scholarly journals Generating contextual embeddings for emergency department chief complaints

JAMIA Open ◽  
2020 ◽  
Vol 3 (2) ◽  
pp. 160-166
Author(s):  
David Chang ◽  
Woo Suk Hong ◽  
Richard Andrew Taylor
Keyword(s):  
Emergency Department ◽  
Short Term Memory ◽  
Language Model ◽  
Chief Complaint ◽  
Language Models ◽  
Free Text ◽  
Set Size ◽  
Chief Complaints ◽  
Ed Visits ◽  
Cutoff Threshold

Abstract Objective We learn contextual embeddings for emergency department (ED) chief complaints using Bidirectional Encoder Representations from Transformers (BERT), a state-of-the-art language model, to derive a compact and computationally useful representation for free-text chief complaints. Materials and methods Retrospective data on 2.1 million adult and pediatric ED visits was obtained from a large healthcare system covering the period of March 2013 to July 2019. A total of 355 497 (16.4%) visits from 65 737 (8.9%) patients were removed for absence of either a structured or unstructured chief complaint. To ensure adequate training set size, chief complaint labels that comprised less than 0.01%, or 1 in 10 000, of all visits were excluded. The cutoff threshold was incremented on a log scale to create seven datasets of decreasing sparsity. The classification task was to predict the provider-assigned label from the free-text chief complaint using BERT, with Long Short-Term Memory (LSTM) and Embeddings from Language Models (ELMo) as baselines. Performance was measured as the Top-k accuracy from k = 1:5 on a hold-out test set comprising 5% of the samples. The embedding for each free-text chief complaint was extracted as the final 768-dimensional layer of the BERT model and visualized using t-distributed stochastic neighbor embedding (t-SNE). Results The models achieved increasing performance with datasets of decreasing sparsity, with BERT outperforming both LSTM and ELMo. The BERT model yielded Top-1 accuracies of 0.65 and 0.69, Top-3 accuracies of 0.87 and 0.90, and Top-5 accuracies of 0.92 and 0.94 on datasets comprised of 434 and 188 labels, respectively. Visualization using t-SNE mapped the learned embeddings in a clinically meaningful way, with related concepts embedded close to each other and broader types of chief complaints clustered together. Discussion Despite the inherent noise in the chief complaint label space, the model was able to learn a rich representation of chief complaints and generate reasonable predictions of their labels. The learned embeddings accurately predict provider-assigned chief complaint labels and map semantically similar chief complaints to nearby points in vector space. Conclusion Such a model may be used to automatically map free-text chief complaints to structured fields and to assist the development of a standardized, data-driven ontology of chief complaints for healthcare institutions.

scholarly journals Monitoring Out-of-State Patients during a 2017 Hurricane Response using ESSENCE

2018 ◽  
Vol 10 (1) ◽  
Author(s):  
Caleb Wiedeman ◽  
Julie Shaffner ◽  
Kelly Squires ◽  
Jeffrey Leegon ◽  
Rendi Murphree ◽  
...  
Keyword(s):  
Emergency Department ◽  
Syndromic Surveillance ◽  
Situational Awareness ◽  
Chief Complaint ◽  
Free Text ◽  
Chief Complaints ◽  
Major Hurricane ◽  
Ed Visits ◽  
Zip Code ◽  
Hurricane Irma

