Can Automated Retrieval of Data from Emergency Department Physician Notes Enhance the Imaging Order Entry Process?
Ontology highlight
ABSTRACT:
SUBMITTER: Rousseau JF
PROVIDER: S-EPMC6426724 | biostudies-literature | 2019 Mar
REPOSITORIES: biostudies-literature
Publications
Can Automated Retrieval of Data from Emergency Department Physician Notes Enhance the Imaging Order Entry Process?
Rousseau Justin F JF Ip Ivan K IK Raja Ali S AS Valtchinov Vladimir I VI Cochon Laila L Schuur Jeremiah D JD Khorasani Ramin R
Applied clinical informatics 20190320 2
<h4>Background</h4>When a paucity of clinical information is communicated from ordering physicians to radiologists at the time of radiology order entry, suboptimal imaging interpretations and patient care may result.<h4>Objectives</h4>Compare documentation of relevant clinical information in electronic health record (EHR) provider note to computed tomography (CT) order requisition, prior to ordering of head CT for emergency department (ED) patients presenting with headache.<h4>Methods</h4>In thi ...[more]
Similar Datasets
Project description:BACKGROUND:Machine learning has been used extensively in clinical text classification tasks. Deep learning approaches using word embeddings have been recently gaining momentum in biomedical applications. In an effort to automate the identification of altered mental status (AMS) in emergency department provider notes for the purpose of decision support, we compare the performance of classic bag-of-words-based machine learning classifiers and novel deep learning approaches. METHODS:We used a case-control study design to extract an adequate number of clinical notes with AMS and non-AMS based on ICD codes. The notes were parsed to extract the history of present illness, which was used as the clinical text for the classifiers. The notes were manually labeled by clinicians. As a baseline for comparison, we tested several traditional bag-of-words based classifiers. We then tested several deep learning models using a convolutional neural network architecture with three different types of word embeddings, a pre-trained word2vec model and two models without pre-training but with different word embedding dimensions. RESULTS:We evaluated the models on 1130 labeled notes from the emergency department. The deep learning models had the best overall performance with an area under the ROC curve of 98.5% and an accuracy of 94.5%. Pre-training word embeddings on the unlabeled corpus reduced training iterations and had performance that was statistically no different than the other deep learning models. CONCLUSION:This supervised deep learning approach performs exceedingly well for the detection of AMS symptoms in clinical text in our environment. Further work is needed for the generalizability of these findings, including evaluation of these models in other types of clinical notes and other environments. The results seem promising for the ultimate use of these types of classifiers in combination with other information derived from the electronic health records as input for clinical decision support.
| S-EPMC6701023 | biostudies-literature
Project description:BackgroundIt is uncertain if computerized physician order entry (CPOE) systems are effective at reducing adverse drug event (ADE) rates in community hospitals, where mainly vendor-developed applications are used.ObjectiveTo evaluate the impact of vendor CPOE systems on the frequency of ADEs.Design and patientsProspective before-and-after study conducted from January 2005 to September 2010 at five Massachusetts community hospitals. Participants were adults admitted during the study period. A total of 2,000 charts were reviewed for orders, medication lists, laboratory reports, admission histories, notes, discharge summaries, and flow sheets.Main measuresThe primary outcome measure was the rate of preventable ADEs. Rates of potential ADEs and overall ADEs were secondary outcomes.Key resultsThe rate of preventable ADEs decreased following implementation (10.6/100 vs. 7.0/100 admissions; p = 0.007) with a similar effect observed at each site. However, the associated decrease in preventable ADEs was balanced against an increase in potential ADEs (44.4/100 vs. 57.5/100 admissions; p < 0.001). We observed a reduction of 34.0% for preventable ADEs, but an increase of 29.5% in potential ADEs following implementation. The overall rate of ADEs increased (14.6/100 vs. 18.7/100 admissions; p = 0.03), which was driven by non-preventable events (4.0/100 vs. 11.7/100 admissions; p < 0.001).ConclusionsAdoption of vendor CPOE systems was associated with a decrease in the preventable ADE rate by a third, although the rates of potential ADEs and overall ADEs increased. Our findings support the use of vendor CPOE systems as a means to reduce drug-related injury and harm. The potential ADE rate could be reduced by making refinements to the vendor applications and their associated decision support.
