https://orcid.org/0000-0002-1469-3608
ABSTRACT
Objective
To assess the accuracy of the AM-PAC “6-Clicks” in predicting discharge dispositions among severely injured patients with an acute traumatic brain injury (TBI).
Methods
We performed a retrospective review of patients with a TBI who presented to our trauma center from 2016 through 2018 and received a “6-Clicks” assessment. Outcomes were hospital length of stay (LOS) and discharge disposition: home, inpatient rehabilitation facility (IRF), subacute location (SL), or death/hospice. Subgroup analyses evaluated patients with concomitant mobility-limiting injuries (CM-LI).
Results
There were 432 patients with a TBI; 42.6% (n = 184) had CM-LI. CM-LI patients had lower “6-Clicks” scores compared to patients with an isolated TBI (9 vs 14, p < .0001) and a longer hospital LOS (16.5 d vs 9 d, p < .0001). Increasing “6-Clicks” scores were associated with a home discharge (OR 1.21, 95% CI 1.15–1.28, p < .0001) while decreasing scores were predictive of an IRF or SL discharge or death/hospice. Increasing scores correlated with decreasing hospital LOS for the cohort (β − 8.93, 95% CI −10.24 – −7.62, p < .0001).
Conclusion
Among patients with an acute TBI, increasing “6 Clicks” scores were associated with a shorter hospital LOS and greater likelihood of home discharge. Decreasing mobility scores correlated with discharge to an IRF, SL, and death/hospice.
Introduction
Traumatic brain injury (TBI) is a leading cause of morbidity and mortality in the U.S. with approximately 2.8 million related emergency department (ED) visits, hospitalizations, and deaths annually (Citation1,Citation2). TBI-associated care is estimated to cost $50 to $60 billion dollars each year in healthcare expenditures and as much as $2 million per injured patient over a lifetime (Citation3–5). Besides these financial encumbrances, more than 5 million Americans suffer from long-term psychosocial disturbances, cognitive impairment, and mobility issues due to a TBI (Citation4). Therefore, early recognition of disability during a patient’s initial hospitalization is essential for identifying a discharge location that best addresses a patient’s functional needs. Unfortunately, Sorensen et al. recently found that 14% of patients with a TBI have significant delays in discharge due to the actual discharge destination (Citation6).
Several tools exist to evaluate the functional status of injured patients, such as the Activity Measure for Post-Acute Care (AM-PAC). The AM-PAC assesses a patient’s basic mobility, daily activity, and applied cognition (Citation7,Citation8). The AM-PAC “6-Clicks” is an abbreviated version of the AM-PAC and specifically evaluates basic mobility through six skills. This tool can be quickly performed and has consistently been shown to assist with post-acute care planning for medical patients (Citation8,Citation9). Another benefit of the “6-Clicks” is that it can be administered by a physical therapist (PT), nurse, and/or clinician while maintaining excellent interrater reliability (Citation8,Citation10,Citation11). Accordingly, scores serve as a standardized lexicon for the care team to effectively communicate discharge options (Citation10).
While the “6-Clicks” has been shown to successfully prognosticate discharge locations in the acute hospital setting, it is unknown if the tool can be used for severely injured trauma patients with an acute TBI. Therefore, we sought to investigate this question and hypothesized that the initial “6-Clicks” scores could be used to predict discharge dispositions among hospitalized patients with a TBI.
Methods
Participants
We performed a retrospective review of patients with a TBI who were presented to our level 1 trauma center (L1TC) from 2016 through 2018. Patients who were dead on arrival to the emergency department (ED) were discharged to legal custody or left against medical advice, or were intubated, comatose, or in a vegetative state were excluded. We further excluded individuals without an AM-PAC “6-Clicks” assessment and patients with an Injury Severity Score (ISS) < 16. Of note, we excluded patients with an ISS < 16 because at our L1TC, all severely injured (ISS 16 to 24) (Citation12) and very severely injured (ISS > 24) patients automatically receive physical therapy (PT) consultations. However, in accord with the Choosing Wisely initiative (Citation13), trauma patients at our L1TC who are mildly (ISS 1 to 8) or moderately (ISS 9 to 15) injured will only receive a PT consultation if specifically requested. Our institutional review board reviewed this study and provided a waiver of consent.
