Integrated Genomic and Social Network Analyses of SARS-CoV-2 Transmission in the Healthcare Setting.

BACKGROUND: Infection prevention (IP) measures are designed to mitigate the transmission of pathogens in healthcare. Using large-scale viral genomic and social network analyses, we determined if IP measures used during the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic were adequate in protecting healthcare workers (HCWs) and patients from acquiring SARS-CoV-2. METHODS: We performed retrospective cross-sectional analyses of viral genomics from all available SARS-CoV-2 viral samples collected at UC San Diego Health and social network analysis using the electronic medical record to derive temporospatial overlap of infections among related viromes and supplemented with contact tracing data. The outcome measure was any instance of healthcare transmission, defined as cases with closely related viral genomes and epidemiological connection within the healthcare setting during the infection window. Between November 2020 through January 2022, 12 933 viral genomes were obtained from 35 666 patients and HCWs. RESULTS: Among 5112 SARS-CoV-2 viral samples sequenced from the second and third waves of SARS-CoV-2 (pre-Omicron), 291 pairs were derived from persons with a plausible healthcare overlap. Of these, 34 pairs (12%) were phylogenetically linked: 19 attributable to household and 14 to healthcare transmission. During the Omicron wave, 2106 contact pairs among 7821 sequences resulted in 120 (6%) related pairs among 32 clusters, of which 10 were consistent with healthcare transmission. Transmission was more likely to occur in shared spaces in the older hospital compared with the newer hospital (2.54 vs 0.63 transmission events per 1000 admissions, P < .001). CONCLUSIONS: IP strategies were effective at identifying and preventing healthcare SARS-CoV-2 transmission.

Nurse and Physician Perceptions and Decision Making During Interdisciplinary Communication: Factors That Influence Communication Channel Selection.

Errors in decision making and communication play a key role in poor patient outcomes. Safe patient care requires effective decision making during interdisciplinary communication through communication channels. Research on factors that influence nurse and physician decision making during interdisciplinary communication is limited. Understanding influences on nurse and physician decision making during communication channel selection is needed to support effective communication and improved patient outcomes. The purpose of the study was to explore nurse and physician perceptions of and decision-making processes for selecting interruptive or noninterruptive interdisciplinary communication channels in medical-surgical and intermediate acute care settings. Twenty-six participants (10 RNs, 10 resident physicians, and six attending physicians) participated in semistructured interviews in two acute care metropolitan hospitals for this qualitative descriptive study. The Practice Primed Decision Model guided interview question development and early data analysis. Findings include a core category, Development of Trust in the Communication Process, supported by three main themes: (1) Understanding of Patient Status Drives Communication Decision Making; (2) Previous Interdisciplinary Communication Experience Guides Channel Selection; and (3) Perceived Usefulness Influences Communication Channel Selection. Findings from this study provide support for future design and research of communication channels within the EHR and clinical decision support systems.

Implementation, Maintenance, and Outcomes of an Electronic Referral to a Tobacco Quitline Across Five Health Systems.

