A Longitudinal Study of Health-Related Quality of Life After Traumatic Brain Injury

Methods 

Participants 

Traumatic brain injury 

The TBI group consisted of adolescent and adult patients who sustained complicated mild to severe traumatic brain injuries and were consecutively admitted to a level I trauma center between 1991 and 1995. These participants are a subsample of the 492 people who were enrolled in the Valproate Seizure Prophylaxis Study within 24 hours of injury. This was a randomized, placebo-controlled, double-blinded investigation of the efficacy of valproate sodium in preventing posttraumatic seizures and its neurobehavioral side effects. Eligibility criteria for this study included computed tomography (CT) evidence of intracranial abnormality including cortical contusion, intracerebral, subdural, or epidural hematoma; depressed skull fracture; or penetrating brain injury. In addition, participants had to be at least 14 years of age at the time of injury and could not have a history of preinjury seizures or other significant brain injury or neurologic condition. Results from the clinical trial found that valproate does not prevent late posttraumatic seizures and does not have any effect, positive or negative, on neuropsychologic functioning.18, 19 The original data collection as well as the additional analyses conducted for this study were approved by the institutional review board at the University of Washington.

The present study focuses on 133 patients who completed the SIP at all 4 of the following time points that were assessed: 1 month, 6 months, 12 months, and 3 to 5 years postinjury. Participants from the larger recruited sample were not included in this study for a variety of reasons including they were found to be ineligible for the study or died at some point during data collection (n=146); were not followed past the 6-month time point because of budgetary constraints (n=61); were untestable at 1 month because of the severity of their injury (n=44); or did not complete the SIP during at least one of the follow-up periods for reasons such as they refused or were lost to follow-up, English was not their first language, or they were physically unable to do so (n=108). In total, 286 subjects completed the SIP at least 1 time. Of those, the 153 patients not included in this study fell into 2 subgroups: 38 did not complete the SIP at 1 month because of severity of their brain injury and associated neuropsychologic deficits, whereas the remaining 115 patients were not included because at 1 or more points they were either lost to follow-up or did not complete the SIP.

A series of 1-way analyses of variance (ANOVAs) with post hoc Tukey tests were conducted to determine how the sample used in this article compared with those 2 groups of participants who completed the SIP at least once but were excluded from this study. Those unable to take the test at 1 month were more severely injured as indicated by both Glasgow Coma Scale (GCS) and time to follow commands (TFCs) than those included in the study and those excluded because of loss of follow-up (P<.001). Severity information for the 3 groups is presented in table 1. The sample used in the study did not differ from either of the excluded subsamples in age or education.

Table 1. Injury Severity Characteristics for Included and Excluded Sample

	Characteristics	Included Sample, n (%)	Excluded because of CNS, n (%)	Excluded for Other Reasons, n (%)
	GCS score
	3–8	9(7)	20(53)	11(10)
	9–12	26(20)	6(16)	23(20)
	13–15	86(65)	7(18)	73(63)
	Unknown	12(9)	5(13)	8(7)
	Total	133	38	115
	TFC
	≤24h	92(69)	5(13)	81(70)
	25h to 6d	23(17)	5(13)	23(20)
	7–13d	13(10)	4(11)	6(5)
	14–28d	5(4)	24(63)	4(4)
	Unknown	0	0	1(1)

Abbreviation: CNS, central nervous system.

Results 

Injury severity for the subjects with TBI is presented in table 1 and demographic data are summarized in table 2. Injuries in the mild range should be considered complicated mild or moderately severe cases because all of those participants, despite GCS scores of 13 to 15, had CT abnormalities.29

Table 2. Demographic Characteristics of TBI, Trauma Control, and Friend Control Groups

	Characteristics	TBI	Trauma	Friend
	n	133	111	87
	Mean age ± SD	37±14	30±13⁎	25±8⁎
	Mean education ± SD	13±2	12±3	12±2
	% Men	81	70	63⁎
	% White	89	77	97⁎

Abbreviation: SD, standard deviation.

⁎ Comparison with TBI group significant at the P<.01 level.