Data points
A TBI was determined based upon a patient’s computed tomography (CT) head obtained as part of his/her initial trauma workup. All imaging was read and reviewed by an attending radiologist and classified by ICD-10 codes: cerebral edema (S06.1), diffuse (S06.2), focal (S06.3), epidural hematoma (S06.4), subdural hematoma (S06.5), and subarachnoid hematoma (S06.6). In addition, we collected data on patient demographics, trauma mechanism (blunt or penetrating), index Glasgow Coma Scale (GCS) score, ISS, and frailty status. We used the 11-item modified frailty index (mFI-11), which has previously been applied to the TBI population (Citation14). Furthermore, we used ICD-10 codes to track whether a patient had a concomitant mobility-limiting injury including a spinal cord injury (S14, S24, S34), lumbar or pelvic fracture, dislocation, or sprain (S32, S33), and an upper or lower extremity fracture, dislocation, or sprain (S42, S43, S52, S53, S82, S83, S92, S93).
The AM-PAC “6-clicks” basic mobility score
At our institution, PTs evaluate trauma patients each day and determine if a patient is clinically appropriate for assessment. The instrument consists of six exercises: (1) turning over in bed, (2) sitting down on and standing up from a chair, (3) moving from supine to sitting, (4) transferring from a bed to a chair, (5) ambulating in a hospital room, and (6) climbing 3 to 5 steps with a railing (Citation10,Citation15). Each task is rated on an ordinal scale from 1 (unable to complete) to 4 (independently completes) and summated to yield a “6-Clicks” score. We recorded each patient’s scores, time from admission to “6-Clicks” assessment (t1), and time to discharge from the assessment (t2).
Outcomes
Our primary outcome was discharge disposition, which we classified as home, inpatient rehabilitation facility (IRF), subacute location (SL), or death/hospice. Subacute locations included intermediate care facilities, long- or short-term care hospitals, and/or skilled nursing facilities. Death/hospice indicated that a patient was admitted to the hospital but subsequently died as an inpatient or that the patient was transitioned to hospice. Secondary outcomes included hospital length of stay (LOS) and in-hospital complications. Complications included unplanned events (unplanned intensive care unit admission, unplanned intubation, unplanned return to the operating room) as well as the following types: cardiovascular (cardiac arrest with cardiopulmonary resuscitation, cerebrovascular accident, myocardial infarction), infectious (central line associated bloodstream infection, sepsis, surgical site infection, urinary tract infection, ventilator associated pneumonia), pulmonary (acute respiratory distress syndrome), renal (acute kidney injury), and thromboembolic (deep venous thrombosis, pulmonary embolus).
Statistical analysis
Continuous variables are presented as medians (interquartile range) and compared using Wilcoxon rank sums tests or the Dunn method for joint ranking. Categorical data are presented as, n (percentage) and compared with Pearson chi-squared tests or Fisher’s exact tests when appropriate. A p-value < .05 was considered statistically significant. Outcomes were first calculated for the entire cohort, then a subgroup analysis was performed to compare patients with a CM-LI and an isolated TBI.
Then, to efficiently determine if a relationship existed between any hospital variables that were collected and discharge disposition, we used a response screening platform. This platform rapidly generates a table of p-values for each pairwise relationship. A variable with a p ≥ .05 and ≤ 0.1 was classified as a potential confounder and a variable with a p < .05 was considered a covariate. Using multivariable logistic regression, we adjusted for these covariates and confounders while evaluating the effect of “6-Clicks” scores on discharge dispositions.
The same variables used to predict discharge disposition were also used to model hospital LOS. Because LOS is a continuous response, we plotted a histogram for the numeric values of LOS and superimposed a Beta, Cauchy, Exponential, Gamma, Lognormal, Normal, and Weibull distribution. We compared the corresponding Akaike information criteria (AIC) for these distributions and the one with the smallest AIC was selected (Citation16). Next, a goodness of fit test was used to confirm if the selected distribution was also statistically appropriate (i.e., p < .05). If so, it was used as the distribution for our model’s predicted responses.
