Relationship Between Depression and Psychosocial Functioning After Traumatic Brain Injury

Transcription

1 S43 Relationship Between Depression and Psychosocial Functioning After Traumatic Brain Injury Mary R. Hibbard, PhD, Teresa A. Ashman, PhD, Lisa A. Spielman, PhD, Doris Chun, PhD, Heather J. Charatz, MA, Seton Melvin, BA From the Department of Rehabilitation Medicine, Mount Sinai School of Medicine, New York, NY. Supported by the National Institute on Disability and Rehabilitation Research, US Department of Education (grant no. H133B980013). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Reprint requests to Mary R. Hibbard, PhD, Dept of Rehabilitation Medicine, Mount Sinai School of Medicine, One Gustave Levy Pl, Box 1240, New York, NY 10029, mary.hibbard@mssm.edu /04/ $30.00/0 doi: /j.apmr ABSTRACT. Hibbard MR, Ashman TA, Spielman LA, Chun D, Charatz HJ, Melvin S. Relationship between depression and psychosocial functioning after traumatic brain injury. Arch Phys Med Rehabil 2004;85(4 Suppl 2):S Objective: To examine the relationship between depression and psychosocial functioning up to 5 years after traumatic brain injury (TBI). Design: Longitudinal cohort study with 2 assessments completed. Setting: Community. Participants: Individuals (N 188) with TBI living in the community. Interventions: Not applicable. Main Outcome Measures: Structured Clinical Interview for Depression, self-reports of depression severity, functional symptoms, quality of life (QOL), unmet important needs, and psychosocial functioning. Results: Based on observed depression patterns at initial and repeat assessment, 4 subgroups were created: no depression, resolved depression, late-onset depression, and chronic depression. Groups were equivalent in terms of demographic and injury-related factors but differed significantly in perceived psychosocial functioning. The no-depression group reported fewer depressive symptoms and higher levels of psychosocial functioning, whereas the chronic-depression group reported the poorest psychosocial functioning, with a further decline in QOL at reassessment. Although the resolved-depression and late-onset depression groups reported similar psychosocial functioning at initial assessment, psychosocial functioning had improved for the resolved-depression group and declined for the late-onset depression group at reassessment. Pre- and postpsychiatric diagnoses were common in all groups, with pre-tbi diagnosis of depression not predictive of post-tbi depression. Conclusions: Findings highlight the need for broad-based assessments and timely interventions for both mood and psychosocial challenges after TBI. Key Words: Brain injuries; Depression; Psychiatry; Rehabilitation by the American Congress of Rehabilitation Medicine THE MOST FREQUENTLY reported psychosocial challenge after traumatic brain injury (TBI) is major depression, 1 with prevalence rates consistently elevated when compared with expected prevalence rates within community-based samples. 2 Earlier studies of post-tbi depression reported wide variations, with prevalence rates ranging between 6% and 77%. 3-9 These variations have been attributed to differences in study methodology, such as use of differing assessment tools, severity of TBI, time since injury, and the different population groups (ie, inpatient, outpatient, community) in the samples Use of empirically based and reliable psychiatric criteria, as defined by the Diagnostic and Statistical Manual of Mental Disorders 13 (DSM-IV), has permitted greater consistency in reporting prevalence rates in cross-sectional studies. 10,12,14,15 Studies that have used DSM-IV criteria to diagnose major depressive disorders after a TBI have suggested various patterns of emergence, resolution, and chronicity of depression over time. For example, the frequency of major depression after TBI has been reported at 13% at 1-year postinjury, 16 38% at 3 years, 10 50% at 5 years, 11 and 61% at 8 years, 12 which suggests that mood disorders may increase over time postinjury. In contrast, studies focused on current rates of major depression suggest more consistent findings across years postinjury, that is, prevalence rates ranging between 26% and 31% at time periods as varied as 1 month to 8 years post- TBI. 10,12,17,18 Robust resolution rates of depression have also been reported, 10,12 highlighting the often time-limited nature of a major depressive episode. 13 Although there is considerable literature documenting the high prevalence of depression after TBI, little empirical information exists about the nature and complexity of post-tbi depression and its psychosocial factors. Several researchers have suggested that psychiatric diagnoses after TBI may vary over time, with their prevalence declining at approximately 2 years postinjury, and then gradually increasing in subsequent years. These studies have relied on findings from cross-sectional samples, an approach limited potentially by cohort effects that may either over- or underestimate prevalence rates. 21 Longitudinal studies tracking potential patterns of major depression after injury are indicated. Potential predictors of post-tbi depression have been examined in an attempt to better define who is at risk of mood disorders after injury. Despite considerable efforts, few predictors have been identified. 14 Demographic factors such as age, gender, and ethnicity, that have been found to be predictors of depression in non-tbi samples, 13 have been inconsistent predictors post-tbi ,16,21-24 Injury-related factors, such as TBI severity 10-12,24-27 and history of TBI, 23 have also produced inconsistent findings. The most consistent predictor of post- TBI depression has been a prior psychiatric disorder, 11,16,21-26,28,29 a finding supported by the high incidence of pre-tbi psychiatric disorders found in individuals with depressive symptoms post-tbi. 11,12 The relationship between a prior psychiatric diagnosis and patterns of post-tbi depression over time has not been examined and is the subject of this study. Major depression has been associated with significant levels of disability in non-tbi samples. 30 The incremental negative impact of depression on psychosocial functioning and adjustment after TBI is significant. 4,17,31-34 Individuals with co-oc-