INTRODUCTION: Electronic referral (e-referral) to quitlines helps connect tobacco-using patients to free, evidence-based cessation counseling. Little has been published about the real-world implementation of e-referrals across U.S. health systems, their maintenance over time, and the outcomes of e-referred patients. AIMS AND METHODS: Beginning in 2014, the University of California (UC)-wide project called UC Quits scaled up quitline e-referrals and related modifications to clinical workflows from one to five UC health systems. Implementation strategies were used to increase site readiness. Maintenance was supported through ongoing monitoring and quality improvement programs. Data on e-referred patients (n = 20 709) and quitline callers (n = 197 377) were collected from April 2014 to March 2021. Analyses of referral trends and cessation outcomes were conducted in 2021-2022. RESULTS: Of 20 709 patients referred, the quitline contacted 47.1%, 20.6% completed intake, 15.2% requested counseling, and 10.9% received it. In the 1.5-year implementation phase, 1813 patients were referred. In the 5.5-year maintenance phase, volume was sustained, with 3436 referrals annually on average. Among referred patients completing intake (n = 4264), 46.2% were nonwhite, 58.8% had Medicaid, 58.7% had a chronic disease, and 48.8% had a behavioral health condition. In a sample randomly selected for follow-up, e-referred patients were as likely as general quitline callers to attempt quitting (68.5% vs. 71.4%; p = .23), quit for 30 days (28.3% vs. 26.9%; p = .52), and quit for 6 months (13.6% vs. 13.9%; p = .88). CONCLUSIONS: With a whole-systems approach, quitline e-referrals can be established and sustained across inpatient and outpatient settings with diverse patient populations. Cessation outcomes were similar to those of general quitline callers. IMPLICATIONS: This study supports the broad implementation of tobacco quitline e-referrals in health care. To the best of our knowledge, no other paper has described the implementation of e-referrals across multiple U.S. health systems or how they were sustained over time. Modifying electronic health records systems and clinical workflows to enable and encourage e-referrals, if implemented and maintained appropriately, can be expected to improve patient care, make it easier for clinicians to support patients in quitting, increase the proportion of patients using evidence-based treatment, provide data to assess progress on quality goals, and help meet reporting requirements for tobacco screening and prevention.

A Novel System to Improve Case Capture for Surgery Morbidity and Mortality Conference.

INTRODUCTION: Complications are often under-reported at surgical morbidity and mortality (M&M) conferences due to the sole reliance on voluntary case submission. While most institutions have databases used for targeted initiatives in quality improvement, these are not routinely used for M&M. We aimed to increase case capture for M&M conferences by developing a novel system that augments the existing case submission system with cases representing complications from quality improvement databases and the electronic health record (EHR). METHODS: We developed and implemented a novel system for increasing the capture rate of complications for M&M conferences by developing custom software that combines data from the following sources: an existing voluntary case submission system for M&M, local quality databases-National Surgical Quality Improvement Program and Vizient, and an EHR-based case capture tool. We evaluated this system on a retrospective cohort of all postoperative complications at a single center in a 32-mo period and in a prospective cohort over a 4-mo period after system implementation. RESULTS: In the retrospective cohort, we identified 433 complications among all data sources. Inclusion of the new system introduced 280 new potential cases for M&M review over the 32-mo period. After implementation, the system provided 31% of cases presented at M&M conference that would have otherwise been omitted. CONCLUSIONS: A novel system that includes complications identified in the EHR and quality improvement databases increased the case capture volume for surgical M&M conference, which provides an objective case referral system that can identify complementary quality improvement opportunities.

Advancing Diagnostic Safety Research: Results of a Systematic Research Priority Setting Exercise.

BACKGROUND: Diagnostic errors are a major source of preventable harm but the science of reducing them remains underdeveloped. OBJECTIVE: To identify and prioritize research questions to advance the field of diagnostic safety in the next 5 years. PARTICIPANTS: Ninety-seven researchers and 42 stakeholders were involved in the identification of the research priorities. DESIGN: We used systematic prioritization methods based on the Child Health and Nutrition Research Initiative (CHNRI) methodology. We first invited a large international group of expert researchers in various disciplines to submit research questions while considering five prioritization criteria: (1) usefulness, (2) answerability, (3) effectiveness, (4) potential for translation, and (5) maximal potential for effect on diagnostic safety. After consolidation, these questions were prioritized at an in-person expert meeting in April 2019. Top-ranked questions were subsequently reprioritized through scoring on the five prioritization criteria using an online questionnaire. We also invited non-research stakeholders to assign weights to the five criteria and then used these weights to adjust the final prioritization score for each question. KEY RESULTS: Of the 207 invited researchers, 97 researchers responded and 78 submitted 333 research questions which were then consolidated. Expert meeting participants (n = 21) discussed questions in different breakout sessions and prioritized 50, which were subsequently reduced to the top 20 using the online questionnaire. The top 20 questions addressed mostly system factors (e.g., implementation and evaluation of information technologies), teamwork factors (e.g., role of nurses and other health professionals in the diagnostic process), and strategies to engage patients in the diagnostic process. CONCLUSIONS: Top research priorities for advancing diagnostic safety in the short-term include strengthening systems and teams and engaging patients to support diagnosis. High-priority areas identified using these systematic methods can inform an actionable research agenda for reducing preventable diagnostic harm.

