Impact of strategies to improve flow and lower hospital admissions for low back pain in the emergency department: an interrupted time-series analysis

Abstract

Background: Low back pain (LBP) is a common, disabling condition that often results in presentation to the emergency department (ED). There is currently a paucity of research examining how to optimise care for this patient group and concurrently improve service efficiency. We assessed the impact of a suite of co-designed strategies to improve the flow and reduce admission rate of people with LBP presenting to the ED.

Methods: A quasi-experimental implementation science study in a 30-bed ED at an Australian tertiary referral hospital. We implemented four improvement strategies: (1) co-design and development of a local ED guideline, (2) patient handout, (3) clinician education programme and (4) rapid-access physiotherapy follow-up clinic. We included all ED presentations from patients ≥16 years old, with a final diagnosis of LBP. Using an interrupted time series analysis, we compared a pre-implementation phase (July 2014-December 2019), a transition phase (July 2020-February 2021) and a post implementation phase (March 2021-November 2021). Our primary outcome was the proportion of patients admitted to hospital, in each phase. Secondary outcomes included ED length of stay (LOS), ED presentation cost, re-presentations within five days, admission LOS and outcome measures for those patients who attended the rapid-access physiotherapy clinic.

Results: There were 5301 LBP presentations pre-implementation, 752 during transition and 732 post implementation. The mean age of patients was 50.3 (20.1) years, 53% were female and 6.5% identified as Aboriginal and/or Torres Strait Islander. We observed a small reduction (2%) in the proportion of admissions per month (post implementation vs pre-implementation slope difference 0.98 ((95% CI 0.928 to 1.035), p=0.467) and a monthly decrease of 1.5 min in the average ED LOS for these patients (post implementation vs pre-implementation slope difference -1.5 ((95% CI -8.1 to 5.1), p=0.653) which was not statistically significant. 121 patients were referred to the follow-up physiotherapy clinic, and 41% attended an appointment.

Conclusion: The improvement strategies did not significantly reduce admissions or LOS for people with LBP presenting to the ED, and based on the CIs, we cannot exclude that these increased.