Development Of Clinical Decision Support Research Articles

Advance Thermography diagnostics: automatic algorithm to find out acute sinusitis

Development of medical imaging techniques (including medical thermography) provided an opportunity to mine data that can be explored for the development of clinical decision support systems.Aim. Find out quantitative thermography criteria for acute rhinosinusitis and implement these criteria into automated analysis protocols to create a diagnostic complex suitable for use by professionals of any medical speciality.Material, methods. After necessary adaptation, facial thermography by mean of thermography camera ТВС300-меd LLC “СТК СИЛАР” Russia, (384x288 pixel, 30 mK) was performed in 100 healthy volunteers and 305 patients with suspected acute rhinosinusitis (in 173 of them diagnosis was supported). Resulted thermograms were processed in “cloud service” TVision (LLC “Dignosys”, Russia). Point and regions of interest (ROI), namely, point in the middle of the line, connected inner edge of eyebrows (Тср) and ROI, covered the projections of maxillary sinuses were marked automatically. Difference between temperature in Тср point and mean temperature in each ROI (ΔТr/ΔТl), as well as mean temperature between ROI were calculated (ΔTпаз).Result. The best results ( sensitivity 82%, specificity 74%, accuracy 78%) were received when complex of ΔТr/ΔТl ≤0 and module |ΔTпаз| ≥0,5°C (both or at least one criterion) was applied. Based on these data, automatic algorithm of thermography acute rhinosinusitis diagnostics was created to use in primary medical care routine as clinical decision support systems.

Deep representation learning identifies associations between physical activity and sleep patterns during pregnancy and prematurity

Preterm birth (PTB) is the leading cause of infant mortality globally. Research has focused on developing predictive models for PTB without prioritizing cost-effective interventions. Physical activity and sleep present unique opportunities for interventions in low- and middle-income populations (LMICs). However, objective measurement of physical activity and sleep remains challenging and self-reported metrics suffer from low-resolution and accuracy. In this study, we use physical activity data collected using a wearable device comprising over 181,944 h of data across N = 1083 patients. Using a new state-of-the art deep learning time-series classification architecture, we develop a ‘clock’ of healthy dynamics during pregnancy by using gestational age (GA) as a surrogate for progression of pregnancy. We also develop novel interpretability algorithms that integrate unsupervised clustering, model error analysis, feature attribution, and automated actigraphy analysis, allowing for model interpretation with respect to sleep, activity, and clinical variables. Our model performs significantly better than 7 other machine learning and AI methods for modeling the progression of pregnancy. We found that deviations from a normal ‘clock’ of physical activity and sleep changes during pregnancy are strongly associated with pregnancy outcomes. When our model underestimates GA, there are 0.52 fewer preterm births than expected (P = 1.01e − 67, permutation test) and when our model overestimates GA, there are 1.44 times (P = 2.82e − 39, permutation test) more preterm births than expected. Model error is negatively correlated with interdaily stability (P = 0.043, Spearman’s), indicating that our model assigns a more advanced GA when an individual’s daily rhythms are less precise. Supporting this, our model attributes higher importance to sleep periods in predicting higher-than-actual GA, relative to lower-than-actual GA (P = 1.01e − 21, Mann-Whitney U). Combining prediction and interpretability allows us to signal when activity behaviors alter the likelihood of preterm birth and advocates for the development of clinical decision support through passive monitoring and exercise habit and sleep recommendations, which can be easily implemented in LMICs.

Implementation of a Nurse-Driven Asthma Pathway in the Pediatric Intensive Care Unit.