| S-EPMC3378729 | biostudies-literature
Project description:BackgroundFreestanding emergency departments (FrEDs) could reduce wait times in overcrowded emergency departments (EDs), but they might also increase usage and overall spending for emergency care. We investigate the relationship between the number of FrEDs entering a local market and overall spending on emergency care.MethodsWe accessed data from Arizona, Florida, North Carolina, and Texas in Blue Cross Blue Shield Axis; a limited data set of deidentified insurance data claims that we linked to Public Use Microdata Area (PUMA) data from the American Community Survey; and lists of licensed FrEDs from state agencies. Regression analysis was used to estimate the association between changes in the number of FrEDs in 495 PUMAs and total spending on emergency care, out-of-pocket spending, utilization, and price per visit from January 2013 to December 2017. Final estimates came from a PUMA-level fixed-effects model, with controls for state, quarter, and PUMA-level demographics.ResultsEntry of an additional FrED in a PUMA was associated with a 3.6 percentage point (pp; CI = 2.4 to 4.9) increase in emergency provider reimbursement per insured beneficiary in Texas, Florida, and North Carolina. There was no change in spending (2.5 pp; CI = -8.2 to 3.1) associated with a FrED's entry in Arizona. Entry of an additional FrED was associated with a 0.18 (CI = 0.12 to 0.23) increase in the number of emergency care visits per 100 enrollees in Texas, Florida, and Arizona. In contrast, entry of another FrED was not associated with a change in utilization (-0.03; CI = -0.09 to 0.02) in North Carolina. Estimated out-of-pocket payments for emergency care increased 3.6 pp (CI = 2.5 to 4.8) with the entry of a FrED in Texas, Florida, and Arizona, but declined by 15.3 pp (CI = -26.8 to -3.7) in North Carolina.ConclusionsRather than functioning as substitutes for hospital-based EDs, FrEDs have increased local market spending on emergency care in three of four states' markets where they have entered. State policy makers and researchers should carefully track spending and utilization of emergency care as FrEDs disseminate to better understand their potential health benefits and cost implications for patients.
| S-EPMC6899627 | biostudies-literature
Project description:A physician's recommendation is the most effective published method of motivating HPV vaccination initiation. The emergency department (ED) is the 'public health safety net', and often the only access to care for underserved populations. Recommendation of the HPV vaccine in the ED is a potential avenue to improve vaccination rates among sub-populations who do not have access to routine medical care. We assessed willingness of EM physicians to recommend the vaccine, target high-risk women, and disclose perceived barriers to vaccination in the ED. A cross sectional study using an 11-item survey, was used to assess physician attitudes toward recommending the HPV vaccine in an ED setting to age eligible patients. 67.4% stated they would recommend the vaccine, 23.9% were neutral, and 8.7% would not recommend the vaccine to age eligible patients in the ED. 41% noted lack of adequate reimbursement for vaccination as a barrier to vaccination in the ED (P<0.05). Physicians were comfortable targeting women at high risk for cervical cancer for vaccination (P<0.05). EM physicians are comfortable targeting high-risk women for HPV vaccination in an ED setting. Support of EM physicians in the national effort to improve HPV vaccine uptake is an important step in eradicating a largely preventable yet lethal cancer.
| S-EPMC4977447 | biostudies-other
Project description:BackgroundAbout 30% of home healthcare patients are hospitalized or visit an emergency department (ED) during a home healthcare (HHC) episode. Novel data science methods are increasingly used to improve identification of patients at risk for negative outcomes.ObjectivesThe aim of the study was to identify patients at heightened risk hospitalization or ED visits using HHC narrative data (clinical notes).MethodsThis study used a large database of HHC visit notes (n = 727,676) documented for 112,237 HHC episodes (89,459 unique patients) by clinicians of the largest nonprofit HHC agency in the United States. Text mining and machine learning algorithms (Naïve Bayes, decision tree, random forest) were implemented to predict patient hospitalization or ED visits using the content of clinical notes. Risk factors associated with hospitalization or ED visits were identified using a feature selection technique (gain ratio attribute evaluation).ResultsBest performing text mining method (random forest) achieved good predictive performance. Seven risk factors categories were identified, with clinical factors, coordination/communication, and service use being the most frequent categories.DiscussionThis study was the first to explore the potential contribution of HHC clinical notes to identifying patients at risk for hospitalization or an ED visit. Our results suggest that HHC visit notes are highly informative and can contribute significantly to identification of patients at risk. Further studies are needed to explore ways to improve risk prediction by adding more data elements from additional data sources.
