Manual Data Entry Research Articles

Exciting Era of Sensor-Based Electronic Monitoring of Adherence in Pediatric Asthma.

* Abbreviations: ED — : emergency department RCT — : randomized controlled trial Adherence to daily inhaled corticosteroids and short-acting β-agonist medications is an important yet complicated component of asthma management. Despite the importance of daily preventive medication, up to 70% of pediatric patients with asthma are nonadherent to their prescribed medical treatments,1–14 which often leads to increased illness exacerbations, health care use, and missed days of school.15,16 With recent advancements in technology such as Bluetooth sensors, sensor-based monitoring of these medications has become increasingly accessible and popular. Most often, these monitors are paired with a mobile health app that provides some combination of self-monitoring; feedback; education about health, asthma, medications and triggers; encouragement; and the potential to data share with the medical team. In this issue of Pediatrics , Gupta et al17 conducted a randomized controlled trial (RCT) evaluating Propeller Health’s sensor-based electronic monitoring system (Propeller Health, Madison, WI). Findings suggest that sensor-based inhaler monitoring with clinical feedback may improve asthma symptom control and caregiver quality of life within diverse populations. It should be noted, however, that although a statistically significant improvement in asthma symptom control was demonstrated, this change did not meet the threshold for a clinically meaningful difference.18 One interesting finding of this study is the increase in health care use in … Address correspondence to Theresa W. Guilbert, MD, MS, Division of Pulmonary Medicine, Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave, MLC 7041, Cincinnati, OH 45229. E-mail: theresa.guilbert{at}cchmc.org

Automatic Detection of Road Types From the Third Military Mapping Survey of Austria-Hungary Historical Map Series With Deep Convolutional Neural Networks

With the increased amount of digitized historical documents, information extraction from them gains pace. Historical maps contain valuable information about historical, geographical and economic aspects of an era. Retrieving information from historical maps is more challenging than processing modern maps due to lower image quality, degradation of documents and the massive amount of non-annotated digital map archives. Convolutional Neural Networks (CNN) solved many image processing challenges with great success, but they require a vast amount of annotated data. For historical maps, this means an unprecedented scale of manual data entry and annotation. In this study, we first manually annotated the Third Military Mapping Survey of Austria-Hungary historical map series conducted between 1884 and 1918 and made them publicly accessible. We recognized different road types and their pixel-wise positions automatically by using a CNN architecture and achieved promising results.

825. An Academic-Information Technology Partnership to Create an Infectious Diseases Translational Science Database

BackgroundTranslational science is the process of turning observations in the laboratory, clinic, and community into interventions that improve human health. The coordinated effort to maintain integrated, validated laboratory and clinical data is often a rate-limiting step for research laboratories, especially for multi-site studies. Previous research shows a rate of error between 2.3 and 5.2% for basic data collection in clinical databases, up to 26.9% for more complex data points. The purpose of this project was to create a translational science database prototype that would be responsive to the unmet needs of the translational research community.MethodsTranslational scientists, IT experts, and lab technicians mapped the workflow of a high-throughput research laboratory including clinical and laboratory data. Database goals were to develop processes that would minimize data entry time, avoid redundancies, and validate data in a secure environment (HIPAA-compliant). Unique to this platform was the ability to map creation of new samples (for example, PCR products) from parent samples (biologic samples). The platform was developed in an iterative process utilizing interviews, workflow study, analysis of supporting artifacts, and mock-ups.ResultsThe current prototype allows for electronic upload or manual data entry of clinical data. In a small controlled study we found the rate of error for basic data entry to be below 1% within it. Pre-populated data entry screens map laboratory work-flow with custom data entry fields produced based on laboratory results earlier in the work flow. Work-flow mapping includes microbiology, phenotypic descriptions (MIC), molecular biology (PCR), and customized experiments. Sequence data, housed separately, has data linkers stored in the database. The launch screen and data entry forms are populated based on specific criteria entered for each user.ConclusionThe Translational Science Database allows for efficient capture of high-quality data with baseline validation enabling seamless linking of translational data for single or multi-site laboratories. Future development work will expand the number of experiments and also incorporate stored biobank information into the database.DisclosuresJeffrey Beairsto, BSc Eng (ME), Populus (Employee, Shareholder) Randal Neptune, BSc., MSc., Populus Global Solutions (Employee) Beth Webster, BSc, MBA, Populus Global Solutions Inc (Employee, Shareholder) John N. Rutter, BscEng, Populus Global Solutions (Board Member, Employee, Shareholder) Tristan Rutter, BA, Populus (Employee) Kevin W. Garey, PharMD, MS, FASHP, Merck & Co. (Grant/Research Support, Scientific Research Study Investigator)