Background:Asthma is one of the most common conditions requiring admission to a pediatric intensive care unit. Dosing and weaning medications, particularly bronchodilators, are highly variable, and evidence-based weaning algorithms for clinicians are lacking in this setting.Methods:Patients admitted to a quaternary pediatric intensive care unit diagnosed with acute severe asthma were evaluated for time spent receiving continuous albuterol therapy, the length of stay in the intensive unit care unit, and the length of stay in the hospital. We developed an asthma pathway and continuous bronchodilator weaning algorithm to be used by bedside nurses. We then implemented two major Plan-Do-Study-Act cycles to facilitate the use of the pathway. They included implementing the algorithm and then integrating it as a clinical decision support tool in the electronic medical record. We used standard statistics and quality improvement methodology to analyze results.Results:One-hundred twenty-six patients met inclusion criteria during the study period, with 32 during baseline collection, 60 after weaning algorithm development and implementation, and 34 after clinical decision support implementation. Using quality improvement methodology, hours spent receiving continuous albuterol decreased from a mean of 43.6 to 28.6 hours after clinical decision support development. There were no differences in length of stay using standard statistics and QI methodology.Conclusion:Protocolized asthma management in the intensive care unit setting utilizing a multidisciplinary approach and clinical decision support tools for bedside nursing can reduce time spent receiving continuous albuterol and may lead to improved patient outcomes.

Cancer specialists in the VA as early adopters of clinical genetic services.

11029 Background: Genetic testing has become essential to delivery of cancer treatment, risk assessment, surveillance, and prevention. We sought to understand the use of genetic tests by clinicians in the Department of Veterans Affairs (VA). Methods: We administered a web-based survey to clinicians at 20 VA facilities with precision oncology programs. We excluded respondents if they were: not at one of the 20 VA facilities; not seeing patients in VA; not a physician, nurse practitioner (NP), physician assistant (PA), or pharmacist; a medical geneticist or specialty was not reported; or if the survey was incomplete. Using multiple logistic regression, we assessed the association between genetic test ordering, genetics referral, and clinician characteristics. Results: There were 909 (909/11,442, 8%) eligible respondents with 61% women and 64% under age 55. There were 571 physicians (63%), 200 NPs (22%), 93 pharmacists (10%), and 45 PAs 5(%). There were 361 (40%) primary care providers (PCPs), 90 (10%) cancer specialists, and 458 (50%) non-cancer specialists. Only 21% of clinicians reported feeling prepared to use genetic tests in their practice. In the past year, only 8% had ordered at least one multi-gene cancer test (germline, tumor or both), 12% a pharmacogenetic test, and 0.2%, an exome. Compared to physicians, NPs were 60% less likely (OR = 0.42, 0.23-0.77, p = 0.005), pharmacists, 80% less likely (OR = 0.22, 0.08-0.62, p = 0.005), and PAs, 90% less likely (OR = 0.08, 0.01-0.58, p = 0.01) to have ordered a genetic test. Compared to PCPs, cancer specialists were almost 5 times more likely to order a genetic test (OR = 4.74, 2.57-8.73, p < 0.0001); there was no difference in genetic test ordering between PCPs and non-cancer specialists. Among clinicians (n = 72) who had ordered cancer genetic tests, only about two-thirds were confident in knowing the indications for testing; discussing the potential benefits, harms and limitations of testing; understanding the test report; and knowing implications of results on disease management and prevention. Clinicians (n = 106) who had ordered pharmacogenetic tests had lower frequencies of confidence in these tasks. About half (52%) of the cancer specialists had referred patients to genetics in the past year; they were 1.8 times more likely than PCPs to refer (OR = 1.82, 1.10-3.03, p = 0.02), and non-cancer specialists were about 50% less likely than PCPs to refer (OR = 0.46, 0.33-0.64, p < 0.0001). Conclusions: In the VA, cancer specialists are integrating genetic testing and genetics referral into their practice more than PCPs and other specialists. However, genetic testing is underutilized, and many clinicians remain unprepared to use genetic tests in their practice. These results will inform workforce planning, clinician education, and development of clinical decision support to facilitate genetic risk assessment, informed consent, and ordering of genetic tests.