Frequency, Risk Factors, Causes, and Consequences of Diagnostic Errors in Critically Ill Medical Patients: A Retrospective Cohort Study.

OBJECTIVE: Diagnostic errors are a source of significant morbidity and mortality but understudied in the critically ill. We sought to characterize the frequency, causes, consequences, and risk factors of diagnostic errors among unplanned ICU admissions. DESIGN: We conducted a retrospective cohort study of randomly selected nonsurgical ICU admissions between July 2015 and June 2016. SETTING: Medical ICU at a tertiary academic medical center. SUBJECTS: Critically ill adults with unplanned admission to the medical ICU. MEASUREMENTS AND MAIN RESULTS: The primary investigator reviewed patient records using a modified version of the Safer Dx instrument, a validated instrument for detecting diagnostic error. Two intensivists performed secondary reviews of possible errors, and reviewers met periodically to adjudicate errors by consensus. For each confirmed error, we judged harm on a 1-6 rating scale. We also collected detailed demographic and clinical data for each patient. We analyzed 256 unplanned ICU admissions and identified 18 diagnostic errors (7% of admissions). All errors were associated with harm, and only six errors (33%) were recognized by the ICU team within the first 24 hours. More women than men experienced a diagnostic error (11.7% vs 2.7%; p = 0.015, χ test). On multivariable logistic regression analysis, female sex remained independently associated with risk of diagnostic error both at admission (odds ratio, 5.18; 95% CI, 1.34-20.08) and at 24 hours (odds ratio, 11.6; 95% CI, 1.37-98.6). Similarly, Quick Sequential Organ Failure Assessment score greater than or equal to 2 at admission was independently associated with diagnostic error (odds ratio, 5.73; 95% CI, 1.72-19.01). CONCLUSIONS: Diagnostic errors may be an underappreciated source of ICU-related harm. Women and higher acuity patients appear to be at increased risk for such errors. Further research is merited to define the scope of error-associated harm and to clarify risk factors for diagnostic errors among the critically ill.

An open access medical knowledge base for community driven diagnostic decision support system development.

INTRODUCTION: While early diagnostic decision support systems were built around knowledge bases, more recent systems employ machine learning to consume large amounts of health data. We argue curated knowledge bases will remain an important component of future diagnostic decision support systems by providing ground truth and facilitating explainable human-computer interaction, but that prototype development is hampered by the lack of freely available computable knowledge bases. METHODS: We constructed an open access knowledge base and evaluated its potential in the context of a prototype decision support system. We developed a modified set-covering algorithm to benchmark the performance of our knowledge base compared to existing platforms. Testing was based on case reports from selected literature and medical student preparatory material. RESULTS: The knowledge base contains over 2000 ICD-10 coded diseases and 450 RX-Norm coded medications, with over 8000 unique observations encoded as SNOMED or LOINC semantic terms. Using 117 medical cases, we found the accuracy of the knowledge base and test algorithm to be comparable to established diagnostic tools such as Isabel and DXplain. Our prototype, as well as DXplain, showed the correct answer as "best suggestion" in 33% of the cases. While we identified shortcomings during development and evaluation, we found the knowledge base to be a promising platform for decision support systems. CONCLUSION: We built and successfully evaluated an open access knowledge base to facilitate the development of new medical diagnostic assistants. This knowledge base can be expanded and curated by users and serve as a starting point to facilitate new technology development and system improvement in many contexts.

Provider views on childhood obesity management in primary care settings: a mixed methods analysis.