ObjectiveTo demonstrate the use of ESSENCE in the BioSense Platform to monitor out-of-State patients seeking emergency healthcare in Tennessee during Hurricanes Harvey and Irma.IntroductionSyndromic surveillance is the monitoring of symptom combinations (i.e., syndromes) or other indicators within a population to inform public health actions. The Tennessee Department of Health (TDH) collects emergency department (ED) data from more than 70 hospitals across Tennessee to support statewide syndromic surveillance activities. Hospitals in Tennessee typically provide data within 48 hours of a patient encounter. While syndromic surveillance often supplements disease- or condition-specific surveillance, it can also provide general situational awareness about emergency department patients during an event or response.During Hurricanes Harvey (continental US landfall on August 25, 2017) and Irma (continental US landfall on September 10, 2017), TDH supported all hazards situational awareness using the Electronic Surveillance System for the Early Notification of Community-based Epidemics (ESSENCE) in the BioSense Platform supported by the National Syndromic Surveillance Program (NSSP). The volume of out-of-state patients in Tennessee was monitored to assess the impact on the healthcare system and any geographic- or hospital-specific clustering of out-of-state patients within Tennessee. Results were included in daily State Health Operations Center (SHOC) situation reports and shared with agency response partners such as the Tennessee Emergency Management Agency (TEMA).MethodsData were monitored from August 18, 2017 through September 24, 2017. A simple query was established in ESSENCE using the Patient Location (Full Details) dataset. Data were limited to hospital ED visits reported by Tennessee (Site = “Tennessee”). To monitor ED visits among residents of Texas before, during, and after Major Hurricane Harvey, data were queried for a patient zip code within Texas (State = “Texas”). ED visits among Florida residents were monitored similarly (State = “Florida”) before, during, and after Major Hurricane Irma. Additionally, a free text chief complaint search was implemented for the terms “Harvey”, “Irma, “hurricane”, “evacuee”, “evacuate”, “Florida”, and “Texas”. Chief complaint search results were then filtered to remove encounters with patient zip codes within Tennessee.ResultsFrom August 18, 2017 through September 24, 2017, Tennessee hospital EDs reported 277 patient encounters among Texas residents and 1,041 patient encounters among Florida residents. The number of encounters among patients from Texas remained stable throughout the monitoring period. In contrast, the number of encounters among patients from Florida exceeded the expected value on September 7, peaked September 10 at 116 patient encounters, and returned to expected levels on September 16 (Figure 1). The increase in patients from Florida was evenly distributed across most of Tennessee, with some clustering around a popular tourism area in East Tennessee. No concerning trends in reported syndromes or chief complaints were identified among Texas or Florida patients.The free text chief complaint query first exceeded the expected value on September 9, peaked on September 11 with 5 patient encounters, and returned to expected levels on September 14. From August 18 through September 24, 21 of 30 visits captured by the query were among Florida residents. One Tennessee hospital appeared to be intentionally using the term “Irma” in their chief complaint field to indicate patients from Florida impacted by the hurricane.ConclusionsThe ESSENCE instance in the BioSense platform provided TDH the opportunity to easily locate and monitor out-of-state patients seen in Tennessee hospital EDs. While TDH was unable to validate whether all patients identified as residents of Florida were displaced because of Major Hurricane Irma, the timing of the rise and fall of patient encounters was highly suggestive. Likewise, seeing no substantial increase ED patients with residence in Texas reassured TDH that the effects of Hurricane Harvey were not impacting hospital emergency departments in Tennessee.TDH used information and charts from ESSENCE to support situational awareness in our SHOC and at TEMA. Use of patient zip code to identify out-of-state residents was more sensitive than chief complaint searches by keyword during this event. ESSENCE allowed TDH to see where out-of-state patients appeared to be concentrating in Tennessee and monitor the need for targeting messaging and resources to heavily affected areas. Additionally, close surveillance of chief complaints among out-of-state patients provided assurance that no unusual patterns in illness or injury were occurring.ESSENCE is the only TDH information source capable of rapidly collecting health information on out-of-state patients. ESSENCE allowed TDH to quickly identify a change within the patient population seen at Tennessee emergency departments and monitor the situation until the patient population returned to baseline levels.

scholarly journals Free-Text Mining to Improve Syndrome Definition Matching Across Emergency Departments

2018 ◽  
Vol 10 (1) ◽  
Author(s):  
Kristin Arkin
Keyword(s):  
Text Mining ◽  
Influenza Season ◽  
Chief Complaint ◽  
Free Text ◽  
Diagnosis Code ◽  
Diagnosis Codes ◽  
Chief Complaints ◽  
Group A ◽  
Ed Visits ◽  
Group B

