Early Versus Later Admission to Postacute Rehabilitation: Impact on Functional Outcome After Traumatic Brain Injury

Effectiveness of Broad Spectrum Programs 

Support for postacute rehabilitation effectiveness comes from studies examining patients with mixed brain injury etiologies who received rehabilitation via broad-spectrum programs. Such studies have demonstrated improved functional outcome at various intervals following discharge with increases seen in productive activity5, 6, 7, 8 as well as improvements in independent living and household management.7, 8 These studies attempted to control for spontaneous recovery, an issue that can confound studies not including a traditional no-treatment control group. Cope et al5, 6 divided patients into groups entering postacute rehabilitation immediately after acute care and those who entered later (ie, >1y postinjury) and reported gains in employment after postacute rehabilitation even for those persons who entered rehabilitation after the critical period of spontaneous recovery. Johnston7, 8 and colleague handled this issue in a slightly different way. Gains in functioning after postacute rehabilitation were still observed after covarying the interval between injury and admission to postacute rehabilitation. It is important to note, however, that the above studies included a mixed etiology population, and thus, the results are generalizable to a general rehabilitation population and not exclusively to TBI. Of greater concern, however, is that these studies did not report or control for severity of injury in TBI, and it is reasonable to assume that slope of recovery may be steeper for less severe injuries.

Vocational Outcomes After Participation in PABIR Programs 

Although comprehensive-integrated PABIR programs focus on both vocational rehabilitation and social integration, most of the studies of the effectiveness of such programs have primarily examined vocational outcomes. In the 1980s, Ben-Yishay et al9 examined improvement in employability and vocational outcome for subjects with severe TBI after participation in an intensive holistic cognitive rehabilitation program. These persons were unemployable at admission to the program and were judged to have reached a neurologic plateau with respect to recovery. After their involvement in the program, 84% of participants were able to engage in some productive endeavor, 63% of them at a competitive level. This program served as a model for many of the programs subsequently studied.

Prigatano et al10 studied persons with severe TBI who were treated in an outpatient neuropsychologic rehabilitation program and a comparison group of persons with TBI matched on age, sex, education, and injury severity, who were referred for treatment but did not enroll in the program. After 6 months in rehabilitation, 50% of the treatment group was working, compared to only 36% in the control group. In a later study, Prigatano et al11 compared persons who participated in outpatient neuropsychologic rehabilitation with a control group and found that the treatment group had a greater percentage of subjects employed or in school at a 3-year follow-up time point. However, outcome was not measured at discharge, limiting the ability to specifically address gains made while in the program.

Malec and Basford12 investigated the effectiveness of a comprehensive-integrated postacute rehabilitation program for persons with varying levels of brain injury severity and other etiology (eg, stroke, anoxia). At discharge from the program, there was a clear decrease in overall disability, and independent living increased from 48% at admission to 90%. Improved work status was also noted, with 61% of persons employed at discharge. However, this statistic included persons employed with job modifications. Unfortunately, this study did not report functional or employment status at a follow-up time point, which would provide valuable information concerning the maintenance of gains made while in the program.

Klonoff et al13 studied persons with mixed etiology and examined outcome after milieu-based rehabilitation. Outcome was defined as level of discharge productivity adjusted for staff ratings of initial functional impairment. At discharge, 90% of patients showed fair or good outcome with 83% being productive in some capacity and 62% gainfully employed or full-time students. Klonoff et al14 examined productive activity (gainful employment, school or volunteer work) and employment in 164 persons with heterogeneous brain injury etiologies (68.9% were TBI) who participated in milieu-based day treatment neurorehabilitation. The cross-sectional sample was evaluated at 3 months and 1, 3, 5, 7, 9, and 11 years postdischarge. Eighty-three percent of patients were productive up to 11 years postdischarge with 67.1% engaged in work or school. No decline in productivity was seen over time from discharge.

