The Effect of Physiatry Involvement for Patients With Acute Traumatic Spinal Cord Injury at a Level 1 Trauma Center

Abstract

Objective

To explore the impact of physiatry on acute traumatic spinal cord injury (ATSCI) outcomes using a propensity score matching (PSM) analysis.

Methods

We retrospectively reviewed all patients with ATSCI at a level 1 trauma center from 2018 to 2019. In a 1:1 fashion, we matched patients who were evaluated by physiatry to those who were not. Our PSM analysis controlled for patient demographics, Glasgow Coma Scale (GCS) score, Injury Severity Score (ISS), comorbidities, mechanism, and presence of a traumatic brain injury (TBI). Outcomes included complications, discharge disposition, and 30-day mortality. Survival analysis was performed using Kaplan-Meier plots.

Results

A total of 102 patients (physiatry 51; no physiatry 51) were matched. Median age was 38 (28–55) years, and median ISS was 25.5 (17–35); 82.4% (n = 84) were male, and 77.5% (n = 79) were bluntly injured. Rates of in-hospital complications were similar between groups. Physiatry involvement was associated with increased odds of discharge to inpatient rehabilitation (odds ratio, 4.6; 95% CI, 2–11.6; p < .001). There was a significant survival benefit seen with physiatry involvement at 30 days (92.6% vs. 78.6%, p = .004) that correlated with a decreased risk of mortality (hazard ratio, 0.2; 95% CI, 0.03–0.7; p = .01).

Conclusion

Incorporating physiatry into the management of patients with ATSCI is associated with improved survival and greater odds of discharge to rehabilitation. In this population, physiatry should be incorporated into the trauma care team to optimize patient outcomes.

Introduction

Approximately 12,000 individuals sustain an acute traumatic spinal cord injury (ATSCI) each year in the United States.¹ These patients are ideally triaged to a designated trauma center where there are resources and experienced clinicians to deliver specialized care.^2–4 Patients with neurological deficits are monitored in an intensive care unit (ICU) and usually admitted to a trauma surgery service because of the presence of other traumatic injuries.^5–7 However, an array of specialists including physiotherapists, occupational therapists, social workers, and physiatrists should comprise a multidisciplinary care team for these patients.^4,8–10

Partnerships between a trauma surgery team and other specialties such as physiatry are linked to better short- and long-term outcomes in this population.^4,9,11 The American College of Surgeons Committee on Trauma recognizes these benefits and stipulates that any verified level 1 trauma center (L1TC) must have in-house rehabilitation consultation services.¹² Although the Committee on Trauma does not specifically mandate a physiatrist, rehabilitation services are invariably coordinated by the physiatry team. Physiatrists have extensive training in diagnosing and treating secondary complications due to ATSCI, facilitating discharge planning, and determining whether a patient is suitable for an inpatient rehabilitation facility (IRF).¹³

According to a national survey of 105 L1TCs, 85% reported having in-house physiatrists available to evaluate trauma patients.¹¹ Unfortunately, due to a poor understanding of the skills provided by a physiatrist, many trauma surgery teams do not consult these physicians for assistance and/or recommendations.^11,14 We sought to determine whether collaboration between a trauma service and physiatry team would be associated with any differences in complications, discharge dispositions, and mortality among patients with ATSCI.

Methods

Population and data points

We performed a retrospective review of patients aged 16 years and older with an ATSCI who presented to Grady Memorial Hospital (L1TC) between 2018 and 2019. We verified whether a consultation for physiatry was placed in the electronic medical record and successfully completed. The decision to place a consultation was at the discretion of the trauma surgery attending physician. Patients with missing data were excluded. This study was approved by the institutional review board.

We collected data on patient demographics (age, sex, race), admission Glasgow Coma Scale (GCS) score, mechanism of injury (blunt or penetrating), Injury Severity Score (ISS), and preinjury medical conditions (alcohol use disorder, chronic obstructive pulmonary disease, current smoker, diabetes, functionally dependent health status, hypertension, and steroid use). We abstracted level of ATSCI (cervical, lumbar, thoracic) and the presence of a concomitant traumatic brain injury (TBI). Time from patient admission to consultation completion was also collected.

