From Big Data to the Bedside: answering big questions in emergency department pain care using artificial intelligence and patient-reported outcomes

Grant ID: EMLE-166R34-2020-CHU

Project Summary

One of the main reasons that acute pain is not well treated in the emergency department (ED) setting is that pain is difficult to measure. While patient-reported outcome measures (PROMS) are commonly used to help guide treatment of pain in settings such as chronic pain care, cancer care and migraine care, there are no similar tools available for patients with acute pain in the ED. Further hampering efforts to provide better ED pain care is poor overall understanding of the numbers and types of patients that experience pain.

Since it is a symptom rather than a diagnosis, information about pain is not systematically collected and is often obscured within free-text clinical notes. The lack of readily-available data makes it difficult to determine who exactly has experienced pain, and to design research studies to evaluate new and existing treatments.

Researchers aim to validate a PROM for pain care in the ED by administering to 400 patients who present with pain to one of two large hospital EDs. The aim is to find out the incidence and characteristics of patients who present with pain to the Royal Brisbane and Women’s Hospital ED, by using novel machine- and deep-learning techniques to process free-text information from clinical notes. This study will provide new knowledge and techniques that are essential for clinician-researchers to design and conduct studies that will ultimately improve pain care in the ED.

---

Leveraged Funds

- Queensland University of Technology, 2022, $20,000

---

Dissemination

Publication:

- Hughes, J.A., Douglas, C., Jones, L., Brown, N.J., Nguyen, A., Jarugula, R., Lyrstedt, A., Hazelwood, S., Wu, Y., Sadat Saleh, F., Chu, K., 2023. Identifying patients presenting in pain to the adult emergency department: A binary classification task and description of prevalence. International Emergency Nursing, 68, 101272. https://doi.org/10.1016/j.ienj.2023.101272.

Conference presentations:

- Hughes, J.A., Wu, Y., Hazelwood, S., Brown, N., Joines, L., Douglas, C., Jarugula, R., Chu, K., Nguyen, A., Identifying changes in emergency department pain care during the COVID-19 pandemic - and application of clinical text deep learning models. 4th Global Conference on Emergency Nursing and Trauma Care, Gothenburg, Sweden, 9-77 November 2023

- Hughes, J.A. Determining changes in pain care during the COVID-19 pandemic using clinical text deep learning models. International Conference for Emergency Nursing, Perth, Australia, 4-6 October 2023.

- Hughes, J.A., Thirty minutes too long - defining time-based targets for pain care in the emergency department in the context of patient-reported outcomes (Winner of best oral presentation). International Conference for Emergency Nursing, Gold Coast, Australia, October 2022.

- Hughes, J.A., Patient-reported outcome measures of pain care in the adult emergency department - Introducing the APS-POQ-RED. International Conference for Emergency Nursing, Gold Coast, Australia, October 2022.

- Wu, Y., Developing robust clinical text deep learning models - a "painless" approach. MedINFO23, Sydney, Australia, July 2023.

- Hughes, J., Did the Patient Present to the ED in Pain? An Application of Artificial Intelligence (Winner of best oral presentation). International Conference for Emergency Nursing, online, 2021.

- Hughes, J.A., Brown, N.J., Vu, T., Nguyen, A., Did the Patient Present to the ED in Pain? An Application of Artificial Intelligence. Herston Health Precinct Symposium, December 2020.

---

SHARE