# Temporal Patterns of Healthcare Provider Referrals in an Emergency Department Research Study

## TL;DR

This study analyzed when and by whom patients were referred to a research study in an emergency department, finding patterns that could improve future recruitment strategies.

## Contribution

The study identifies temporal and provider-specific referral patterns in an ED research context, offering actionable insights for optimizing engagement strategies.

## Key findings

- Referral rates varied by month, with higher rates in December and lower in September.
- Registered nurses were the primary referrers, with increased activity after engagement initiatives.
- Physician referrals increased during overnight shifts in later years.

## Abstract

Background: Clinical research in the emergency department (ED) relies on healthcare providers to identify and refer eligible patients, yet the ED’s fast-paced and unpredictable workflow can limit provider capacity for research participation. To inform more efficient engagement strategies for the Evaluating Microdosing in the Emergency Department (EMED) study, which is evaluating methods of initiating buprenorphine/naloxone from EDs, we examined temporal and provider-specific patterns of healthcare provider referrals (HCPRs).

Objective: To quantify temporal variation in referral activity by comparing referral rates per hour across months, days of the week, and work shifts, adjusting for research assistant availability.

Methods: We conducted a retrospective analysis of HCPRs to the EMED study at the Vancouver General Hospital's ED from July 23, 2021, to December 31, 2023. Referral records captured provider type and referral timing and were analyzed during periods of research assistant coverage (morning 07:00-14:59, evening 15:00-22:59, and occasional overnight 23:00-06:59). We calculated referral frequency density (referrals per research-assistant hour), and compared rates across months, days of the week, and shifts using Poisson regression with an offset for research-assistant availability.

Results: From July 2021 to December 2023, we received 2,244 HCPRs, primarily from registered nurses (1,928; 85.9%), followed by physicians (230; 10.2%), other providers (56; 2.5%), pharmacists (26; 1.2%), and social workers (4; 0.2%). After adjusting for research assistant availability, referral rates varied by month (compared to January, higher in December (incidence rate ratio (IRR) = 1.38); lower in September (IRR = 0.77)), by day of the week (compared to Monday, higher on Thursday (IRR = 1.39) and Friday (IRR = 1.20)), and by shift (compared to evenings, higher in mornings (IRR = 1.29); lower in overnights (IRR = 0.37)). Across all months, nurses consistently had the highest referral frequency density, with an apparent increase beginning in October 2022 that coincided with enhanced engagement activities, and a temporary decrease in November 2023 during a brief pause in EMED activity. Referral activity was most commonly observed during morning shifts for nurses, pharmacists, and other providers, whereas physician referrals shifted over time, with increased overnight referral frequency in later years.

Conclusion: HCPRs to EMED were temporally clustered by month, day of the week, and shift after accounting for research assistant availability, and were predominantly contributed by registered nurses, with an apparent increase following enhanced engagement activities. These patterns may help inform the timing and tailoring of future engagement approaches, particularly for provider groups with consistently low referral activity.