All regression models were applied to the entire patient cohort and then to patients with an isolated TBI. The accuracy of these models was determined using a generalized-R2. The generalized-R2 is conceptually identical to R2 in that it reflects the degree of variation in a response variable that can be attributed to the model rather than to random error (Citation17). Values closer to 0 suggest that there are other variables unaccounted for in the model while values closer to 1 imply that the chosen variables appropriately explain the variation in response (Citation18). However, because our model’s input variables were heteroscedastic by nature, we used the generalized-R2 (Citation17). Data analyses were performed with JMP® Pro software, Version 15.1 of the SAS® System for Windows®. Copyright © 2019 SAS Institute Inc., SAS Campus Drive, Cary, North Carolina, 27513, USA.
Results
The cohort was comprised of 432 severely injured patients with an acute TBI (). The median age was 52 (31–69) years, median ISS 24 (17–29), and median GCS score 14 (7–15). The most common TBI types were a subdural hemorrhage (56.7%, n = 245) and subarachnoid hemorrhage (52.3%, n = 226). Of the cohort, 42.6% (n = 184) patients had CM-LI ().
Table 1. Cohort (N = 432) characteristics.
Figure 1. Flow diagram for patients included in study.AMA against medical advice; ISS injury severity score.
The “6-Clicks” scores were significantly higher for patients discharged home compared to the remaining discharge locations (p < .0001) (). The hospital LOS for patients discharged home was also shorter compared to patients discharged to an IRF and SL (p < .0001) as well as death/hospice (p < .05). The median time until “6-Clicks” assessment for the cohort was 4 (3–8) days with a median score of 12 (7–18) (). The median time until discharge from assessment was 7 (2–16) days and median hospital LOS was 12 (6–25) days. Compared to patients with an isolated TBI, individuals with a CM-LI had significantly lower “6-Clicks” scores (9 vs 14, p < .0001) and a longer hospital LOS (16.5 d vs 9 d, p < .0001). Most patients in the cohort were discharged home (53.7%, n = 232) and there was no difference in rates of complications or discharge locations between patients with a CM-LI or isolated TBI ().
Table 2. Comparison of “6-Clicks” scores and LOS amongst discharge dispositions (N = 432).
Table 3. Study outcomes for the cohort and subgroups.
The response screening platform revealed a significant relationship between discharge disposition and age, complication(s), ISS, GCS score, mFI-11, t1, and t2. To prevent multicollinearity, we evaluated the correlation between the “6-Clicks” score and t1 (Spearman’s ρ − 0.45) as well as the “6-Clicks “score and t2 (Spearman’s ρ − 0.68). Because of the stronger correlation between t2 and mobility scores, we selected t1 to be used in the regression analyses.
An increasing “6-Clicks” scores correlated with a greater odds of discharge home for the entire cohort (OR 1.21, 95% CI 1.15–1.28, p < .0001) and for patients with an isolated TBI (OR 1.18, 95% CI 1.11–1.26, p < .0001) (). On the contrary, an increasing “6-Clicks” score was associated with a decreased odds of discharge to an IRF for both the overall cohort (OR 0.86, 95% CI 0.81–0.91, p < .0001) and patients with an isolated TBI (OR 0.89, 95% CI 0,83–0.95, p < .0001). Amongst the overall cohort, the ISS was significantly associated with a home discharge (OR 0.96, 95% CI 0.93–0.99, p = .01) and IRF discharge (OR 1.03, 95% CI 1.01–1.06, p = .01), yet lost significance when evaluating only patients with a TBI ().
Table 4. Multivariable logistic regression predicting discharge disposition.
Increasing age was associated with an increased odds of discharge to a SL and death/hospice for the overall cohort and for patients with an isolated TBI. The “6-Clicks” score was associated with a decreased odds of an SL discharge for the entire cohort (OR 0.9, 95% CI 0.83–0.96, p < .01), but lost significance when evaluating patients with an isolated TBI (). While the “6-Clicks” did not correlate with death/hospice for the overall cohort, it was predictive of a decreased chance of death/hospice in patients with an isolated TBI (OR 0.77, 95% CI 0.62–0.94, p = .01). A complication increased the odds of death/hospice for the entire cohort (OR 3.57, 95% CI 1.11–11.5, p = .03), yet lost significance in patients with an isolated TBI (). The generalized-R2 for the multivariable logistic regression model predicting discharge disposition for the entire cohort and for patients with an isolated TBI were 0.245 and 0.24, respectively.
