Making the most of injury surveillance data: Using narrative text to identify exposure information in case-control studies

doi:10.1016/j.injury.2014.11.012

Injury

Volume 46, Issue 5, May 2015, Pages 891-897

https://doi.org/10.1016/j.injury.2014.11.012 Get rights and content

Abstract

Introduction

Free-text fields in injury surveillance databases can provide detailed information beyond routinely coded data. Additional data, such as exposures and covariates can be identified from narrative text and used to conduct case-control studies.

Methods

To illustrate this, we developed a text-search algorithm to identify helmet status (worn, not worn, use unknown) in the U.S. National Electronic Injury Surveillance System (NEISS) narratives for bicycling and other sports injuries from 2005 to 2011. We calculated adjusted odds ratios (ORs) for head injury associated with helmet use, with non-head injuries representing controls. For bicycling, we validated ORs against published estimates. ORs were calculated for other sports and we examined factors associated with helmet reporting.

Results

Of 105,614 bicycling injury narratives reviewed, 14.1% contained sufficient helmet information for use in the case-control study. The adjusted ORs for head injuries associated with helmet-wearing were smaller than, but directionally consistent, with previously published estimates (e.g., 1999 Cochrane Review). ORs illustrated a protective effect of helmets for other sports as well (less than 1).

Conclusions

This exploratory analysis illustrates the potential utility of relatively simple text-search algorithms to identify additional variables in surveillance data. Limitations of this study include possible selection bias and the inability to identify individuals with multiple injuries. A similar approach can be applied to study other injuries, conditions, risks, or protective factors. This approach may serve as an efficient method to extend the utility of injury surveillance data to conduct epidemiological research.

Introduction

Administrative health and injury surveillance databases often include narrative text fields that provide additional detailed information beyond routinely coded data. Researchers use free text to validate coded data retrospectively or to obtain supplemental information about patients, illnesses, injuries, comorbidities, outcomes, or health services received [1], [2], [3], [4], [5], [6]. Secondary use of data from free text illustrates one way to extend the value of electronic health information for application in clinical research, quality improvement, or public health surveillance [7]. To our knowledge, however, researchers have not used narratives or free text to obtain additional information for a case-control study.

The National Electronic Injury Surveillance System (NEISS) is an electronic database of injury information from a national probability sample of U.S. emergency departments (EDs), managed by the Consumer Product Safety Commission (CPSC). NEISS provides coded information on body part injured and injury type and has been widely used in epidemiological studies to analyze injury mechanisms for many sports. NEISS does not systematically capture information on known risk factors (e.g. alcohol use) or protective factors (e.g. helmets or other protective equipment). However, since January 1, 2002, NEISS data have included 142-character narratives that provide additional detail on the injury and circumstances around its occurrence. For some injuries such as bicycling, narrative text may provide important etiological information about the injury and whether protective gear was used. Although some studies have used NEISS narrative data to ascertain the activity at the time of injury, they have not been widely used to evaluate the prevalence or effectiveness of protective devices, such as helmets.

Many head injuries can be prevented through improved use of protective gear, especially helmets. Several landmark case-control studies have shown that in bicycling, appropriate helmet use reduces head injuries by up to 85% [8], [9], [10]. These studies provide “gold standards” for estimates of the protective effect of helmets.

The primary objective of this study was to validate the use of narrative-derived exposure information by comparing odds ratios (ORs) for the association between bicycle helmet use and ED-reported head injuries obtained from NEISS with ORs previously reported in the literature. As an exploratory analysis, we also estimated ORs for head injuries associated with helmet use for sports-related injuries with NEISS narratives that report helmet use. Finally, we investigated the factors associated with report of helmet information in the narrative text.

Section snippets

Methods

We obtained NEISS data from the Consumer Product Safety Commission for sports-related injuries that presented to NEISS hospital emergency departments from 2005 to 2011. Injury information provided by NEISS includes age, sex, ethnicity, body part, diagnosis, discharge disposition, location of incident, consumer product(s) associated with injury, and 142-character narrative text field. Diagnosis code refers to the “most severe and specific diagnosis” given by the attending physician and is

Bicycling injuries

There were 105,614 bicycling injury narratives reported in NEISS from 2005 to 2011, of which 14,925 (14.1%) referenced helmet use. Of narratives that referenced helmets, 5270 (35.3%) were categorized as helmeted, 7287 (48.8%) as unhelmeted, and 2368 (15.9%) with helmet mentioned but use unknown.