DESIGN OF SYSTEM FOR MORE EFFECTIVE DATA ACTUALISATION IN SIMULATION

The main object of the article is a design system for faster data actualisation in Tecnomatix Plant Simulation software through the AttributeExplorer function and Microsoft Excel. Many companies that use simulation as the optimisation tool is meet the problem of the models up to date and actualisation time demands. The designed system that uses the AttributeExplorer function of the Tecnomatix Plant Simulation reduces the actualisation time demands for activities. The end part of the article contains a comparison of the normal manual data entry and time demands for using the designed system.

Assessment of Electronic Health Record Use Between US and Non-US Health Systems

Understanding how the electronic health record (EHR) system changes clinician work, productivity, and well-being is critical. Little is known regarding global variation in patterns of use. To provide insights into which EHR activities clinicians spend their time doing, the EHR tools they use, the system messages they receive, and the amount of time they spend using the EHR after hours. This cross-sectional study analyzed the deidentified metadata of ambulatory care health systems in the US, Canada, Northern Europe, Western Europe, the Middle East, and Oceania from January 1, 2019, to August 31, 2019. All of these organizations used the EHR software from Epic Systems and represented most of Epic Systems's ambulatory customer base. The sample included all clinicians with scheduled patient appointments, such as physicians and advanced practice practitioners. Clinician EHR use was tracked by deidentified and aggregated metadata across a variety of clinical activities. Descriptive statistics for clinician EHR use included time spent on clinical activities, note documentation (as measured by the percentage of characters in the note generated by automated or manual data entry source), messages received, and time spent after hours. A total of 371 health systems were included in the sample, of which 348 (93.8%) were located in the US and 23 (6.2%) were located in other countries. US clinicians spent more time per day actively using the EHR compared with non-US clinicians (mean time, 90.2 minutes vs 59.1 minutes; P < .001). In addition, US clinicians vs non-US clinicians spent significantly more time performing 4 clinical activities: notes (40.7 minutes vs 30.7 minutes; P < .001), orders (19.5 minutes vs 8.75 minutes; P < .001), in-basket messages (12.5 minutes vs 4.80 minutes; P < .001), and clinical review (17.6 minutes vs 14.8 minutes; P = .01). Clinicians in the US composed more automated note text than their non-US counterparts (77.5% vs 60.8% of note text; P < .001) and received statistically significantly more messages per day (33.8 vs 12.8; P < .001). Furthermore, US clinicians used the EHR for a longer time after hours, logging in 26.5 minutes per day vs 19.5 minutes per day for non-US clinicians (P = .01). The median US clinician spent as much time actively using the EHR per day (90.1 minutes) as a non-US clinician in the 99th percentile of active EHR use time per day (90.7 minutes) in the sample. These results persisted after controlling for organizational characteristics, including structure, type, size, and daily patient volume. This study found that US clinicians compared with their non-US counterparts spent substantially more time actively using the EHR for a wide range of clinical activities or tasks. This finding suggests that US clinicians have a greater EHR burden that may be associated with nontechnical factors, which policy makers and health system leaders should consider when addressing clinician wellness.

AI technology for remote clinical assessment and monitoring.

