If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Sady MD, Sander AM, Clark AN, Sherer M, Nakase-Richardson R, Malec JF. Relationship of preinjury caregiver and family functioning to community integration in adults with traumatic brain injury.
To investigate the relationship of preinjury caregiver and family functioning to community integration outcomes in persons with traumatic brain injury (TBI).
Three TBI Model Systems inpatient rehabilitation facilities.
Persons with TBI (N=141) and their caregivers admitted to inpatient rehabilitation and followed up at 1 to 2 years after injury.
Main Outcome Measures
Community Integration Questionnaire and the Social and Occupation scales of the Craig Handicap Assessment and Reporting Technique.
There were significant interactions of several preinjury caregiver and family variables with injury severity. For persons with complicated mild/moderate injury, better family functioning was associated with greater home integration, and less caregiver distress was associated with better social integration. For persons with severe injuries, greater caregiver perceived social support was associated with better outcomes in productivity and social integration.
Preinjury caregiver and family characteristics interact with injury severity to affect outcomes in persons with injury. Research on outcomes should include measures of caregiver and family functioning. Early interventions targeted toward decreasing caregiver distress, increasing support, and improving family functioning may have a positive impact on later outcomes.
Although TBI results in decreased participation for many persons, there is variability in outcome for individual patients. Variability in outcomes may be related in part to differences in injury severity. Measures of injury severity, including duration of coma and posttraumatic amnesia, have been inversely related to employment outcomes,
One such environmental variable is family functioning. Studies of pediatric patients with TBI emphasize the importance of the family environment in determining outcome after injury. In children with TBI, preinjury family functioning has been predictive of global adaptive functioning, social competence, behavior problems, academic performance, and cognitive functioning at 1 year after injury.
Specifically, children from families with greater cohesion, more positive relationships, and greater flexibility showed better outcomes. Family functioning showed a stronger relationship to children's behavioral outcomes than to academic or cognitive outcomes. Although injury severity was most predictive of the rate of children's recovery during the first year after injury, the family environment was more predictive of their absolute level of outcome.
The role of the family environment in determining outcomes for adults with TBI has received less emphasis. Research has documented that family members experience significant distress after injury.
Family functioning certainly has the potential to affect the outcome of adults with TBI because family members often bear primary responsibility for assisting persons with injury to reintegrate into their homes and communities.
Using semistructured interviews, a qualitative study of social integration showed that persons with TBI cited family members as having key roles in community reintegration after rehabilitation services ended.
from the time of admission to 1 month postdischarge from a post–acute TBI rehabilitation program. Family functioning accounted for significant additional variance after accounting for injury severity, admission DRS scores, and time from admission to follow-up. Although this study was a first step in documenting the impact of the family environment on outcomes, it was limited by small sample size and inclusion of only a single site. Furthermore, the study took place in a postacute setting, with participants' average time since injury of 607 days, and family functioning was assessed at the time of postacute admission. In a prior study, Sander et al
found that 37% of caregivers of persons with TBI admitted to inpatient rehabilitation reported a high level of emotional distress during the month before injury. Between 25% and 33% of caregivers reported unhealthy family functioning in 1 or more areas during the month before injury. Because studies in the pediatric brain injury literature have indicated a relationship between preinjury family environment and outcomes, investigation of this relationship for adults with TBI seems warranted.
The purpose of the present study was to determine the contribution of preinjury caregiver and family functioning to community integration outcomes at 1 to 2 years after injury. Multidimensional assessments of preinjury caregiver and family functioning and community integration outcomes were used. The hypothesis was that healthier preinjury caregiver and family functioning would be associated with better community integration outcomes for the person with injury. Interactions of injury severity with caregiver and family variables also were investigated.
