Beyond KABCO: Improving Our Understanding of Pedestrian and Bicyclist Injuries with Hospital Data
BACKGROUND Active transportation, including walking and biking, has considerable health, social, and environmental benefits [1]. However, many United States (US) communities lack a comprehensive infrastructure to support safe active transportation and the political will to accelerate investment in safe network development. Even as the US is undergoing a pedestrian traffic fatality crisis, in which the pedestrian fatality rate has increased by 37% from 2009-2017 [2], the health and other impacts of these crashes and key crash prediction variables that can be leveraged for injury prevention remain unclear. Improved and integrated data are vital to inform stakeholder actions and build collaborative intra-agency partnerships to address safety barriers that limit active travel. We developed a strategic implementation plan to create an integrated crash-health outcome surveillance system for NC. As a first step, we linked statewide crash and hospital encounter data for pedestrians and bicyclists. The results can inform future planning, injury prevention, and policy decisions.

PURPOSE
Linking crash and health data has the potential to create a more comprehensive understanding of the nature and severity of pedestrian and bicyclist injuries and broader health outcomes. This, in collaboration with key stakeholders, will inform the development of a transportation system that encompasses the needs of all road users.

METHODOLOGY
To demonstrate the feasibility of data integration, we linked 2017 NC pedestrian/bicyclist crash data with hospital encounter data from the North Carolina Healthcare Association (NCHA). Since hospital encounter data contain personally identifiable information protected under HIPAA, NCHA performed the linkage. NCHA matched records based on crash/hospital admission date, age, sex, five-digit residential ZIP code, and residential street address, using deterministic methods. Throughout the linkage process, we solicited feedback and provided updates to stakeholders through regularly scheduled facilitated discussions.

RESULTS
NCHA linked hospital encounter records for 27% of pedestrians (N=810) and bicyclists (N=295) identified in NC crash reports in 2017. We used health metrics, such as diagnosis codes, hospital admittance status, length of stay, and discharge disposition, to describe the nature and severity of injury. These metrics yielded valuable insights. For example, while most of the injured pedestrians were <60 years of age. the majority of patients this age were discharged home from the emergency department, while older adults, who made up a relatively small proportion of the patient population, were much more likely to be admitted to the hospital for their injuries (Figures 1 and 2).

Figure 1. Frequency of pedestrians treated at NC hospitals, stratified by age group: 2017 (N=810) (see abstract).

Figure 2. Frequency of pedestrians admitted to NC hospitals, stratified by age group: 2017 (N=810) (see abstract).

CONCLUSION
This project demonstrates a widely replicable stakeholder engaged process for linking crash and health outcomes data for the study of pedestrian and bicyclist injuries. It also contributes to the science and best practices around pedestrian and bicycle data linkage methods and demonstrates the knowledge gains produced by linked data systems.

Presenter: Katherine Peticolas
Agency Affiliation: Carolina Center for Health Informatics

Presenter Biographical Statement: [biography]