To report the clinical validation of an innovative, artificial intelligence (AI)-powered, portable and non-invasive medical device called Wound Viewer. The AI medical device uses dedicated sensors and AI algorithms to remotely collect objective and precise clinical data, including three-dimensional (3D) wound measurements, tissue composition and wound classification through the internationally recognised Wound Bed Preparation (WBP) protocol; this data can then be shared through a secure General Data Protection Regulation (GDPR)- and Health Insurance Portability and Accountability Act (HIPAA)-compliant data transfer system. This trial aims to test the reliability and precision of the AI medical device and its ability to aid health professionals in clinically evaluating wounds as efficiently remotely as at the bedside. This non-randomised comparative clinical trial was conducted in the Clinica San Luca (Turin, Italy). Patients were divided into three groups: (i) patients with venous and arterial ulcers in the lower limbs; (ii) patients with diabetes and presenting with diabetic foot syndrome; and (iii) patients with pressure ulcers. Each wound was evaluated for area, depth, volume and WBP wound classification. Each patient was examined once and the results, analysed by the AI medical device, were compared against data obtained following visual evaluation by the physician and research team. The area and depth were compared with a Kruskal-Wallis one-way analysis of variations in the obtained distribution (expected p-value>0.1 for both tests). The WBP classification and tissue segmentation were analysed by directly comparing the classification obtained by the AI medical device against that of the testing physician. A total of 150 patients took part in the trial. The results demonstrated that the AI medical device's AI algorithm could acquire objective clinical parameters in a completely automated manner. The AI medical device reached 97% accuracy against the WBP classification and tissue segmentation analysis compared with that performed in person by the physician. Moreover, data regarding the measurements of the wounds, as analysed through the Kruskal-Wallis technique, showed that the data distribution proved comparable with the other methods of measurement previously clinically validated in the literature (p=0.9). These findings indicate that remote wound assessment undertaken by physicians is as effective through the AI medical device as bedside examination, and that the device was able to assess wounds and provide a precise WBP wound classification. Furthermore, there was no need for manual data entry, thereby reducing the risk of human error while preserving high-quality clinical diagnostic data.

Data entry quality of double data entry vs automated form processing technologies: A cohort study validation of optical mark recognition and intelligent character recognition in a clinical setting.

Background and AimsPatient‐reported outcome measures (PROMs) are increasingly used in health services. Paper forms are still often used to register such data. Manual double data entry (DDE) has been defined as the gold standard for transferring data to an electronic format but is laborious and costly. Automated form processing (AFP) is an alternative, but validation in a clinical context is warranted. The study objective was to examine and validate a local hospital AFP setup.MethodsPatients over 18 years of age who were scheduled for knee or hip replacement at Stavanger University Hospital from 2014 to 2017 who answered PROMs were included in the study and contributed PROM data. All paper PROMs were scanned using the AFP techniques of optical mark recognition (OMR) and intelligent character recognition (ICR) and were processed by DDE by health secretaries using a data entry program. OMR and ICR were used to capture different types of data. The main outcome was the proportion of correctly entered numbers, defined as the same response recorded in AFP and DDE or by consulting the original paper questionnaire at the data field, item, and PROM level.ResultsA total of 448 questionnaires from 255 patients were analyzed. There was no statistically significant difference in error proportions per 10 000 data fields between OMR and DDE for data from check boxes (3.52 95% confidence interval (CI) 2.17 to 5.72 and 4.18 (95% CI 2.68‐6.53), respectively P = .61). The error proportion for ICR (nine errors) was statistically significantly higher than that for DDE (two errors), that is, 3.53 (95% CI 1.87‐6.57) vs 0.78 (95% CI 0.22‐2.81) per 100 data fields/items/questionnaires; P = .033. OMR (0.04% errors) outperformed ICR (3.51% errors; P < .001), Fisher's exact test.ConclusionsOMR can produce an error rate that is comparable to that of DDE. In our setup, ICR is still problematic and is highly dependent on manual validation. When AFP is used, data quality should be tested and documented.

Text-mining in electronic healthcare records can be used as efficient tool for screening and data collection in cardiovascular trials: a multicenter validation study

ObjectiveThis study aimed to validate trial patient eligibility screening and baseline data collection using text-mining in electronic healthcare records (EHRs), comparing the results to those of an international trial. Study Design and SettingIn three medical centers with different EHR vendors, EHR-based text-mining was used to automatically screen patients for trial eligibility and extract baseline data on nineteen characteristics. First, the yield of screening with automated EHR text-mining search was compared with manual screening by research personnel. Second, the accuracy of extracted baseline data by EHR text mining was compared to manual data entry by research personnel. ResultsOf the 92,466 patients visiting the out-patient cardiology departments, 568 (0.6%) were enrolled in the trial during its recruitment period using manual screening methods. Automated EHR data screening of all patients showed that the number of patients needed to screen could be reduced by 73,863 (79.9%). The remaining 18,603 (20.1%) contained 458 of the actual participants (82.4% of participants).In trial participants, automated EHR text-mining missed a median of 2.8% (Interquartile range [IQR] across all variables 0.4–8.5%) of all data points compared to manually collected data. The overall accuracy of automatically extracted data was 88.0% (IQR 84.7–92.8%). ConclusionAutomatically extracting data from EHRs using text-mining can be used to identify trial participants and to collect baseline information.