P0539 / #801: DEVELOPING CLINICAL INFORMATION SYSTEM IN A PAEDIATRIC INTENSIVE CARE UNIT – SIMPLICITY AND EFFICIENCY MATTER FOR PHYSICIANS

Aims & Objectives: Clinical information system supports electronic health record, computerized physician order entry, clinical decision support, and data science application. We described its early development in a paediatric intensive care unit. Methods: A case study and qualitative analysis of a unit as part of a new children’s hospital. Period was from opening on 27th March to 31st December 2019. Physician’s team included 8 seniors and 2 rotating residents. Bedspaces expanded from 4 to 6. Metavision (iMDsoft) was used, and integrated with the medication ordering function of the hospital’s clinical management system. The feedbacks, monthly updates, and meeting notes were reviewed; and the process was summarised. Results: The first phase focused on fine-tuning with daily workflow. Effort to improve the comprehensiveness of physician order entries, and incorporate computerized entry into doctor’s routine. Graphical representation and trending was promoted to facilitate data interpretation. The second phase focused on addressing doctor’s acceptance, as cases grew in complexity. Entries that were inefficient were replaced by free-text. Frames were re-positioned to reduce time lag. Education was prioritised, as rotating residents showed low incentive to familiarize and were stressed by the system requirement. The third phase focused on improving quality of documentation, as the system was simplified. Redundant notes from “copy and paste” were discouraged. The next steps were: develop clinical decision support for specific conditions, and start automated data extraction for audits. Conclusions: A simple and efficient system for documentation and orders appeals to new users. However, clinical information system’s potentials are to be fully utilized through development of clinical decision support and data science application.

Multidisciplinary Teams are Essential for Developing Clinical Decision Support to Improve Pediatric Health Outcomes: An Exemplar

Clinical decision support with individualized patient education information can facilitate the translation of evidence-based guidelines into practice to improve pediatric patient outcomes. Interdisciplinary teams are required to develop and implement this technology support into practice. Engineering expertise with attention to three components is required: backend (e.g., data science, predictions), front end (e.g., user interface), and integration (e.g., workflow) must be addressed to achieve useful technology that will be adopted. An engineering framework, Technology Acceptance Model, can be used to guide the development of clinical decision support with patient education materials and includes a partnership with end users, both clinicians and patients.

To Share is Human! Advancing Evidence into Practice through a National Repository of Interoperable Clinical Decision Support.

Background Healthcare systems devote substantial resources to the development of clinical decision support (CDS) largely independently. The process of translating evidence-based practice into useful and effective CDS may be more efficient and less duplicative if healthcare systems shared knowledge about the translation, including workflow considerations, key assumptions made during the translation process, and technical details. Objective Describe how a national repository of CDS can serve as a public resource for healthcare systems, academic researchers, and informaticists seeking to share and reuse CDS knowledge resources or “artifacts.” Methods In 2016, the Agency for Healthcare Research and Quality (AHRQ) launched CDS Connect as a public, web-based platform for authoring and sharing CDS knowledge artifacts. Researchers evaluated early use and impact of the platform by collecting user experiences of AHRQ-sponsored and community-led dissemination efforts and through quantitative/qualitative analysis of site metrics. Efforts are ongoing to quantify efficiencies gained by healthcare systems that leverage shared, interoperable CDS artifacts rather than developing similar CDS de novo and in isolation. Results Federal agencies, academic institutions, and others have contributed over 50 entries to CDS Connect for sharing and dissemination. Analysis indicates shareable CDS resources reduce team sizes and the number of tasks and time required to design, develop, and deploy CDS. However, the platform needs further optimization to address sociotechnical challenges. Benefits of sharing include inspiring others to undertake similar CDS projects, identifying external collaborators, and improving CDS artifacts as a result of feedback. Organizations are adapting content available through the platform for continued research, innovation, and local implementations. Conclusion CDS Connect has provided a functional platform where CDS developers are actively sharing their work. CDS sharing may lead to improved implementation efficiency through numerous pathways, and further research is ongoing to quantify efficiencies gained.

User stories as lightweight requirements for agile clinical decision support development.