| S-EPMC7606545 | biostudies-literature
Project description:IntroductionLaw enforcement officers (LEO) interact with patients and clinicians in the emergency department (ED) for many reasons. There is no current consensus on what should comprise, or how to best enact, guidelines that ideally balance LEO activities in the service of public safety with patient health, autonomy, and privacy. The purpose of this study was to explore how a national sample of emergency physicians (EP) perceives activities of LEOs during the delivery of emergency medical care.MethodsMembers of the Emergency Medicine Practice Research Network (EMPRN) were recruited via an email-delivered, anonymous survey that elicited experiences, perceptions, and knowledge of policies that guide interactions with LEOs in the ED. The survey included multiple-choice items, which we analyzed descriptively, and open-ended questions, which we analyzed using qualitative content analysis.ResultsOf 765 EPs in the EMPRN, 141 (18.4%) completed the survey. Respondents represented diverse locations and years in practice. A total of 113 (82%) respondents were White, and 114 (81%) were male. Over a third reported LEO presence in the ED on a daily basis. A majority (62%) perceived LEO presence as helpful for clinicians and clinical practice. When asked about the factors deemed highly important in allowing LEOs to access patients during care, 75% reported patients' potential as a threat to public safety. A small minority of respondents (12%) considered the patients' consent or preference to interact with LEOs. While 86% of EPs felt that information-gathering by LEO was appropriate in the ED setting, only 13% were aware of policy to guide these decisions. Perceived barriers to implementation of policy in this area included: issues of enforcement; leadership; education; operational challenges; and potential negative consequences.ConclusionFuture research is warranted to explore how policies and practices that guide intersections between emergency medical care and law enforcement impact patients, clinicians, and the communities that health systems serve.
| S-EPMC10047729 | biostudies-literature
Project description:BackgroundComputerized provider order entry (CPOE) is a technology with potential to transform care delivery. While CPOE systems have been studied in adult populations, less is known about the implementation of CPOE in the neonatal intensive care unit (NICU) and perceptions of nurses and physicians using the system.ObjectiveTo examine perceptions of clinicians before and after CPOE implementation in the NICU of a pediatric hospital.MethodsA cross-sectional survey of clinicians working in a Level III NICU was conducted. The survey was distributed before and after CPOE implementation. Participants were asked about their perception of CPOE on patient care delivery, implementation of the system, and effect on job satisfaction. A qualitative section inquired about additional concerns surrounding implementation. Responses were tabulated and analyzed using the Chi-square test.ResultsThe survey was distributed to 158 clinicians with a 47% response rate for pre-implementation and 45% for post-implementation. Clinicians understood why CPOE was implemented, but felt there was incomplete technical training. The expectation for increased job satisfaction and ability to recruit high-quality staff was high. However, there was concern about the ability to deliver appropriate treatments before and after implementation. Physicians were more optimistic about CPOE implementation than nurses who remained concerned that workflow may be altered.ConclusionsIntroducing CPOE is a potentially risky endeavor and must be done carefully to mitigate harm. Although high expectations of the system can be met, it is important to attend to differing expectations among clinicians with varied levels of comfort with technology. Interdisciplinary collaboration is critical in planning a functioning CPOE to ensure that efficient workflow is maintained and appropriate supports for individuals with a lower degree of technical literacy is available.
| S-EPMC6241742 | biostudies-literature
Project description:BACKGROUND:Computerized physician order entry (CPOE) has grown since the early 1990s. While many systems serve adult patients, systems for pediatric and neonatal populations have lagged. Adapting adult CPOE systems for pediatric use may require significant modifications to address complexities associated with pediatric care such as daily weight changes and small medication doses. OBJECTIVE:This article aims to review the neonatal intensive care unit (NICU) CPOE literature to characterize trends in the introduction of this technology and to identify potential areas for further research. METHODS:Articles pertaining to NICU CPOE were identified in MEDLINE using MeSH terms "medical order entry systems," "drug therapy," "intensive care unit, neonatal," "infant, newborn," etc. Two physician reviewers evaluated each article for inclusion and exclusion criteria. Consensus judgments were used to classify the articles into five categories: medication safety, usability/alerts, clinical practice, clinical decision Support (CDS), and implementation. Articles addressing pediatric (nonneonatal) CPOE were included if they were applicable to the NICU setting. RESULTS:Sixty-nine articles were identified using MeSH search criteria. Twenty-two additional articles were identified by hand-searching bibliographies and 6 articles were added after the review process. Fifty-five articles met exclusion criteria, for a final set of 42 articles. Medication safety was the focus of 22 articles, followed by clinical practice (10), CDS (10), implementation (11), and usability/alerts (4). Several addressed more than one category. No study showed a decrease in medication safety post-CPOE implementation. Within clinical practice articles, CPOE implementation showed no effect on blood glucose levels or time to antibiotic administration but showed conflicting results on mortality rates. Implementation studies were largely descriptive of single-hospital experiences. CONCLUSION:CPOE implementation within the NICU has demonstrated improvement in medication safety, with the most consistent benefit involving a reduction in medication errors and wrong-time administration errors. Additional research is needed to understand the potential limitations of CPOE systems in neonatal intensive care and how CPOE affects mortality.