ObjectiveWe sought to use free text mining tools to improve emergency department (ED) chief complaint and discharge diagnosis data syndrome definition matching across facilities with differing robustness of data in the Electronic Surveillance System for the Early Notification of Community-based Epidemics (ESSENCE) application in Idaho’s syndromic surveillance system.IntroductionStandard syndrome definitions for ED visits in ESSENCE rely on chief complaints. Visits with more words in the chief complaint field are more likely to match syndrome definitions. While using ESSENCE, we observed geographic differences in chief complaint length, apparently related to differences in electronic health record (EHR) systems, which resulted in disparate syndrome matching across Idaho regions. We hypothesized that chief complaint and diagnosis code co-occurrence among ED visits to facilities with long chief complaints could help identify terms that would improve syndrome match among facilities with short chief complaints.MethodsThe ESSENCE-defined influenza-like illness (ILI) chief complaint syndrome was used as the base syndrome for this analysis. Syndrome-matched visits were defined as visits that match the syndrome definition.We assessed chief complaints and diagnosis code co-occurrence of syndrome-matched visits using the RCRAN TidyText package and developed a bigram network from normalized, concatenated chief complaint and diagnosis code (CCDD) fields and normalized diagnosis code (DD) fields per previously described methodologies.1 Common connections were defined by a natural break in frequency of pair occurrence for CCDD pairs (30 occurrences) and DD pairs (5 occurrences).The ESSENCE syndrome was revised by adding relevant bigram network clusters and logic operators. We compared time series of the percent of ED visits matched to the ESSENCE syndrome with those matched to the revised syndrome. We stratified the time series by facilities grouped by short (average < 4 words, “Group A”) and long (average ≥ 4 words, “Group B”) chief complaint fields (Figure 1). Influenza season start was defined as two consecutive weeks above baseline, or the 95% upper confidence limit of percent syndrome-matched visits outside of the CDC ILI surveillance season. Season trends and influenza-related deaths in Idaho residents were compared.ResultsDuring August 1, 2016 through July 31, 2017, 1,587 (1.17%) of 135,789 ED visits matched the ESSENCE syndrome. Bigram networks of CCDD fields produced clusters already included by the ESSENCE syndrome. The bigram network of DD fields (Figure 2) produced six clusters. The revised syndrome definition included the ESSENCE syndrome, 3 single DD terms, and 3 two DD terms combined. The start of influenza season was identified as the same week for both ILI syndrome definitions (ESSENCE baseline 0.70%; revised baseline 2.21%). The ESSENCE syndrome indicated the season peaked during Morbidity and Mortality Weekly Report (MMWR) week 2017-05 with the season ending MMWR week 2017-14. The revised syndrome indicated 2017-20 as the season end. Multiple peaks seen with the revised syndrome during MMWR weeks 2017-02, 2017-05, and 2017-10 mirrored peaks in influenza-related deaths during MMWR weeks 2017-03, 2017-06, and 2017-11.ILI season onset was five weeks earlier with the revised syndrome compared with the ESSENCE syndrome in Group A facilities, but remained the same in Group B. The annual percentage of ED visits related to ILI was more uniform between facility groups under the revised syndrome than the ESSENCE syndrome. Unlike the trend seen with the ESSENCE syndrome, the revised syndrome shows low-level ILI activity in both groups year-round.ConclusionsIn Idaho, dramatic differences in ED visit chief complaint word counts were seen between facilities; bigram networks were found to be an important tool to identify diagnosis codes and logical operators that built more inclusive syndrome definitions when added to an existing chief complaint syndrome. Bigram networks may aid understanding the relationship between chief complaints and diagnosis codes in syndrome-matched visits.Use of trade names and commercial sources is for identification only and does not imply endorsement by the Centers for Disease Control and Prevention, the Public Health Service, or the U.S. Department of Health and Human Services.References1. Silge, J., Robinson, D. (2017). “Text Mining with R”. O’Reilly.

Abstract P317: Sex and Age Differences in a typical Chief Complains for Heart Failure in Emergency Department Visits

Circulation ◽  
2020 ◽  
Vol 141 (Suppl_1) ◽  
Author(s):  
Michelle L Meyer ◽  
Montika Bush ◽  
Jason J Bischof ◽  
Anna E Waller ◽  
Timothy F Platts-Mills
Keyword(s):  
Heart Failure ◽  
Emergency Department ◽  
Chest Pain ◽  
Age Differences ◽  
Chief Complaint ◽  
Age Group ◽  
Chief Complaints ◽  
Ed Visits ◽  
Respiratory Complaints ◽  
Sex And Age Differences