ObjectiveWe sought to demonstrate applicability of user stories, progressively elaborated by testable acceptance criteria, as lightweight requirements for agile development of clinical decision support (CDS).Materials and MethodsUser stories employed the template: As a [type of user], I want [some goal] so that [some reason]. From the “so that” section, CDS benefit measures were derived. Detailed acceptance criteria were elaborated through ensuing conversations. We estimated user story size with “story points,” and depicted multiple user stories with a use case diagram or feature breakdown structure. Large user stories were split to fit into 2-week iterations.ResultsOne example user story was: As a rheumatologist, I want to be advised if my patient with rheumatoid arthritis is not on a disease-modifying anti-rheumatic drug (DMARD), so that they receive optimal therapy and can experience symptom improvement. This yielded a process measure (DMARD use), and an outcome measure (Clinical Disease Activity Index). Following implementation, the DMARD nonuse rate decreased from 3.7% to 1.4%. Patients with a high Clinical Disease Activity Index improved from 13.7% to 7%. For a thromboembolism prevention CDS project, diagrams organized multiple user stories.DiscussionUser stories written in the clinician’s voice aid CDS governance and lead naturally to measures of CDS effectiveness. Estimation of relative story size helps plan CDS delivery dates. User stories prove to be practical even on larger projects.ConclusionsUser stories concisely communicate the who, what, and why of a CDS request, and serve as lightweight requirements for agile development to meet the demand for increasingly diverse CDS.

A Quality Improvement Initiative to Decrease Inappropriate Intravenous Acetaminophen Use at an Academic Medical Center.

Background: Following availability in the United States in 2011, intravenous acetaminophen (IV APAP) was added to many hospital formularies for multimodal pain control. In 2014, the price of IV APAP increased from $12/g to $33/g and became a top 10 medication expenditure at our institution. Objective: To promote appropriate IV APAP prescribing and reduce costs. Design, Setting, Participants: Quality improvement project at a 562-bed academic medical center involving all inpatient admissions from 2010 to 2017. Interventions: Using Plan-Do-Study-Act (PDSA) methodology, our Pharmacy & Therapeutics (P&T) committee aimed to reduce inappropriate use of IV APAP by refinement of restriction criteria, development of clinical decision support in the electronic medical record, education of clinical staff on appropriate use, and empowerment of hospital pharmacists to enforce restrictions. Measurements: Monthly IV APAP utilization and spending were assessed using statistical process control charts. Balancing measures included monthly usage of IV opioid, IV ketorolac, and oral ibuprofen. Results: Five PDSA cycles were conducted during the study period. Monthly spending on IV APAP decreased from the highest average of $56 038 per month to $5822 per month at study conclusion. Interventions resulted in an 80% annual cost savings, or an approximate savings of $600 000 per year. Usage of IV opioids, IV ketorolac, and oral ibuprofen showed no major changes during the study period. Conclusions: IV APAP can be restricted in a safe and cost effective manner without concomitant increase in IV opioid use.

Clinical Practice Guidelines for Pain Management in Acute Musculoskeletal Injury

We aimed to produce comprehensive guidelines and recommendations that can be utilized by orthopaedic practices as well as other specialties to improve the management of acute pain following musculoskeletal injury. A panel of 15 members with expertise in orthopaedic trauma, pain management, or both was convened to review the literature and develop recommendations on acute musculoskeletal pain management. The methods described by the Grading of Recommendations Assessment, Development, and Evaluation Working Group were applied to each recommendation. The guideline was submitted to the Orthopaedic Trauma Association (OTA) for review and was approved on October 16, 2018. We present evidence-based best practice recommendations and pain medication recommendations with the hope that they can be utilized by orthopaedic practices as well as other specialties to improve the management of acute pain following musculoskeletal injury. Recommendations are presented regarding pain management, cognitive strategies, physical strategies, strategies for patients on long term opioids at presentation, and system implementation strategies. We recommend the use of multimodal analgesia, prescribing the lowest effective immediate-release opioid for the shortest period possible, and considering regional anesthesia. We also recommend connecting patients to psychosocial interventions as indicated and considering anxiety reduction strategies such as aromatherapy. Finally, we also recommend physical strategies including ice, elevation, and transcutaneous electrical stimulation. Prescribing for patients on long term opioids at presentation should be limited to one prescriber. Both pain and sedation should be assessed regularly for inpatients with short, validated tools. Finally, the group supports querying the relevant regional and state prescription drug monitoring program, development of clinical decision support, opioid education efforts for prescribers and patients, and implementing a department or organization pain medication prescribing strategy or policy. Balancing comfort and patient safety following acute musculoskeletal injury is possible when utilizing a true multimodal approach including cognitive, physical, and pharmaceutical strategies. In this guideline, we attempt to provide practical, evidence-based guidance for clinicians in both the operative and non-operative settings to address acute pain from musculoskeletal injury. We also organized and graded the evidence to both support recommendations and identify gap areas for future research.