Hospital length of stay
The median hospital LOS for the cohort was 12 (6–25) days and LOS data were best fit using a lognormal distribution (Anderson-Darling test, p < .01). On multivariable linear regression, increasing AM-PAC “6-Clicks” scores significantly correlated with a decreasing LOS (β − 8.93, 95% CI −10.24 – −7.62, p < .0001) for the entire cohort and for patients with isolated TBIs (−7.3, 95% CI −8.71 – −5.9, p < .0001) (). The generalized-R2 for the models were 0.65 and 0.64, respectively.
Table 5. Multivariable linear regression predicting hospital length of stay.
Discussion
Discharge planning for patients with a TBI is a complex process, which requires clear communication between the entire care team to identify an appropriate discharge location for each patient (Citation19). Tools that measure functional status can help with decision making, but these instruments are not always intuitive (Citation8,Citation10,Citation11). Fortunately, the AM-PAC “6-Clicks” has been lauded for its simplicity, reliability, and understandability(Citation10).
We observed that increasing “6-clicks” mobility scores predicted a shorter hospital LOS for the cohort and for patients with only a TBI. Furthermore, patients discharged home had higher mobility scores and a shorter hospital LOS, while patients discharged to an SL or IRF had lower scores and a longer hospital LOS. Menendez et al. similarly demonstrated that patients with lower “6-Clicks” scores had significantly longer hospitalizations following total joint arthroplasty (Citation20). In addition, Sorensen et al. demonstrated that patients with a TBI who were discharged to non-home locations had at least a 10-times increased odds of discharge delay and a corresponding longer hospital LOS (Citation6). They also noted that among patients with discharge delays, 15% experienced a complication (Citation6). Likewise, we observed that patients experiencing any in-hospital complication had a significantly longer hospital LOS and a decreased odds of home discharge.
In addition to its ability to predict hospital LOS, we demonstrated that mobility scores can predict discharge disposition as well. The median “6-Clicks” score for the cohort was 17, which was significantly higher than scores for the remaining three dispositions. Similarly, in a cohort of 673 patients with cardiovascular illness, Fernandez et al determined that a “6-Clicks” score of 18.5 was the most sensitive and specific value for predicting a home discharge (Citation21). Furthermore, increasing “6-Clicks” scores were predictive of an increased likelihood of home discharge for the cohort and patients with an isolated TBI. Interestingly, the ISS was not associated with home discharge in patients with an isolated TBI but was for entire cohort. This finding suggests that discharge planning should not only use “6-Clicks” scores for polytrauma patients with a TBI, but also account for injury burden.
Similarly, the ISS was not associated with predicting IRF discharge for patients with an isolated TBI, but a worsening injury burden was significantly predictive of an increased odds of IRF discharge for the entire cohort. Foreman et al. suggest that extracranial injuries, which are measured by the ISS, significantly correlate with functional outcomes in patients with a TBI (Citation22). It seems appropriate that more physically injured patients would be discharged to an IRF because of its emphasis on strength training, exercise, and other mobility-related activities (Citation23). In addition, lower mobility scores were associated with a increased likelihood of an IRF discharge for both patient groups. Menendez et al. examined 744 patients following total joint arthroplasty and also noted that lower AM-PAC “6-Clicks” Mobility scores were associated with nonroutine discharge (Citation20,Citation24). Moreover, Pfoh et al. found in their retrospective analysis of 17,022 medical inpatients found that patients who scored ≤ 12 had a > 80% chance of non-home discharge.
Of note, we did not identify a difference in “6-Clicks” scores between patients discharged to an IRF versus an SL. This finding is important to consider because TBI rehabilitation focuses on physical, occupational, and speech therapy to optimize or regain function (Citation25). Conversely, SLs provide medical care, typically for chronic medical conditions, and do not necessarily promote functional independence (Citation23,Citation25). Thus, we do not advise choosing between these two discharge locations using only “6-Clicks” scores. Nevertheless, we did identify other predictors that could be used in conjunction with the “6-Clicks” to help differentiate between the two dispositions. For example, older age was associated with an increased odds of SL discharge for the cohort and isolated TBI patients, whereas age was not predictive of an IRF discharge for either group. Geriatric patients have reduced brain plasticity, less cognitive reserve, and lower life satisfaction after TBI compared to their younger counterparts (Citation23,Citation26). These factors, in addition to a decreased resilience to injury, may explain why age was associated with SL discharge and not an IRF discharge.