In contrast to the above studies, Salazar et al15 found no difference between rates of employment in a study of 120 persons with moderate to severe TBI, who were randomly assigned to either an inpatient milieu-oriented cognitive rehabilitation program or a home-based program involving education and structured exercise and recreational activities. No discharge assessment was reported. Although this study is among the first randomized controlled studies concerning the effectiveness of rehabilitation, there are several caveats that need to be addressed. First, injury severity was defined differently than that employed in other studies. Moderate to severe TBI was defined as Glasgow Coma Scale (GCS) score of 13 or less and posttraumatic amnesia (PTA) of at least 24 hours or focal findings on computed tomography or magnetic resonance imaging. Generalizability is limited because TBI patients had to be “active military duty, not pending medical separation,” which might exclude more severe injuries. Furthermore, this study was implemented in an inpatient setting, whereas most previous studies have been conducted in outpatient programs designed to reintegrate patients with moderate to severe injuries into their community in as real-world a setting as possible. Finally, the return to work rate was much higher than base rates reported in other outcome studies of TBI,4, 5, 6, 7, 8, 11, 16, 17 indicating that the severity of injury for participants in the Salazar15 study was less than for participants in these other studies.

Social Integration After Participation in PABIR Programs 

Recent studies have examined social integration and community participation as well as vocational outcome after participation in comprehensive PABIR programs. Seale et al18 used the Community Integration Questionnaire (CIQ)19 to evaluate the effectiveness of a residential postacute rehabilitation program for persons with TBI. Participants were categorized by the amount of time since injury prior to starting the rehabilitation program. Participants admitted to the program less than 1 year postinjury formed one group. Participants admitted to the program between 1 to 5 years postinjury formed the other group. No differences were found between the groups on measures of severity of initial injury, age, education, sex, race, or drug and alcohol abuse. Participants were assessed at admission to the rehabilitation program and then again at about 1 month after discharge. Both groups showed significant improvements from admission to follow-up. The group admitted earlier after injury showed the most gains on the CIQ. However, significant improvement was also observed for the group admitted later. Gains in community integration, including productivity, could not be accounted for by spontaneous recovery.

Other studies have more carefully studied continued maintenance of gains made in postacute rehabilitation at long-term follow-up time points. These studies are especially important in determining the generalizability of strategies and skills learned in the rehabilitation setting. Sander et al20 examined functional status and disability ratings in 34 subjects with mild to severe TBI. Assessments were conducted at admission and discharge from postacute rehabilitation and also at 2 separate follow-up time points approximately 1 and 4 years postdischarge. Significant improvements in functional status and disability were seen from admission to discharge on all measures. Additionally, the functional levels did not change from discharge to either of the 2 follow-up time points, indicating that functional gains were maintained over time. However, examination of individual cases indicated that not all persons maintained the functional gains acquired in rehabilitation. Individual changes from discharge to the follow-up time points were noted in both positive and negative directions.

A more recent study by Cicerone et al21 compared the effectiveness of an intensive cognitive rehabilitation program to a standard rehabilitation program. The standard program consisted primarily of physical therapy, occupational therapy, speech therapy, and neuropsychologic treatment. Duration of treatment was approximately 4 months for both groups. While both groups showed significant improvement on the CIQ, participants in the intensive cognitive rehabilitation program were twice as likely to show significant clinical benefit on the measure.

Focus of the Current Study 

The literature on the effectiveness of comprehensive PABIR programs remains incomplete. While most studies have supported the conclusion that functional outcomes, in particular, vocational outcomes, improve after PABIR, the only randomized study15 examining the effectiveness of comprehensive PABIR produced negative results. It is still unclear what the characteristics are of persons with TBI who may benefit from postacute rehabilitation. In addition, while a number of studies have examined vocational outcomes after comprehensive PABIR, relatively few have examined changes in community participation or the utility of standard measures for measuring changes in outcome. Few studies have examined the effectiveness of PABIR begun at various time points after injury and few studies have looked at the stability of gains made during PABIR.