Monitoring clinical decision support in the electronic health record.

Hospitals utilize healthcare information technology such as electronic health records (EHRs) to optimize patient care and the development of clinical decision support (CDS) systems to further that goal. CDS systems aid the clinician with decision-making suggestions using relevant content at the

Drug-Drug Interactions Among Hospitalized Children Receiving Chronic Antiepileptic Drug Therapy.

Children treated with chronic medications are at risk of drug-drug interactions (DDIs) when hospitalized with an acute illness and prescribed new medications. We aimed to measure the prevalence of potential DDIs (pDDIs) among hospitalized children treated with antiepileptic drugs (AEDs) and to evaluate the impact of computerized physician order entry (CPOE) on pDDIs. We analyzed a national sample of pediatric hospitalizations from 2005 to 2012 associated with administration of an AED and identified those prescribed a second medication with risk of a DDI. The prevalence of hospitalizations associated with a pDDI was calculated for each AED. We identified the drugs most commonly implicated in pDDIs and factors associated with pDDIs. Rates of pDDIs were measured in pre- and post-CPOE implementation periods. A pDDI was identified in 181 380 (41.7%) hospitalizations associated with the use of an AED, with 117 880 (27.1%) classified as severe pDDIs. AEDs most often implicated with a pDDI were phenobarbital, valproic acid, and phenytoin. Hospitalizations with pDDIs were associated with increased length of stay and a greater number of medications, ICU admissions, and operating room procedures. The implementation of CPOE did not result in a change in the rate of pDDIs (42.7% before versus 40.8% after; P = .48). Children treated with AEDs are at risk of pDDIs while hospitalized. The use of CPOE has not been associated with a significant decrease in the rate of pDDIs. Additional investigation to better define the impact of pDDIs and to advance development of clinical decision support within CPOE systems is warranted.

Pharmacists critical to managing antimicrobials, developing infection control procedures

Infectious diseases kill more people worldwide than any other single cause, according to the U.S. National Library of Medicine. With Ebola raising its ugly head, combined with the mysterious enterovirus D68 and the fact that we're heading into cold and influenza season, hospital pharmacists continue to play key roles in infectious disease control, antimicrobial stewardship programs, and managing antibiotics.

Emerging roles for pharmacists in clinical implementation of pharmacogenomics.

Pharmacists are uniquely qualified to play essential roles in the clinical implementation of pharmacogenomics. However, specific responsibilities and resources needed for these roles have not been defined. We describe roles for pharmacists that emerged in the clinical implementation of genotype-guided clopidogrel therapy in the University of Florida Health Personalized Medicine Program, summarize preliminary program results, and discuss education, training, and resources needed to support such programs. Planning for University of Florida Health Personalized Medicine Program began in summer 2011 under leadership of a pharmacist, with clinical launch in June 2012 of a clopidogrel-CYP2C19 pilot project aimed at tailoring antiplatelet therapies for patients undergoing percutaneous coronary intervention and stent placement. More than 1000 patients were genotyped in the pilot project in year 1. Essential pharmacist roles and responsibilities that developed and/or emerged required expertise in pharmacy informatics (development of clinical decision support in the electronic medical record), medication safety, medication-use policies and processes, development of group and individual educational strategies, literature analysis, drug information, database management, patient care in targeted areas, logistical issues in genetic testing and follow-up, research and ethical issues, and clinical precepting. In the first 2 years of the program (1 year planning and 1 year postimplementation), a total of 14 different pharmacists were directly and indirectly involved, with effort levels ranging from a few hours per month, to 25-30% effort for the director and associate director, to nearly full-time for residents. Clinical pharmacists are well positioned to implement clinical pharmacogenomics programs, with expertise in pharmacokinetics, pharmacogenomics, informatics, and patient care. Education, training, and practice-based resources are needed to support these roles and to facilitate the development of financially sustainable pharmacist-led clinical pharmacogenomics practice models.