For this prospective study, participants were recruited from among consecutive admissions to the comprehensive inpatient rehabilitation programs at The Institute for Rehabilitation and Research, Methodist Rehabilitation Center in Mississippi, or the Mayo Clinic between February 1999 and June 2002. Caregivers of persons admitted to 1 of the units with a diagnosis of TBI were approached for participation if they met the following criteria for participation in the TBI Model Systems database: medically documented TBI, treatment at the emergency department of a level I trauma center within 24 hours of injury, receipt of inpatient rehabilitation within the TBI Model Systems, age at least 16 years at the time of injury, and provision of informed consent by the person with injury or a legal proxy. In addition to TBI Model Systems criteria, caregivers included in the study had to be at least 18 years old and identify themselves as the primary person responsible for the care of the person with injury. Measures of preinjury caregiver and family functioning were completed by the caregivers within 2 weeks of admission to the rehabilitation unit, and outcomes of the person with injury were assessed at 1 to 2 years after injury.
Persons with TBI (N=289) and their caregivers were enrolled and completed assessments within 2 weeks of admission to the rehabilitation unit. Of these, 96 were excluded for absence of outcome data at 1 to 2 years after injury. An additional 52 were excluded because of a change in caregiver between the time of initial assessment and follow-up. The final sample consisted of 141 dyads of caregivers and persons with TBI.
Most persons with injury were white men with a high school education or higher. Almost two-thirds of participants had sustained a severe TBI, defined as a GCS score of 8 or less at emergency department admission.
Although some participants had mild TBI based on emergency department GCS scores, their cognitive impairments were significant enough for them to be admitted to the inpatient brain injury unit of a TBI Model Systems rehabilitation facility. Caregivers of the person with injury were primarily parents (50%) or spouses (40%). Although half the participants with mild to moderate injury were cared for by spouses, more than half of those with severe injury were cared for by parents. The mild/moderate injury group was significantly older at the time of injury, and there was a longer period between injury and caregiver assessment in the severe injury group. The latter finding is not unexpected because patients with more severe injuries are more likely to stay in the hospital longer and enter rehabilitation later. Demographic characteristics of the sample, categorized by injury severity, are listed in table 1.
Table 1Demographics and Injury Characteristics for Participants
Table 2 lists scores for patients included (n=141) compared with those who were excluded (n=148) on the preinjury caregiver and family variables. Those included had significantly better preinjury family functioning (lower mean FAD scores, t=−2.66; P<.01) and greater preinjury perceived social support (higher mean MSPSS scores, t=2.60; P<.01) than patients who were not included. Included patients did not differ from those excluded on preinjury BSI scores. As listed in table 2, the 2 groups also did not differ with regard to relationships between the caregiver and person with TBI, sex of the caregiver, or sex, age at injury, or injury severity of the person with TBI.
Table 2Comparison of Participants to Persons Excluded on Preinjury Caregiver Variables
Persons Excluded (n=148)
Preinjury GSI score from BSI
Preinjury General Functioning Scale score from FAD
Injury severity was coded as complicated mild/moderate (n=47) or severe (n=89) based on GCS scores on admission to the emergency department, with mild/moderate injuries corresponding to GCS scores of 9 or higher, and severe injuries, to GCS scores of 8 or lower. Complicated mild injuries and moderate injuries were grouped together for analysis because research has shown that outcomes of persons with GCS scores of 13 to 15 who have abnormalities on neuroimaging have outcomes similar to those of persons with moderate injury (GCS score, 9–12).
The General Functioning Scale of the FAD contains 12 items related to overall family dynamics and interactions. Each item is rated on a scale ranging from 0 (strongly disagree) to 4 (strongly agree). An advantage of the FAD is the availability of normative data and cutoff scores to classify dysfunction.
For the General Functioning Scale, an average rating of 2.0 or higher indicates unhealthy family functioning. Researchers have shown the validity of the overall FAD for use with family members of persons with TBI.
The degree of distress caused by each symptom is rated on a scale ranging from 0 (not at all) to 4 (extremely). Adequate internal consistency has been shown by Cronbach α coefficients ranging from .71 to .85 for the subscales.
In addition to 9 subscale scores, the BSI yields a GSI, which can be converted to a T score. The GSI has shown sensitivity to the types of difficulties often experienced by family members of persons with TBI.