A 10% random sample of bicycling injury narratives containing the word helmet resulted in 1493 narratives, all of which were reviewed for validation of the text-search algorithm. Among the reviewed

Discussion

This study utilized oft-overlooked narrative text fields in a national injury surveillance database to identify and code supplemental information (in this case exposure data) for use in an illustrative case-control study. This approach can potentially be used on any free text data, such as administrative data or medical records, to extend the utility of this information in an efficient and economical manner. This exploratory case-control study of bicycling head injuries and helmet use resulted

Conflict of interest statement

Authors have no conflicts of interest to declare.

Acknowledgements

Research reported in this publication was supported by the National Institute of Child Health and Human Development of the U.S. National Institutes of Health under award number T32HD057822 (Rivara). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Dr. Rivara holds the Seattle Children's Guild Endowed Chair in Paediatrics. Dr. Hagel holds the Alberta Children's Hospital Foundation Professorship in

References (17)

K. McKenzie et al.
The use of narrative text for injury surveillance research: a systematic review
Accid Anal Prev Mar
(2010)
H.J. Murff et al.
Automated identification of postoperative complications within an electronic medical record using natural language processing
J Am Med Assoc
(2011)
Z. Wang et al.
Extracting diagnoses and investigation results from unstructured text in electronic health records by semi-supervised machine learning
PLoS ONE
(2012)
Y. Stahl et al.
Psychosocial health information in free text notes of Swedish children's health records
Scand J Caring Sci
(August 2012)
R.A. Wilke et al.
Use of an electronic medical record for the identification of research subjects with diabetes mellitus
Clin Med Res
(2007)
M. Steidl et al.
Data for free – can an electronic medical record provide outcome data for incontinence/prolapse repair procedures?
J Urol
(2013)
W.R. Hersh
Adding value to the electronic health record through secondary use of data for quality assurance, research, and surveillance
Am J Manag Care
(2007)
R.S. Thompson et al.
A case-control study of the effectiveness of bicycle safety helmets
N Engl J Med
(1989)

There are more references available in the full text version of this article.

Cited by (23)