Background: Around 1 million United States emergency department (ED) visits per year are due to exacerbation of heart failure (HF) symptoms, with ~80% of those patients admitted to the hospital. However, sex and age differences in HF symptom presentation in the ED have not been thoroughly investigated. Objectives: To describe sex and age differences in chief complaints of ED patients with a HF diagnosis. Methods: We included patients ≥18 years old with an ED diagnosis of HF in NC DETECT, a statewide syndromic surveillance system. We defined a HF diagnosis using ICD-9-CM and ICD-10-CM codes from ED visits between 2010 and 2016. We classified the ED chief complaints into categories by symptom groups (e.g. respiratory complaint includes hypoxia, respiratory distress, breathing difficulties). Chief complaint categories are not mutually exclusive. We calculated frequencies of chief complaint categories for ED visits by sex and age (18-44 (n=55,216), 45-64 (n=260,397), ≥65 (n=578,313) years old) and evaluated for a 10% standardized difference between groups. Results: There were 422,720 patients with 893,950 total unique visits (1.6 average visits/person). Of these visits, 55.0% were by women and 59.5% patients were admitted. Overall, the top chief complaint categories were dyspnea (19.1%), chest pain (13.5%), and respiratory complaints (13.4%), and were similar by sex and by ED disposition (admitted or discharged) and sex. When stratified by sex and age group, in those 18-44 years old, women had more reports of nausea/vomiting (6.7%) compared with men (4.1%) and headache (4.2%) compared with men (2.0%). In those 45-64 and ≥65 years old, chief complaint categories were similar between women and men. When stratified by age group alone, reports of chest pain decreased with age (21.4% in 18-44, 17.7% in 45-64, and 10.8% in ≥65 year olds), whereas reports of balance issues (1.2% in 18-44, 2.4% in 45-64, and 6.0% in ≥65 year olds), weakness (1.7% in 18-44, 2.7% in 45-64, and 5.5% in ≥65 year olds), and confusion (0.8% in 18-44, 2.1% in 45-64, and 4.5% in ≥65 year olds) increased with age. Compared to those ≥65 years old, those 18-44 years old had fewer respiratory complaints (10.0% vs. 13.9%), but more reports of headache (3.2% vs. 0.8%) and nausea/vomiting (5.5% vs. 3.2%). Conclusion: In a state-wide population of ED patients with HF diagnoses, sex differences in chief complaint categories that are less obvious symptoms of HF were observed for those 18-44 years old, with women reporting more nausea/vomiting and headache compared to men. Chief complaint categories that are less obvious symptoms of HF were more common among patients 18-44 (nausea/vomiting, headache) and ≥65 (balance issues, confusion, weakness) years old. Characterizing the variation of symptoms of HF patients in the ED may help inform the identification of ED patients with HF and the outpatient management of HF-related symptoms.

scholarly journals DICE: A Drug Indication Classification and Encyclopedia for AI-Based Indication Extraction

2021 ◽  
Vol 4 ◽  
Author(s):  
Arjun Bhatt ◽  
Ruth Roberts ◽  
Xi Chen ◽  
Ting Li ◽  
Skylar Connor ◽  
...  
Keyword(s):  
Language Processing ◽  
Clinical Decision Making ◽  
Short Term Memory ◽  
Drug Repositioning ◽  
External Validation ◽  
Clinical Decision ◽  
Language Models ◽  
Free Text ◽  
Drug Labeling ◽  
Real World Evidence

Drug labeling contains an ‘INDICATIONS AND USAGE’ that provides vital information to support clinical decision making and regulatory management. Effective extraction of drug indication information from free-text based resources could facilitate drug repositioning projects and help collect real-world evidence in support of secondary use of approved medicines. To enable AI-powered language models for the extraction of drug indication information, we used manual reading and curation to develop a Drug Indication Classification and Encyclopedia (DICE) based on FDA approved human prescription drug labeling. A DICE scheme with 7,231 sentences categorized into five classes (indications, contradictions, side effects, usage instructions, and clinical observations) was developed. To further elucidate the utility of the DICE, we developed nine different AI-based classifiers for the prediction of indications based on the developed DICE to comprehensively assess their performance. We found that the transformer-based language models yielded an average MCC of 0.887, outperforming the word embedding-based Bidirectional long short-term memory (BiLSTM) models (0.862) with a 2.82% improvement on the test set. The best classifiers were also used to extract drug indication information in DrugBank and achieved a high enrichment rate (&gt;0.930) for this task. We found that domain-specific training could provide more explainable models without performance sacrifices and better generalization for external validation datasets. Altogether, the proposed DICE could be a standard resource for the development and evaluation of task-specific AI-powered, natural language processing (NLP) models.

