Development Of Decision Support Tool Research Articles

Development of comprehensive decision support tools in distance learning quality management processes

The object of this study is the quality of distance learning. The need for procedures to assess the quality of this form of education was manifested most clearly in connection with the COVID-19 pandemic, wars, and other global problems, which predetermine the relevance of the study. The study considers the construction of a decision support model for assessing the quality of distance learning. Underlying the method is a combination of the method of expert assessments and the criterion model of data analysis, the basic method for analyzing the data obtained is the method of hierarchy analysis. Structural and functional schemes of the quality management system for distance learning are proposed. During the study, 10 criteria and 52 indicators were selected, and the weight of each indicator was calculated. Based on the weight values obtained, a scheme of the criteria model of decision support was built to assess the quality of distance learning. During the expert evaluation of the criteria and indicators, it was determined that the weight of indicators within the criterion ranges from 0.09953 to 0.34262. Such a difference in weight values indicates the optimality of the set of indicators within the criterion. Due to the combination of a criteria-based approach to data analysis in combination with the method of expert assessments, the model can be easily adapted for a point assessment of individual components and finding problem areas in the implementation of distance learning and management decision-making. The results of the study reported here may be of interest to both heads of educational institutions and employees of services involved in processing information about the organization and reporting for strategic decision-making.

Development and Utilization of a Diagnostic Support Tool for Asthma within the Electronic Medical Record.

Asthma in childhood is a common and costly chronic disease. Quality asthma care can lead to better control of asthma thus decreasing use of health services. The gold standard for pediatric asthma diagnosis and management is the National Heart, Lung and Blood Institute (NHLBI) guidelines for Diagnosis and Management of Asthma which center on precisely establishing the severity of asthma, as this precise classification delineates appropriate therapy. However, navigating these guidelines is a challenge for primary care providers that creates a barrier to providing quality care. We aim to improve precision in asthma severity classification in the community healthcare setting through the development of an electronic asthma decision support tool (eADST) incorporating NHLBI guidelines embedded within the electronic health record system. We developed an algorithm for the eADST to guide the health care provider to the appropriate classification and subsequent therapy. We engaged our health system's electronic health record informatics team and together developed and revised the tool. We launched the tool in three academic community clinics and measured precision in asthma classification in the twelve months prior to the availability of the tool and the twelve months following the launch. We found a significant improvement in precision of asthma severity classification following the launch, a necessary first step in improvement of asthma care. The next step will be to evaluate the impact of the tool on asthma outcomes.

A comparison of tree planting prioritization frameworks: i-Tree Landscape versus spatial decision support tool

Different models as well as planting prioritization and optimization schemes based upon diverse ecological, social, and economic goals and preferences have been used to develop more efficient and effective tree planting schemes. We compare tree planting prioritization scenarios identified from i-Tree Landscape’s priority planting index to optimal scenarios identified from a spatially explicit multi-objective decision support framework at the census block group level in the Bronx, NY. We explore four scenarios with varying objectives considering populations below the poverty line, avoided runoff, and PM2.5 air pollutant removal monetary benefits. Results show that when prioritizing single objectives (e.g., PM2.5 air pollutant removal) using the same per area of tree canopy benefits from the spatially distributed modeling of ecosystem services, the two approaches recommend similar block groups for additional tree cover. Scenarios considering multiple objectives, however, result in different optimal solutions, with the decision support framework generally recommending more block groups for increased tree cover than i-Tree Landscape’s methodology. When the per area of tree canopy benefits from i-Tree Landscape are used as input in the i-Tree Landscape prioritization scenarios and the spatially distributed benefits used in the decision support framework scenarios, different optimal solutions are identified between the two approaches across all four scenarios, with i-Tree Landscape recommending fewer block groups for increased tree cover. Such a comparison will help inform the development of flexible multi-objective decision support tools to guide future greening initiatives towards prioritizing planting locations that maximize multiple objectives, as well as areas to preserve urban forests.

