Background

Musculoskeletal pain is a leading cause of disability globally and a common complaint in patients presenting to the emergency department (ED). [1,2] It is estimated that 13.8% of ED visits are attributable to primary musculoskeletal disorders. [3] Back pain, the most common of these disorders, accounts for 4.3 million ED visits per annum in the US. [2,4] When faced with managing non-traumatic musculoskeletal pain in the ED, physicians often have limited treatment options available. ED providers may find themselves trapped between conflicting mandates to alleviate pain and maintain patient satisfaction while practicing in accordance with national guidelines that recommend reduced opioid prescription and imaging. [5–8] Patients presenting to the ED with musculoskeletal pain may also have culturally conditioned expectations, such as an expectation to receive imaging to assist in their diagnosis and medication to ameliorate their symptoms. [7]

While the busy ED environment is not conducive to a physician spending the amount of time that may be required to provide individualized, patient-specific education and reassurance, this role matches the education and skill set of a physical therapist. [9] Physical therapists have been practicing internationally in EDs for over 20 years and have recently become more prevalent in EDs in the United States. [9,10] As ED patient volumes continue to increase, the number of patients who may potentially benefit from ED-based physical therapy (ED-PT) increased proportionally. [11]

Existing data suggest that ED-PT has a positive effect on a number of clinical and operational outcomes and is viewed positively by both patients and physicians. [12–14] In the primary care setting, early physical therapy (PT) appears to be associated with lower utilization of advanced imaging, lower rates of lumbar spinal injections and lumbar spinal surgery, and improved patient satisfaction. [15–19] In the ED, a primary contact physical therapy model has been shown to reduce wait times and improve patient flow, directing patients to early effective care, and freeing emergency physicians to focus on other emergent cases. [20–24] A primary contact model involves the Physical Therapist acting as primary provider and autonomous practitioner. This practice is now commonplace in the UK, Canada and Australia. It contrasts with the predominant model seen in US EDs in which the Physical Therapist acts as a ‘secondary contact practitioner’ after initial examination by an ED Physician and medical referral. [22,25] Furthermore, ED-PT has been suggested as an alternative to opioid prescribing for pain management in the ED. [26,27] In the primary care setting, early PT referral has been shown to offer reductions in longer-term opioid use and resulted in lower-intensity opioid use for patients with atraumatic lower back pain, although studies to date have failed to show benefit in the context of the ED. [28]

The purpose of this study was to determine the impact of ED-PT on overall rates of imaging, rates of opioid prescribing and length of stay when compared with patients not receiving PT. We hypothesize that patients who receive ED-PT would demonstrate reduced imaging rates, opioid prescribing and length of stay.

Methods

Results

Over the 18-month period, we identified 524 patients who presented to the ED with musculoskeletal pain and were included in the study. A total of 381 patients (72.7%) received ED-PT, whereas 143 patients (27.3%) did not receive PT. The PT and non-PT groups were similar in average age (42.8 years vs. 45.1 years, p = 0.155), gender (percent female: 53% vs. 46.9%, p-0.209), race and average presenting pain score. [Table 1] Primary presenting chief complaint was also similar between groups (cervical, thoracic, or lumbar pain: 57.7% vs. 53.1%, p = 0.345). [Table 2]

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics of study population.

https://doi.org/10.1371/journal.pone.0231476.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Comparison of presenting chief complaint.

https://doi.org/10.1371/journal.pone.0231476.t002

Average ED length of stay was significantly lower in the PT group when compared to the non-PT group (4 hours vs. 6.2 hours, p<0.001, range 1–26 hours vs. 1–28 hours). [Overall rates of imaging were also significantly lower in the PT group (38.3% vs. 51% p = 0.009). PT patients were less likely to have an x-ray performed in the ED (30.0% vs 43.4%, p = 0.00452). However, the rates of CT scanning (8.1% vs 11.2%, p = 0.27572) and MRI (4.7% vs 4.2%, p = 0.79486) did not differ significantly between the two groups. [Fig 1]

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Imaging and opioid administration rates.

https://doi.org/10.1371/journal.pone.0231476.g001

PT patients were less likely to receive and opioid pain medication in the ED (17.5% vs. 32.9%, p<0.001). PT patients were also less likely to receive acetaminophen in the ED (5.8% vs 11.9%, p = 0.01732) but had comparable rates of NSAID administration (21.5% vs 28.7%, p = 0.08544). Rates of outpatient opioid prescriptions were similar between the two groups (16% vs. 21.7%, p = 0.129). [Fig 1]

ED-PT patients received fewer benzodiazepines in the ED (92% vs 16.8%, p = 0.0217); however, benzodiazepine prescribing at the point of discharge was not significantly different (3.9% vs 4.2%, p = 0.9203). Topical Lidocaine preparations were infrequently prescribed in both ED-PT and non-PT groups both in the ED (0.3% vs 0%, p = 0.6033) and at discharge (0.3% vs 2.1%, p = 0.1124). Gabapentin was prescribed to fewer patients in the ED-PT group at discharge (1.8% vs 7.7%, p = 0.0023).

