Law enforcement officers (LEOs) frequently intersect with healthcare settings, especially emergency departments (EDs) [1]. LEOs co-respond with emergency medical services (EMS) to many 9-1-1 calls (e.g., gunshot wounds, mental health crises, unresponsive child) and play an integral prehospital role. LEOs survey a scene for safety, relay key information to clinicians at the hospital, and often administer initial first aid measures when they are the first to arrive. However, once LEOs arrive in EDs or other hospital settings, their role is less defined. Patients may appreciate the efficiency of filing a report with LEOs while waiting for medical care in the ED; staff may feel safer with LEO presence when there is an acute threat of violence. However, sometimes the role of LEOs and their public safety priorities may not align with patient or family needs and the priorities of the medical team [2]. The pediatric setting presents unique criminal justice and ethical concerns as minors may not have legal or medical autonomy and often lack the developmental maturity to advocate for themselves [2]. LEO presence raises questions about patient privacy, medical and legal priorities, and the rights of unaccompanied children seeking medical care. Existing literature signals LEO presence in treatment spaces may impact clinical care and health outcomes [3,4].

Existing qualitative data reports LEO presence during medical care can violate patient privacy and weaken the shared decision-making relationship between providers and patients [1,3]. One study revealed that when ED nurses perceive patient criminality, they justify delays in medical care, and LEO accompaniment often leads to criminal stigma [[5], [6], [7]]. LEO interrogations in treatment spaces increase adult trauma victims' sense of vulnerability as patients [8]. When medical team members are seen speaking with LEOs, patients also report a deepened sense of mistrust, even withholding information from clinicians when they fear it will be shared with LEOs [8]. This incomplete information exchange can lead to missed or delayed diagnosis and suboptimal care. While many outcomes studies are observational and do not lead to causative conclusions, some of these qualitative studies begin to point towards mechanisms through which LEOs may cause harm. These studies signal potential impact on outcomes from ED length of stay to pain management to return visits for missed diagnoses [2,8,9]. Despite this growing evidence, quantitative analysis of LEO prevalence and potential negative impact on the key outcomes identified in the qualitative literature (e.g., delays in medication administration, prolonged length of stay, delayed or missed diagnoses due to lack of information sharing, etc.) is currently nonexistent.

The lack of quantitative data is in part due to the lack of reliable and standardized electronic health record (EHR) documentation of LEO presence; LEO presence is documented in narrative form in provider notes, making extraction from discrete data fields difficult. Structured EHR fields such as arrival method are unreliable sources to accurately identify LEO presence; LEOs frequently accompany EMS when needed but do not often physically transport pediatric patients who require medical attention. Further, LEOs often arrive after the patient when called by hospital staff or a parent, or after co-responding with EMS. Without means to accurately identify exposure to LEOs during ED encounters, there cannot be thorough examination of prevalence and impact on related health and public safety outcomes. EHR-based methods, such as keyword searches and automated multivariable models have the potential to identify LEO presence accurately and reliably without discrete data fields, overcoming the biggest barrier to quantitative analysis.

The primary objective of this study was to create a reliable and generalizable method to identify LEO presence during ED encounters using an EHR-based multivariable predictive model. This scalable EHR-based approach will 1) facilitate large scale analysis of LEO prevalence in EDs and related impact on clinical outcomes, as opposed to small sample subsets requiring manual review 2) enable real-time surveillance of changes in LEO prevalence and impact after intervention (policy or guideline) implementation, and 3) facilitate comparisons of results among multiple institutions to identify similarities and differences that will inform tailored interventions.