The scale consists of 12 items assessing satisfaction with support from family members, friends, and a spouse or significant other. Respondents rate every item on a scale of 1 (very strongly disagree) to 7 (very strongly agree). Three distinct factors have emerged from factor analytic analyses, corresponding to perceived support from (1) friends, (2) family, and (3) a significant other. A total mean score also is obtained by summing all item ratings and dividing by the number of items. Good internal consistency has been shown by Cronbach α coefficients ranging from .85 to .94 for total and individual scale scores. Concurrent validity has been shown through significant correlations between the MSPSS and the Social Support Behaviors Scale, another scale of perceived social support.
The CHART consists of 5 scales: Physical Independence, Mobility, Occupation, Social Integration, and Economic Self-Sufficiency. The CHART initially was developed for assessing outcomes in persons with spinal cord injury and has shown reliability and validity for this population.
Most items are rated on a scale of 0 to 2, with 2 representing greater independence and integration. Initial factor analysis using a sample of 49 persons with TBI yielded a 3-factor structure that the authors labeled Home Integration, Social Integration, and Productive Activity.
The revised factor structure was used in the present analysis. A total score also is calculated, with a range of 0 to 29 points possible. Two versions of the CIQ are available, 1 for completion by patients and another for completion by a family member. Test-retest reliability for both versions has ranged from .83 to .97 for the CIQ total and subscale scores.
Both the CHART and CIQ were used because each assesses productivity and social activities in a different way. For example, the CHocc includes recreational and self-improvement activities, as well as employment, academic pursuits, and homemaking. Conversely, the CIQ separates homemaking from other productive activities and does not include self-improvement and recreational activities. The CHsi contributes information about the number of people in the participant's social network, whereas the CIQ focuses primarily on frequency of social activities. Both pieces of information are important in assessing participation after TBI.
The study procedures, including informed consent, were approved by the institutional review boards at each of the 3 participating centers. Caregivers of persons with injury completed measures of caregiver and family functioning within 2 weeks of admission to inpatient rehabilitation, and outcomes assessment of the person with injury took place 1 to 2 years after injury. Average ± SD time from injury to caregiver assessment was 1.25±.86 months (range, .30–6.93mo), and average time from injury to follow-up was 13.38±2.54 months (range, 9.40–28.10mo).
During the initial caregiver assessment, trained examiners administered a structured interview and the self-report questionnaires. For each caregiver measure (BSI, FAD, MSPSS), the examiners instructed caregivers to complete the questionnaire with regard to functioning during the month before injury. Most (96%) caregivers completed the questionnaires within 3 months of the injury; therefore, difficulty recalling functioning before the injury should have been minimized. There was no relationship of injury severity (mild/moderate vs severe) or time between injury and caregiver assessment to scores on the caregiver measures; therefore, it is unlikely that responses were biased by either of these factors. Although it is impossible to determine whether caregivers were truly responding based on preinjury functioning, implementation of these quality control measures should have maximized the validity of responses.
When possible, the CHART and CIQ were completed by both the person with injury and the family member, and the examiner used clinical judgment to determine which response was more reliable for each measure. The most reliable response was used for data analysis. Although this method of using examiner judgment raises the potential of bias, prior research has indicated that CIQ scores calculated on the basis of self-report of persons with TBI are not significantly different from those resulting from responses of significant others.
found good overall agreement between persons with TBI and their family members using the CHART. When both sources of data were considered to be reliable in the present study, preference was given to data provided by the person with injury. Most outcomes were provided by the person with injury (81% of CHART responses and 95% of CIQ responses).
To compare participants with those who were excluded in regard to demographics, injury severity, and scores on the family and caregiver variables, t tests or chi-square analyses were used.
