adding a limited panel of SDOH factors - income, insurance status and occupation, increased the AUC to 0.859, but this improvement was only statistically significant at the 0.10 level (p=0.093). The model adding the full 6 SDOH risk panel to the PCE, saw the highest AUC at 0.863 (p=0.029). Conclusions: Adding SDOH risk factors to the PCE improved risk prediction. These findings should be explored in a larger database to determine if adding SDOH to the PCE can improve risk prediction for ASCVD events among vulnerable populations and guide more optimal use of prevention strategies.

‘Fake News’ in Seventeenth-Century France: The Case of Le Mercure galant

Abstract Le Mercure galant, one of France’s first newspapers, is notable for its diverse content: politics and foreign affairs, court news, science and medicine, the arts and literature. Directed by Jean Donneau de Visé from its inception in 1672 until his death in 1710, this influential and innovative monthly publication circulated throughout France and beyond its borders. The Mercure’s tendency to blur the lines between truth and fiction, between history and propaganda, and between information and entertainment, makes it an instructive case study in early modern ‘fake news’. Donneau de Visé, a self-styled royal historiographer and the beneficiary of a generous royal pension, dedicated his periodical to the Dauphin and published abundant praise of Louis XIV. The Mercure’s news reporting included distortions and propaganda intended to bolster and further entrench the king’s foreign and domestic policies. The Mercure’s nouvelles, short stories presented as true recent events, constituted another type of ‘fake news’, one that often had a different effect, inviting the re-examination of social norms. The nouvelles appealed to the Mercure’s sizeable community of women readers by accentuating female agency and providing a vehicle for the exploration of scenarios of female empowerment.

A deep-learning approach for automatically detecting gait-events based on foot-marker kinematics in children with cerebral palsy-Which markers work best for which gait patterns?

Neuromotor pathologies often cause motor deficits and deviations from typical locomotion, reducing the quality of life. Clinical gait analysis is used to effectively classify these motor deficits to gain deeper insights into resulting walking behaviours. To allow the ensemble averaging of spatio-temporal metrics across individuals during walking, gait events, such as initial contact (IC) or toe-off (TO), are extracted through either manual annotation based on video data, or through force thresholds using force plates. This study developed a deep-learning long short-term memory (LSTM) approach to detect IC and TO automatically based on foot-marker kinematics of 363 cerebral palsy subjects (age: 11.8 ± 3.2). These foot-marker kinematics, including 3D positions and velocities of the markers located on the hallux (HLX), calcaneus (HEE), distal second metatarsal (TOE), and proximal fifth metatarsal (PMT5), were extracted retrospectively from standard barefoot gait analysis sessions. Different input combinations of these four foot-markers were evaluated across three gait subgroups (IC with the heel, midfoot, or forefoot). For the overall group, our approach detected 89.7% of ICs within 16ms of the true event with a 18.5% false alarm rate. For TOs, only 71.6% of events were detected with a 33.8% false alarm rate. While the TOE|HEE marker combination performed well across all subgroups for IC detection, optimal performance for TO detection required different input markers per subgroup with performance differences of 5-10%. Thus, deep-learning LSTM based detection of IC events using the TOE|HEE markers offers an automated alternative to avoid operator-dependent and laborious manual annotation, as well as the limited step coverage and inability to measure assisted walking for force plate-based detection of IC events.

Partisan bias in false memories for misinformation about the 2021 U.S. Capitol riot

ABSTRACT Memory for events can be biased. For example, people tend to recall more events that support than oppose their current worldview. The present study examined partisan bias in memory for events related to the January 6, 2021, Capitol riot in the United States. Participants rated their memory for true and false events that were either favourable to their political party or the other major political party in the United States. For both true and false events, participants remembered more events that favoured their political party. Regression analyses showed that the number of false memories that participants reported was positively associated with their tendency to support conspiracy beliefs and with their self-reported engagement with the Capitol riot. These results suggest that Democrats and Republicans remember the Capitol Riot differently and that certain individual difference factors can predict the formation of false memories in this context. Misinformation played an influential role in the Capitol riot and understanding differences in memory for this event is beneficial to avoiding similar tragedies in the future.