In the following study, persons with TBI were carefully characterized by demographic characteristics, initial severity of injury, resulting level of disability, level of independence, and level of community participation. Persons with TBI who started a holistic comprehensive PABIR program at a variety of intervals from time since injury were studied to determine who would benefit most from rehabilitation services and whether gains made were retained over time, and to determine if persons with TBI would benefit from rehabilitation services even after reaching a stable neurologic recovery more than a year after injury. Reliable and valid clinical scales and self-report measures were used to assess multiple functional domains at admission to and discharge from postacute rehabilitation to determine if participation led to gains in functional status and productivity. It was hypothesized (1) that persons with disability resulting from severe TBI would benefit from participation in a postacute rehabilitation program regardless of when they started the program. It was anticipated (2) that gains made during postacute rehabilitation would be retained over time. Finally, it was expected (3) that persons with TBI who started the program early in their recovery process would show the most functional gains.

Methods 

The current study assessed gains in functioning after participation in a comprehensive-integrative postacute rehabilitation program for a group of subjects with TBI. Other etiologies of brain injury were not included. Standard clinical scales and self-report measures were utilized to assess multiple functional domains (ie, global disability, productivity, independence, community integration) at admission and discharge. Additionally, data were collected at a follow-up time point to assess maintenance of these gains. Participants were divided into groups based how long after injury the intervention took place to determine whether the timing of the intervention affected the amount of functional gain.

Results 

Characterization of the Sample 

Detailed demographic characteristics of the participants are presented in table 1. There was a higher percentage of African Americans who entered PABIR greater than 12 months after injury than in the other 2 groups. Otherwise, the 3 groups did not differ with respect to demographic variables including age, level of education, sex, and marital status. In terms of injury severity, the >12 month group took longer to follow commands and had longer periods of PTA than the <6 month group. The >12 month group also had longer acute and rehabilitation hospitalization stays than the <6 month group. The 6–12 month group generally fell between the other 2 groups with respect to injury severity and length of hospital stay. Most participants were male and white. The majority of participants had completed high school and were either working or enrolled in school at the time of injury. Highest GCS score21 within the first 24 hours of injury was available for approximately 86% (n=144) of the sample and was used as an indicator of injury severity (severe injury, GCS score ≤8). For the remaining 14% (n=23) of the sample, GCS score at admission to the emergency department or duration of unconsciousness (severe injury, >24h to follow commands) were used as indicators of injury severity. Injury severity was missing for 6 patients due to incomplete medical records. On the basis of these criteria, 73 (45%) of the 161 patients with available injury severity data were classified as having sustained a severe TBI. The remainder of the sample was classified as having sustained a mild to moderate injury. However, all participants had cognitive impairments that impacted their ability to return to independence or productivity.

Table 1. Demographic and Injury Severity Characteristics

	Variables	Total	>6mo Group 1 (n=115)	6–12mo Group 2 (n=23)	>12mo Group 3 (n=29)	P
	Age (n=167)	30.9±11.5	31.5±12.2	32.8±10.6	27.2±8.9	NS
	Education (y) (n=164)	12.8±2.5	12.8±2.4	13.3±2.8	12.3±2.6	NS
	Sex (% male)	70.7	74.8	60.9	62.1	NS
	Race (%)					<.01
	White	72.3	78.3	68.2	51.7
	African American	13.3	7.0	18.2	34.5
	Hispanic	12.0	13.9	4.5	10.3
	Other	2.4	.9	9.1	3.4
	Days to follow commands (n=141)	11.2±18.8	7.6±11.8	16.1±21.1	23.2±33.3	1<2 (P=.051)
						1<3 (P<.01)
	GCS score (highest in first 24h) (n=144)	9.6±3.8	9.7±3.6	10.3±3.7	8.7±4.3	NS
	PTA duration (d) (n=92)	43.0±36.6	39.1±30.8	48.3±38.6	62.8±60.3	1<3 (P<.05)

NOTE. Values are mean ± standard deviation or as indicated. Abbreviation: NS, not significant.