BACKGROUND: Pediatric providers are key players in the treatment of childhood obesity, yet rates of obesity management in the primary care setting are low. The goal of this study was to examine the views of pediatric providers on conducting obesity management in the primary care setting, and identify potential resources and care models that could facilitate delivery of this care. METHODS: A mixed methods approach was utilized. Four focus groups were conducted with providers from a large pediatric network in San Diego County. Based on a priori and emerging themes, a questionnaire was developed and administered to the larger group of providers in this network. RESULTS: Barriers to conducting obesity management fell into four categories: provider-level/individual (e.g., lack of knowledge and confidence), practice-based/systems-level (e.g., lack of time and resources), parent-level (e.g., poor motivation and follow-up), and environmental (e.g., lack of access to resources). Solutions centered around implementing a team approach to care (with case managers and health coaches) and electronic medical record changes to include best practice guidelines, increased ease of documentation, and delivery of standardized handouts/resources. Survey results revealed only 23.8% of providers wanted to conduct behavioral management of obesity. The most requested support was the introduction of a health educator in the office to deliver a brief behavioral intervention. CONCLUSION: While providers recognize the importance of addressing weight during a well-child visit, they do not want to conduct obesity management on their own. Future efforts to improve health outcomes for pediatric obesity should consider implementing a collaborative care approach.

Physician activity during outpatient visits and subjective workload.

We describe methods for capturing and analyzing EHR use and clinical workflow of physicians during outpatient encounters and relating activity to physicians' self-reported workload. We collected temporally-resolved activity data including audio, video, EHR activity, and eye-gaze along with post-visit assessments of workload. These data are then analyzed through a combination of manual content analysis and computational techniques to temporally align streams, providing a range of process measures of EHR usage, clinical workflow, and physician-patient communication. Data was collected from primary care and specialty clinics at the Veterans Administration San Diego Healthcare System and UCSD Health, who use Electronic Health Record (EHR) platforms, CPRS and Epic, respectively. Grouping visit activity by physician, site, specialty, and patient status enables rank-ordering activity factors by their correlation to physicians' subjective work-load as captured by NASA Task Load Index survey. We developed a coding scheme that enabled us to compare timing studies between CPRS and Epic and extract patient and visit complexity profiles. We identified similar patterns of EHR use and navigation at the 2 sites despite differences in functions, user interfaces and consequent coded representations. Both sites displayed similar proportions of EHR function use and navigation, and distribution of visit length, proportion of time physicians attended to EHRs (gaze), and subjective work-load as measured by the task load survey. We found that visit activity was highly variable across individual physicians, and the observed activity metrics ranged widely as correlates to subjective workload. We discuss implications of our study for methodology, clinical workflow and EHR redesign.

Predictive value of the present-on-admission indicator for hospital-acquired venous thromboembolism.

BACKGROUND: Hospital-acquired venous thromboembolic (HA-VTE) events are an important, preventable cause of morbidity and death, but accurately identifying HA-VTE events requires labor-intensive chart review. Administrative diagnosis codes and their associated "present-on-admission" (POA) indicator might allow automated identification of HA-VTE events, but only if VTE codes are accurately flagged "not present-on-admission" (POA=N). New codes were introduced in 2009 to improve accuracy. METHODS: We identified all medical patients with at least 1 VTE "other" discharge diagnosis code from 5 academic medical centers over a 24-month period. We then sampled, within each center, patients with VTE codes flagged POA=N or POA=U (insufficient documentation) and POA=Y or POA=W (timing clinically uncertain) and abstracted each chart to clarify VTE timing. All events that were not clearly POA were classified as HA-VTE. We then calculated predictive values of the POA=N/U flags for HA-VTE and the POA=Y/W flags for non-HA-VTE. RESULTS: Among 2070 cases with at least 1 "other" VTE code, we found 339 codes flagged POA=N/U and 1941 flagged POA=Y/W. Among 275 POA=N/U abstracted codes, 75.6% (95% CI, 70.1%-80.6%) were HA-VTE; among 291 POA=Y/W abstracted events, 73.5% (95% CI, 68.0%-78.5%) were non-HA-VTE. Extrapolating from this sample, we estimated that 59% of actual HA-VTE codes were incorrectly flagged POA=Y/W. POA indicator predictive values did not improve after new codes were introduced in 2009. CONCLUSIONS: The predictive value of VTE events flagged POA=N/U for HA-VTE was 75%. However, sole reliance on this flag may substantially underestimate the incidence of HA-VTE.