The devil you know: A population-level analysis of tumor-based next-generation sequencing of high-grade epithelial ovarian cancer (151)

<b>Objectives:</b> Clinical use of tumor-based next-generation sequencing (tbNGS) is expanding, but the indications and implications remain vague. We aimed to assess the clinical relevance of tbNGS data in a large cohort of patients with high-grade epithelial ovarian cancer. <b>Methods:</b> Electronic medical records (both internal notes and external scanned documents) of 905 patients with high-grade serous or endometrioid ovarian cancer were queried for tbNGS data using optical character recognition (OCR) and natural language processing (NLP) technologies. Identified results and clinical-demographic information were entered into an institutional database. Data were analyzed with descriptive statistics and log-rank tests comparing selected genetic alterations to progression-free survival (PFS). <b>Results:</b> Three hundred and twelve (34.4%) patients had their tbNGS results identified. Tests used had variable breadth, ranging from 46 genes to whole-exome sequencing. The most common was an institutional panel that evaluated 146 genes. Nearly all (99.7%) patients had their genomic alterations (mutations or copy number variations) identified. On average, each tumor had 2.38 (±2.46) aberrations. Two hundred and fifty-two (80.8%) patients had actionable alterations, defined as relaying eligibility for targeted treatment or clinical trial, according to the Precision Oncology Decision Support System (PODSS) database. Twenty-seven (10.7% of those with actionable abnormalities) enrolled in relevant clinical trials. In total, 630 mutations were detected in the population, with half of these (58.6%) considered actionable. The most common type of abnormality encountered was missense substitution (52.9%), followed by amplification (14.7%) and a frameshift mutation (9.72%). The most frequently mutated genes were <i>TP53</i> (93.0%), <i>PIK3CA</i> (9.35%), <i>BRCA1/2</i> (8.30%), <i>NF1</i> (4.90%), and <i>KRAS</i> (4.52%). Among these, only <i>BRCA1/2</i> was associated with a significant improvement in PFS (20.1 vs 14.1 months, p=0.046). <i>CCNE1</i> amplification, the most prevalent copy number variant, was present in 9.4% of those tested. Specific mutation types or copy number variations were not correlated with PFS. <b>Conclusions:</b> Clinical use of tbNGS is prevalent in ovarian cancer. Of patients undergoing tbNGS at our institution, four out of five gained the opportunity for an FDA-approved treatment or genotypically relevant clinical trial. Description of population-level tumor genomics could help identify therapeutic targets and guide the development of clinical decision support tools.

A geography-based decision support tool to quantify the circular bioeconomy and financial performance in the forest-based sector (r.forcircular)

This study focuses on the development, implementation and application of a spatial-based decision support tool—named r.forcircular—aimed at quantifying both the level of the circular bioeconomy and the financial performance in the forest-based sector. The methodology merges a set of indicators based on the 4R framework (Reduce, Reuse, Recycle, Recover) of circular economy in a multicriteria approach. Outputs are computed for different scenarios and are calibrated based on variations in the selling price of raw materials and on higher mechanisation of production process phases (felling, processing, extraction and chipping of wood residues). The increase in wood assortment value leads to an improvement in the sustainability of the forest-wood supply chain in circular bioeconomy and financial terms. The application of a higher level of mechanisation seems to have conflicting results compared to those of other scenarios. The r.forcircular model was tested in an Italian case study (in the Municipality Union of Valdarno and Valdisieve in the Tuscany region, Italy) with the aim of understanding its applicability and replicability in other contexts. The results of the test showed that, in the study area, superior outcomes were observed for high forests than for coppices due to the low value of wood products obtainable from coppices.

Fuzzy Multi-Objective Chance-Constrained Portfolio Optimization Under Uncertainty Considering Investment Return, Investment Risk, and Sustainability

Effective portfolio management can increase business value by maximizing investment return, reducing possible risk and, making the best use of limited resources. Achieving this goal is constrained by multiple conflicting objectives and uncertainty, which are inherent in all projects. The investment return, investment risk, and sustainability have been simultaneously evaluated in this study by fuzzy multi-objective chance-constrained portfolio optimization, aligned with an organization’s strategic direction. Through the fuzzy chance-constrained programming, the obtained portfolio under uncertainties can be optimized with a specified confidence level imposed by decision makers. With such an NP-hard problem, a Genetic Algorithm (GA) is embedded in the developed decision support tool for sustainable investment portfolio selection and optimization. The applicability of the proposed approach and the constructed tool are illustrated through an example.

