The development of a machine learning algorithm to identify occupational injuries in agriculture using pre-hospital care reports

Leveraging Administrative Health Databases to Address Health Challenges in Farming Populations: Scoping Review and Bibliometric Analysis (1975-2024)

BACKGROUND

Although agricultural health has gained importance, to date, much of the existing research relies on traditional epidemiological approaches that often face limitations related to sample size, geographic scope, temporal coverage, and the range of health events examined. To address these challenges, a complementary approach involves leveraging and reusing data beyond its original purpose. Administrative health databases (AHDs) are increasingly reused in population-based research and digital public health, especially for populations such as farmers, who face distinct environmental risks.

OBJECTIVE

We aimed to explore the reuse of AHDs in addressing health issues within farming populations by summarizing the current landscape of AHD-based research and identifying key areas of interest, research gaps, and unmet needs.

METHODS

We conducted a scoping review and bibliometric analysis using PubMed and Web of Science. Building upon previous reviews of AHD-based public health research, we conducted a comprehensive literature search using 72 terms related to the farming population and AHDs. To identify research hot spots, directions, and gaps, we used keyword frequency, co-occurrence, and thematic mapping. We also explored the bibliometric profile of the farming exposome by mapping keyword co-occurrences between environmental factors and health outcomes.

RESULTS

Between 1975 and April 2024, 296 publications across 118 journals, predominantly from high-income countries, were identified. Nearly one-third of these publications were associated with well-established cohorts, such as Agriculture and Cancer and Agricultural Health Study. The most frequently used AHDs included disease registers (158/296, 53.4%), electronic health records (124/296, 41.9%), insurance claims (106/296, 35.8%), population registers (95/296, 32.1%), and hospital discharge databases (41/296, 13.9%). Fifty (16.9%) of 296 studies involved >1 million participants. Although a broad range of exposure proxies were used, most studies (254/296, 85.8%) relied on broad proxies, which failed to capture the specifics of farming tasks. Research on the farming exposome remains underexplored, with a predominant focus on the specific external exposome, particularly pesticide exposure. A limited range of health events have been examined, primarily cancer, mortality, and injuries.

CONCLUSIONS

The increasing use of AHDs holds major potential to advance public health research within farming populations. However, substantial research gaps persist, particularly in low-income regions and among underrepresented farming subgroups, such as women, children, and contingent workers. Emerging issues, including exposure to per- and polyfluoroalkyl substances, biological agents, microbiome, microplastics, and climate change, warrant further research. Major gaps also persist in understanding various health conditions, including cardiovascular, reproductive, ocular, sleep-related, age-related, and autoimmune diseases. Addressing these overlooked areas is essential for comprehending the health risks faced by farming communities and guiding public health policies. Within this context, promoting AHD-based research, in conjunction with other digital data sources (eg, mobile health, social health data, and wearables) and artificial intelligence approaches, represents a promising avenue for future exploration.

Estimating occupation-related crashes in light and medium size vehicles in Kentucky: A text mining and data linkage approach

Occupational motor vehicle (OMV) crashes are a leading cause of occupation-related injury and fatality in the United States. Statewide crash databases provide a good source for identifying crashes involving large commercial vehicles but are less optimal for identifying OMV crashes involving light or medium vehicles. This has led to an underestimation of OMV crash counts across states and an incomplete picture of the magnitude of the problem. The goal of this study was to develop and pilot a systematic process for identifying OMV crashes in light and medium vehicles using both state crash and health-related surveillance databases. A two-fold process was developed that included: 1) a machine learning approach for mining crash narratives and 2) a deterministic data linkage effort with crash state data and workers compensation (WC) claims records and emergency medical service (EMS) data, independently. Overall, the combined process identified 5,302 OMV crashes in light and medium vehicles within one year's worth of crash data. Findings suggest the inclusion of multi-method approaches and multiple data sources can be implemented and used to improve OMV crash surveillance in the United States.