Impact of an automated patient outcome feedback system on emergency medicine resident patient follow-up: An interrupted time series analysis.

Objectives: Emergency medicine (EM) residents desire, but often lack, reliable feedback of patient outcomes following handoffs to other providers. This gap is a substantial barrier to calibrating their diagnostic decision making and learning. To address this educational priority, we developed and evaluated the Post-Handoff Reports of Outcomes (PHAROS) system-an automated system within our electronic health record (EHR) to deliver provider-specific patient outcome feedback. Methods: PHAROS includes: (1) individualized lists of patients seen and brief summaries of each case, (2) flags for important posthandoff events, and (3) links to charts to facilitate review. Starting June 2020, we coupled PHAROS with a resident educational session and individualized emails every 2 weeks outlining patients seen, number of posthandoff events, and instructions on how to access the PHAROS system. Results: From July 2017 through April 2022, we measured the proportion of handoffs followed by reaccessing patients' charts between 2 and 14 days posthandoff-a proxy for following up on the patient's outcomes. We performed an interrupted time series analysis on this outcome to determine if PHAROS was associated with a significant change in the trend of our outcome over time. Our secondary outcome was the number of times PHAROS was viewed each month. Our primary outcome had a significant increase in the slope over time (+0.13%/month, p = 0.03) after the introduction of the personalized reports and a nonsignificant change (-1.6%, p = 0.07) at the time of the intervention. The median (IQR) number of views of PHAROS per month was 33.2 (23.75-38.75). Conclusions: The PHAROS system was associated with a significant increase in the rate of posthandoff chart reaccess among EM residents over time. The PHAROS project demonstrated the feasibility of harnessing the capabilities of the EHR to create an automated system to support EM trainee feedback of patient outcomes-a key component of diagnostic calibration and learning.

An Automated System for Physician Trainee Procedure Logging via Electronic Health Records.

Importance: Procedural proficiency is a core competency for graduate medical education; however, procedural reporting often relies on manual workflows that are duplicative and generate data whose validity and accuracy are difficult to assess. Failure to accurately gather these data can impede learner progression, delay procedures, and negatively impact patient safety. Objective: To examine accuracy and procedure logging completeness of a system that extracts procedural data from an electronic health record system and uploads these data securely to an application used by many residency programs for accreditation. Design, Setting, and Participants: This quality improvement study of all emergency medicine resident physicians at University of California, San Diego Health was performed from May 23, 2023, to June 25, 2023. Exposures: Automated system for procedure data extraction and upload to a residency management software application. Main Outcomes and Measures: The number of procedures captured by the automated system when running silently compared with manually logged procedures in the same timeframe, as well as accuracy of the data upload. Results: Forty-seven residents participated in the initial silent assessment of the extraction component of the system. During a 1-year period (May 23, 2022, to May 7, 2023), 4291 procedures were manually logged by residents, compared with 7617 procedures captured by the automated system during the same period, representing a 78% increase. During assessment of the upload component of the system (May 8, 2023, to June 25, 2023), a total of 1353 procedures and patient encounters were evaluated, with the system operating with a sensitivity of 97.4%, specificity of 100%, and overall accuracy of 99.5%. Conclusions and Relevance: In this quality improvement study of emergency medicine resident physicians, an automated system demonstrated that reliance on self-reported procedure logging resulted in significant procedural underreporting compared with the use of data obtained at the point of performance. Additionally, this system afforded a degree of reliability and validity heretofore absent from the usual after-the-fact procedure logging workflows while using a novel application programming interface-based approach. To our knowledge, this system constitutes the first generalizable implementation of an automated solution to a problem that has existed in graduate medical education for decades.

Impact of a deep learning sepsis prediction model on quality of care and survival.