MSProDiscuss™ Clinical Decision Support Tool for Identifying Multiple Sclerosis Progression.

This article describes the rationale for the development of the MSProDiscuss™ clinical decision support (CDS) tool, its development, and insights into how it can help neurologists improve care for patients with multiple sclerosis (MS). MS is a progressive disease characterized by heterogeneous symptoms and variable disease course. There is growing consensus that MS exists on a continuum, with overlap between relapsing–remitting and secondary progressive phenotypes. Evidence demonstrates that neuroaxonal loss occurs from the outset, that progression can occur independent of relapse activity, and that continuous underlying pathological processes may not be reflected by inflammatory activity indicative of the patient’s immune response. Early intervention can benefit patients, and there is a need for a tool that assists physicians in rapidly identifying subtle signs of MS progression. MSProDiscuss, developed with physicians and patients, facilitates a structured approach to patient consultations. It analyzes multidimensional data via an algorithm to estimate the likelihood of progression (the MSProDiscuss score), the contribution of various symptoms, and the impact of symptoms on daily living, enabling a more personalized approach to treatment and disease management. Data from CDS tools such as MSProDiscuss offer new insights into disease course and facilitate informed decision-making and a holistic approach to MS patient care.

Usefulness, acceptation and feasibility of electronic medical history tool in reflux disease.

To investigate usefulness, feasibility, and patient satisfaction of an electronic pre-consultation medical history tool (EPMH) in laryngopharyngeal reflux (LPR) work-up. Seventy-five patients with LPR were invited to complete electronic medical history assessment prior to laryngology consultation. EPMH collected the following parameters: demographic and epidemiological data, medication, medical and surgical histories, diet habits, stress and symptom findings. Stress and symptoms were assessed with perceived stress scale and reflux symptom score. Duration of consultation, acceptance, and satisfaction of patients (feasibility, usefulness, effectiveness, understanding of questions) were evaluated through a 9-item patient-reported outcome questionnaire. Seventy patients completed the evaluation (93% participation rate). The mean age of cohort was 51.2 ± 15.6years old. There were 35 females and 35 males. Patients who refused to participate (N = 5) were > 65years old. The consultation duration was significantly lower in patients who used the EPMH (11.3 ± 2.7min) compared with a control group (18.1 ± 5.1min; p = 0.001). Ninety percent of patients were satisfied about EPMH easiness and usefulness, while 97.1% thought that EPMH may improve the disease management. Patients would recommend similar approach for otolaryngological or other specialty consultations in 98.6% and 92.8% of cases, respectively. The use of EPMH is associated with adequate usefulness, feasibility, and satisfaction outcomes in patients with LPR. This software is a preliminary step in the development of an AI-based diagnostic decision support tool to help laryngologists in their daily practice. Future randomized controlled studies are needed to investigate the gain of similar approaches on the traditional consultation format.

Developing a Multilevel Decision Support Tool for Urban Mobility

Decisions on transport policy measures have long-term and important impacts on the economy, environment and society. Transport policy measures can lock up capital for decades and cause manifold external effects. In order to allow policymakers to evaluate transport policies, the developed decision support tool facilitates the evaluation of the multidimensional impacts of the implementation of transport policies. The objective of the decision support toolset presented in this paper is to support transportation planning and design practices based on an integrated transportation analysis of the area of examination to determine the most applicable combination of mobility services. This paper provides a comprehensive description of the interactive decision support tool implemented to help cities and decision makers design their strategies and shape the urban mobility of the future.