scholarly journals COVID-19 sentiment analysis via deep learning during the rise of novel cases

PLoS ONE ◽  
2021 ◽  
Vol 16 (8) ◽  
pp. e0255615 ◽  
Author(s):  
Rohitash Chandra ◽  
Aswin Krishna
Keyword(s):  
Deep Learning ◽  
Sentiment Analysis ◽  
Short Term Memory ◽  
Language Model ◽  
Deep Understanding ◽  
Language Models ◽  
Catastrophic Events ◽  
Social Scientists ◽  
Significant Group ◽  
Global Vector

Social scientists and psychologists take interest in understanding how people express emotions and sentiments when dealing with catastrophic events such as natural disasters, political unrest, and terrorism. The COVID-19 pandemic is a catastrophic event that has raised a number of psychological issues such as depression given abrupt social changes and lack of employment. Advancements of deep learning-based language models have been promising for sentiment analysis with data from social networks such as Twitter. Given the situation with COVID-19 pandemic, different countries had different peaks where rise and fall of new cases affected lock-downs which directly affected the economy and employment. During the rise of COVID-19 cases with stricter lock-downs, people have been expressing their sentiments in social media. This can provide a deep understanding of human psychology during catastrophic events. In this paper, we present a framework that employs deep learning-based language models via long short-term memory (LSTM) recurrent neural networks for sentiment analysis during the rise of novel COVID-19 cases in India. The framework features LSTM language model with a global vector embedding and state-of-art BERT language model. We review the sentiments expressed for selective months in 2020 which covers the major peak of novel cases in India. Our framework utilises multi-label sentiment classification where more than one sentiment can be expressed at once. Our results indicate that the majority of the tweets have been positive with high levels of optimism during the rise of the novel COVID-19 cases and the number of tweets significantly lowered towards the peak. We find that the optimistic, annoyed and joking tweets mostly dominate the monthly tweets with much lower portion of negative sentiments. The predictions generally indicate that although the majority have been optimistic, a significant group of population has been annoyed towards the way the pandemic was handled by the authorities.

scholarly journals Syndromic Surveillance for Situation Awareness: Understanding Syndrome Performance

2018 ◽  
Vol 10 (1) ◽  
Author(s):  
Kristin Arkin
Keyword(s):  
Emergency Department ◽  
Situation Awareness ◽  
Syndromic Surveillance ◽  
Chief Complaint ◽  
Surveillance Program ◽  
Initial Visit ◽  
Word Count ◽  
Data Submission ◽  
Daily Frequency ◽  
Ed Visits