Sepsis remains a major cause of mortality and morbidity worldwide. Algorithms that assist with the early recognition of sepsis may improve outcomes, but relatively few studies have examined their impact on real-world patient outcomes. Our objective was to assess the impact of a deep-learning model (COMPOSER) for the early prediction of sepsis on patient outcomes. We completed a before-and-after quasi-experimental study at two distinct Emergency Departments (EDs) within the UC San Diego Health System. We included 6217 adult septic patients from 1/1/2021 through 4/30/2023. The exposure tested was a nurse-facing Best Practice Advisory (BPA) triggered by COMPOSER. In-hospital mortality, sepsis bundle compliance, 72-h change in sequential organ failure assessment (SOFA) score following sepsis onset, ICU-free days, and the number of ICU encounters were evaluated in the pre-intervention period (705 days) and the post-intervention period (145 days). The causal impact analysis was performed using a Bayesian structural time-series approach with confounder adjustments to assess the significance of the exposure at the 95% confidence level. The deployment of COMPOSER was significantly associated with a 1.9% absolute reduction (17% relative decrease) in in-hospital sepsis mortality (95% CI, 0.3%-3.5%), a 5.0% absolute increase (10% relative increase) in sepsis bundle compliance (95% CI, 2.4%-8.0%), and a 4% (95% CI, 1.1%-7.1%) reduction in 72-h SOFA change after sepsis onset in causal inference analysis. This study suggests that the deployment of COMPOSER for early prediction of sepsis was associated with a significant reduction in mortality and a significant increase in sepsis bundle compliance.

Developing a highly-reliable learning health system.

Multiple independent frameworks to support continuous improvement have been proposed to guide healthcare organizations. Two of the most visible are High-reliability Health care, (Chassin et al., 2013) which is emphasized by The Joint Commission, and Learning Health Systems, (Institute of Medicine, 2011) highlighted by the National Academy of Medicine. We propose that organizations consider tightly linking these two models, creating a "Highly-reliable Learning Health System." We describe several efforts at our organization that has resulted from this combined model and have helped our organization weather the COVID-19 pandemic. The organizational changes created using this framework will enable our health system to support a culture of quality across our teams and better fulfill our tripartite mission of high-quality care, effective education of trainees, and dissemination of important innovations.

Physicians differ in their perceptions of sensitive medical records: Survey and interview study.

Physician categorizations of electronic health record (EHR) data (e.g., depression) into sensitive data categories (e.g., Mental Health) and their perspectives on the adequacy of the categories to classify medical record data were assessed. One thousand data items from patient EHR were classified by 20 physicians (10 psychiatrists paired with ten non-psychiatrist physicians) into data categories via a survey. Cluster-adjusted chi square tests and mixed models were used for analysis. 10 items were selected per each physician pair (100 items in total) for discussion during 20 follow-up interviews. Interviews were thematically analyzed. Survey item categorization yielded 500 (50.0%) agreements, 175 (17.5%) disagreements, 325 (32.5%) partial agreements. Categorization disagreements were associated with physician specialty and implied patient history. Non-psychiatrists selected significantly (p = .016) more data categories than psychiatrists when classifying data items. The endorsement of Mental Health and Substance Use categories were significantly (p = .001) related for both provider types. During thematic analysis, Encounter Diagnosis (100%), Problems (95%), Health Concerns (90%), and Medications (85%) were discussed the most when deciding the sensitivity of medical information. Most (90.0%) interview participants suggested adding additional data categories. Study findings may guide the evolution of digital patient-controlled granular data sharing technology and processes.

Impact of automated alerts on follow-up of post-discharge microbiology results: a cluster randomized controlled trial.