Development and Evaluation of an Automated Algorithm to Estimate the Nutrient Intake of Infants from an Electronic Complementary Food Frequency Questionnaire

Background: We previously validated a four-day complementary food frequency questionnaire (CFFQ) to estimate the nutrient intake in New Zealand infants aged 9-12 months. However, manual entry of the CFFQ data into nutritional analysis software was time-consuming. Therefore, we developed an automated algorithm and evaluated its accuracy by comparing the nutrient estimates with those obtained from the nutritional analysis software. Methods: We analysed 50 CFFQ completed at 9- and 12-months using Food Works nutritional analysis software. The automated algorithm was programmed in SAS by multiplying the average daily consumption of each food item by the nutrient content of the portion size. We considered the most common brands for commercially prepared baby foods. Intakes of energy, macronutrients, and micronutrients were compared between methods using Bland-Altman analysis. Results: The automated algorithm did not have any significant bias for estimates of energy (kJ) (MD 15, 95% CI -27, 58), carbohydrate (g) (MD -0.1, 95% CI -1.2,1.0), and fat (g) (-0.1, 95% CI -0.3,0.1), but slightly underestimated intake of protein (MD -0.4 g, 95% CI -0.7,-0.1), saturated fat, PUFA, dietary fibre, and niacin. The algorithm provided accurate estimates for other micronutrients. The limits of agreement were relatively narrow. Conclusion: This automated algorithm is an efficient tool to estimate the nutrient intakes from CFFQ accurately. The small negative bias observed for few nutrients was clinically insignificant and can be minimised. This algorithm is suitabl

Real-World Evidence of User Engagement With Mobile Health for Diabetes Management: Longitudinal Observational Study.

BackgroundPatient support apps have risen in popularity and provide novel opportunities for self-management of diabetes. Such apps offer patients to play an active role in monitoring their condition, thereby increasing their own treatment responsibility. Although many health apps require active user engagement to be effective, there is little evidence exploring engagement with mobile health (mHealth).ObjectiveThis study aims to analyze the extent to which users engage with mHealth for diabetes and identify patient characteristics that are associated with engagement.MethodsThe analysis is based on real-world data obtained by Novo Nordisk’s Cornerstones4Care Powered by Glooko diabetes support app. User engagement was assessed as the number of active days and using measures expressing the persistence, longevity, and regularity of interaction within the first 180 days of use. Beta regressions were estimated to assess the associations between user characteristics and engagement outcomes for each module of the app.ResultsA total of 9051 individuals initiated use after registration and could be observed for 180 days. Among these, 55.39% (5013/9051) used the app for one specific purpose. The average user activity ratio varied from 0.05 (medication and food) to 0.55 (continuous glucose monitoring), depending on the module of the app. Average user engagement was lower if modules required manual data entries, although the initial uptake was higher for these modules. Regression analyses further revealed that although more women used the app (2075/3649, 56.86%), they engaged significantly less with it. Older people and users who were recently diagnosed tended to use the app more actively.ConclusionsStrategies to increase or sustain the use of apps and availability of health data may target the mode of data collection and content design and should take into account privacy concerns of the users at the same time. Users’ engagement was determined by various user characteristics, indicating that particular patient groups should be targeted or assisted when integrating apps into the self-management of their disease.