ObjectiveIn August 2017, a large influx of visitors was expected to view the total solar eclipse in Idaho. The Idaho Syndromic Surveillance program planned to enhance situation awareness during the event. In preparation, we sought to examine syndrome performance of several newly developed chief complaint and combination chief complaint and diagnosis code syndrome definitions to aid in interpretation of syndromic surveillance data during the event.IntroductionThe August 21, 2017 total solar eclipse in Idaho was anticipated to lead to a large influx of visitors in many communities, prompting a widespread effort to assure Idaho was prepared. To support these efforts, the Idaho Syndromic Surveillance program (ISSp) developed a plan to enhance situation awareness during the event by conducting syndromic surveillance using emergency department (ED) visit data contributed to the National Syndromic Surveillance Program’s BioSense platform by Idaho hospitals. ISSp sought input on anticipated threats from state and local emergency management and public health partners, and selected 8 syndromes for surveillance.Ideally, the first electronic message containing information on an emergency department visit is sent to ISSp within 24 hours of the visit and includes the chief complaint for the visit. Data on other variables, such as diagnosis codes, are updated by subsequent messages for several days after the visit. Chief complaint (CC) text and discharge diagnosis (DD) codes are the primary variables used for syndrome match; delay in reporting these variables adversely affects timely syndrome match of visits. Because our plan included development of new syndrome definitions and querying data within 24 hours of visits, earlier than ISSp had done previously for trend analysis, we sought to better understand syndrome performance.MethodsWe defined messages with completed CC and DD as the last message regarding a visit where term count increased from previous messages regarding that visit, indicating new information was added to the field. We retrospectively assessed the total number of ED visits and calculated the daily frequency of completed CC and DD by days since visit date for visits during June 1–July 31, 2017. Additionally, we calculated facility mean word count in CC fields by averaging the word count of parsed, complete CC fields for visits occurring June 1–July 31, 2017 for each facility.During July 10–24, 2017, we calculated the daily frequency of visits occurring in the previous 90 days for total ED visits and syndrome-matched visits for 8 selected syndromes (heat-related illness; cold exposure; influenza-like-illness; nausea, vomiting, and diarrhea; animal/bug bites and stings; drowning/submersion; alcohol/drug intoxication; and medication replacement). Syndrome-matched visits were defined as visits with CC or DD that match the syndrome definition. We calculated the percent of syndrome-matched visits by syndromes defined with CC or CC and DD combined (CCDD) over time. Syndromes with fewer than 5 matched visits were excluded from analysis.ResultsComplete CCs were received for 99.1% of visits and complete DDs were received for 89.8% of visits. Complete CCs were submitted for 58.2% of visits within 1 day of the visit, 88.9% of visits within 3 days, and 98.9% of visits within 7 days. In contrast, complete DDs were submitted for 24.3% of visits within 1 day, 38.7% of visits within 3 days, and 53.7% of visits within 7 days (Table 1).During the observation period, data submission from facilities representing approximately 33% of visits was interrupted for 5 (36%) of 14 days. Heat-related illness, cold exposure, and drowning/submersion, were excluded from syndrome-match analysis. During the 9 days of uninterrupted data submission, 100% syndrome-matched visits for syndromes defined by CC alone and 69.1% syndrome-matched visits for syndromes defined by CCDD were identified within 6–7 days of initial visit. Facilities with interrupted data submission contributed 75% of CC syndrome-matched visits and 33% of CCDD syndrome-matched visits. The facility mean word count in CC fields from these facilities was >15 compared with 2–4 from other facilities.ConclusionsExamination of syndrome performance prior to a known event quantitated differences in timeliness of CC and DD completeness and syndrome match. CCs and DDs in visit messages were not complete within 24 hours of initial visit. CC completion was nearly 34 percentage points greater than DD completeness 1 day after initial visit and did not converge until ≥15 days after initial visit. Higher percentages of syndrome match within 6–7 days of initial visit were seen by CC alone than CCDD defined syndromes. Facilities using longer CCs contributed disproportionately to syndrome matching using CC, but not CCDD syndrome definitions. Syndromic surveillance system characteristics, including timeliness of CCs and DDs, length of CCs, and characteristics of facilities from which data transmission is interrupted should be considered when building syndrome definitions that will be used for surveillance within 7 days of emergency department visits and when interpreting syndromic surveillance findings.

scholarly journals Validation of a Syndromic Case Definition for Detecting Emergency Department Visits Potentially Related to Marijuana

2017 ◽  
Vol 132 (4) ◽  
pp. 471-479 ◽  
Author(s):  
Kathryn DeYoung ◽  
Yushiuan Chen ◽  
Robert Beum ◽  
Michele Askenazi ◽  
Cali Zimmerman ◽  
...  
Keyword(s):  
Public Health ◽  
Emergency Department ◽  
Health Impact ◽  
Case Definition ◽  
Emergency Department Visits ◽  
Free Text ◽  
Search Term ◽  
Diagnosis Codes ◽  
Ed Visits ◽  
The Impact

Objectives: Reliable methods are needed to monitor the public health impact of changing laws and perceptions about marijuana. Structured and free-text emergency department (ED) visit data offer an opportunity to monitor the impact of these changes in near-real time. Our objectives were to (1) generate and validate a syndromic case definition for ED visits potentially related to marijuana and (2) describe a method for doing so that was less resource intensive than traditional methods. Methods: We developed a syndromic case definition for ED visits potentially related to marijuana, applied it to BioSense 2.0 data from 15 hospitals in the Denver, Colorado, metropolitan area for the period September through October 2015, and manually reviewed each case to determine true positives and false positives. We used the number of visits identified by and the positive predictive value (PPV) for each search term and field to refine the definition for the second round of validation on data from February through March 2016. Results: Of 126 646 ED visits during the first period, terms in 524 ED visit records matched ≥1 search term in the initial case definition (PPV, 92.7%). Of 140 932 ED visits during the second period, terms in 698 ED visit records matched ≥1 search term in the revised case definition (PPV, 95.7%). After another revision, the final case definition contained 6 keywords for marijuana or derivatives and 5 diagnosis codes for cannabis use, abuse, dependence, poisoning, and lung disease. Conclusions: Our syndromic case definition and validation method for ED visits potentially related to marijuana could be used by other public health jurisdictions to monitor local trends and for other emerging concerns.