Identifying prehospital trauma patients from ambulance patient care records; comparing two methods using linked data in New South Wales, Australia

BACKGROUND

Linked datasets for trauma system monitoring should ideally follow patients from the prehospital scene to hospital admission and post-discharge. Having a well-defined cohort when using administrative datasets is essential because they must capture the representative population. Unlike hospital electronic health records (EHR), ambulance patient-care records lack access to sources beyond immediate clinical notes. Relying on a limited set of variables to define a study population might result in missed patient inclusion. We aimed to compare two methods of identifying prehospital trauma patients: one using only those documented under a trauma protocol and another incorporating additional data elements from ambulance patient care records.

METHODS

We analyzed data from six routinely collected administrative datasets from 2015 to 2018, including ambulance patient-care records, aeromedical data, emergency department visits, hospitalizations, rehabilitation outcomes, and death records. Three prehospital trauma cohorts were created: an Extended-T-protocol cohort (patients transported under a trauma protocol and/or patients with prespecified criteria from structured data fields), T-protocol cohort (only patients documented as transported under a trauma protocol) and non-T-protocol (extended-T-protocol population not in the T-protocol cohort). Patient-encounter characteristics, mortality, clinical and post-hospital discharge outcomes were compared. A conservative p-value of 0.01 was considered significant RESULTS: Of 1 038 263 patient-encounters included in the extended-T-population 814 729 (78.5 %) were transported, with 438 893 (53.9 %) documented as a T-protocol patient. Half (49.6 %) of the non-T-protocol sub-cohort had an International Classification of Disease 10th edition injury or external cause code, indicating 79644 missed patients when a T-protocol-only definition was used. The non-T-protocol sub-cohort also identified additional patients with intubation, prehospital blood transfusion and positive eFAST. A higher proportion of non-T protocol patients than T-protocol patients were admitted to the ICU (4.6% vs 3.6 %), ventilated (1.8% vs 1.3 %), received in-hospital transfusion (7.9 vs 6.8 %) or died (1.8% vs 1.3 %). Urgent trauma surgery was similar between groups (1.3% vs 1.4 %).

CONCLUSION

The extended-T-population definition identified 50 % more admitted patients with an ICD-10-AM code consistent with an injury, including patients with severe trauma. Developing an EHR phenotype incorporating multiple data fields of ambulance-transported trauma patients for use with linked data may avoid missing these patients.

Understanding Stakeholder Dissemination Preferences for an Agriculture, Forestry, and Fishing Injury Surveillance System

Researchers and epidemiologists are working to improve the capture of agriculture, forestry, and fishing (AgFF) injuries in a variety of ways. A critical component of any surveillance system is the dissemination of information. The purpose of this paper is to report on a survey conducted with AgFF injury surveillance stakeholders to understand preferred dissemination strategies. The survey was distributed using REDCap via web link to organizational stakeholders, which included advisory board members, safety trainers, industry managers and workers, and research collaborators. In total, there were 75 respondents (21% response rate). Occupation and industry influenced preference in update methods. Regarding the length and breadth of updates, 63% of respondents prefer reports (one to five pages), followed by 57% desiring a summary (less than one page), while only 24% wanted a detailed analysis. Social media and news preferences were also different among stakeholders. Surveillance data were desired for 1) trend analysis, 2) tailoring activities and solutions for education, training, outreach and interventions and 3) for research purposes such as grant proposals and evaluation. The dissemination of injury surveillance data should be tailored to the intended audience. Greater attention needs to be paid to the ways in which we share our findings.

A comparative study of electronic and pen-paper safety inspections: A mixed method study design for assessing ergonomic parameters

BACKGROUND

Workplace inspections are applied to facilitate the adherence to the occupational health and safety regulations. The Iranian Ministry of Health introduced a new software system for tablets to inspect workplaces.