Data Science &amp; Engineering into Food Science: A novel Big Data Platform for Low Molecular Weight Gelators’ Behavioral Analysis

The objective of this article is to introduce a comprehensiveend-to-end solution aimed at enabling the applicationof state-of-the-art Data Science and Analyticmethodologies to a food science related problem. Theproblem refers to the automation of load, homogenization,complex processing and real-time accessibility tolow molecular-weight gelators (LMWGs) data to gaininsights into their assembly behavior, i.e. whether agel can be mixed with an appropriate solvent or not.Most of the work within the field of Colloidal andFood Science in relation to LMWGs have centered onidentifying adequate solvents that can generate stablegels and evaluating how the LMWG characteristics canaffect gelation. As a result, extensive databases havebeen methodically and manually registered, storingresults from different laboratory experiments. Thecomplexity of those databases, and the errors causedby manual data entry, can interfere with the analysisand visualization of relations and patterns, limiting theutility of the experimental work.Due to the above mentioned, we have proposed ascalable and flexible Big Data solution to enable theunification, homogenization and availability of the datathrough the application of tools and methodologies.This approach contributes to optimize data acquisitionduring LMWG research and reduce redundant data processingand analysis, while also enabling researchersto explore a wider range of testing conditions and pushforward the frontier in Food Science research.

Evaluating Errors and Inefficiencies in Brachytherapy: An Approach Toward Process Redesign and Patient Safety

Despite extensive literature on patient safety (PS) and operating room crew resource management, the intersection between critical events during brachytherapy (BT) and fundamentals of PS are not well described. Radiation oncology (RO) quality and safety are commonly approached from a physics and dosimetry perspective, creating a need to better understand issues impacting quality of care and safety in BT. This study uses a mixed-methods approach to incident analysis of PS events and process improvement (PI) opportunities, reported during BT, to identify and quantify errors, inefficiencies and workflow interruptions. The purpose is to describe the current landscape and allow for multidisciplinary development of adaptive interventions toward more efficient and safer BT practices. Safety reports [SR: PS events] and glitch logs [GL: PI opportunities], submitted by multidisciplinary team members, were retrospectively reviewed from a prospectively collected database at a large, tertiary cancer center providing high-volume BT services. The study spanned from 9/2014 - 12/2019. Events were quantified by patterns of occurrence. Review of SR/GL narratives were done to qualitatively analyze events for recurrent themes. A total of 2,455 patients received 7,606 fractions between 2014 and 2019. There were 105 events, accounting for 4.28% of all patients and 1.38% of all fractions (none were state/NRC reportable). Four predominant recurring themes were seen (Table 1): (1) equipment concerns (inappropriate use, failure, missing/broken), (2) human factors (poor communication, handoffs, cross coverage, unclear roles/responsibilities, training/experience), (3) workflow inefficiencies (scheduling, coordination, new processes, transport, room turnover) and (4) electronic documentation (software, planning, manual data entry, information transfer). Human factors were the most common category of events documented: SR (42.9%) and GL (37.1%). Number of events reported was higher from July through December in every year (in aggregate, these represented 70% of SR and 60% of GL). During reporting, two safety systems were put into place from sentinel events: (1) a checklist to prevent retained foreign objects and (2) a team checklist for high-risk significant bleeding following interstitial implant removal. In-depth evaluation of errors and inefficiencies by multidisciplinary teams in BT may lead to better understanding and implementation of PS procedures and PI strategies for BT delivery. Multidisciplinary teams are critical in providing perspectives for addressing common workflow issues and innovative solutions for patient care, with the potential for implementation at other BT capable institutions.Abstract 2449; TableEvent Characteristics & Predominant Themes by Category NumberPredominant Themes in SR/GLCombined Events n = 105119 (18.1%) [11.3%,26.8%]243 (41.0%) [31.5%,51.0%]328 (26.7%) [18.5%,36.2%]415(14.3%) [8.2%-22.5%] Open table in a new tab

Deep Learning With Electronic Health Records for Short-Term Fracture Risk Identification: Crystal Bone Algorithm Development and Validation.