We performed a subgroup analysis of patients presenting with low back pain and matched patients by age and gender. We identified 112 patients with atraumatic low back pain for this matched comparison (PT 56 vs. non-PT 56). Average age in each group was 42 years (range PT 18–82 vs. non-PT 18–81) and 50% of each group was female. Average ED length of stay was similar between groups (PT 3.7 hours vs. non-PT 4.6 hours, p = 0.21). ED-PT patients received fewer imaging studies (PT 25% vs. non-PT 57%, p = 0.029) but had similar rates of opioid administration in the ED (PT 36% vs. non-PT 43%, P = 0.792) and outpatient opioid prescriptions (17.9%. vs 17.9%, p = 1.0). [Fig 2]

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Subgroup comparison of matched patients with low back pain.

https://doi.org/10.1371/journal.pone.0231476.g002

Discussion

We found that patients who received ED-PT had a shorter ED length of stay, lower overall imaging rates, and fewer opioids administered in the ED. This study is the first to focus on the utilization of imaging at the discrete point of the ED visit, and it identifies an association between PT and decreased diagnostic imaging rates and therefore cost.

A number of large cohort and descriptive studies have demonstrated an association between the early initiation of PT in lower back pain and reduced downstream healthcare utilization. [18] Fritz et. al., using data extracted from a large database of employee sponsored health plans, retrospectively analyzed the effects of early (< 14 days) versus delayed (>14 days) PT on downstream healthcare utilization in patients with a first presentation of low back pain (LBP) to primary care providers. LBP-related healthcare costs were analyzed over an 18-month follow up period. Early physical therapy timing was associated with decreased risk of advanced imaging (odds ratio [OR] = 0.34, 95% confidence interval [CI]: 0.29, 0.41), additional physician visits (OR = 0.26, 95% CI: 0.21, 0.32), surgery (OR = 0.45, 95% CI: 0.32, 0.64), injections (OR = 0.42, 95% CI: 0.32, 0.64), and opioid medications (OR = 0.78, 95% CI: 0.66, 0.93) compared with delayed physical therapy. [19] This study is more recently supported by the work of Liu et. al. who demonstrated, in a retrospective cohort study of 6668 patients with LBP receiving PT in a community primary care setting, an association between early PT initiation (< 3 days) and downstream healthcare utilization costs. [30]

Of note, in our subgroup analysis of patients presenting with low back pain, the differences in opioid administration rates and ED length of stay were not present. The difference in imaging rates was more pronounced in this subgroup of patients than in the overall population, which may speak to the potential impact of PT consultation on imaging rates among patients with low back pain. It is worth noting that PT patients did have shorter ED length of stay and lower opioid administration rates during the ED stay, but these differences were not statistically significant. This may be due to the smaller size of the groups, and the analysis may have been underpowered to detect these differences in this subset of patients.

Initiating PT in the ED provides patients with meaningful intervention at the earliest possible stage of their presentation. [16,31] The ED often represents the first point of contact with the healthcare system for a given patient. Seeing a physical therapist in the ED offers a unique opportunity to provide guideline-adherent care for musculoskeletal pain from the outset, with the potential to realize the benefits of education, early mobility, and reduced exposure to unnecessary imaging. Some patients who may benefit from seeing a physical therapist may not otherwise have access to PT, potentially due to regulatory and health insurance restrictions or simply due to patient awareness of this resource. [18] Having PT available in ED may mitigate some of these barriers and increase the likelihood that patients have the opportunity to receive this care.

