Background The prevalence of back pain is rising, as is the use of high-cost imaging in the ED. The objective of our study was to determine if an MRI in the ED for patients with back pain resulted in a lower incidence of ED return visit and to determine if these patients had longer ED length of stay (LOS) and use of ED observation.
Methods A retrospective cohort study of consecutive patients seen with back pain was conducted at an urban, university-affiliated ED between 1 January 2012 and 11 July 2014. The association of MRI on return within 7 days was assessed using a χ2 test and a multivariable logistic regression model and the difference in median ED LOS was compared using a Wilcoxon rank-sum test.
Results During the study period, 6094 patients were evaluated in the ED with back pain as the primary diagnosis. Of these, 797 (13%) received an MRI. Among all patients with back pain, 277 (4.5%) returned within 7 days. Univariate analysis found that patients who received an MRI were no less likely to return within 7 days than patients who did not (4.3% vs 4.6%; p=0.68). Patients who had an MRI were more likely to be admitted to observation (74.2% vs 10.8%; p<0.0001) and had a longer ED LOS (median 4.8 hours vs 2.7; p<0.0001). Multivariable regression confirmed that MRI did not decrease the rate of a 7-day return visit (OR=0.98; 95% CI 0.68 to 1.42).
Conclusions In patients with uncomplicated back pain, performing an MRI will not mitigate their likelihood of return; however, it leads to a longer ED LOS and more ED observation admissions.
- imaging, CT/MRI
- Emergency Department utilisation
- cost- effectiveness
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Back pain accounts for a significant number of ED visits annually, is associated with substantial healthcare cost and is a leading cause of disability globally.1–3 Despite clear guidelines, utilisation of high-cost imaging, such as MRI, in the ED is steadily rising without an increasing yield in identifying pathology.4–8
This is likely related to incomplete dissemination of guidelines, incomplete availability of prior studies, a desire to meet patients’ expectations and defensive practice.9 10 We suspect that an additional motivation for advanced imaging may lie in the belief that it will prevent further ED return visits.
The primary objective of our study was to determine if an MRI in the ED for patients with back pain resulted in a lower incidence of ED return visit. Our secondary objective was to determine if patients who received an MRI had longer ED length of stay (LOS) and use of ED observation.
This retrospective cohort study of consecutive patients seen with back pain was conducted at an urban, university-affiliated ED with 106 000 annual visits. The hospital’s institutional review board approved the study (IRB: 2016 P000136).
Study setting and population
All patients who visited the ED from 1 January 2012 to 11 July 2014 with a principal discharge ICD-9 (International Classification of Diseases Version 9.0) code related to low-risk back pain constituted the study population. There were no age limitations.
Data collection and processing
All study variables were selected a priori for analysis. The dataset was extracted by a trained abstractor onto predetermined data forms from our department’s electronic health record.
Our primary outcome measure was per cent of patients returning to the ED within 7 days for back pain. Secondary outcome measures were ED LOS, MRI utilisation during ED return visit and ED discharge disposition. The ED LOS does not include time spent in ED observation.
Continuous variables were summarised using mean with SD or median with interquartiles, and compared using two-sample t-tests or Wilcoxon rank-sum tests between those who did and those who did not have an MRI. Categorical variables were summarised using frequency with percentage and compared using χ2 tests. A multivariable logistic regression analysis was used to examine the association between the measured variables of age, gender, language, prior ED visit in the last 30 days, insurance type and return within 7 days. A p value of <0.05 was considered statistically significant. All analyses were conducted using SAS V.9.4.
Characteristics of study population
During the study period, 6094 patients seen in the ED had back pain as the primary diagnosis. Of these patients, 44% were female, with a mean age of 46 (+/−17). Median ED LOS was 2.9 (IQR 1.8–4.5) hours. Of these, 797 patients (13%) received an MRI. Patients who had an MRI were older (mean 50.8 vs 45.8; p<0.0001), predominantly female (51.1% vs 42.5%; p<0.0001), English speaking (92.6% vs 87.8%; p<0.0001) and were more likely to have private insurance (42.4% vs 36.1%; p<0.0001) (table 1). Patients who had an MRI were more likely to be admitted to observation (74.2% vs 10.8%; p<0.0001), and more likely to be admitted to the hospital (9.4% vs 3.3%; p<0.0001). Patients who had an MRI had a longer ED LOS (median 4.8 hours vs 2.7; p<0.0001) (table 1).
Return visit to ED within 7 days
Overall, 277 patients returned within 7 days (4.5%). Univariate analysis found that patients who received an MRI on their index visit were no less likely to return to the ED within 7 days than patients who did not (4.3% vs 4.6%; p=0.68). Additionally, univariate analysis found that male (p<0.0001), English-speaking patients (p=0.001), and patients with public insurance (p<0.0001) had higher 7-day return rates (table 2). Patients who did not receive an MRI during their index visit had the same rate of MRI use during their return visit for the same complaint (p=0.43).
Multivariate regression confirmed the univariate results and found that MRI was not significantly associated with a lower risk of 7-day return visit in the logistic regression model controlling for age, gender, language and insurance type (OR=0.98; 95% CI 0.68 to 1.42). On the other hand, the multivariate model found that male gender (OR=2.07; 95% CI 1.57 to 2.72), English speaking (OR=2.34; 95% CI 1.38 to 3.98) and possession of a public insurance versus private insurance (OR=1.89; 95% CI 1.41 to 2.53) remained significant predictors of 7-day return visit. Patients who were seen in the ED who had a visit 30 days prior to the index visit were no more likely to have an MRI than those who did not (7.4% vs 7.4%; p=0.95).
This study is limited by its inability to capture patients who returned to another hospital. This was a single-centre study at a tertiary care centre in the USA with 24-hour access to MRI. As a result, the findings may not be generalisable to a global healthcare setting. Additionally, as a retrospective study, we are unable to determine why the testing was ordered, and our ability to draw causal relationships is limited.
Our study suggests that the use of MRI for patients presenting to the ED with low-risk back pain increased ED LOS, increased ED observation admissions and did not result in a lower 7-day return visits.
Patients who received an MRI were more likely to be female, speak English and have private insurance. We suspect that this is related to these groups’ ability to advocate for themselves and obtain the testing they desire.
Several studies have shown that MRI has limited or no impact on outcomes,11 and that higher MRI utilisation has been associated with increased ED LOS and crowding.12 13 Our data support several of these assertions. We suspect that some physicians may be driven by a belief that there is a psychological effect on patients’ symptoms following an MRI, and that definitive imaging may prevent return ED visits. Our study demonstrates that this latter assumption is not the case, and that MRI is not associated with lower odds of 7-day return. Rather, in an era of increasing pressure to optimise throughput, our findings show that patients who receive an MRI require more ED resources with extended LOS and more ED observation admissions.
Contributors All authors listed have contributed sufficiently to the project to be included as authors, and all those who are qualified to be authors are listed in the author byline. To the best of our knowledge, no conflict of interest, financial or other, exists.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.