Increasing age was also predictive of death/hospice for both patient groups, yet unlike SL discharges, “6-Clicks” scores for the entire cohort were not predictive of death. On the contrary, a decreasing “6-Clicks” score was predictive of death/hospice in patients with an isolated TBI. This distinction could indicate that the instrument is only helpful in predicting death/hospice in patients with an isolated TBI. In addition, we excluded intubated, comatose, or vegetative patients, so it is possible that there was a selection bias or that this tool is limited to patients with an isolated, but less severe, TBI. Finally, other studies exploring the “6-Clicks” for discharge prognostication purposes exclude patients who died during their hospitalization, making it difficult to draw relevant comparisons (Citation21).
Limitations
This study has several limitations. The research was based on a heterogeneous population of patients with a TBI who were admitted to a L1TC. Therefore, we caution extrapolation of our findings to non-L1TCs and a homogeneous population of patients with a TBI. In addition, although we controlled for ISS, we did not include other descriptors of TBI severity. It is important to note that we were unable to control for the qualifications of the PTs performing each assessment and we also relied on them to determine when a patient was ready for an assessment. Additionally, we did not include whether subsequent “6-Clicks” assessments were performed because the intent of this study was to investigate the index assessment. Finally, a patient’s discharge location can be influenced by other factors not tracked by our registry, such as social issues, financial constraints, patient preference, inter-family dynamics, and bed availability (Citation2,Citation5,Citation27).
Conclusion
Among patients with an acute TBI, increasing AM-PAC “6 Clicks” scores were associated with a shorter hospital LOS and a greater likelihood of home discharge. Decreasing mobility scores correlated with a greater odds of discharge to an IRF, SL, and death/hospice. The “6-Clicks” should not replace clinical judgment, rather it should account for concomitant mobility-limiting injuries, complications, and age to best facilitate discharge planning.
Acknowledgments
None
Disclosure statement
No potential conflict of interest was reported by the author(s).
Additional information
Funding
References
- Taylor CA, Bell JM, Breiding MJ, et al. Traumatic brain injury–related emergency department visits, hospitalizations, and deaths — United States, 2007 and 2013. MMWR Surveill Summ. 2017;66(9):1–16. doi:10.15585/mmwr.ss6609a1.
- Rogers S, Richards KC, Davidson M, et al. Description of the moderate brain injured patient and predictors of discharge to rehabilitation. Arch Phys Med Rehabil. 2015;96(2):276–82. doi:10.1016/j.apmr.2014.09.018.
- Centers for Disease Control and Prevention (CDC). CDC grand rounds: reducing severe traumatic brain injury in the United States. MMWR Morb Mortal Wkly Rep. 2013;62(27):549–52.
- Faul M, Xu L, Wald MM, et al. Traumatic brain injury in the United States: national estimates of prevalence and incidence, 2002-2006. Inj Prev. 2010;16(Supplement 1):A268–A268. doi:10.1136/ip.2010.029215.951.
- Cuthbert JP, Corrigan JD, Harrison-Felix C, et al. Factors that predict acute hospitalization discharge disposition for adults with moderate to severe traumatic brain injury. Arch Phys Med Rehabil. 2011;92(5):721–730.e3. doi:10.1016/j.apmr.2010.12.023.
- Sorensen M, Sercy E, Salottolo K, et al. The effect of discharge destination and primary insurance provider on hospital discharge delays among patients with traumatic brain injury: a multicenter study of 1,543 patients. Patient Saf Surg. 2020;14(1):2. doi:10.1186/s13037-019-0227-z.
- Haley SM, Coster WJ, Andres PL, et al. Activity outcome measurement for postacute care. Med Care. 2004;42(1):I–49. doi:10.1097/01.mlr.0000103520.43902.6c.
- Jette DU, Stilphen M, Ranganathan VK, et al. AM-PAC “6-clicks” functional assessment scores predict acute care hospital discharge destination. Phys Ther. 2014;94(9):1252–61. doi:10.2522/ptj.20130359.
- Haley SM, Ni P, Lai J-S, et al. Linking the activity measure for post acute care and the quality of life outcomes in neurological disorders. Arch Phys Med Rehabil. 2011;92(10):S37–S43. doi:10.1016/j.apmr.2011.01.026.