BackgroundFractures as a result of osteoporosis and low bone mass are common and give rise to significant clinical, personal, and economic burden. Even after a fracture occurs, high fracture risk remains widely underdiagnosed and undertreated. Common fracture risk assessment tools utilize a subset of clinical risk factors for prediction, and often require manual data entry. Furthermore, these tools predict risk over the long term and do not explicitly provide short-term risk estimates necessary to identify patients likely to experience a fracture in the next 1-2 years.ObjectiveThe goal of this study was to develop and evaluate an algorithm for the identification of patients at risk of fracture in a subsequent 1- to 2-year period. In order to address the aforementioned limitations of current prediction tools, this approach focused on a short-term timeframe, automated data entry, and the use of longitudinal data to inform the predictions.MethodsUsing retrospective electronic health record data from over 1,000,000 patients, we developed Crystal Bone, an algorithm that applies machine learning techniques from natural language processing to the temporal nature of patient histories to generate short-term fracture risk predictions. Similar to how language models predict the next word in a given sentence or the topic of a document, Crystal Bone predicts whether a patient’s future trajectory might contain a fracture event, or whether the signature of the patient’s journey is similar to that of a typical future fracture patient. A holdout set with 192,590 patients was used to validate accuracy. Experimental baseline models and human-level performance were used for comparison.ResultsThe model accurately predicted 1- to 2-year fracture risk for patients aged over 50 years (area under the receiver operating characteristics curve [AUROC] 0.81). These algorithms outperformed the experimental baselines (AUROC 0.67) and showed meaningful improvements when compared to retrospective approximation of human-level performance by correctly identifying 9649 of 13,765 (70%) at-risk patients who did not receive any preventative bone-health-related medical interventions from their physicians.ConclusionsThese findings indicate that it is possible to use a patient’s unique medical history as it changes over time to predict the risk of short-term fracture. Validating and applying such a tool within the health care system could enable automated and widespread prediction of this risk and may help with identification of patients at very high risk of fracture.

VALIDATION OF TROPICAL RAINFALL MEASURING MISSION (TRMM) DATA IN THE UPPER KAPUAS RIVER BASIN

Rainfall is a difficult parameter to measure, due to large spatial and temporal variations. Lack of data availability, data incompletely, less spreading of station, less observer, and manual data entry are other problems for rainfall predicting. To encourage these problems rainfall satellite can be used, because it has high temporal and spatial resolution, widely coverage, near real-time and fast accessibility. This research was conducted in the upper Kapuas River Basin, West Kalimantan, to determine how TRMM satellite-derived rainfall compares with ground-measured values and the possibility of using it to complement ground-measured rainfall. The statistical analyses and correction factor development for TRMM data are conducted to validate and correct the TRMM data on eleven sub basin in Kapuas River basin. Validation showed high correlation between TRMM and gauge data. Validation shows a high correlation and lowest RMSE between TRMM and gauge data in the sub basin adjacent to the gauge station (r= 0.76-0.8, RMSE 0,84-0,92). The results of the analysis also show that after correction, the corrected TRMM data errors were reduced for the eleven rainfall

Creating Hardware Component Knowledge Bases with Training Data Generation and Multi-task Learning

Hardware component databases are vital resources in designing embedded systems. Since creating these databases requires hundreds of thousands of hours of manual data entry, they are proprietary, limited in the data they provide, and have random data entry errors. We present a machine learning based approach for creating hardware component databases directly from datasheets. Extracting data directly from datasheets is challenging because: (1) the data is relational in nature and relies on non-local context, (2) the documents are filled with technical jargon, and (3) the datasheets are PDFs, a format that decouples visual locality from locality in the document. Addressing this complexity has traditionally relied on human input, making it costly to scale. Our approach uses a rich data model, weak supervision, data augmentation, and multi-task learning to create these knowledge bases in a matter of days. We evaluate the approach on datasheets of three types of components and achieve an average quality of 77 F1 points—quality comparable to existing human-curated knowledge bases. We perform application studies that demonstrate the extraction of multiple data modalities including numerical properties and images. We show how different sources of supervision such as heuristics and human labels have distinct advantages that can be utilized together to improve knowledge base quality. Finally, we present a case study to show how this approach changes the way practitioners create hardware component knowledge bases.