Data-driven automated predictions of the avalanche danger level for dry-snow conditions in Switzerland

Abstract. Even today, the assessment of avalanche danger is by and large a subjective yet data-based decision-making process. Human experts analyse heterogeneous data volumes, diverse in scale, and conclude on the avalanche scenario based on their experience. Nowadays, modern machine learning methods and the rise in computing power in combination with physical snow cover modelling open up new possibilities for developing decision support tools for operational avalanche forecasting. Therefore, we developed a fully data-driven approach to assess the regional avalanche danger level, the key component in public avalanche forecasts, for dry-snow conditions in the Swiss Alps. Using a large data set of more than 20 years of meteorological data measured by a network of automated weather stations, which are located at the elevation of potential avalanche starting zones, and snow cover simulations driven with these input weather data, we trained two random forest (RF) classifiers. The first classifier (RF 1) was trained relying on the forecast danger levels published in the official Swiss avalanche bulletin. To reduce the uncertainty resulting from using the forecast danger level as target variable, we trained a second classifier (RF 2) that relies on a quality-controlled subset of danger level labels. We optimized the RF classifiers by selecting the best set of input features combining meteorological variables and features extracted from the simulated profiles. The accuracy of the models, i.e. the percentage of correct danger level predictions, ranged between 74 % and 76 % for RF 1 and between 72 % and 78 % for RF 2. We assessed the accuracy of forecasts with nowcast assessments of avalanche danger by well-trained observers. The performance of both models was similar to the agreement rate between forecast and nowcast assessments of the current experience-based Swiss avalanche forecasts (which is estimated to be 76 %). The models performed consistently well throughout the Swiss Alps, thus in different climatic regions, albeit with some regional differences. Our results suggest that the models may well have potential to become a valuable supplementary decision support tool for avalanche forecasters when assessing avalanche hazard.

Informatics Research on Mental Health Functioning: Decision Support for the Social Security Administration Disability Program.

The disability determination process of the Social Security Administration's (SSA's) disability program requires assessing work-related functioning for individual claimants alleging disability due to mental impairment. This task is particularly challenging because the determination process involves the review of a large file of information, including objective medical evidence and self-reports from claimants, families, and former employers. To improve this decision-making process, SSA entered an interagency agreement with the Rehabilitation Medicine Department, Epidemiology and Biostatistics Section, in the Clinical Center of the National Institutes of Health, intending to use data science and informatics to develop decision support tools. This collaborative effort over the past decade has led to the development of the Work Disability-Functional Assessment Battery and has initiated an approach to applying natural language processing to the review of claimants' files for information on mental health functioning. This informatics research collaboration holds promise for improving the process of disability determination for individuals with mental impairments who make claims at the SSA.

Clinical implications of tumor-based next-generation sequencing in ovarian cancer.

5545 Background: Epithelial ovarian cancer is genetically heterogeneous, both among and within histologic subtypes. Advances in next-generation sequencing have made it feasible to ascertain the somatic genetic signature of each patient, however, critical analysis of population-level sequencing results is required to maximize the potential of this technology. Here, we aimed to assess the clinical relevance of tumor-based next-generation sequencing (tbNGS) in a large cohort of patients with high-grade epithelial ovarian cancer. Methods: Our study population comprised patients with high-grade serous (n = 972), clear cell (n = 33), endometrioid (n = 28), mucinous (n = 4), and mixed (n = 34) or unspecified (n = 21) epithelial ovarian carcinoma diagnosed between April 2013 and September 2021. tbNGS results were identified within the electronic medical record using optical character recognition and natural language processing. Genetic, clinical, and demographic information was collected for patients who had undergone tbNGS. Progression-free survival (PFS) and overall survival (OS) were calculated from date of first treatment to date of first recurrence and date of death, respectively. Data were analyzed using descriptive statistics, univariate and multivariate Cox regression models, and clustering analyses. Results: Of 1092 patients in the described population, 409 (37.5%) had tbNGS results identified. Nearly all (96.1%) revealed one or more genetic aberrations. Most patients (74.6%) had an actionable mutation, defined as relaying eligibility for a targeted treatment or clinical trial. The most frequent alterations were TP53, PIK3CA, and NF1 mutations; and CCNE1 amplification. Ten different targeted institutional and commercial panels were employed, covering a range of 35 to 600+ gene loci. The median time from diagnosis to testing was 14.5 months, likely corresponding to time of recurrence. Though no standalone alterations were significantly related to survival, multivariate and clustering analyses identified several genetic patterns which corresponded to patient outcomes. Mutation of BRAF, PIK3R1, NOTCH3, MET, and/or ATR was correlated with shorter PFS (HR 1.84, p = 0.001); mutation of ATM, RB1, CDKN2A, FGFR1, and/or FGFR2 was associated with improved PFS (HR 0.64, p = 0.04), as was mutation of NBN and/or ATRX (HR 0.54, p &lt; 0.05). MYC, NOTCH3, and/or CREBBP mutations were significantly correlated with worse OS (HR 1.95, p = 0.02). In our population, 40 patients (9.78%) were enrolled in genotypically-relevant clinical trials. Conclusions: tbNGS is prevalent at our institution, and often yields actionable information. We identified several mutational patterns that correlate to patient survival. Detailed analysis of population-level tumor genomics may help to identify therapeutic targets and guide development of clinical decision support tools.