BACKGROUND: Failure to follow up microbiology results pending at the time of hospital discharge can delay diagnosis and treatment of important infections, harm patients, and increase the risk of litigation. Current systems to track pending tests are often inadequate. OBJECTIVE: To design, implement, and evaluate an automated system to improve follow-up of microbiology results that return after hospitalized patients are discharged. DESIGN: Cluster randomized controlled trial. SUBJECTS: Inpatient and outpatient physicians caring for adult patients hospitalized at a large academic hospital from February 2009 to June 2010 with positive and untreated or undertreated blood, urine, sputum, or cerebral spinal fluid cultures returning post-discharge. INTERVENTION: An automated e-mail-based system alerting inpatient and outpatient physicians to positive post-discharge culture results not adequately treated with an antibiotic at the time of discharge. MAIN MEASURES: Our primary outcome was documented follow-up of results within 3 days. Secondary outcomes included physician awareness and assessment of result urgency, impact on clinical assessments and plans, and preferred alerting scenarios. KEY RESULTS: We evaluated the follow-up of 157 post-discharge microbiology results from patients of 121 physicians. We found documented follow-up in 27/97 (28%) results in the intervention group and 8/60 (13%) in the control group [aOR 3.2, (95% CI 1.3-8.4); p=0.01]. Of all inpatient physician respondents, 32/82 (39%) were previously aware of the results, 45/77 (58%) felt the results changed their assessments and plans, 43/77 (56%) felt the results required urgent action, and 67/70 (96%) preferred alerts for current or broader scenarios. CONCLUSION: Our alerting system improved the proportion of important post-discharge microbiology results with documented follow-up, though the proportion remained low. The alerts were well received and may be expanded in the future.

An approach to improve LOINC mapping through augmentation of local test names.

Mapping medical test names into a standardized vocabulary is a prerequisite to sharing test-related data between health care entities. One major barrier in this process is the inability to describe tests in sufficient detail to assign the appropriate name in Logical Observation Identifiers, Names, and Codes (LOINC®). Approaches to address mapping of test names with incomplete information have not been well described. We developed a process of "enhancing" local test names by incorporating information required for LOINC mapping into the test names themselves. When using the Regenstrief LOINC Mapping Assistant (RELMA) we found that 73/198 (37%) of "enhanced" test names were successfully mapped to LOINC, compared to 41/191 (21%) of original names (p=0.001). Our approach led to a significantly higher proportion of test names with successful mapping to LOINC, but further efforts are required to achieve more satisfactory results.

Enhancing Diagnosis Through Technology: Decision Support, Artificial Intelligence, and Beyond.

Patient care in intensive care environments is complex, time-sensitive, and data-rich, factors that make these settings particularly well-suited to clinical decision support (CDS). A wide range of CDS interventions have been used in intensive care unit environments. The field needs well-designed studies to identify the most effective CDS approaches. Evolving artificial intelligence and machine learning models may reduce information-overload and enable teams to take better advantage of the large volume of patient data available to them. It is vital to effectively integrate new CDS into clinical workflows and to align closely with the cognitive processes of frontline clinicians.

Electronic health record machine learning model predicts trauma inpatient mortality in real time: A validation study.

INTRODUCTION: Patient outcome prediction models are underused in clinical practice because of lack of integration with real-time patient data. The electronic health record (EHR) has the ability to use machine learning (ML) to develop predictive models. While an EHR ML model has been developed to predict clinical deterioration, it has yet to be validated for use in trauma. We hypothesized that the Epic Deterioration Index (EDI) would predict mortality and unplanned intensive care unit (ICU) admission in trauma patients. METHODS: A retrospective analysis of a trauma registry was used to identify patients admitted to a level 1 trauma center for >24 hours from October 2019 to July 2020. We evaluated the performance of the EDI, which is constructed from 125 objective patient measures within the EHR, in predicting mortality and unplanned ICU admissions. We performed a 5 to 1 match on age because it is a major component of EDI, then examined the area under the receiver operating characteristic curve (AUROC), and benchmarked it against Injury Severity Score (ISS) and new injury severity score (NISS). RESULTS: The study cohort consisted of 1,325 patients admitted with a mean age of 52.5 years and 91% following blunt injury. The in-hospital mortality rate was 2%, and unplanned ICU admission rate was 2.6%. In predicting mortality, the maximum EDI within 24 hours of admission had an AUROC of 0.98 compared with 0.89 of ISS and 0.91 of NISS. For unplanned ICU admission, the EDI slope within 24 hours of ICU admission had a modest performance with an AUROC of 0.66. CONCLUSION: Epic Deterioration Index appears to perform strongly in predicting in-patient mortality similarly to ISS and NISS. In addition, it can be used to predict unplanned ICU admissions. This study helps validate the use of this real-time EHR ML-based tool, suggesting that EDI should be incorporated into the daily care of trauma patients. LEVEL OF EVIDENCE: Prognostic, level III.