The preservation of cause and effect in the rock record

AbstractEvolutionary events may impact the geological carbon cycle via transient imbalances in silicate weathering, and such events have been implicated as causes of glaciations, mass extinctions, and oceanic anoxia. However, suggested evolutionary causes often substantially predate the environmental effects to which they are linked—problematic when carbon cycle perturbations must be resolved in less than a million years to maintain Earth's habitability. What is more, the geochemical signatures of such perturbations are recorded as they occur in widely distributed marine sedimentary rocks that have been densely sampled for important intervals in Earth history, whereas the fossil record—particularly on land—is governed by the availability of sedimentary basins that are patchy in both space and time, necessitating lags between the origination of an evolutionary lineage and its earliest occurrence in the fossil record. Here, we present a simple model of the impact of preservational filtering on sampling to show that an evolutionary event that causes an environmental perturbation via weathering imbalance should not appear earlier in the rock record than the perturbation itself and, if anything, should appear later rather than simultaneously. The Devonian Hangenberg glaciation provides an example of how evolutionary events might be more fruitfully considered as potential causes of environmental perturbations. Just as the last samplings of species lost in mass extinction are expected to come before the true environmental event, first appearance should be expected to postdate the geological expression of a lineage's environmental impact with important implications for our reading of Earth history.

TRADISI KEMBAR MAYANG DALAM PROSES PERNIKAHAN ADAT JAWA DI KAPANEWON TEMPEL, KABUPATEN SLEMAN

The Javanese traditional wedding ceremony is still the community's preference because it includes a series of sacred processions, even though it is complicated since there are so many of them. One of these series, namely the Kembar Mayang, is considered a sign to the ancestors that a wedding ceremony is taking place. Therefore, this study focuses on Kembar Mayang research as a still-preserved culture or custom. This is a field study based on true events that was analyzed using an anthropological perspective; data was collected through observation, interviews, and literature review. This study reveals that residents in Kapanewon Tempel, Sleman, continue to consider the Kembar Mayang tradition, including its complementary series of offerings, janur, and tarub. According to the customs of the people, this tradition has become a necessity for traditional Javanese weddings, as it is believed to facilitate the process and prevent supernatural interference. Despite what might be expected, the Kembar Mayang tradition is not contrary with religion. Likewise, this practice or tradition provides numerous benefits. The people of Kapanewon Tempel, Sleman Regency agree that this ritual is not troublesome, as it has become a habit for every citizen that chooses for Javanese wedding traditions.
 
 Keywords : Customs, Kembar Mayang Traditions, Weddings, Culture

Transient acceleration events in LISA Pathfinder data: Properties and possible physical origin

We present an in depth analysis of the transient events, or glitches, detected at a rate of about one per day in the differential acceleration data of LISA Pathfinder. We show that these glitches fall in two rather distinct categories: fast transients in the interferometric motion readout on one side, and true force transient events on the other. The former are fast and rare in ordinary conditions. The second may last from seconds to hours and constitute the majority of the glitches. We present an analysis of the physical and statistical properties of both categories, including a cross-analysis with other time series like magnetic fields, temperature, and other dynamical variables. Based on these analyses we discuss the possible sources of the force glitches and identify the most likely, among which the outgassing environment surrounding the test-masses stands out. We discuss the impact of these findings on the LISA design and operation, and some risk mitigation measures, including experimental studies that may be conducted on the ground, aimed at clarifying some of the questions left open by our analysis.9 MoreReceived 24 May 2022Accepted 23 August 2022DOI:https://doi.org/10.1103/PhysRevD.106.062001© 2022 American Physical SocietyPhysics Subject Headings (PhySH)Research AreasExperimental studies of gravityGravitational wavesTechniquesGravitational wave detectorsGravitation, Cosmology & Astrophysics

Evaluation of the response of plastic scintillator bars and measurement of neutron capture time in non-reactor environment for the ISMRAN experiment

We present a detailed study on detector response to different radioactive sources and the measurements of non-reactor environmental backgrounds with the Indian Scintillator Matrix for Reactor Anti-Neutrinos (ISMRAN) detector setup consisting of 9 × 10 Plastic Scintillator Bars (PSBs) array at BARC, Mumbai. These measurements are useful in the context of the ISMRAN detector setup, which will be used to detect the reactor anti-neutrinos (ν¯e) and measure its energy spectra through the inverse beta decay (IBD) process. A GEANT4 based Monte Carlo (MC) simulation is used to understand the optical transmission, energy resolution and energy non-linearity of the ISMRAN detector. A detailed analysis procedure has been developed to understand the natural radioactive, cosmogenic and cosmic muon-induced backgrounds with the ISMRAN detector setup in a non-reactor environment, based on their energy deposition, number of bars hit as well as topological event selection criteria in position and time for triggered events. Data and MC simulated distributions of reconstructed sum energy and number of bars hit has been compared for the radioactive γ + positron source such as 22Na placed at the center of the ISMRAN array. Fast neutron energy response and capture time distribution in ISMRAN array has been studied using a novel technique involving Time of Flight (TOF) of the measured fast neutrons. The observed characteristic neutron capture time (τ) of 68.29±9.48μs is in good agreement with ∼65μs obtained from MC simulation. These experimentally measured results will be useful for discriminating the correlated and uncorrelated background events from the true IBD events in reactor ON and OFF conditions inside the reactor hall.