Development of a Novel Preoperative Venous Thromboembolism Risk Assessment Model

Standardization of preoperative venous thromboembolism (VTE) risk assessment remains challenging due to variation in risk assessment models (RAMs) and the cumbersome workflow addition that most RAMs represent. We aimed to develop a parsimonious RAM that is automatable and actionable within the preoperative workflow. We performed a case-controlled review of all 18 VTE cases reported over a 12-month period and 171 matched controls included in an institutional National Surgical Quality Improvement Project (NSQIP) data set. We examined the predictive value of the Caprini, Padua, and NSQIP RAMs. We identified the 5 most impactful risk factors in VTE development by contribution to the known RAMs. We compared the predictive ability of cancer, age, body mass index, black race, and American Society of Anesthesiologists Physical Status (ASA-PS) score, to the Caprini, Padua, and NSQIP RAMs for VTE outcomes. Finally, we evaluated concordance between each of the models. The Caprini Score was found to be 88.9% sensitive and 32.7% specific using a threshold of 5. The Padua score was found to be 61.1% sensitive and 47.4% specific using a threshold of 4. The novel 5-factor RAM was found to be 94.4% sensitive and 38.0% specific using a threshold of 4. The Caprini and Padua models were discordant in 26% of patients. Cumbersome manual data entry contributes to the ongoing challenge of standardized VTE risk assessment and prophylaxis. Universally documented information and patient demographics can be utilized to create clinical decision support tools that can improve the efficiency of perioperative workflow and improve the quality of care.

Decision analysis and reinforcement learning in surgical decision-making

BackgroundSurgical patients incur preventable harm from cognitive and judgment errors made under time constraints and uncertainty regarding patients’ diagnoses and predicted response to treatment. Decision analysis and techniques of reinforcement learning theoretically can mitigate these challenges but are poorly understood and rarely used clinically. This review seeks to promote an understanding of decision analysis and reinforcement learning by describing their use in the context of surgical decision-making. MethodsCochrane, EMBASE, and PubMed databases were searched from their inception to June 2019. Included were 41 articles about cognitive and diagnostic errors, decision-making, decision analysis, and machine-learning. The articles were assimilated into relevant categories according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews guidelines. ResultsRequirements for time-consuming manual data entry and crude representations of individual patients and clinical context compromise many traditional decision-support tools. Decision analysis methods for calculating probability thresholds can inform population-based recommendations that jointly consider risks, benefits, costs, and patient values but lack precision for individual patient-centered decisions. Reinforcement learning, a machine-learning method that mimics human learning, can use a large set of patient-specific input data to identify actions yielding the greatest probability of achieving a goal. This methodology follows a sequence of events with uncertain conditions, offering potential advantages for personalized, patient-centered decision-making. Clinical application would require secure integration of multiple data sources and attention to ethical considerations regarding liability for errors and individual patient preferences. ConclusionTraditional decision-support tools are ill-equipped to accommodate time constraints and uncertainty regarding diagnoses and the predicted response to treatment, both of which often impair surgical decision-making. Decision analysis and reinforcement learning have the potential to play complementary roles in delivering high-value surgical care through sound judgment and optimal decision-making.

Highlights from the clinical trials in organ transplantation (CTOT)-20 and CTOT-22 Consortium studies in lung transplant.

Long-term survival after lung transplant lags behind that of other commonly transplanted organs, reflecting the current incomplete understanding of the mechanisms involved in the development of posttransplant lung injury, rejection, infection, and chronic allograft dysfunction. To address this unmet need, 2 ongoing National Institute of Allergy and Infectious Disease funded studies through the Clinical Trials in Organ Transplant Consortium (CTOT) CTOT-20 and CTOT-22 were dedicated to understanding the clinical factors and biological mechanisms that drive chronic lung allograft dysfunction and those that maintain cytomegalovirus polyfunctional protective immunity. The CTOT-20 and CTOT-22 studies enrolled 800 lung transplant recipients at 5 North American centers over 3years. Given the number and complexity of subjects included, CTOT-20 and CTOT-22 utilized innovative data transfers and capitalized on patient-entered data collection to minimize site manual data entry. The data were coupled with an extensive biosample collection strategy that included DNA, RNA, plasma, serum, bronchoalveolar lavage fluid, and bronchoalveolar lavage cell pellet. This Special Article describes the CTOT-20 and CTOT-22 protocols, data and biosample strategy, initial results, and lessons learned through study execution.