Pearson product-moment correlation coefficients were calculated to investigate relationships among the 5 outcome scales. Correlations are shown in table 3. The CHocc and CIQprod scales showed the strongest correlations (r=.66; P<.0001). The CIQhi also was related to these (r=.45; P<.0001 with CHocc; r=.41; P<.0001 with CIQprod). All 3 include some combination of work, school, and volunteer activities, as well as housework and parenting. CHsi and CIQsi scores were significantly correlated (r=.40; P<.001), and these scales share assessment of social interaction and activities. The 5 dependent variables were entered into 2 multivariate multiple regression models, divided according to content (work/productivity vs social integration), with this division supported by the finding that correlations were highest between scales assessing similar outcomes. Thus, the dependent variables in the first model were scores on the CHocc, CIQprod, and CIQhi. The dependent variables in the second model were CHsi and CIQsi scores. Multivariate regression was chosen, as opposed to using 5 separate regression equations, because it is a more conservative technique resulting in a smaller chance of type I error. Planned contrasts were conducted to investigate the contribution of the predictor variables for each dependent variable.
Table 3Pearson Product-Moment Correlations Between Outcome Measures
In both models, caregiver and family functioning variables were entered after controlling for injury severity (coded 0 = complicated mild/moderate, 1 = severe). In each model, 2-way interactions between injury severity and caregiver or family functioning also were tested. The nature of significant interactions was investigated by running analyses and testing predictors for significance separately within each severity group.
One-tailed tests were used for each model because there were specific predictions that more positive preinjury caregiver ratings (eg, fewer symptoms, higher perceived social support) would predict better patient outcomes (more social integration, greater productivity). These predictions were based on a prior study by Sander et al
reporting that family functioning at the time of admission to postacute rehabilitation was predictive of outcomes after discharge. The appropriate use of 1-tailed tests to investigate directional hypotheses is documented in the statistical literature.
Muthén & Muthén, 3463 Stoner Ave, Los Angeles, CA 90066.
a using full information maximum likelihood estimation. This method allows all available data to be used in estimation rather than eliminating observations case-wise. The method is 1 of the most powerful for analyzing samples with missing data, similar to using multiple imputation but allowing estimation with only true data.
The proportion of data present for each variable ranged from 83% to 100%, and 75% of cases (n=105) had data for all variables.
Relationship of Preinjury Caregiver and Family Functioning to Outcome
Descriptive statistics for the independent and dependent measures are listed in table 4, whereas results for the 2 regression models are listed in Table 5, Table 6. Injury severity was not a significant predictor of any dependent variables, but there were several significant interactions of severity with family environment variables.
Table 4Descriptive Statistics for Preinjury Caregiver Functioning and Outcomes in the Person With Injury
Preinjury caregiver measures
FAD General Functioning
Outcomes in person with injury
NOTE. Values expressed as mean ± SD. There were no significant differences between injury severity groups for any variable. Injury severity information was unavailable for 5 persons with injury.
The entire model accounted for 24% of the variance in CHocc scores (R2=.24). None of the caregiver or family variables or their interactions contributed significantly to the variance. The model accounted for 7% of the variance in CIQprod scores (R2=.075). There was a significant interaction between injury severity and caregiver perceived social support (P=.036). For participants with mild/moderate injuries, there was no association of caregiver social support with productivity outcome (β=−.075; P=.316). For participants with severe injuries, greater caregiver perceived social support was associated with improved productivity outcome (β=.29; P=.009). The interaction is shown in figure 1. Persons with severe injuries who had caregivers with higher levels of perceived social support had better productivity outcomes than those with severe injuries and caregivers who had lower levels of support.
For CIQhi, the entire model accounted for 4% of the variance (R2=.037). There was a main effect of FAD General Functioning scores such that unhealthy family functioning was associated with lower CIQhi scores (β=−.36; P=.03). However, this effect was moderated by injury severity. There was a significant interaction between injury severity and family functioning (P=.029). For patients with mild/moderate injuries, more healthy family functioning was associated with improved independence in home activities (β=−.31; P=.033), but for patients with severe injuries, there was no association of family functioning and home integration outcome (β=.084; P=.29). The interaction is shown in figure 2.