Detection of Physical Impacts of Shipping Containers during Handling Operations Using the Impact Detection Methodology

The transportation of cargo inside shipping containers is a risky operation that requires constant monitoring activities and real-time operational actions. Yet, the detection of the real dynamics of the container and the surrounding infrastructure and extraction of true subsequent critical events is still an unresolved issue among engineers. In this paper, we analyze the new physical impact detection method, namely the Impact Detection Methodology (IDM), to detect the most obvious and force-dependent impacts from acceleration data, using the IoT sensor in an experimental environment using the heavy machinery of a seaport. By variating the threshold level, we have observed the changes in the number of impacts detected within three separate case studies. Results suggest that the optimal parameters tend to provide an adequate number of events, yet even the slightest change in the threshold level can increase or decrease the number of detected impacts in a non-linear fashion, making the detection harder, due to unforeseen external impacts on the dataset, the filtering of which is still the main priority of our future research.

WAMS-based two-level robust detection methodology of power system events

With the advance of Wide-Area Measurement Systems (WAMS), power system operators have direct access to a large amount of data with valuable information about the power system dynamic performance. As a result, there is a clear need for new data-driven methodologies capable of extracting relevant information from this collected data. One of the key challenges is correctly detecting power system disturbances to avoid false alarms during real-time operation as well as off-line disturbance analysis. This paper proposes a two-level robust event detection methodology aiming to reduce false disturbance detection (false positives/alarms) and validate true events. The methodology is divided into two-levels: (i) signal processing analysis (ii) deep neural network (DNN) classification. In the first level, we apply a widely used spectral analysis based on the Discrete Wavelet Transform (DWT) to event detection. In the second level, the events detected by the DWT are processed by a DNN to check if they are real events or false alarms. Finally, the proposed methodology is evaluated using real synchrophasor event records from the Brazilian Interconnected Power System (BIPS).

Accuracy of ICD-10-CM encounter diagnoses from health records for identifying self-harm events.

Assess the accuracy of ICD-10-CM coding of self-harm injuries and poisonings to identify self-harm events. In 7 integrated health systems, records data identified patients reporting frequent suicidal ideation. Records then identified subsequent ICD-10-CM injury and poisoning codes indicating self-harm as well as selected codes in 3 categories where uncoded self-harm events might be found: injuries and poisonings coded as undetermined intent, those coded accidental, and injuries with no coding of intent. For injury and poisoning encounters with diagnoses in those 4 groups, relevant clinical text was extracted from records and assessed by a blinded panel regarding documentation of self-harm intent. Diagnostic codes selected for review include all codes for self-harm, 43 codes for undetermined intent, 26 codes for accidental intent, and 46 codes for injuries without coding of intent. Clinical text was available for review for 285 events originally coded as self-harm, 85 coded as undetermined intent, 302 coded as accidents, and 438 injury events with no coding of intent. Blinded review of full-text clinical records found documentation of self-harm intent in 254 (89.1%) of those originally coded as self-harm, 24 (28.2%) of those coded as undetermined, 24 (7.9%) of those coded as accidental, and 48 (11.0%) of those without coding of intent. Among patients at high risk, nearly 90% of injuries and poisonings with ICD-10-CM coding of self-harm have documentation of self-harm intent. Reliance on ICD-10-CM coding of intent to identify self-harm would fail to include a small proportion of true self-harm events.