OBJECTIVES

The aim of this study was to take measurements of the usability, mental workload, and mood of inspectors.

METHODS

Inspectors used both pen-and-paper and tablet methods to inspect the automotive industry in a mixed-method procedure. The NASA-TLX score, QUIS score, I-PANAS (SF) situation, inspection time, and number of errors were collected throughout the procedure. The differences were investigated using a paired sample and the Wilcoxon signed ranks test.

RESULTS

In terms of efficacy, using the tablet resulted in lower error rates, but it took longer to complete the inspection task (P < 0.001). Participants perceived a lower workload when inspecting with a tablet rather than the traditional method. (Mental Demand: p < 0.002, Performance: p < 0.009, Effort: p < 0.012, TLX: p < 0.002 based on various subcomponents of NASA-TLX). The newly introduced system's usability was insufficient.

CONCLUSION

Although the use of tablets has improved safety inspections, ergonomic redesign of the system and consideration of a user-centered approach, as well as inspector training, can make the system more likely to succeed.

Ensuring data quality and maximizing efficiency in coding agricultural and forestry injuries: Lessons to improve occupational injury surveillance

INTRODUCTION

Specialized occupational injury surveillance systems are filling the gap in the undercount of work-related injuries in industries such as agriculture and forestry. To ensure data quality and maximize efficiency in the operation of a regional occupational injury surveillance system, the need for continued dual coding of occupational injury records was assessed.

METHODS

Kappa scores and percent agreement were used to compare interrater reliability for assigned variables in 1,259 agricultural and forestry injuries identified in pre-hospital care reports. The variables used for the comparison included type of event, source of injury, nature of injury, part of body, injury location, intentionality, and farm and agriculture injury classification (FAIC).

RESULTS

Kappa (κ) ranged from 0.2605 for secondary source to 0.8494 for event and exposure. Individual coder accuracy ranged from medium to high levels of agreement. Agreement beyond the first digit of OIICS coding was measured in percent agreement, and type of event or exposure, body part, and primary source of injury continued to meet levels of accord reaching 70% or greater agreement between all coders and the final choice, even to the most detailed 4th digit of OIICS.

CONCLUSIONS

This research supports evidence-based decision making in customizing an occupational injury surveillance system, ultimately making it less costly while maintaining data quality. We foresee these methods being applicable to any surveillance system where visual inspection and human decisions are levied.

PRACTICAL APPLICATIONS

Assessing the rigor of occupational injury record coding provides critical information to tailor surveillance protocols, especially those targeted to make the system less costly. System administrators should consider evaluating the quality of coding, especially when dealing with free-text narratives before deciding on single coder protocols. Further, quality checks should remain a part of the system going forward.

Occupational Injury Risk Mitigation: Machine Learning Approach and Feature Optimization for Smart Workplace Surveillance

Forecasting the severity of occupational injuries shall be all industries' top priority. The use of machine learning is theoretically valuable to assist the predictive analysis, thus, this study attempts to propose a feature-optimized predictive model for anticipating occupational injury severity. A public database of 66,405 occupational injury records from OSHA is analyzed using five sets of machine learning models: Support Vector Machine, K-Nearest Neighbors, Naïve Bayes, Decision Tree, and Random Forest. For model comparison, Random Forest outperformed other models with higher accuracy and F1-score. Therefore, it highlighted the potential of ensemble learning as a more accurate prediction model in the field of occupational injury. In constructing the model, this study also proposed the feature optimization technique that revealed the three most important features; 'nature of injury', 'type of event', and 'affected body part' in developing model. The accuracy of the Random Forest model was improved by 0.5% or 0.895 and 0.954 for the prediction of hospitalization and amputation, respectively by redeveloping and optimizing the model with hyperparameter tuning. The feature optimization is essential in providing insight knowledge to the Safety and Health Practitioners for future injury corrective and preventive strategies. This study has shown promising potential for smart workplace surveillance.