HRQOL Consequences of TBI 

Preliminary analyses were conducted to examine SIP scores according to valproate status. There was no difference between scores of those on valproate or placebo among participants who were blinded to valproate status at either 1 or 6 months postinjury. Therefore, the participants in the TBI group are treated as 1 group in the remaining analyses. A series of 2 MANOVAs were conducted to compare the responses of the TBI group with the trauma control and friend control groups at 1 month and 1 year postinjury. The overall main effect for group was significant for both MANOVAs (P<.001). At 1 month postinjury, the effect for group was significant on each of the 15 dependent variables (SIP subscales and factor scores) (P<.001). At 1 year postinjury, the effect for group was also significant for all of the dependent variables (all P<.001, except communication [P=.002] and eating [P=.013]). Post hoc comparisons were conducted on each of the subscales and factor scores and are presented in table 3. To facilitate easy inspection, the pattern of problem endorsement of these groups is also shown in Fig 1, Fig 2. The friend control group reported minimal to no problems across all subscales. The TBI group endorsed significantly greater difficulties on all scales at 1 month postinjury as compared with their healthy peers. A similar pattern was seen at 12 months, although both groups reported similarly low levels of difficulty on the body care and movement subscale. In contrast, the trauma control and TBI groups followed a similar pattern of endorsement at both time points. In fact, the reported difficulties of the TBI and trauma control participants did not significantly differ on 8 of the 12 subscales at 1 month postinjury. On the remaining 4 subscales (sleep and rest, mobility, recreation and pastimes, work), the TBI participants endorsed significantly more difficulties than the trauma control participants.

Table 3. Mean Scores on SIP for TBI, Trauma Control, and Friend Control Groups

	SIP	TBI		Trauma Control		Friend Control
		1mo	1y	1mo	1y	1mo	1y
	Sleep and rest	36±29	13±22	24±25‡	11±18	1±6‡	1±3‡
	Emotional behavior	14±17	10±17	15±17	10±14	2±7‡	1±4‡
	Body care and movement	13±17	3±6	11±12	3±6	0±1‡	0±1†
	Home management	37±31	7±15	26±25†	8±17	0±2‡	0±1‡
	Mobility	19±21	4±10	11±16‡	2±7	0±1‡	0±0‡
	Social interaction	18±19	10±18	12±12⁎	7±15	1±4‡	0±2‡
	Ambulation	18±20	4±8	14±16	6±11	0±1‡	0±1‡
	Alertness behavior	18±26	17±27	11±19	9±19†	0±1‡	0±1‡
	Communication	8±14	5±12	8±12	5±13	0±2‡	0±2‡
	Work	56±27	21±30	39±33‡	13±25	3±12‡	0±1‡
	Recreation and pastimes	40±26	18±22	25±21‡	12±17⁎	2±6‡	1±4‡
	Eating	5±9	1±3	3±4	1±4	0±2‡	0±1‡
	Psychosocial	15±16	11±16	12±11	8±13	1±3‡	1±1‡
	Physical	16±16	3±6	12±12	4±6	0±1‡	0±0‡
	Total	20±14	8±11	15±11‡	6±10	1±2‡	0±1‡

NOTE: Values are mean ± SD. Trauma control and friend control scores at each time point were compared with TBI group at that time point.

⁎ Significant at the P<.05 level. † Significant at the P<.01 level. ‡ Significant with Bonferroni adjustment for 15 tests (P<.003).

View Large Image Download to PowerPoint

Fig 1. SIP scores at 1 month postinjury. Abbreviations: AB, alertness behavior; AM, ambulation; BCM, body care and movement; C, communication; EAT, eating; EB, emotional behavior; FC, friend controls; HM, home management; MO, mobility; PHYS, physical factor; PSYC, psychosocial factor; RP, recreation and pastimes; SI, social interaction; SR, sleep and rest subscale; TC, trauma controls; TOT, total score; W, work.

View Large Image Download to PowerPoint

Fig 2. SIP scores at 1 year postinjury. Abbreviations: see fig 1. *OSI, other system injury.