Outcomes

Our primary outcome was 30-day mortality. Secondary outcomes included in-hospital complications, hospital length of stay (LOS), mechanical ventilation duration, and discharge disposition. Complications were classified as physiatry relevant (acute kidney injury [AKI], deep venous thrombosis [DVT], pneumonia [PNA], pulmonary embolism [PE], unplanned intubation, and urinary tract infection [UTI])^15,16 and non-physiatry relevant (cardiopulmonary arrest, cerebrovascular accident [CVA], myocardial infarction [MI], surgical site infection [SSI], unplanned intensive care unit [ICU] admission, and unplanned return to the operating room). Discharge dispositions included IRF, home, or subacute location (skilled nursing facility, long-term acute care hospital).

Statistical analyses

We performed propensity score matching (PSM) to create two balanced patient groups for comparison. A logistic regression model was used to generate a propensity score for each patient based upon demographics, GCS score, ISS, mechanism, preinjury medical conditions, and presence of a TBI. We did not match for the Abbreviated Injury Score (AIS)-spine because AIS is used to calculate ISS. Patients treated by physiatry were matched with patients not treated by physiatry using a 1-to-1 nearest neighbor technique with calipers of 0.2.^17,18 After PSM, continuous nonparametric data were compared using Wilcoxon tests, and continuous parametric data were compared using t tests. Categorical data were compared using chi-square tests or Fisher’s exact tests when appropriate. All p values were from two-sided tests, and a p < .05 was considered statistically significant.

Survival analysis was performed using a nonparametric Kaplan-Meier estimator. The corresponding percentage of patients surviving at 30 days in each group was compared using a log-rank test. Cox proportional hazards regression was then used to determine the risk of 30-day mortality. Logistic regression was used to identify associations between physiatry involvement and discharge disposition. Data analyses were performed with JMP Pro software (Version 15.1 of the SAS System for Windows; SAS Institute Inc., Cary, NC).

Results

A total of 102 patients were matched: 51 physiatry and 51 no physiatry (Figure 1). The two groups were well balanced after PSM (Table 1). For the overall cohort, the median age was 38 (28–55) years, and median ISS was 25.5 (17–35). Most patients were male (82.4%, n = 84) and 77.5% (n = 79) were injured via a blunt mechanism. The median time to physiatry evaluation was 1 (0–1) day and 82.4% (n = 42/51) of patients were evaluated within ≤1 day following admission.

Figure 1.

Study population flow diagram. ATSCI = acute traumatic spinal cord injury; PMR = physiatry.

Table 1.

Cohort details stratified by PMR involvement after matching (N = 102)

Variable	No PMR (n = 51)	Yes PMR (n = 51)	p
Age, years	39 (28–57)	38 (26–52)	.5
Gender			.6
Male	43 (42.2)	41 (40.2)
Female	8 (7.8)	10 (9.8)
Race			.9
Black	29 (28.4)	31 (30.4)
White	20 (19.6)	18 (17.7)
Other	2 (2.0)	2 (2.0)
Comorbidities
Current smoker	15 (14.7)	18 (17.7)	.5
Hypertension	10 (9.8)	10 (9.8)	1.0
Alcohol abuse	5 (4.9)	5 (4.9)	1.0
COPD	2 (2.0)	1 (1.0)	1.0
FDHS	1 (1.0)	1 (1.0)	1.0
Steroid use	1 (1.0)	1 (1.0)	1.0
Diabetes	0 (0.0)	1 (1.0)	1.0
Mechanism			.8
Blunt	40 (39.2)	39 (38.2)
Penetrating	11 (10.8)	12 (11.8)
Glasgow Coma Scale score	15 (3–15)	15 (13–15)	.5
Injury Severity Score	21 (14–35)	26 (21–35)	.1
Spinal cord injury level			.2
Cervical	30 (29.4)	22 (21.6)
Thoracic	12 (11.8)	21 (20.6)
Lumbar	9 (8.8)	8 (7.8)
Traumatic brain injury	12 (11.8)	13 (12.8)	.8

Note: Categorical data are presented as n (%) and continuous data are presented as median (interquartile range). COPD = chronic obstructive pulmonary disease; FDHS = functionally dependent health status; PMR = physiatry.

Outcomes

For the entire cohort (N = 102), 18.6% (n = 19) experienced a non-physiatry-relevant complication and 22.5% (n = 23) experienced a physiatry-relevant complication (Table 2). The physiatry group had a significantly higher rate of unplanned ICU admissions (8.8% vs. 0%, p < .01). There was no difference in rates of any other complication between the two groups (Table 2). Compared to patients without complications, the hospital LOS was significantly longer for patients with non-physiatry-relevant (30 vs. 10 days, p < .0001) and physiatry-relevant complications (61 vs. 11 days, p < .0001).