1132-P: Classification and Prediction of Diabetes Using Electronic Health Records and Wearable Devices for Clinical Decision Support

Introduction: Diagnosis of T2DM necessitates clinical tests that are time-consuming and expensive. Machine learning (ML) techniques can accelerate the diagnosis and classification of T2DM and allow clinicians to personalize treatments based on blood glucose concentrations (BGC) , physical fitness (PF) , and diabetes distress patterns observed in daily life. Analyzing electronic health records (EHR) , physiological variables collected with wearable devices, and patient-reported outcomes (PROs) using ML techniques can lead to the development of clinical decision support tools that provide a comprehensive picture of an individual’s diabetes management needs. Methods: Clinical experimental data (n=85, F:40/M:45, HbA1c: 7.83±2.16; age: 57±7.72: BMI: 33.kg/m2±6.72; means and SDs) were used to identify clusters of subjects based on medical tests, and ML models were developed using readily measured data to classify new subjects to the identified clusters. Latent variable methods and k-means clustering were used to identify clusters based on HbA1c, physical performance tests, and PROs. ML models, including logistic regression (LR) and support vector machines (SVM) , were developed to assign new subjects to the identified clusters using the readily measured input variables from wearable devices, EHR and PROs. Results: Three distinct subject clusters were identified within the study cohort based on the descriptive variables. New subjects were assigned to the identified clusters with 87% and 91% accuracy for LR and SVM, respectively. Conclusion: EHRs, wearable device data, and PROs can be used to accurately and conveniently identify a person’s overall BGC, PF, and diabetes distress to aid in clinical decision-making. In the future, clinical decision support tools can be developed for personalized treatment suggestions based on cluster membership. Disclosure A.Shahidehpour: None. C.Fritschi: None. M.Rashid: None. A.Cinar: None. L.T.Quinn: n/a.

Clinical Reasoning among Registered Nurses in Emergency Medical Services: A Case Study

In emergency medical services (EMS), the clinical reasoning (CR) of registered nurses (RNs) working in ambulance care plays an important role in providing care and treatment that is timely, accurate, appropriate and safe. However, limited existing knowledge about how CR is formed and influenced by the EMS mission hinders the development of service provision and decision support tools for RNs that would further enhance patient safety. To explore the nature of CR and influencing factors in this context, an inductive case study examined 34 observed patient–RN encounters in an EMS setting focusing on ambulance care. The results reveal a fragmented CR approach involving several parallel decision-making processes grounded in and led by patients’ narratives. The findings indicate that RNs are not always aware of their own CR and associated influences until they actively reflect on the process, and additional research is needed to clarify this complex phenomenon.

A novel LOPCOW‐DOBI multi‐criteria sustainability performance assessment methodology: An application in developing country banking sector

Banks do not directly contribute to sustainable development, yet they indirectly play an undeniable role. The sustainability of the banking sector, which is the heart of finance, is crucial both for this sector and economy as well as for the realization of the 2030 sustainable development goals. Despite the growing interest of practitioners and scholars on this issue, previous works are rare for developing countries. With a novel multi-criteria framework developed, therefore, this paper assesses banks in terms of sustainability and clarifies their levels of corporate sustainability performances. For the first time, this work introduces an objective weighting method named LOgarithmic Percentage Change-driven Objective Weighting (LOPCOW) with the advantages of eliminating the gap due to the size of the data, generating more reasonable weightings, and considering both positive and negative data in the weighting process. Further, a new multi-criteria method named DOmbi Bonferroni (DOBI) is developed and utilized for deciding the ranking of the alternatives. The technique provides flexible decision-making, considering the risk attitude of decision-makers. The applicability of the proposed framework is demonstrated by a case study of Turkish banks. Banks are analyzed regarding the three pillars of sustainability and their seventeen drivers. The environmental pillar indicates the top significant impact on banks' sustainability. The “average return on shareholders' equity”, “electricity consumption”, “number of branches,” and “number of staff” are the top four key drivers for banks' sustainability. As a result of ranking the nine banks that release sustainability reports in Turkey concerning their sustainability performances, it is highlighted that Garanti BBVA outperforms the others. With a comprehensive sensitivity check, the robustness and applicability of the model are further tested. The developed decision support tool could allow different organizations to conduct sustainability assessments under the main pillars and help monitor their improvement over time. Some political and managerial implications are as well mentioned.