Patients were enrolled in PABIR for an average of 4.3 months (range, 1–17.87mo). However, duration of enrollment in PABIR was not necessarily an indication of the amount of PABIR the person with TBI received. Participants in the PABIR program were often not formally discharged from the program until some weeks or months after their most intensive treatment phase. In many cases, participants actually returned to work for some period of time before formal discharge. This allowed staff to continue to work with the participant as they integrated into the workplace or other community setting. Services received were tapered at a time and rate determined on a case-by-case basis. This treatment strategy, however, made it difficult to quantify the amount of PABIR services the participant received. Only the formal discharge date was available in the dataset.

Disability Rating Scale 

Of the 141 persons available for outcome analysis, 128 had available injury severity data and total DRS scores available at all 3 time points. Repeated measures using multivariate analyses were used to examine changes in functional outcome across the 3 time points with injury severity as a covariate. A significant group by time interaction was observed (Pillai, F4,248=5.5, P<.001). Therefore, planned comparisons were examined separately for each group. Within the <6 month group, the first planned contrast revealed a significant difference between admission and discharge total DRS scores (P<.001) indicating a decrease in disability. Planned comparisons between total DRS scores at discharge and the follow-up time point revealed no significant changes in functional status, indicating that functional gains were maintained over time (P=.76). Between group post hoc follow-up tests revealed that the 3 groups did not differ with respect to level of disability at admission to PABIR (all P>.05), although the <6 month group showed less disability than the other 2 groups at discharge and follow-up. The 6–12 month group and >12 month groups did not differ from each other at any time point (fig 1).

View Large Image Download to PowerPoint

Fig 1. DRS scores at admission, discharge, and long-term follow-up from a comprehensive PABIR program. DRS scores are adjusted (Adj) for the covariate days to follow commands. Error bars represent standard errors (SEs). Participants began the program at less than 6 months postinjury, 6 to 12 months postinjury, or more than 12 months postinjury.

Supervision Rating Scale 

Of the 141 persons available for outcome analysis, 112 of them had available injury severity data and SRS ratings available at all 3 time points. Repeated measures using multivariate analyses were used to examine changes in functional outcome across the 3 time points with injury severity as a covariate. A significant group by time interaction was observed (Pillai, F4,216=4.5, P=.002); thus, planned comparisons were examined separately for each group. Within the <6 month group, planned contrasts revealed a decrease in need for supervision between admission and discharge (P<.001) and a further decrease in need for supervision between discharge and follow-up (P<.001). Decreases in the level of supervision needed were seen between admission and discharge in the 6–12 month (P=.001) and >12 month groups (P=.002), and level of supervision did not differ between discharge and follow-up (all P>.05) for either group. Between group post hoc follow-up tests revealed that the >12 month group required less supervision than the <6 month (P=.001) and the 6–12 month (P=.05) groups at admission to PABIR, but the 3 groups required similar levels of supervision at discharge and follow-up time points (fig 2).

View Large Image Download to PowerPoint

Fig 2. SRS scores at admission, discharge, and long-term follow-up from a comprehensive PABIR program. SRS scores are adjusted for the covariate days to follow commands. Error bars represent SEs.

Community Integration Questionnaire 

Analyses were conducted separately for the total CIQ score and each of the 3 CIQ scale scores: home competency, social integration, and productive activity. Of the 141 persons available for outcome analysis, 90 had injury severity data and CIQ data available at all 3 time points.

CIQ total score 

First, repeated measures using multivariate analyses were used to examine changes in functional outcome across the 3 time points for the CIQ total score with injury severity as a covariate. The group by time interaction was not statistically significant. There was a main effect of time (Pillai, F2,85=20.1, P<.001) but no main effect of group (P>.05) indicating that all groups displayed a similar pattern of change across time points. Follow-up tests revealed an increase in CIQ total scores between admission and discharge (P<.001) with further increased CIQ total scores between discharge and follow-up (P=.003) (fig 3).

View Large Image Download to PowerPoint

Fig 3. CIQ total scores at admission, discharge, and long-term follow-up from a comprehensive PABIR program. Scores are adjusted for the covariate days to follow commands. Error bars represent SEs.