Endorsed difficulties of the 2 groups were even more similar at 1 year postinjury, with the difference emerging being a trend toward TBI participants reporting more limitations on the alertness behavior subscale. Thus, by this point in recovery, both the trauma control and TBI groups continue to report difficulties in all areas relative to healthy people, but people with TBI were reporting more cognitive difficulties. To try to better understand the effects attributable to the head injury versus other system injuries, the group was split according to whether they had (n=94) or had not (n=39) sustained other system injuries. Figure 2 includes an illustration of the pattern of endorsement for those who sustained only a TBI (with no other system injury) compared with the trauma control and friend control groups at 1 year postinjury. This shows that TBI alone also leads to significant difficulties in functioning at this point in recovery.

Recovery of HRQOL in TBI Participants 

To test the significance of change over time for each of the SIP subscales and factor scores, GEE analyses were conducted. The results of these analyses are presented in table 4. In general, significant improvement is seen between 1 month and 6 months postinjury, and then the recovery levels off. However, examination of the change over time on the psychosocial and physical factor scores revealed that these domains have quite different recovery paths. Although there was a similar level of endorsement of physical and psychosocial concerns at 1 month postinjury, by 6 months the physical complaints had declined, whereas the psychosocial difficulties remained high. This discrepancy is shown in figure 3. The significant difference between the psychosocial and physical factor scores over time was confirmed by the GEE analysis that showed an interaction between domain and time (P<.001).

Table 4. Results of GEE Analysis

	SIP	Difference Scores ± SD			Overall Significance
		1mo to 6mo	6mo to 12mo	12mo to 3–5y	P
	Sleep and rest	−18±33‡	−4±24⁎	0±22	<.001
	Emotional behavior	−2±20	−2±18	2±20	.259
	Body care and movement	−10±15‡	−1±5	1±6	<.001
	Home management	−28±29‡	−3±14⁎	3±15⁎	<.001
	Mobility	−15±21‡	−1±8	2±11	<.001
	Social interaction	−7±18‡	−1±14	1±16	<.001
	Ambulation	−13±18‡	−1±7⁎	0±8	<.001
	Alertness behavior	1±24	−2±21	4±23	.250
	Communication	−2±16	−1±12	1±12	.209
	Work	−23±33‡	−12±30‡	−3±30	<.001
	Recreation and pastimes	−20±28‡	−1±20	−2±20	<.001
	Eating	−3±9‡	0±3	0±4	<.001
	Psychosocial factor	−3±15⁎	−1±11	2±14	.022
	Physical factor	−11±15‡	−1±4⁎	1±6	<.001
	Total score	−10±12‡	−2±7†	1±9	<.001

⁎ Significant at the P<.05 level. † Significant at the P<.01 level. ‡ Significant with Bonferroni adjustment for 15 tests (P<.003).

View Large Image Download to PowerPoint

Fig 3. Physical and psychosocial factor scores over time in TBI sample.

In fact, the level of difficulties endorsed on 3 of the 4 subscales that contribute to the psychosocial factor score (emotional behavior, alertness behavior, communication) did not change significantly from 1 month to 3 to 5 years postinjury. In contrast, a significant decline in reported difficulties was noted between 1 and 6 months postinjury on the other 9 SIP subscales, including all of the physical subscales. Further examination of the responses of the persons with TBI revealed a difference between the 3 scales that did not change over the time and the remaining subscales. In particular, approximately half of the participants reported no problems on the emotional behavior, alertness behavior, and communication subscales at 1 month postinjury, and this percentage remains around that level across the other 3 time points. In contrast, the percentage of persons with TBI reporting no problems on the other scales tended to be lower at 1 month but then quickly rose by 6 months postinjury.

As can be seen in figure 4, the pattern of endorsed difficulties at 6 months, 1 year, and 3 to 5 years postinjury was nearly the same for all of the SIP subscales except the work subscale. This scale had the highest level of endorsement at 1, 6, and 12 months, reflecting the significant negative effects of brain injuries on employment. Furthermore, it was the only subscale that continued to show a significant decline in reported difficulties from 6 to 12 months postinjury. Three other subscales (ambulation, home management, sleep and rest) as well as the physical factor score and the total score showed a trend of decreasing reported limitations between 6 and 12 months. However, the mean change over time is not statistically significant at the level chosen for multiple comparisons, nor is this change large enough to have clinical significance.