Solar Irradiance Ramp Forecasting Based on All-Sky Imagers

Solar forecasting constitutes a critical tool for operating, producing and storing generated power from solar farms. In the framework of the International Energy Agency’s Photovoltaic Power Systems Program Task 16, the solar irradiance nowcast algorithms, based on five all-sky imagers (ASIs), are used to investigate the feasibility of ASIs to foresee ramp events. ASIs 1–2 and ASIs 3–5 can capture the true ramp events by 26.0–51.0% and 49.0–92.0% of the cases, respectively. ASIs 1–2 provided the lowest (<10.0%) falsely documented ramp events while ASIs 3–5 recorded false ramp events up to 85.0%. On the other hand, ASIs 3–5 revealed the lowest falsely documented no ramp events (8.0–51.0%). ASIs 1–2 are developed to provide spatial solar irradiance forecasts and have been delimited only to a small area for the purposes of this benchmark, which penalizes these approaches. These findings show that ASI-based nowcasts could be considered as a valuable tool for predicting solar irradiance ramp events for a variety of solar energy technologies. The combination of physical and deep learning-based methods is identified as a potential approach to further improve the ramp event forecasts.

Diversification of landslide areas as a means for reducing noise in dendrogeomorphic dating

The result of tree-ring-based reconstruction of past landslide events is often the development of a single total chronology. This approach can be very effective for small homogeneous landslides. However, compiling chronological data from heterogeneous (often independent) zones of large complex landslide areas into one chronology can induce over- or underestimation of some events, resulting in lowered reliability of the reconstruction. The solution for elimination of this effect can lie in the diversification of complex landslide areas into homogeneous zones with separate analyses. The aim of this study was to quantify the effect of this separation on detected slope movement events and to define parameters whose investigation could distinguish events (sliding) from noise (creeping). For this purpose, 412 tree-ring series from 206 disturbed common spruce ( Picea abies (L.) Karst.) occupying complex landslide areas were dendrogeomorphically analysed. The landslide area was divided into five homogeneous zones using geomorphic mapping, LiDAR-based DEM and geophysical sounding (ERT). Five events (verified in individual zones) were detected in the total chronology. Two extra events in the total chronology (28.6%) were considered noise. Moreover, two zonal events were detected but not recorded in the total chronology. This indicates that the noise in the total chronology of the complex landslide area could reach more than a quarter of dated events. Next, true slide events and noise (caused by creep) were differentiated in the structure of growth disturbances (reaction wood vs. abrupt growth suppression) and their proportion in event reconstruction, spatial patterns of trees containing slope movement signals, and the character of triggers. Thus, for better filtering of noise from signals in tree-ring-based chronologies of landslides, not only observations of dendrogeomorphic index values but also the morphology of landslides and characteristics of dated processes must be considered.

Optimization of detector modules for measuring gamma-ray polarization in Positron Emission Tomography

Detection of γ-ray polarization in Positron Emission Tomography (PET) is as yet an unexploited feature that could be used as an additional handle to improve signal-to-background ratio in this imaging modality. The γ polarization is related to the azimuthal angle in the Compton scattering process, so the initial correlation of polarizations of the annihilation quanta translates to the correlation of the azimuthal angles in events where both annihilation photons undergo Compton scattering in respective detectors. This results in a modulated distribution of the azimuthal angle difference for true events, while this modulation is lacking for the background events. We present a comprehensive experimental study of five detector configurations based on scintillator matrices and silicon photomultipliers, suitable for measuring the azimuthal modulation. The modules consist of either GaGG:Ce or LYSO:Ce pixels with sizes varying from 1.9×1.9×20mm3 to 3×3×20mm3. The distinctive feature of the modules is that they can reconstruct the Compton scattering by detecting the recoil electron and the scattered γ in a single detector layer, which simplifies extension to larger systems. The amplitude modulation of the azimuthal angles’ difference is clearly observable in all configurations ranging from 0.26±0.01 to 0.34±0.02 depending on the event selection criteria. The results suggest that finer detector segmentation plays a leading role in achieving higher modulation factors.

A robust machine learning structure for driving events recognition using smartphone motion sensors

Driving behavior monitoring by smartphone sensors is one of the most investigated approaches to ameliorate road safety. Various methods are adopted in the literature; however, to the best of our knowledge, their robustness to the prediction of new unseen data from different drivers and road conditions is not explored. In this paper, a two-phase Machine Learning (ML) method with taking advantage of high-pass, low-pass, and wavelet filters is developed to detect driving brakes and turns. In the first phase, accelerometer and gyroscope filtered time series are fed into Random Forest and Artificial Neural Network classifiers, and the suspicious intervals are extracted by a high recall. Following that, in the next phase, statistical features calculated based on the obtained intervals are used to determine the false and true positive events. To compare the predicted and real labels of the recorded events and calculate the accuracy, a method that covers the limitations of previous sliding windows is also employed. Real-world experimental result shows that the proposed method can predict new unseen datasets with average F1-scores of 71% in brake detection and 82% in turn detection which is comparable with previous works. Moreover, by sensitivity analysis of our proposed model, it is proven that implementing high-pass and low-pass filters can affect the accuracy for turn detection up to 30%.