scholarly journals Revisits within 48 Hours to a Thai Emergency Department

2016 ◽  
Vol 2016 ◽  
pp. 1-5 ◽  
Author(s):  
Jiraporn Sri-on ◽  
Adisak Nithimathachoke ◽  
Gregory Philip Tirrell ◽  
Sataporn Surawongwattana ◽  
Shan Woo Liu
Keyword(s):  
Emergency Department ◽  
Chart Review ◽  
Tertiary Care ◽  
Related Factor ◽  
Contributing Factors ◽  
Related Factors ◽  
Chief Complaints ◽  
Ed Visits ◽  
The Mean ◽  
Patient Related Factor

Objective. Emergency department (ED) revisits are a common ED quality measure. This study was undertaken to ascertain the contributing factors of revisits within 48 hours to a Thai ED and to explore physician-related, illness-related, and patient-related factors behind those revisits.Methods. This study was a chart review from one tertiary care, urban Thai hospital from October 1, 2009, to September 31, 2010. We identified patients who returned to the ED within 48 hours for the same or related complaints after their initial discharge. Three physicians classified revisit as physician-related, illness-related, and patient-related factors.Results. Our study included 172 ED patients’ charts. 86/172 (50%) were male and the mean age was38 ± 5.6(SD) years. The ED revisits contributing factors were physician-related factors [86/172 (50.0%)], illness-related factors [61/172 (35.5%)], and patient-related factor [25/172 (14.5%)], respectively. Among revisits classified as physician-related factors, 40/86 (46.5%) revisits were due to misdiagnosis and 36/86 (41.9%) were due to suboptimal management. Abdominal pain [27/86 (31.4%)] was the majority of physician-related chief complaints, followed by fever [16/86 (18.6%)] and dyspnea [15/86 (17.4%)].Conclusion. Misdiagnosis and suboptimal management contributed to half of the 48-hour repeat ED visits in this Thai hospital.

scholarly journals Using Real-Time Syndromic Surveillance to Analyze the Impact of a Cold Weather Event in New Mexico

2018 ◽  
Vol 2018 ◽  
pp. 1-6
Author(s):  
Victoria F. Dirmyer
Keyword(s):  
Real Time ◽  
Syndromic Surveillance ◽  
Chief Complaint ◽  
Cold Weather ◽  
Weather Event ◽  
Diagnosis Codes ◽  
Chief Complaints ◽  
Icd 10 ◽  
Ed Visits ◽  
The Impact

Objective. This report describes the development of a novel syndromic cold weather syndrome for use in monitoring the impact of cold weather events on emergency department attendance. Methods. Syndromic messages from seven hospitals were analyzed for ED visits that occurred over a 12-day period. A cold weather syndrome was defined using terms in the self-reported chief complaint field as well as specific ICD-10-CM codes related to cold weather. A κ statistic was calculated to assess the overall agreement between the chief complaint field and diagnosis fields to further refine the cold weather syndrome definition. Results. Of the 3,873 ED visits that were reported, 487 were related to the cold weather event. Sixty-three percent were identified by a combination of diagnosis codes and chief complaints. Overall agreement between chief complaint and diagnosis codes was moderate (κ=0.50; 95% confidence interval = 0.48–0.52). Conclusion. Due to the near real-time reporting of syndromic surveillance data, analysis results can be acted upon. Results from this analysis will be used in the state’s emergency operations plan (EOP) for cold weather and winter storms. The EOP will provide guidance for mobilization of supplies/personnel, preparation of roadways and pedestrian walkways, and the coordination efforts of multiple state agencies.

scholarly journals Developing and Validating a Fireworks-Related Syndrome Definition in Kansas