- Hoyer EH, Young DL, Klein LM, et al. Toward a common language for measuring patient mobility in the hospital: Reliability and construct validity of interprofessional mobility measures. Phys Ther. 2018;98(2):133–42. doi:10.1093/ptj/pzx110.
- Nichol AD, Higgins AM, Gabbe BJ, et al. Measuring functional and quality of life outcomes following major head injury: Common scales and checklists. Injury. 2011;42(3):281–87. doi:10.1016/j.injury.2010.11.047.
- Copes WS, Champion HR, Sacco WJ, et al. The injury severity score revisited. J Trauma Inj Infect Crit Care. 1988;28(1):69–77. doi:10.1097/00005373-198801000-00010.
- Probasco JC, Lavezza A, Cassell A, et al. Choosing wisely together: physical and occupational therapy consultation for acute neurology inpatients. The Neurohospitalist. 2018;8(2):53–59. doi:10.1177/1941874417729981.
- Tracy BM, Carlin MN, Tyson JW, et al. The 11-item modified frailty index as a tool to predict unplanned events in traumatic brain injury. Am Surg. 2020;86(11):1596–601. doi:10.1177/0003134820942196.
- Jette DU, Stilphen M, Ranganathan VK, et al. Interrater reliability of AM-PAC “6-clicks” basic mobility and daily activity short forms. Phys Ther. 2015;95(5):758–66. doi:10.2522/ptj.20140174.
- Wagenmakers E-J, Farrell S. AIC model selection using Akaike weights. Psychon Bull Rev. 2004;11(1):192–96. doi:10.3758/BF03206482.
- Wang X, Jiang B, Liu JS. Generalized R-squared for detecting dependence. Biometrika. 2017;104(1):129–39. doi:10.1093/biomet/asw071.
- Schober P, Boer C, Schwarte LA. Correlation Coefficients. Anesth Analg. 2018;126(5):1763–68. doi:10.1213/ANE.0000000000002864.
- Dutton RP, Cooper C, Jones A, et al. Daily multidisciplinary rounds shorten length of stay for trauma patients. J Trauma Inj Infect Crit Care. 2003;55(5):913–19. doi:10.1097/01.TA.0000093395.34097.56.
- Menendez ME, Schumacher CS, Ring D, Freiberg AA, Rubash HE, Kwon Y-M, et al. Does “6-clicks” day 1 postoperative mobility score predict discharge disposition after total hip and knee arthroplasties? J Arthroplasty. 2016;31(9):1916–20. doi:10.1016/j.arth.2016.02.017.
- Fernandez N, Gore S, Benson S, et al. Use of AM-PAC “6 click” scores to predict discharge location post-hospitalization in adults with cardiovascular disease: a retrospective cohort study. Cardiopulm Phys Ther J. 2020;31(4):152–58. doi:10.1097/CPT.0000000000000128.
- Foreman BP, Caesar RR, Parks J, et al. Usefulness of the abbreviated injury score and the injury severity score in comparison to the Glasgow coma scale in predicting outcome after traumatic brain injury. J Trauma Inj Infect Crit Care. 2007;62(4):946–50. doi:10.1097/01.ta.0000229796.14717.3a.
- Khan F, Amatya B, Galea MP, et al. Neurorehabilitation: applied neuroplasticity. J Neurol. 2017;264(3):603–15. doi:10.1007/s00415-016-8307-9.
- Pfoh ER, Hamilton A, Hu B, et al. The six-clicks mobility measure: a useful tool for predicting discharge disposition. Arch Phys Med Rehabil. 2020;101(7):1199–203. doi:10.1016/j.apmr.2020.02.016.
- Sanford AM, Orrell M, Tolson D, et al. An international definition for “nursing home. J Am Med Dir Assoc. 2015;16(3):181–84. doi:10.1016/j.jamda.2014.12.013.
- Dams-O’Connor K, Cuthbert JP, Whyte J, et al. Traumatic brain injury among older adults at level I and II trauma centers. J Neurotrauma. 2013;30(24):2001–13. doi:10.1089/neu.2013.3047.
- Bonds BW, Dhanda A, and Wade C, et al. Prognostication of mortality and long term functional outcomes following traumatic brain injury: can we do better? J Neurotrauma. 2021;38(8):1168–1176.