View Large Image Download to PowerPoint

Fig 4. SIP scores over time for the TBI sample. Abbreviations: see fig 1.

Discussion 

The results of this longitudinal study show that TBI is associated with broad and significant perceived limitations in most aspects of everyday life, including both physical and psychosocial functioning. There was substantial improvement from 1 to 6 months, particularly in the physical domains. Improvement beyond this point was modest, occurring primarily in the area of work up to 12 months. In interpreting this finding, one must keep in mind that this sample consisted largely of subjects who had sustained complicated mild or moderate TBIs and who had recovered enough to take the SIP at 1 month after injury.

In contrast to the marked improvement in the physical areas, recovery in psychosocial domains was smaller and, in some cases, nonsignificant up to 3 to 5 years postinjury. Comparisons with friend controls indicate that these limitations are not present in their healthy peers and thus are most likely related to the injury and its consequences. Interestingly, comparison of reported difficulties between the trauma control and TBI groups revealed a similar pattern of difficulties, especially by 1 year postinjury. An exception to this pattern was, as expected, persons with TBI reported greater complaints in the area of cognitive functioning. It is interesting that the persons in the trauma control group did report some cognitive difficulties, even though they were screened for evidence of head injury. The source of these difficulties is unclear, and potential contributing factors, such as current medications and pain, were not assessed in the current study. It should be pointed out, however, that it is not unusual for even young healthy peers to complain of symptoms, such as cognitive problems, that are typically considered to be related to head injury.30

This overall similarity between the trauma control and TBI groups in reported limitations was not expected and is difficult to explain. Although injury to other parts of the body clearly has an effect, the study was not able to equate the type and extent of other injuries between those with brain injury and trauma controls. One should not take the similarity of pattern and level to mean that brain injury of this severity is inconsequential. The TBI group did report greater cognitive problems. Also, the 29% of TBI cases who had no other system injuries reported similar limitations to the trauma control group and the TBI cases with other system injuries, all of which were significantly higher than the friend control group. It is possible that the SIP is not sensitive enough to capture the differences in limitations that may be present between the TBI and trauma control groups. Regardless of cause, it can be stated that both of these groups are experiencing a range of emotional and cognitive difficulties 12 months after injury, and these deficits persist in this TBI sample up to 3 to 5 years postinjury.

Examination of the level of endorsement on subscales over time reveals that limitations in work and leisure/recreation activities are a prominent concern for TBI patients at 1 month postinjury and remain that way over time. Furthermore, limitations in home-management abilities are initially prominent relative to the other subscales but improved considerably by 6 months postinjury. However, from 6 months up to 3 to 5 years postinjury, cognitive difficulties emerged as one of the most endorsed areas of difficulty. In contrast, patients reported minimal limitations in eating and in subscales related to physical functioning (eg, body care and movement, ambulation, mobility) from 6 months on.

The greater improvement over time in physical functioning relative to psychosocial functioning noted in this study is consistent with previous findings.7, 31, 32, 33, 34 However, these results revealed that the magnitude of reported limitations and pattern of recovery is quite variable across the psychosocial and other functional outcome categories. For instance, the subjects in this study did not perceive a significant change in the magnitude of cognitive and emotional problems they were experiencing from 1 month to 3 to 5 years postinjury, at least at the level captured by the SIP. Communication also remained the same over all time periods assessed, although its initial elevation was lower than alertness behavior and emotional behavior. In contrast, these subjects reported fewer limitations in numerous functional activity based domains, such as work, social interaction, home management, and recreation and pastimes from 1 to 6 months. Improvement in these psychosocial role activities but not in cognitive and emotional functioning is intriguing but also is difficult to explain. One possible explanation might be that the alertness behavior and emotional behavior scales examine “impairments,” whereas the other scales examine “participation” in role activities. This line of thinking may further suggest that despite perceived impairments in the areas of cognition and emotions, persons with TBI learn to manage and adjust to their role activities and responsibilities.