Table 2.

In-hospital complications (N = 102)

Complication	No PMR (n = 51)	Yes PMR (n = 51)	p
Non-PMR relevant	6 (5.9)	13 (12.8)	.1
Cardiopulmonary arrest	4 (3.9)	3 (2.9)	1.0
Stroke	1 (1.0)	1 (1.0)	1.0
Surgical site infection	1 (1.0)	0 (0.0)	1.0
Unplanned return to OR	1 (1.0)	2 (2.0)	1.0
Unplanned admission to ICU	0 (0.0)	9 (8.8)	<.01
PMR-relevant	11 (10.8)	12 (11.8)	1.0
Acute kidney injury	5 (4.9)	3 (2.9)	.7
Pneumonia	5 (4.9)	6 (5.9)	1.0
Deep venous thrombosis	1 (1.0)	1 (1.0)	1.0
Pulmonary embolism	3 (2.9)	4 (3.9)	1.0
Unplanned intubation	1 (1.0)	3 (2.9)	.6
Urinary tract infection	1 (1.0)	0 (0.0)	1.0

Note: Categorical data are presented as n (%). ICU = intensive care unit; OR = operating room; PMR = physiatry.

The hospital LOS was significantly shorter for patients in the non-physiatry group (7 vs. 20 days, p < .0001) (Table 3). Amongst all patients discharged home (n = 34), hospital LOS for the non-physiatry group was shorter compared to the physiatry group (5 vs. 10.5 days, p = 0.003). The duration of mechanical ventilation was also not different between the two groups (14.5 vs. 13 days, p = .5). More patients in the physiatry group were discharged to an IRF (52.9% vs. 19.6%, p < .01) (Table 3). After excluding patients who expired, were discharged to prison, or left against medical advice (n = 21), the PMR group had a significantly greater odds of discharge to an IRF (odds ratio [OR], 4.6; 95% CI, 2–11.6; p = .0004).

Table 3.

Hospital length of stay and discharge disposition ^a

Disposition	No PMR (n = 51)	Yes PMR (n = 51)	p
Hospital length of stay, days	7 (3–18)	20 (11–33)	<.0001
Mechanical ventilation duration, days	14.5 (3–46.3)	13 (2–23.5)	.5
Discharge location			<.01
Home	22 (21.6)	12 (11.8)
IRF	10 (19.6)	27 (52.9)
Subacute	5 (9.8)	5 (9.8)

Note: Categorical data are presented as n (%) and continuous data are presented as median (interquartile range). IRF = inpatient rehabilitation facility; PMR = physiatry.

^aAnalysis excludes 21 patients who died, were discharged to prison, or left against medical advice.

At 30 days following admission, more patients in the physiatry group survived (92.7% vs. 78.6%, p = .004). There was also a significantly lower risk of mortality associated with the physiatry group (hazard ratio [HR], 0.2; 95% CI, 0.03–0.7; p = .01). For the 14 patients who died, the median time until death was 3 (0–17) days, and it was significantly shorter in the non-physiatry group (2 vs. 26 days, p = .03). After excluding the three patients from the physiatry group who died, there was no significant difference between time until death and time until physiatry consultation completion (2 vs. 1 days, p = .3).

Discussion

The clinical impact of physiatry and trauma services providing care to patients with an ATSCI in the acute setting is currently unknown.^4,10,19 Greiss et al.²⁰ showed that an acute physiatry consultation was associated with better postrehabilitation functional outcomes, but the study only included patients with a TBI. Other studies have demonstrated an improvement in short-term outcomes when patients with an ATSCI are treated in a specialized SCI unit, but they do not examine how physiatry contributed to these improvements.⁴

We discovered that there was a significant survival benefit for patients in the physiatry group. Furthermore, there was no difference in time until death for patients in the non-physiatry group compared to time until completion of a physiatry consultation. These findings may suggest that mortality in the non-physiatry group was not related to patients succumbing to the severity of their ATSCI before a physiatry consultation could be performed. Instead, we believe the survival advantage could be attributed to a physiatrist’s swift recognition of impending decompensation as they have specialized training in identifying and triaging complications specific to ATSCI.^13,15 Moreover, in an attempt to prevent further deterioration with mortality (i.e., failure to rescue), the physiatry team may have recommended escalation of care, which is why only patients in the physiatry group had unplanned ICU admissions.