Home competency scale 

Similar results were observed for the home competency scale. The group by time interaction was not statistically significant. However, there was a main effect of time (Pillai, F2,85=10.9, P<.001), but no main effect of group (P>.05), indicating that all groups displayed a similar pattern of change across time points. Follow-up tests revealed an increase in functioning between admission and discharge (P=.007) with further increased scores between discharge and follow-up (P=.001) (fig 4).

View Large Image Download to PowerPoint

Fig 4. CIQ home competency score at admission, discharge, and long-term follow-up from a comprehensive PABIR program. Scores are adjusted for the covariate days to follow commands. Error bars represent SEs.

Social integration scale 

Likewise, repeated measures using multivariate analyses revealed a main effect of time (Pillai, F2,85=4.0, P=.02), but no main effect of group (P>.05), again indicating that all groups displayed a similar pattern of change across time points. Although no significant differences were seen between admission and discharge or between discharge and follow-up, a significant increase was seen between admission and the follow-up time point (P=.014), indicating a gradual improvement over the longer time interval (fig 5).

View Large Image Download to PowerPoint

Fig 5. CIQ social integration score at admission, discharge, and long-term follow-up from a comprehensive PABIR program. Scores are adjusted for the covariate days to follow commands. Error bars represent SEs.

Productive activity scale 

For the productive activity scale, a significant group by time interaction was observed (Pillai, F4,172=3.1, P=.02). Planned comparisons revealed increased productive activity between admission and discharge for the <6 month (P<.001) and >12 month (P=.006) groups. After the Bonferroni adjustment, the 6–12 month group showed only a marginally significant increase in productive activity between admission and discharge (P=.04). Increases were maintained at follow-up, evidenced by no change in productivity scores for any group between discharge and follow-up time points (all P>.05). Between-group post hoc follow-up tests revealed that the 3 groups were equivalent with respect to productive activity at admission to PABIR (all P>.05). The <6 month group demonstrated higher productive activity scores than the 6–12 month (P=.018) and >12 month (P=.049) groups at discharge and higher than the 6–12 month group at the follow-up time point (P=.004). The 6–12 month group and >12 month groups did not differ from each other at any time point (fig 6).

View Large Image Download to PowerPoint

Fig 6. CIQ productive activity score at admission, discharge, and long-term follow-up from a comprehensive PABIR program. Scores are adjusted for the covariate days to follow commands. Error bars represent SEs.

Scores on the productive activity scale primarily reflect employment but include credit for volunteer work and educational activities. Figure 7 shows that gains were made for all 3 groups for percentage of participants employed. These employment gains were largely maintained at long-term follow-up. The largest gains were seen by the participants who began PABIR earliest (<6mo). The least gain was shown by the persons who began PABIR latest (>12mo) but the gains were statistically and practically significant (χ2 test, P<.05).

View Large Image Download to PowerPoint

Fig 7. Percentage of participants employed at admission, discharge, and long-term follow-up from a comprehensive PABIR program.

Discussion 

Previous studies have focused primarily on vocational outcomes after participation in comprehensive PABIR programs. This study replicates the vocational gains shown by other studies,9, 10, 11, 12, 13, 14 but also demonstrates gains made during postacute rehabilitation in the areas of independence and community participation. This study shows that gains can be effectively measured on 3 different reliable and valid measures of functional outcome. The study shows that community participation and level of independence are measurable outcomes that improve during PABIR. In addition, improvement in level of community participation and level of independence was maintained for more than 1 year after discharge. Furthermore, significant gains were made even for very severely injured persons with persisting disability who participated in the program more than 1 year after injury. The largest gains were demonstrated for persons participating in postacute rehabilitation less than 6 months after their injury. This was evidenced by the significant interactions between time and group on the DRS, SRS, and productive activity scale of the CIQ. Smaller, but significant gains were made by persons with more chronic disability greater than 1 year after injury. The above results partially replicate and extend the findings of Sander,20 Seale,18 and Cicerone21 and colleagues.