ED-PT as part of a primary contact provider model has not been shown to increase ED length of stay. [12,32] However, there remains concern that secondary contact provider models, such as that in our institution, may prolong lengths of stay. [13,33] In our study, ED-PT was associated with shorter ED length of stay. The reason for this finding is likely multifactorial; fewer imaging studies in the PT group and our institutional emphasis on early PT consultation were likely the greatest contributors. A significant contributor to length of stay in some ED programs is likely delayed referral to PT. Providers in our institution are encouraged to refer to PT early and avoid using ED-PT as a ‘last resort’ after exhausting alternative measures (opioid analgesia, diagnostic imaging). As EDs see increasing volumes of patients, alternative and creative ways to reduce wait times and improve patient flow become increasingly important. [32] Our study demonstrates that early referral and PT involvement may provide a strategy to achieve this goal in specific patient populations.

Finally, identifying treatments and therapies that reduce the need for opioid prescribing is invaluable in the midst of the current opioid epidemic. [26,27,34] Back pain is the most frequent diagnosis for which opioids and benzodiazepines are prescribed in the ED. [34] PT has been suggested as an alternative management strategy for managing both acute and chronic pain. [17] A recent study by Kim et. al. was the first to focus on the effect of ED-PT on opioid prescribing in back and neck pain in the ED setting. [28] They found that individuals presenting to the ED with back and neck pain who were seen by a physical therapist during the visit were just as likely to receive an opioid prescription. Similarly, we found no significant difference in outpatient opioid prescribing but did note a substantial difference in opioid administration while patients were in the ED. In our ED PT program, we have encouraged physicians to consult PT prior to initiating pharmaceutical therapy, which may be a factor in the differing opioid administration levels within the ED itself. However, the present study and that of Kim et al demonstrate no association between ED-PT and reduced opioid prescribing at the point of discharge. This contrasts with a number of claims-based studies indicating lower opioid prescribing in those patients engaging in PT. [35,36] However, these studies focus on outpatients, in whom our population of patients in the ED may significantly differ, both in terms of urgency of encounter and acute pain score. It may also be the case that PT has a more protective effect against opioid prescribing long-term, and is less effective as an opioid substitution therapy acutely at the point of ED discharge Further research may clarify the comparable rates of outpatient opioid prescribing between those who receive ED-PT and those who do not.

Limitations

Our study is limited to a single center, and all patients were seen by a single physical therapist. As such, our findings may be unique to this center, to the musculoskeletal diagnostic evaluations, and to the management strategies at this site.

Our methodology to identify study patients meant that the scope of the PT patients included in the study was comprehensive and represented all patients for whom ED-PT was provided during the 18-month study period. However, our control group represented a convenience sample of patients presenting with a musculoskeletal complaint. Certain limitations may be inherent to the methodology employed to identify appropriate patients for comparison. Given that our physical therapist did not assess the non-PT group, it is possible that our non-PT group is fundamentally different from the group that received PT, and that the non-PT group contains a sub-set of patients who would not have been amenable to ED initiated PT and for whom the physical therapist would not have been consulted if available. Given the limited presence of the physical therapist in the ED, we felt that this offered an appropriate arrangement to recruit a group of similar patients who presented to the ED on days in which the physical therapist was not present and, as such, would form an appropriate control group. Still, this methodology does not utilize a robust randomization process in assigning patient to PT and may have selected for a more complex or unique group of patients among those who did not have PT in the ED.

Additional limitations relate to the reliance upon the electronic medical record for patient demographic, testing, and medication information. While we identified patients prospectively either through the research associates or through the ED physical therapist, we retrieved relevant outcomes through review of the medical record. As such, this methodology is susceptible to bias through mistakes within the medical record or error in the review and transcription process.

This study focuses on overall rates of imaging and does not assess the appropriateness or inappropriateness of the imaging studies performed. We know that many patients with musculoskeletal complaints receive unnecessary imaging investigations, however only overall rates of imaging are compared in this study, and therefore no conclusions can be made about whether PT reduces unnecessary imaging.

In focusing solely on outcomes at the time of the ED encounter, our study is further limited by the lack of ongoing, longitudinal patient follow-up after the ED visit. It is unclear whether the benefits of PT seen in the ED translate into ongoing benefits to the patient or the community, such as reduced use of healthcare resources and the ongoing reduced use of opioids. Such longitudinal endpoints may form the basis for future inquiry.

Conclusion

Our study found that patients who had ED-PT had a shorter length of stay, less overall imaging, and reduced rates of ED opioid administration. However, we found no significant difference in outpatient opioid prescriptions. This study is the first ED-based study to demonstrate an impact of ED-PT on opioid administration and may provide future direction for research assessing ED opioid reduction strategies. Similarly, the impact on overall healthcare utilization through imaging and length of stay suggests added benefits of ED-PT beyond the therapeutic relationship established and care provided within the ED setting.