Testing the Impact of Layered Health and Nutrition Interventions in Rajasthan, India: Baseline Findings on Birthweight

Abstract Objectives The Children's Investment Fund Foundation (CIFF) in partnership with the Government of Rajasthan seeks to reduce prevalence of low birthweight through multi-faceted interventions to strengthen existing delivery of services during pregnancy continuing into early childhood. We employed a quasi-experimental matched case design with matching at district, block, and sub-center level to enable a difference-in-difference estimation of impact of the interventions at endline. Methods We conducted a retrospective baseline survey of mothers with children under three years in 2018 to assess prevalence of low birthweight in six districts of Rajasthan. Primary household survey data was collected on 8583 mothers with children under the age of three years and 2406 currently pregnant women. Birthweight records of 5702 children under three years were collected from public facilities where the children were reported to be delivered by the mothers surveyed. Results Matching facility birthweight records of children born to mothers in the sample confirmed heaping of data, consistent with studies from developing countries that rely on administrative or secondary sources of birthweight records. We observed heaping of birthweight readings around multiples of 500 gms; 23% of the new-borns weighed 2.5 kgs and 19% weighed 3 kgs. Possible reasons for misreporting include malfunctioning or mis calibrated spring scales, manual data entry errors in the labor room registers, poor knowledge of optimal weighing practices and misclassification of low birthweight new-borns as normal birthweight. Conclusions Concerns with reliance on facility records for the key impact indicator have highlighted the need for standardisation of birthweight measurement at public delivery facilities. All public delivery facilities in the evaluation's treatment and control areas will receive digital weighing scales in labor rooms, along with training and mentoring of labor room staff based on a standardised protocol. This has implications for endline assessment of impact - average birthweight (continuous variable) will be used as the primary outcome for difference-in-difference impact estimation at endline. Changes in prevalence of low birthweight will only be tested using endline sample conditional on baseline equivalence. Funding Sources Children's Investment Fund Foundation.

The development an artificial intelligence algorithm for early sepsis diagnosis in the intensive care unit

BackgroundSevere sepsis and septic shock are still the leading causes of death in Intensive Care Units (ICUs), and timely diagnosis is crucial for treatment outcomes. The progression of electronic medical records (EMR) offers the possibility of storing a large quantity of clinical data that can facilitate the development of artificial intelligence (AI) in medicine. However, several difficulties, such as poor structure and heterogenicity of the raw EMR data, are encountered when introducing AI with ICU data. Labor-intensive work, including manual data entry, personal medical records sorting, and laboratory results interpretation may hinder the progress of AI. In this article, we introduce the developing of an AI algorithm designed for sepsis diagnosis using pre-selected features; and compare the performance of the AI algorithm with SOFA score based diagnostic method. Materials and methodsThis is a prospective open-label cohort study. A specialized EMR, named TED_ICU, was implemented for continuous data recording. One hundred six clinical features relevant to sepsis diagnosis were selected prospectively. A labeling work to allocate SEPSIS or NON_SEPSIS status for each ICU patient was performed by the in-charge intensivist according to SEPSIS-3 criteria, along with the automatic recording of selected features every day by TED_ICU. Afterward, we use de-identified data to develop the AI algorithm. Several machine learning methods were evaluated using 5-fold cross-validation, and XGBoost, a decision-tree based algorithm was adopted for our AI algorithm development due to best performance. ResultsThe study was conducted between August 2018 and December 2018 for the first stage of analysis. We collected 1588 instances, including 444 SEPSIS and 1144 NON-SEPSIS, from 434 patients. The 434 patients included 259 (59.6%) male patients and 175 female patients. The mean age was 67.6-year-old, and the mean APACHE II score was 13.8. The SEPSIS cohort had a higher SOFA score and increased use of organ support treatment. The AI algorithm was developed with a shuffle method using 80% of the instances for training and 20% for testing. The established AI algorithm achieved the following: accuracy = 82% ± 1%; sensitivity = 65% ± 5%; specificity = 88% ± 2%; precision = 67% ± 3%; and F1 = 0.66 ± 0.02. The area under the receiver operating characteristic curve (AUROC) was approximately 0.89. The SOFA score was used on the same 1588 instances for sepsis diagnosis, and the result was inferior to our AI algorithm (AUROC = 0.596). ConclusionUsing real-time data, collected by EMR, from the ICU daily practice, our AI algorithm established with pre-selected features and XGBoost can provide a timely diagnosis of sepsis with an accuracy greater than 80%. AI algorithm also outperforms the SOFA score in sepsis diagnosis and exhibits practicality as clinicians can deploy appropriate treatment earlier. The early and precise response of this AI algorithm will result in cost reduction, outcome improvement, and benefit for healthcare systems, medical staff, and patients as well.