Helling et al.²¹ examined clinical deterioration episodes in a heterogenous group of surgical patients who required an unplanned admission to an ICU. Similar to research by Smith et al.,²² the researchers noted that error in recognition by the surgery team was the primary cause of preventable and potentially preventable failure events. Fry et al.²³ evaluated mortality trends among more than 700,000 Medicare beneficiaries from 3404 US hospitals. They showed that hospitals with the greatest mortality reduction over 10 years had a significant decrease in failure to rescue events, while complication rates were unchanged. The authors recommend that physicians improve the management of complications as a better method of lowering mortality rates,²³ which is what we believe the physiatry service may have done.

It should be noted, however, that patients in the physiatry group had a significantly longer hospital LOS. This finding could imply that survivors in the non-physiatry group had a milder injury burden. Indeed, more patients in the non-physiatry group were discharged home, and these patients had a significantly shorter hospital LOS compared to patients in the physiatry group. The longer hospital LOS could also be due to more unplanned ICU admissions in the physiatry group. For instance, Rubano et al.²⁴ found that among 5411 acutely injured trauma patients, 3.9% (n = 212) required an unplanned ICU admission. These 212 patients had a longer hospital LOS and significantly more complications, including thromboembolic events and pneumonia. The authors cite a failure of initial admission triage as well as more in-hospital complications as reasons for longer hospitalizations.²⁴ In addition, Wagner et al.²⁵ found that a physiatry consultation was associated with a longer hospital LOS in a cohort of patients with a TBI. They similarly ascribed the longer hospitalization to these patients experiencing more in-hospital complications and having more severe initial injuries.

Wagner et al.²⁵ found that a physiatry evaluation was coupled with a significant increase in discharge to an IRF, which supports our finding that patients in the physiatry group had 4.6-times greater odds of discharge to an IRF. It is well-known that a physiatry team understands hospital and community resources for injured patients and is therefore poised to coordinate a transfer to an IRF.^10,13 Musick et al.¹⁹ demonstrated that among a heterogeneous group of 22,378 hospitalized patients, 8% (n = 1874) were evaluated by a physiatrist and these specific patients had a greater rate of non-home discharge. Furthermore, 99% of the physiatry consultations in this study were completed within a median of 1 day of request.¹⁹ Not only did we observe higher rates of IRF discharge for patients in the physiatry group but we also saw that more than 80% of patients were evaluated within 1 day of request. This expeditious response time has been associated with a timelier determination of a patient’s potential for functional recovery and less delay to rehabilitation.²⁶

Limitations

This study has several limitations. Our research was retrospective and not randomized; however, we attempted to overcome these problems by using PSM. This methodology allowed us to successfully balance two patient groups with respect to several measured confounders. However, we recognize that there could be other variables we did not include in our PSM, which could have possibly influenced our analysis.¹⁸ Also, our L1TC is in a large metropolitan area, so we caution the extrapolation of our findings to non-urban centers. However, the majority of patients in our study were middle-aged males following a blunt trauma, which mirrors the current national demographic of individuals experiencing an ATSCI.²⁷

Other limitations stem from using a trauma registry as our data source. Not only is a registry inherently at risk for error during data collection and entry,²⁸ but it also does not track all variables specific to an ATSCI. The registry did not track all physiatry-relevant complications, such as heterotopic ossification or autonomic dysreflexia, and we only captured the level of cord injury and ISS, which do not unequivocally describe the severity of ATSCI. The reason why a physiatry consultation was or was not placed was not discernible from the registry. We could not obtain patient American Spinal Injury Association Impairment Scale scores, which are known determinants of functional status and can affect a patient’s discharge disposition.²⁹ Finally, a trauma patient’s discharge location is often multifactorial and can be influenced by social issues, insurance status, bed availability, and/or family or individual preferences.³⁰

Conclusion

Our study emphasizes the positive effect of a physiatry team in assisting with the management of patients with ATSCI. We found that physiatry’s collaboration with the trauma surgery service correlated with a lower mortality risk and a greater chance of discharge to an IRF. Although future studies are warranted to examine the external validity of our findings, we encourage the early integration of physiatry into the clinical care team that treats patients with an ATSCI.