Natural and virtual environments for the training of emergency medicine personnel

EMERGENZA (emergency in Italian) is a `serious' game, developed in the context of the RIMSI project and designed to improve the training of emergency medicine operators. To achieve this goal, it adopts natural interaction paradigms in immersive environments. Virtual reality systems have been used recently in combination with natural interface systems for enhancing patients' rehabilitation procedures though are proving especially effective in the development of clinical decision support and medical training systems. The use of immersive simulations in medical training is extremely useful to confront emergency operators with scenarios that range from usual (e.g. unconscious person on the ground) to extreme (car accident with several injured people) without exposing the simulation participants in any harm. It is critical to exploit 3D virtual worlds in order to provide as much contextual information as possible to the operators. In fact, each emergency procedure needs to be adapted depending on the environmental threats and the presence of multiple injured people in need of assistance or bystanders. EMERGENZA allows to simulate a first-aid scenario with a configurable virtual environment using interactive 3D graphics. Users can interact through a natural interface for navigation and interaction with the virtual environment. In order to evaluate the prototype, several heuristics have been chosen and tested to measure the overall system usability. Results show that the adoption of natural interaction in immersive virtual environments receives good feedback from users.

A mathematical model for interpretable clinical decision support with applications in gynecology.

BackgroundOver time, methods for the development of clinical decision support (CDS) systems have evolved from interpretable and easy-to-use scoring systems to very complex and non-interpretable mathematical models. In order to accomplish effective decision support, CDS systems should provide information on how the model arrives at a certain decision. To address the issue of incompatibility between performance, interpretability and applicability of CDS systems, this paper proposes an innovative model structure, automatically leading to interpretable and easily applicable models. The resulting models can be used to guide clinicians when deciding upon the appropriate treatment, estimating patient-specific risks and to improve communication with patients.Methods and FindingsWe propose the interval coded scoring (ICS) system, which imposes that the effect of each variable on the estimated risk is constant within consecutive intervals. The number and position of the intervals are automatically obtained by solving an optimization problem, which additionally performs variable selection. The resulting model can be visualised by means of appealing scoring tables and color bars. ICS models can be used within software packages, in smartphone applications, or on paper, which is particularly useful for bedside medicine and home-monitoring. The ICS approach is illustrated on two gynecological problems: diagnosis of malignancy of ovarian tumors using a dataset containing 3,511 patients, and prediction of first trimester viability of pregnancies using a dataset of 1,435 women. Comparison of the performance of the ICS approach with a range of prediction models proposed in the literature illustrates the ability of ICS to combine optimal performance with the interpretability of simple scoring systems.ConclusionsThe ICS approach can improve patient-clinician communication and will provide additional insights in the importance and influence of available variables. Future challenges include extensions of the proposed methodology towards automated detection of interaction effects, multi-class decision support systems, prognosis and high-dimensional data.

Functional Data Analysis Applied to a Randomized Controlled Clinical Trial in Hemodialysis Patients Describes the Variability of Patient Responses in the Control of Renal Anemia

The achievement of desirable hemoglobin levels in renal anemia that is treated with epoetins is often incomplete and subject to much variation of outcome values and applied dosage. The further development of clinical decision support for renal anemia requires the characterization of patient responses and an analysis of the dynamics of the dosage and response variables. In this methodologic article, the extended data of a randomized, controlled clinical trial comparing two epoetins were examined by the techniques of functional data analysis to establish how precisely the patterns of treatment response might be described and analyzed. The description of the trajectory of hemoglobin values in each patient as a mathematical function allowed the characterization of individual responses, with a wide variety of patterns being revealed. An analysis of the degree of system control in the management of the anemia was then possible through phase plotting. The analysis also allowed an expression of the dynamic characteristics of the entire experimental system, analyzed in summary group waveforms with standard statistical properties. In addition, a quantification of the notional instability of patient responses enabled the determination of a subset of patients for whom control might be improved in a modified management system. It is concluded that functional data analysis does provide the basis for further characterization and experimental study of the control of renal anemia.