The epidemiology of crib-related head injuries: A ten-year nationwide analysis
2023, American Journal of Emergency Medicine
Falls from cribs resulting in head injury are understudied and poorly characterized. The purpose of this study was to advance current understanding of the prevalence, descriptive characteristics of injury victims, and the types of crib fall-related head injuries (CFHI) using queried patient cases from the National Electronic Injury Surveillance System (NEISS) database.
Using the US Consumer Product Safety Commission's System NEISS database, we queried all CFHIs among children from over 100 emergency departments (EDs). Patient information regarding age, race, sex, location of the incident, diagnoses, ED disposition, and sequelae were analyzed. The number of CFHI from all US EDs during each year was also collected from the database.
There were an estimated 54,799 (95% CI: 30,228–79,369) total visits to EDs for CFHIs between 2012 and 2021, with a decrease in incidence of approximately 20% during the onset of the COVID-19 pandemic (2019: 5616 cases, 2020: 4459 cases). The annual incidence of injuries showed no significant trend over the 10-year study period. An available subset of 1782 cases of head injuries from approximately 100 EDs was analyzed, and 1442 cases were included in final analysis. Injuries were sorted into three primary categories: unspecified closed head injury (e.g., closed head injury, blunt head trauma, or traumatic brain injury), concussion, or open head injury and skull fracture. Unspecified closed head injuries were the most common of all head injuries (95.4%, 1376/1442). Open head injuries (14/1442, 0.97%) and concussions 3.6% (52/1442, 3.6%) were rare. Most injuries involved children under the age of 1 (42.6%) compared to children who were 1, 2, 3, or 4-years old. About a fourth of patients had other diagnoses in addition to their primary injury including scalp/forehead hematomas, emesis, and contusions. Female patients were more likely to present with other diagnoses in addition to their primary head injury (Difference: 12.3%, 95% CI: 9.87%–15.4%, p < .0001).
Despite minimum rail height requirements set by the Consumer Safety Product Commission (CPSC), head injuries associated with crib falls are prevalent in the United States. However, most injuries were minor with a vast majority of patients being released following examination and treatment.
A transformative solution for construction safety: Blockchain-based system for accident information management
2023, Journal of Industrial Information Integration
Effective management of accident information is a crucial component of safety management within the construction industry, as it reflects the safety performance of the company and allows them to identify the root causes of accidents and prevent similar accidents in the future. However, existing safety information systems provide self-owned, isolated, and centralized environments and fail to present a secure, transparent, and trustworthy platform for monitoring and management of accident information. To address these issues, this paper presents a novel decentralized blockchain-based system for accident/incident information management of construction projects. The proposed system leverages the benefits and advantages of blockchain, smart contracts, and decentralized IPFS storage to address the security transparency, tampering, and trustworthiness issues of the conventional approaches. The proposed system is simulated by using real-world construction accident data to demonstrate how blockchain technology can provide a novel solution to assure security, transparency, authenticity, availability, and immutability of the accident/incident data for improving safety management.
Epidemiology and healthcare utilization for rectal foreign bodies in United States adults, 2012–2021
2023, American Journal of Emergency Medicine
Presentations to the emergency department for rectal foreign bodies are common, but there is little epidemiologic information on this condition. This limits the ability to provide evidence-based education to trainees regarding the populations affected, the types and frequency of foreign bodies, and factors associated with hospitalization. To address this, we analyzed national estimates of emergency department presentations for rectal foreign bodies from 2012 to 2021 in the US.
We queried the National Electronic Injury Surveillance System for any injury to the ‘pubic region’ or ‘lower trunk’ with an accompanying diagnosis of foreign body, puncture, or laceration. Two authors manually reviewed all clinical narratives to identify cases of rectal foreign bodies. National estimates were determined using weighting and strata variables, incidence rates calculated using census data, trends assessed by linear regression, and factors associated with hospitalization identified by multivariable logistic regression.
From 885 cases, there were an estimated 38,948 (95% CI, 32,040-45,856) emergency department visits for rectal foreign bodies among individuals ≥15 years from 2012 to 2021. The average age was 43, 77.8% were male, 55.4% of foreign bodies were sexual devices, and 40.8% required hospitalization. The annual incidence of presentations for rectal foreign bodies increased from 1.2 in 2012 to 1.9 per 100,000 persons in 2021 (R² = 0.84, p < 0.01). Males have a bimodal age distribution peaking in the fifth decade, while females have a right-skewed age distribution peaking in the second decade. Female sex (odds ratio [OR] 0.4; 95% confidence interval [CI], 0.2–0.6) and, compared to sexual devices, balls/marbles (OR 0.2; 95% CI, 0.05–0.6) or drugs/paraphernalia (OR 0.1; 95% CI, 0.05–0.4) are associated with a reduced odds of hospitalization.
Presentations to the emergency department for rectal foreign bodies increased for males and females from 2012 to 2021 in the United States. These epidemiologic estimates for a complex form of anorectal trauma provide preclinical information for emergency medicine, surgery, and radiology trainees.
Pediatric all-terrain vehicle (ATV) related head injury rates and patterns: A 10-year nationwide analysis
2023, American Journal of Emergency Medicine
The use of all-terrain vehicles (ATVs) carries significant risk of permanent injury and death, disproportionately affecting children. These injuries commonly affect the head and are especially severe among children as they are often unhelmeted and more likely than adults to experience rollover injuries. Many studies examining patients with ATV-related injuries are single-center cohort studies, with few focusing specifically on head injuries. In the present study, we aimed to characterize the annual incidence of ATV-related head injuries between 2012 and 2021, classify and compare head injury types, and identify descriptive characteristics of ATV-related head injury victims.