Lack of improvement reported in the area of cognitive functioning requires further comment. Numerous studies have used neuropsychologic measures to document improvements in cognitive functioning after TBI, especially in the first 6 months to a year after injury.6, 30, 35, 36, 37 In contrast, the current study used a self-report measure and asked participants to report difficulties in their cognitive functioning that are a result of their state of health or injury. Thus, the comparison is against preinjury baseline, whereas the comparison point used to assess recovery in formal neuropsychologic testing is typically an assessment conducted at some other point postinjury. Thus, these approaches to assessment of cognitive functioning are measuring 2 different aspects of cognitive recovery; the current results are capturing the idea that, despite improvement, some of the persons with TBI may still feel that they are not back to their preinjury baselines. It is also possible that the magnitude of these cognitive deficits is decreasing over time, but the SIP alertness behavior scale is not sensitive enough to capture the improvement. One element that may contribute to this lack of sensitivity is the absence of severity ratings for the difficulties that are endorsed on the individual items of the SIP.

Similarly, an increased level of emotional difficulties was reported across time. The presence of emotional sequelae after TBI has been well documented,38, 39, 40 but the literature regarding the recovery trajectory of these difficulties is mixed. Some studies indicate that, in nonselect cases, emotional functioning improves over time.41, 42 However, there is also evidence that some problems, such as irritability, may increase over time.17, 43, 44 It is possible that the type of problems being endorsed on the emotional behavior subscale in this study is changing over time, but the overall level of endorsement remains the same. Another possibility is that because this scale is not assessing the degree of difficulty, it is not sensitive enough to detect change over time.

Nonetheless, this finding of persistent cognitive and emotional difficulties is particularly notable given that this sample is composed of subjects who were enrolled at the time of injury based on the characteristics of their injury. Thus, these were not persons who were selected because they were involved in rehabilitation or had presented for follow-up because of a complicated recovery or specific complaints about functioning. So, in this sample of subjects with primarily complicated mild to moderate injuries, persisting cognitive and emotional difficulties are perceived to be pervasive by many persons with TBI. These results suggest that persons might benefit from treatment for years beyond their injury. One potential treatment is a scheduled telephone intervention that offers both counseling and education, which has been shown to be effective in improving overall outcome among persons with a recent TBI.45

There are several limitations to take into account when evaluating these results, including the lack of change in reported limitations on the alertness behavior, emotional behavior, and communication subscales. The SIP was chosen for this study because it is a well-validated measure that provides a comprehensive assessment of injury-related difficulties. However, this measure focuses on negative health impacts and thus potentially positive changes in HRQOL were not addressed in this study.46 Furthermore, difficulties on the SIP are endorsed as present or absent, but the degree of impairment or difficulty is not assessed. Thus, as noted previously, the SIP might not be very sensitive to change. It is also important to note that the more severely injured persons were excluded from this study because they were not able to complete the SIP at 1 month postinjury. A trend of perceived ongoing recovery may have been evident if the more severely injured had been included. Future research should explore the differences in quality of life limitations reported by persons with TBI and those with other types of injuries to gain a better understanding of TBI-related difficulties and the variables that contribute to perceived limitations among persons with TBI. This will provide important information that can be used to explore possible treatment interventions.

Conclusions 

This study provides information about the longitudinal course of recovery across various physical and psychosocial domains from 1 month up to 3 to 5 years postinjury. In this sample of persons with mostly complicated mild to moderate TBI, significant early limitations were reported across all domains, with significant improvements over the first 6 months evident in the physical domains. In contrast, the early psychosocial difficulties, which were as prevalent as the physical problems, showed less recovery. Particularly noteworthy was the relatively consistent level of emotional and cognitive complaints from 1 month to the 3 to 5 year period after injury. Difficulties in work and leisure activities were also prominent across all time points assessed. Although persons with TBI had clear and extensive difficulties compared with their healthy peers, they were very similar to those with injuries to body parts other than the brain. An exception was a trend toward greater cognitive complaints by those with TBI. These findings should not be taken to mean that TBI is inconsequential because TBI cases with no other system injuries endorsed problems similar to TBI cases with other system injuries. Rather, these findings highlight the pervasiveness and persistence of injury-related difficulties that compromise the quality of living and the need for efficacious interventions that will reduce the magnitude and duration of such limitations.