Combat-Related Invasive Fungal Wound Infections.

During Operation Enduring Freedom in Afghanistan, an outbreak of combat-related invasive fungal wound infections (IFIs) emerged among casualties with dismounted blast trauma and became a priority issue for the Military Health System. In 2011, the Trauma Infectious Disease Outcomes Study (TIDOS) team led the Department of Defense IFI outbreak investigation to describe characteristics of IFIs among combat casualties and provide recommendations related to management of the disease. To support the outbreak investigation, existing IFI definitions and classifications utilized for immunocompromised patients were modified for use in epidemiologic research in a trauma population. Following the conclusion of the outbreak investigation, multiple retrospective analyses using a population of 77 IFI patients (injured during June 2009 to August 2011) were conducted to evaluate IFI epidemiology, wound microbiology, and diagnostics to support refinement of Joint Trauma System (JTS) clinical practice guidelines. Following cessation of combat operations in Afghanistan, the TIDOS database was comprehensively reviewed to identify patients with laboratory evidence of a fungal infection and refine the IFI classification scheme to incorporate timing of laboratory fungal evidence and include categories that denote a high or low level of suspicion for IFI. The refined IFI classification scheme was utilized in a large-scale epidemiologic assessment of casualties injured over a 5.5-year period. Among 720 combat casualties admitted to participating hospitals (2009-2014) who had histopathology and/or wound cultures collected, 94 (13%) met criteria for an IFI and 61 (8%) were classified as high suspicion of IFI. Risk factors for development of combat-related IFIs include sustaining a dismounted blast injury, experiencing a traumatic transfemoral amputation, and requiring resuscitation with large-volume (>20 units) blood transfusions. Moreover, TIDOS analyses demonstrated the adverse impact of IFIs on wound healing, particularly with order Mucorales. A polymerase chain reaction (PCR)-based assay to identify filamentous fungi and support earlier IFI diagnosis was also assessed using archived formalin-fixed, paraffin-embedded tissue specimens. Although the PCR-based assay had high specificity (99%), there was low sensitivity (63%); however, sensitivity improved to 83% in tissues collected from sites with angioinvasion. Data obtained from the initial IFI outbreak investigation (37 IFI patients) and subsequent TIDOS analyses (77 IFI patients) supported development and refinement of a JTS clinical practice guideline for the management of IFIs in war wounds. Furthermore, a local clinical practice guideline to screen for early tissue-based evidence of IFIs among blast casualties at the Landstuhl Regional Medical Center was critically evaluated through a TIDOS investigation, providing additional clinical practice support. Through a collaboration with the Uniformed Services University Surgical Critical Care Initiative, findings from TIDOS analyses were used to support development of a clinical decision support tool to facilitate early risk stratification. Combat-related IFIs are a highly morbid complication following severe blast trauma and remain a threat for future modern warfare. Our findings have supported JTS clinical recommendations, refined IFI classification, and confirmed the utility of PCR-based assays as a complement to histopathology and/or culture to promote early diagnosis. Analyses underway or planned will add to the knowledge base of IFI epidemiology, diagnostics, prevention, and management.