The entire model accounted for 12% of the variance in CHsi scores (R2=.125). There were 2 significant interactions: severity with BSI scores (P=.039) and severity with social support (P=.031). For the severity-by-BSI interaction, lower caregiver distress was associated with improved social integration in persons with mild/moderate injury (β=−.27; P=.031). There was no effect of caregiver distress on social integration for persons with severe injuries (β=.08; P=.22). For the severity-by-MSPSS interaction, there was no relationship of caregiver perceived social support to social integration in persons with mild/moderate injury (β=.098; P=.24). In contrast, higher perceived social support in caregivers was associated with improved social integration in persons with severe injuries (β=.41; P<.001). The relationship is similar to that found with CIQprod scores, for which the level of caregiver social support did not affect productivity outcome for persons with mild to moderate injuries, whereas higher levels of caregiver social support were associated with improved productivity outcomes for persons with severe injuries.
For CIQsi, the entire model accounted for 7% of the variance (R2=.069). There was a similar interaction of injury severity and social support (P=.041). There was no relationship of social support to integration in persons with mild/moderate injury (β=−.058; P=.35), but higher caregiver perceived social support was associated with improved social integration in persons with severe injuries (β=.28; P=.013).
Post Hoc Examination of Potentially Confounding Variables
Differences between injury severity groups for other variables potentially could drive the interactive influence of injury severity and caregiver or family variables on outcomes for the person with injury. Marital status is 1 such variable, but there was no significant difference in marital status between severity groups. As listed in table 1, persons with milder injuries were more likely to have a spouse as the primary caregiver, whereas persons with more severe injuries were more likely to have a parent as the caregiver. To investigate whether this potential confound influenced the apparent severity effect, we conducted additional regression analyses investigating relationship (spouse vs parent, with “other” excluded from these post hoc analyses), as well as interactions between relationship and the family variables, as potential predictors. These analyses did not indicate that caregiver relationship was the source of the significant interactions between severity and family variables. Two other potential confounds were age at injury, because the mild/moderate group was significantly older at injury, and time since injury, because the mild/moderate group had shorter chronicity on average. The interactions of age and time since injury with family environment variables were each tested separately, and none was significant. This suggests that none of these potential confounds (relationship of caregiver, age at injury, or time since injury) explains the moderating effect of injury severity on the effects of family variables on patient outcomes.
Results of the present study make a unique contribution to the existing research by showing that aspects of the preinjury caregiver and family environment contribute to community integration outcomes of persons with TBI. Although the contribution of the preinjury family environment to outcomes for children with TBI has been long established,
this is the first study to show such a relationship for adults with TBI. However, the relationship of caregiver and family environment and outcomes are complex because they differ for various aspects of the family environment and are dependent on injury severity.
For severe injury, a greater level of perceived social support for the caregiver before injury was associated with increased social integration for the person with injury at 1 to 2 years. This association was present based on both CHsi and CIQsi scores. Greater caregiver social support also was related to greater productivity outcomes on the CIQ. There may be several explanations for these relationships. One possible explanation is that caregivers with better social support have greater access to assistance with aspects of caregiving, have better social networks that can benefit the person with TBI, and have more opportunities to expose the person with TBI to activities outside the home. However, the relationship may be less direct. Prior research has found that caregivers who perceived greater social support were less likely to be distressed and dissatisfied.
Thus, caregivers with good social support may be better adjusted overall and thus better able to assist the person with injury in achieving community integration.
There was no relationship between caregivers' preinjury social support and social integration outcomes for patients with mild to moderate injury. Instead, lower caregiver emotional distress was associated with increased social integration, as assessed by using the CHART. One possible explanation is that caregivers' social support has less of a role for persons with mild to moderate injury because their spontaneous recovery is greater than that of persons with severe TBI. They are likely to be less dependent on caregivers and thus less affected by caregivers' social support. However, they may still be affected negatively when their caregivers are emotionally distressed. Emotional distress in the caregiver may affect their relationship with the person with injury, and may render him/her less capable of assisting the patient in participating in social activities or forming or maintaining social relationships. Thus, for persons with mild to moderate injury, caregivers' social support may be so far removed from the person with injury that the impact is minimal. In contrast, emotional distress in the caregiver is more likely to affect the daily life of the person with TBI.