Using the US Consumer Product Safety Commission's National Electronic Injury Surveillance System (NEISS) database, we queried all head injuries associated with operating or riding an ATV in children under 18 years-old from over 100 emergency departments (EDs). Patient information regarding age, race, sex, location of incident, diagnoses, and sequelae were analyzed. We also collected the estimated number of ATV-related head injuries from all US EDs using the NEISS algorithm provided by the database.
Using the NEISS algorithm we identified 67,957 (95% CI: 43,608 - 92,305) total pediatric ATV-related head injuries between 2012 and 2021. The annual incidence of ATV-related head injury was similar throughout this study period except for a 20% increase during the COVID-19 pandemic period of 2019–2021 (2019: 6382 injuries, 2020: 6757 injuries, 2021: 7600 injuries). A subset of 1890 cases from approximately 100 EDs were then analyzed. Unspecified closed head injuries were the prevailing type of injury (38%, 900/1890), followed by concussions (27%, 510/1890). More severe injuries included intracranial hemorrhages in 91 children (3.8%, 91/1890). Injuries of all types were predominantly seen in 14–17 year-old's (780/1890, 41%) and in males (64.1%, 1211/1890). In addition, ATV-associated injuries were significantly more common in those coded as white (58.0%, 1096/1890) than any other racial group. ATV-associated accidents among children younger than 9 more commonly occurred at the home compared to accidents involving children older than 9 (57% vs. 32%, p < 0.0001).
ATV-related head injuries cause a significant annual burden among children, with growing incidence in recent years. Further research may wish to explore potential benefits of helmet use and supervision of younger children in possible prevention of these accidents and their associated economic and non-economic costs.
Sectoral patterns of accident process for occupational safety using narrative texts of OSHA database
2021, Safety Science
Citation Excerpt :
More studies on narrative texts analysis can be already found in the review papers (McKenzie et al., 2010; Vallmuur, 2015). First, several studies on identifying and classifying the accident process were conducted using various method such as fuzzy method (Taylor et al., 2014), statistics (Graves et al., 2015), topic modeling (Tanguy et al., 2016), and machine learning (Marucci-Wellman et al., 2017; Goh and Ubeynarayana, 2017; Zhang et al., 2019). Along with development of machine learning, unsupervised and supervised methods have been especially applied to categorize the taxonomy of accident and to predict industries where accidents generated.
The narrative text analytics has recently focused on identifying an accident process in the various fields of safety such as manufacturing, construction, chemicals, and service. In particular, narrative texts allow finding multiple accident factors and types of accident process including industry, hazard, work activity, and accident result. To present similarity and difference of accident process by categorizing those multiple accident factors shared across industries, identifying sectoral patterns of accidents are useful. In this respect, this study aims to identify the sectoral patterns of accident process using narrative texts information contained in accident reports. For this, the textmining and latent Dirichlet allocation (LDA) algorithms are used to extract topics of accidents and their main factors, matched with class of industries. As a result of the case study for the Occupational Safety and Health Administration (OSHA) in the United States, the five sectoral patterns of accident process are identified: scale-intensive, facility-intensive, supplier-dominated, market-dominated, and service-dominated patterns. According to these sectoral patterns, managers and policy makers in the fields of safety take a look at the management issues related to the industry, source, activity, and accident result, considering respective characteristics of industrial sites.
Applying interpretable machine learning to classify tree and utility pole related crash injury types
2021, IATSS Research
In spite of enormous improvements in vehicle safety, roadway design, and operations, there is still an excessive amount of traffic crashes resulting in injuries and major productivity losses. Despite the many studies on factors of crash frequency and injury severity, there is still further research to be conducted. Tree and utility pole/other pole related (TUOP) crashes present approximately 12 to 15% of all roadway departure (RwD) fatal crashes in the U.S. The count of TUOP crashes comprise nearly 22% of all fatal crashes in Louisiana. From 2010 to 2016, there were 55,857 TUOP crashes reported in Louisiana. Individually examining each of these crash reports is not a realistic option to investigate crash factors. Therefore, this study employed text mining and interpretable machine learning (IML) techniques to analyze all TUOP crashes (with available crash narratives) that occurred in Louisiana from 2010 to 2016. This study has two major goals: 1) to develop a framework for applying machine learning models to classify injury levels from unstructured textual content, and 2) to apply an IML framework that provides probability measures of keywords and their association with the injury classification. The present study employed three machine learning algorithms in the classification of injury levels based on the crash narrative data. Of the used modeling techniques, the eXtreme gradient boosting (XGBoost) model shows better performance, with accuracy ranging from 0.70 to 24% for the training data and from 0.30% to 16% for the test data.

View all citing articles on Scopus

View full text

Making the most of injury surveillance data: Using narrative text to identify exposure information in case-control studies

Abstract

Introduction

Methods

Results

Conclusions

Introduction

Section snippets

Methods

Bicycling injuries

Discussion

Conflict of interest statement

Acknowledgements

The use of narrative text for injury surveillance research: a systematic review

Accid Anal Prev Mar

Automated identification of postoperative complications within an electronic medical record using natural language processing

J Am Med Assoc

Extracting diagnoses and investigation results from unstructured text in electronic health records by semi-supervised machine learning

PLoS ONE

Psychosocial health information in free text notes of Swedish children's health records

Scand J Caring Sci

Use of an electronic medical record for the identification of research subjects with diabetes mellitus

Clin Med Res

Data for free – can an electronic medical record provide outcome data for incontinence/prolapse repair procedures?

J Urol

Adding value to the electronic health record through secondary use of data for quality assurance, research, and surveillance

Am J Manag Care

A case-control study of the effectiveness of bicycle safety helmets

N Engl J Med