Design, adaptation, and diffusion of an innovative tool to support contraceptive decision-making: Balanced Counseling Strategy Plus

Contraceptive decision-making is highly complex, and family planning (FP) clients choose methods according to a host of personal, interpersonal, and context-specific considerations. These include concerns about side effects, confidence in their ability to adhere to daily or monthly use, efficacy of methods, partner support, and HIV vulnerability. FP decision support tools can support FP clients and providers to engage in a joint decision-making process to ensure clients make informed choices about contraception. For more than two decades, the Balanced Counseling Strategy (BCS) and Balanced Counseling Strategy Plus (BCS+) decision support tools have been used in lower- and middle-income countries, informed by implementation science research and iterative updates and refinements to reflect new developments in contraceptive technology and HIV prevention strategies. To inform the development and scale-up of future FP decision support tools, this article describes the development, evaluation, and proliferation of BCS and BCS+.

Guiding Efficient, Effective, and Patient-Oriented Electrolyte Replacement in Critical Care: An Artificial Intelligence Reinforcement Learning Approach.

Both provider- and protocol-driven electrolyte replacement have been linked to the over-prescription of ubiquitous electrolytes. Here, we describe the development and retrospective validation of a data-driven clinical decision support tool that uses reinforcement learning (RL) algorithms to recommend patient-tailored electrolyte replacement policies for ICU patients. We used electronic health records (EHR) data that originated from two institutions (UPHS; MIMIC-IV). The tool uses a set of patient characteristics, such as their physiological and pharmacological state, a pre-defined set of possible repletion actions, and a set of clinical goals to present clinicians with a recommendation for the route and dose of an electrolyte. RL-driven electrolyte repletion substantially reduces the frequency of magnesium and potassium replacements (up to 60%), adjusts the timing of interventions in all three electrolytes considered (potassium, magnesium, and phosphate), and shifts them towards orally administered repletion over intravenous replacement. This shift in recommended treatment limits risk of the potentially harmful effects of over-repletion and implies monetary savings. Overall, the RL-driven electrolyte repletion recommendations reduce excess electrolyte replacements and improve the safety, precision, efficacy, and cost of each electrolyte repletion event, while showing robust performance across patient cohorts and hospital systems.

Clinical Decision Support Tool for Elevated Pediatric Blood Pressures.

Under-diagnosis of pediatric hypertension remains pervasive due to difficulty recognizing elevated systolic blood pressures (SBPs). We performed a retrospective review comparing recognition of and response to elevated SBPs ≥95th percentile before and after development of a clinical decision support tool (CDST) in an academic pediatric system. Of 44,351 encounters, 477 had elevated SBPs with documented recognition of an elevated SBP in 17.9% of encounters pre-CDST that increased to 33.7% post-CDST (P = .001). Post-CDST, 75.5% of elevated SBPs had repeat measurement, with 90.8% of initially elevated SBPs normalizing to <95th percentile. If repeat measurement was obtained and SBP remained elevated, documented recognition increased from 14.0 to 83.3% (P < .0001). These data support using the CDST is associated with increased identification of elevated SBPs in children with greatest improvements associated with repeat SBP measurement. This suggests targeted training and support systems at medical intake would be high yield for increasing recognition of elevated SBP.

Dynamic algorithmic conversion of compressed sward height to dry matter yield by a rising plate meter

The strategic allocation of pasture grazing area to dairy cows is essential for optimal management and increased outputs. Rising plate meters are frequently used to estimate pasture herbage mass, i.e. dry matter yield per hectare, by employing simple regression equations that relate compressed sward height to herbage mass. However, to improve the accuracy and precision of these equations, so that inherent variation of grasslands is captured, there is a need to incorporate differences in grass types and seasonal growth. Using a total of 308 grass plots, the variation of growth for both perennial ryegrass and hybrid ryegrass was recorded over the seven-month growing season, i.e. March–September. From these data three dynamic equations were derived. The models showed reduced levels of error in comparison to most other conventional equations. As such, the derived models represent a considerable advance for predictive assessment of herbage mass and will support more efficient grassland utilisation by farmers. Although all equations were found to be highly accurate and precise, only a single equation was considered the most effective (R2 = 0.7; RMSE = 248.05), allowing herbage mass to be predicted reliably from compressed sward height data in relation to ryegrass type and calendar month. Although further research will be required, the results presented allow farm operators to calculate herbage mass, as well as support the development of decision support tools to improve on-farm grassland management, particularly at the local paddock rather than national level.