2018 ◽  
Vol 10 (1) ◽  
Author(s):  
Zachary M Stein
Keyword(s):  
Syndromic Surveillance ◽  
Public Information ◽  
False Positives ◽  
Chief Complaint ◽  
Surveillance Program ◽  
Anatomical Location ◽  
Free Text ◽  
Data Set ◽  
Ed Visits ◽  
Key Terms

ObjectiveTo develop a syndrome definition and analyze syndromic surveillance data usefulness in surveillance of firework-related emergency department visits in Kansas. Introduction Across the U.S.A., multiple people seek treatment for fireworks-related injuries around the July 4th holiday. Syndromic surveillance in Kansas allows for near real-time analysis of the injuries occurring during the firework selling season. During the 2017 July 4thholiday, the Kansas Syndromic Surveillance Program (KSSP) production data feed received data from 88 EDs at excellent quality and timeliness. Previous and current firework safety messaging in Kansas is dependent on voluntary reporting from hospitals across the state. With widespread coverage of EDs by KSSP, data can be more complete and timely to better drive analysis and public information Methods:KSSP data was queried through the Electronic Surveillance System for the Early Notification of Community-based Epidemics (ESSENCE) v.1.20 provided by the National Syndromic Surveillance Program. Data between June 12, 2017 and August 13, 2017 were queried. The first query (Query A, Table 1.) searched the Discharge Diagnosis History field for the “W39” ICD-10 Diagnosis code, “Discharge of firework.” These records were searched for common firework terms contained in the Chief Complaint History field. These firework-related free text terms (Query B, Table 1.) were then combined with other potential firework-related terms to create a preliminary free text query (Query C, Table 1.). This preliminary query was run on the Chief Complaint History field. Data were then searched for false positive cases and appropriate negation terms were included to accommodate this. The new query with negation terms (Query D, Table 1.) was run on the Chief Complaint History field, combined with the results from the Discharge Diagnosis History field, and then combined records were de-duplicated based on a unique visit identifier. The final data set was then classified by the anatomical location of the injury and the gender and age group of the patient. Results:The initial query (Query A, Table 1.) for the diagnosis code “W39” returned 101 unique ED visits. Of these 101 unique ED visits, the following terms were identified in the Chief Complaint History field: shell, artillery, bomb, sparkler, grenade, fire cracker, firework, and firework show. These key terms were translated into Query B, Table 1. Other key terms deemed likely to capture specific firework-related exposures were then included into Query C, Table 1. , including roman, candle, lighter, M80, and punk. Query C was then used to query the Chief Complaint History field, returning 144 unique ED visits. Cases captured by Query C were then reviewed by hand for false positives and the negation terms, lighter fluid, fish, nut, and pistachio, were incorporated the Query D, Table 1. The previous process for Query C was then repeated on Query D, leaving a remaining 136 unique cases. Query A’s 101 unique ED visits was then combined with the 136 unique ED visits captured by Query D and de-duplicated. The de-duplicated data set contained 170 unique ED visits which were then reviewed by hand for false positives. The final removal of false positives from the combined and de-duplicated data set left a remaining 154 unique ED visits for firework-related injuries during this time period.For these data, the most common victims of firework injuries were males, accounting for 65.5% of all firework related ED visits and children ages 0 to 19 accounting for 44.2% of these visits. At every age breakout, male injuries exceeded female injuries. The most common anatomical location of the injury was one or both hands with 38.3% of all injuries mentioned hands as their primary injury. Injuries to the eyes, face, and head accounted for the second most injuries (28.6% of all patients). Conclusions: The selling of fireworks will be a yearly occurrence of a specific exposure that can potentially lead to injuries. Utilizing syndromic surveillance to review the holiday firework injuries is a very rapid method to assess the impact of these injuries and may allow for future direction of public information during the holiday. Having a syndrome definition that builds on knowledge from previous years will allow for quicker case identification as well.State public information regarding firework safety can be significantly bolstered by accurate and rapid data assessment. Developing a firework injury syndrome definition that is accurate and returns information rapidly has allowed for increased buy-in to the Kansas Syndromic Surveillance Program from public information offices, fire marshal’s offices, and other program fields.

Sign in / Sign up

Export Citation Format

Share Document