It is important to note that the group that began PABIR latest (>12mo postinjury) was also more severely injured as indicated by the number of days it took for them to follow simple motor commands. For this reason, the number of days until the person was able to follow simple motor commands was covaried for each analysis. However, despite the fact that the group starting the postacute rehabilitation program more than a year postinjury was more severely injured, significant functional gains were observed. Given that their disability had persisted at least a year after injury, functional gains made during the postacute rehabilitation program cannot be reasonably attributed to spontaneous neurologic recovery. The fact that these clients were more severely injured only makes their functional gains more noteworthy.

For all groups and all measures, functional gains made during postacute rehabilitation were maintained at an average of 1.5 years after discharge. For the DRS and the CIQ productivity score, functional gains were maintained. On the SRS, additional gains were observed after discharge for the group that began PABIR less than 6 months after injury. On the CIQ total and home competency scores, additional gains were observed after discharge for all groups. The results indicate that while productive activity tends to plateau after discharge from a postacute rehabilitation program, the need for supervision may continue to decrease (at least for the more recently injured persons) and home competency continues to increase.

The reason for the continued improvement in supervision level and home competency is unclear. The lack of a comparison group that did not participate in postacute rehabilitation makes it impossible to say whether these continued improvements were necessarily related to participation in the PABIR program. However, the continued improvement shown by the group who began the program more than 12 months after injury cannot be attributed to additional neurologic recovery. It may be that the continued improvement is simply the result of additional time at home adapting to the disability. It is also possible that participation in postacute rehabilitation has provided them with the necessary strategies to help them continue adapting to their home environments. Further research using comparison groups who have not participated in comprehensive PABIR but are studied at similar time intervals after injury are needed to answer this question.

It should be noted that the randomized trial by Salazar et al15 did not include a no-treatment comparison group. Rather, the comparison group received an alternative home-based treatment that may have been quite extensive. The authors indicate that 90% or more of both groups returned to work at 1 year after injury. Because this rate of return to work is so much higher than that reported in the current study and others,9, 10, 11, 12, 13, 14 it is quite possible that participants in this study were less severely injured or at least had a lower level of initial disability than persons typically treated in the other programs. Because the outcomes were so uniformly high, it is possible that ceiling effects made it difficult to distinguish between the treatments. It is impossible to compare the findings across studies if the case mix of the participants is not precisely specified. GCS scores, duration of impaired consciousness, and duration of PTA should be specified whenever possible. In addition, the resulting level of impairment, disability, independence, and community participation should be measured using standard reliable and valid measures prior to treatment, following treatment, and at a reasonable follow-up point.

Appropriate comparison groups are usually difficult to obtain in the context of PABIR. Enough evidence exists for the efficacy of PABIR programs that most practitioners would have ethical difficulties with creating a no-treatment comparison group. The literature previously reviewed indicates evidence for the effectiveness of both broad spectrum rehabilitation programs and comprehensive holistic PABIR. Only the study by Cicerone et al21 has compared the effectiveness of these 2 types of programs. No studies have considered the cost effectiveness of the 2 types of programs. This is clearly an area that warrants future research.

Besides the lack of a comparison group, this study has other limitations. Another limitation of the study is the relatively small size of the groups beginning the postacute program more than 6 to 12 months or more than 12 months postinjury. The findings need to be replicated in larger samples. However, the significant findings for these relatively small sample sizes indicate that instruments used were sufficiently sensitive to detect differences within and between groups.

Patients with missing data are a problem for all studies in this area. Although subjects with and without follow-up data did not differ with respect to demographic and injury severity variables, including no differences between groups with respect to age, ethnicity, years of education, length of coma, and highest GCS score in first 24 hours, it is possible that patients with missing data may be different in other systematic ways that were not measured. In addition, because the amount of rehabilitation services received by each group was not precisely quantified, it is possible that there were variations between the groups that affect the results in unknown ways.

Conclusions 

The results point toward the effectiveness of postacute rehabilitation in improving functional outcome after TBI. Gains made during postacute rehabilitation are maintained more than a year after participation in the program. The greatest gains were made by persons admitted to postacute rehabilitation less than 6 months after injury. However, postacute rehabilitation was effective for even very severely injured persons who began PABIR more than a year after injury.