| S-EPMC6609270 | biostudies-literature
Project description:BACKGROUND:Reliably abstracting outcomes from free-text electronic health records remains a challenge. While automated classification of free text has been a popular medical informatics topic, performance validation using real-world clinical data has been limited. The two main approaches are linguistic (natural language processing [NLP]) and statistical (machine learning). The authors have developed a hybrid system for abstracting computed tomography (CT) reports for specified outcomes. OBJECTIVES:The objective was to measure performance of a hybrid NLP and machine learning system for automated outcome classification of emergency department (ED) CT imaging reports. The hypothesis was that such a system is comparable to medical personnel doing the data abstraction. METHODS:A secondary analysis was performed on a prior diagnostic imaging study on 3,710 blunt facial trauma victims. Staff radiologists dictated CT reports as free text, which were then deidentified. A trained data abstractor manually coded the reference standard outcome of acute orbital fracture, with a random subset double-coded for reliability. The data set was randomly split evenly into training and testing sets. Training patient reports were used as input to the Medical Language Extraction and Encoding (MedLEE) NLP tool to create structured output containing standardized medical terms and modifiers for certainty and temporal status. Findings were filtered for low certainty and past/future modifiers and then combined with the manual reference standard to generate decision tree classifiers using data mining tools Waikato Environment for Knowledge Analysis (WEKA) 3.7.5 and Salford Predictive Miner 6.6. Performance of decision tree classifiers was evaluated on the testing set with or without NLP processing. RESULTS:The performance of machine learning alone was comparable to prior NLP studies (sensitivity = 0.92, specificity = 0.93, precision = 0.95, recall = 0.93, f-score = 0.94), and the combined use of NLP and machine learning showed further improvement (sensitivity = 0.93, specificity = 0.97, precision = 0.97, recall = 0.96, f-score = 0.97). This performance is similar to, or better than, that of medical personnel in previous studies. CONCLUSIONS:A hybrid NLP and machine learning automated classification system shows promise in coding free-text electronic clinical data.
| S-EPMC3898888 | biostudies-literature
Project description:Background. While often first treated in the emergency department (ED), identification of sepsis is difficult. Electronic medical record (EMR) clinical decision tools offer a novel strategy for identifying patients with sepsis. The objective of this study was to test the accuracy of an EMR-based, automated sepsis identification system. Methods. We tested an EMR-based sepsis identification tool at a major academic, urban ED with 64,000 annual visits. The EMR system collected vital sign and laboratory test information on all ED patients, triggering a "sepsis alert" for those with ?2 SIRS (systemic inflammatory response syndrome) criteria (fever, tachycardia, tachypnea, leukocytosis) plus ?1 major organ dysfunction (SBP ? 90 mm Hg, lactic acid ?2.0 mg/dL). We confirmed the presence of sepsis through manual review of physician, nursing, and laboratory records. We also reviewed a random selection of ED cases that did not trigger a sepsis alert. We evaluated the diagnostic accuracy of the sepsis identification tool. Results. From January 1 through March 31, 2012, there were 795 automated sepsis alerts. We randomly selected 300 cases without a sepsis alert from the same period. The true prevalence of sepsis was 355/795 (44.7%) among alerts and 0/300 (0%) among non-alerts. The positive predictive value of the sepsis alert was 44.7% (95% CI [41.2-48.2%]). Pneumonia and respiratory infections (38%) and urinary tract infection (32.7%) were the most common infections among the 355 patients with true sepsis (true positives). Among false-positive sepsis alerts, the most common medical conditions were gastrointestinal (26.1%), traumatic (25.7%), and cardiovascular (20.0%) conditions. Rates of hospital admission were: true-positive sepsis alert 91.0%, false-positive alert 83.0%, no sepsis alert 5.7%. Conclusions. This ED EMR-based automated sepsis identification system was able to detect cases with sepsis. Automated EMR-based detection may provide a viable strategy for identifying sepsis in the ED.
| S-EPMC3994640 | biostudies-literature