For persons with mild to moderate injury, healthy family functioning was associated with increased independence in home activities, assessed by using the CIQ. Thus, persons with healthy family functioning were more likely to be independent in such activities as meal preparation, everyday housework, shopping for necessities, financial management, and activity planning. One possible explanation is that families with healthy functioning allowed persons with injury to attempt more activities on their own, supporting patients in their efforts to achieve independence and tolerating resulting errors. However, an alternative explanation relates to the nature of the CIQ. For each item on the CIQhi scale, the person with TBI receives the highest number of points if an activity is performed alone with no assistance. Responsibilities that are shared with another family member earn a lower number of points. It is possible that family members of persons with mild to moderate injuries were more likely to work outside the home because the person with injury needed less supervision. Therefore, the person with injury may have been performing more household tasks without assistance. For persons with severe TBI, there was no relationship between family functioning and independence in home activities. It is possible that more persons with severe TBI may require a family member to be present at all times, and the family member may assist with all activities. Even when a family member is working outside the home, the severity of cognitive deficits in a person with severe TBI may preclude his/her taking over daily household responsibilities.
Results of the present study indicate that the contribution of the family variables can differ depending on the outcome measure used. Preinjury caregiver social support was related to CIQprod scores, but not to CHocc scores. The CIQprod scale is composed of items assessing involvement in work, school, and volunteer work. In contrast, items on the CHocc scale assess participation in work, school, household activities, home maintenance activities, and recreational activities. The results could imply that caregivers' social support has a greater impact on productivity outside of the home as opposed to within the home; however, this should be investigated further in future studies. Furthermore, the CHART requests information about number of hours spent in various activities rather than simply whether persons were participating in these activities independently. The differences in focus and scope between the CIQ and the CHART may have affected results. Thus, it is important to include multiple measures of outcomes when assessing the impact of the family environment.
Although the present study provides important information about the role of the preinjury family environment on community integration outcomes, the results must be interpreted with caution. Caregivers of persons available for follow-up reported greater preinjury social support compared with caregivers of persons lost to follow-up. It is possible that inclusion of more caregivers with lower levels of social support may have resulted in different findings. A further limitation of the study relates to drawing the sample from persons who had received inpatient rehabilitation. Persons who have access to inpatient rehabilitation after TBI may differ in important ways from those who do not have access. Such differences may include education, socioeconomic status, and cultural beliefs regarding family. Such differences may affect the generalizability of the present findings. Future research should focus on replicating our findings with a nonrehabilitation sample. Although center effects were not investigated in the present study, it is possible that the impact of caregiver and family variables may differ by geographic region or center, possibly because of case-mix differences in the clients served (eg, racial/ethnic make-up, socioeconomic status). These potential effects should be investigated in future studies with a larger sample size. Finally, the effects of preinjury family variables on outcomes were small. However, it is possible that even these small effects were associated with meaningful changes in quality of life.
Despite limitations, the present results provide information that can be used to target certain caregivers and families for treatment during inpatient rehabilitation. Persons with TBI who have caregivers reporting low preinjury social support, greater emotional distress, and unhealthy family functioning are at risk for poor community integration at 1 year after injury. Inclusion of treatments to address these issues during rehabilitation may improve the later outcome of persons with TBI. However, a multitude of factors contribute to community integration outcomes, including demographic characteristics, injury severity, and environmental factors other than family. Because the present study was the first to investigate the contribution of preinjury caregiver and family functioning to outcomes in adults with TBI, factors other than injury severity that may affect outcomes and may interact with caregiver and family variables were not included in the models. Post hoc analyses indicated that age of the person with injury, time since injury, and relationship of the caregiver to the person with injury were not interacting with family variables. However, future research could use more sophisticated models to determine the potentially complex relationships among these potential predictors of community integration.
aMuthén & Muthén, 3463 Stoner Ave, Los Angeles, CA 90066.
No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated.
In Sady MD, Sander AM, Clark AN, Sherer M, Nakase-Richardson R, Malec JF. Relationship of preinjury caregiver and family functioning to community integration in adults with traumatic brain injury. Arch Phys Med Rehabil 2010:91;1542-50, the authors regret that the following acknowledgment was omitted from the initial publication.