Bounded Rationality in Byzantine Sensors Under Attacks

In this article, we investigate the behavior of agents with bounded rationality, attacking a set of stochastic sensors measuring the state of a binary event. The coordination problem between the attackers is formulated as a multiplayer non-zero-sum one-shot game. The objective of each attacker is to maximize the probability that a certain detector will produce an erroneous estimate of the true event, while they remain stealthy. To better predict the outcome of this game, we categorize the players based on the number of strategic thinking steps they will take. Each level- <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$k$</tex-math></inline-formula> attacker behaves based on subjective beliefs of the others’ behaviors, which are quantified via a Poisson distribution over the lower levels. The expected best responses for the type of each attacker are derived. The limiting conditions as the cognitive levels increase, as well as when the attackers fully coordinate, are shown to converge to the Nash equilibrium.

An imaging flow cytometry-based methodology for the analysis of single extracellular vesicles in unprocessed human plasma

Extracellular vesicles (EVs) are tissue-specific particles released by cells containing valuable diagnostic information in the form of various biomolecules. To rule out selection bias or introduction of artefacts caused by EV isolation techniques, we present a clinically feasible, imaging flow cytometry (IFCM)–based methodology to phenotype and determine the concentration of EVs with a diameter ≤400 nm in human platelet-poor plasma (PPP) without prior isolation of EVs. Instrument calibration (both size and fluorescence) were performed with commercial polystyrene beads. Detergent treatment of EVs was performed to discriminate true vesicular events from artefacts. Using a combination of markers (CFSE & Tetraspanins, or CD9 & CD31) we found that >90% of double-positive fluorescent events represented single EVs. Through this work, we provide a framework that will allow the application of IFCM for EV analysis in peripheral blood plasma in a plethora of experimental and potentially diagnostic settings. Additionally, this direct approach for EV analysis will enable researchers to explore corners of EVs as cellular messengers in healthy and pathological conditions.

The making of box-office collection: qualitative insights from Bollywood

PurposeThis study explores the making of box-office collection using the Indian film industry, Bollywood, as a case.Design/methodology/approachThis study conducts in-depth interviews with cinematic experts in the Indian film industry and analyzes the interview transcripts using thematic analysis.FindingsThis study uncovers several noteworthy findings. First, films that drew both general (MASS audience) and niche (CLASS audience) viewers dominate the box office. Second, viewers prefer to see films that are based on true events, and their engagement will be deeper if the subject of the film resonates with them. Third, stakeholder share is variable and changes over time. Fourth, the marketing budget for a film is typically higher than its production budget, and it is determined by the producer's financial resources. Fifth, the dominance of big over small banner films motivates the latter to pursue online rather than cinematic releases. Finally, Internet access creates value and returns on investment through sales of satellite and musical rights, while strategic promotion and distribution reap maximum benefit for box-office collection.Originality/valueUnlike past studies that rely on secondary data, this study uses primary qualitative data to explore the making of box-office collection. This study also focuses on an alternative film industry, Bollywood, as it is a vast context that remains underexplored.

Semi-supervised approach to event time annotation using longitudinal electronic health records.

Large clinical datasets derived from insurance claims and electronic health record (EHR) systems are valuable sources for precision medicine research. These datasets can be used to develop models for personalized prediction of risk or treatment response. Efficiently deriving prediction models using real world data, however, faces practical and methodological challenges. Precise information on important clinical outcomes such as time to cancer progression are not readily available in these databases. The true clinical event times typically cannot be approximated well based on simple extracts of billing or procedure codes. Whereas, annotating event times manually is time and resource prohibitive. In this paper, we propose a two-step semi-supervised multi-modal automated time annotation (MATA) method leveraging multi-dimensional longitudinal EHR encounter records. In step I, we employ a functional principal component analysis approach to estimate the underlying intensity functions based on observed point processes from the unlabeled patients. In step II, we fit a penalized proportional odds model to the event time outcomes with features derived in step I in the labeled data where the non-parametric baseline function is approximated using B-splines. Under regularity conditions, the resulting estimator of the feature effect vector is shown as root-n consistent. We demonstrate the superiority of our approach relative to existing approaches through simulations and a real data example on annotating lung cancer recurrence in an EHR cohort of lung cancer patients from Veteran Health Administration.