2 S44 DEPRESSION AND PSYCHOSOCIAL FUNCTIONING AFTER TBI, Hibbard curring depression and TBI report reduced perceptions of health, 10 lower levels of social functioning and support, 22,26,35 elevated depressive symptoms, 17,22 poorer quality of life 36 (QOL), and decreased community integration. 36,37 These findings suggest a wide array of psychosocial impacts associated with onset of post-tbi depression. To date, studies have not examined whether psychosocial functioning varies by emergence, resolution, or chronicity of depression after TBI. As a result, a contextual foundation on which to understand the nature of post-tbi depression is lacking. In this study, we examined the relationship between depression and psychosocial functioning in individuals who were up to 5 years post-tbi by assessing the emergence, chronicity, or resolution of depression over 2 assessment intervals. We hypothesized that (1) individuals without depression at either of the assessment would report higher levels of psychosocial functioning than those who experienced depression at anytime post-tbi, (2) individuals who were depressed at both assessments would report the lowest levels of psychosocial functioning, (3) individuals whose depression had resolved at the second assessment would report better psychosocial functioning, (4) individuals who experienced a delayed onset of depression would report worse psychosocial functioning at reassessment, (5) the frequency of post-tbi depression would be highest among individuals with a preinjury psychiatric disorder, and (6) the frequency of post-tbi co-occurring psychiatric disorders would be highest among individuals who experienced post-tbi depression. METHODS Participants Participants were 188 adults with TBI who were enrolled in a larger longitudinal research study that assessed mood disorders and psychosocial functioning after TBI. The study was conducted by the Research and Training Center (RTC) on Community Integration of Individuals with TBI, Department of Rehabilitation Medicine, Mount Sinai School of Medicine, New York. A description of the recruitment process for individuals with TBI seen in this study can be found in Ashman et al. 21 A total of 188 individuals were recruited for this study. The research protocol was approved by our institutional review board. Procedures Potential participants were screened by telephone to ensure their eligibility to participate. If they met study criteria, they were scheduled for an initial assessment (T1) that included signing the consent form, and undergoing 2 interviews: a Structured Clinical Interview for DSM-IV 13 (SCID) to assess mood and a structured interview to assess overall QOL and extent of community reintegration. Individuals underwent both interviews again approximately 12 months later (T2) to assess stability of mood and psychosocial functioning. To promote clinical validity, diagnostic interviews were conducted by clinicians with a minimum of 3 years of clinical experience in psychology (ie, predoctoral psychology students, postdoctoral psychology fellows, licensed psychologists). To ensure interrater reliability, interviewers were observed by the research supervisor who rated the patient s DSM-IV diagnosis independently until an interrater reliability of greater than.80 was obtained. Measures Structured Clinical Interview for DSM-IV. The SCID 13 is a semistructured interview through which trained clinicians can determine lifetime and current DSM-IV diagnoses, including mood, anxiety, and substance use disorders. The interview follows a triage approach: screening questions are asked for each diagnosis of interest; if the initial screening criteria are positive, diagnostic-specific questions are asked. Depending on the subject s responses and the interviewer s judgment, the interview continues to the next disorder. SCID questions were modified to permit evaluation of the onset of specific disorders relative to the onset of the TBI. The specific timeline allows a clinician to determine whether a psychiatric diagnosis existed before the TBI, or if it developed after the TBI, and whether it still exists. These modifications have been used in other TBI studies Although the SCID was administered at both T1 and T2, pre-tbi diagnoses were queried only at T1. Beck Depression Inventory. The Beck Depression Inventory 38 (BDI) is a 21-item self-report measure of depression severity, with scores ranging from 0 to 63. Higher scores indicate greater depressive symptoms. The BDI was administered at T1 and T2 to assess change in severity of self-reported depressive symptoms over time. Scores from T1 and T2 were used to validate the groupings in this study that were based on patterns of post-tbi depression. Living Life After TBI. Living Life After TBI 39 (LLATBI) is a structured interview designed to collect information about community integration and overall QOL. The LLATBI, an adaptation of an interview developed in earlier research by the RTC on Community Integration of Individuals with TBI, 40 includes assessments of social role functioning, environmental barriers and supports, perceived health, and QOL after TBI. Several existing QOL and community integration instruments were used either in part or in whole LLATBI items we selected for analysis included the following. Unmet Important Needs. The Unmet Important Need (UIN) Scale was adapted from a measure of 15 need areas identified by Flanagan, 41 with 1 additional item religious and spiritual needs added to the original list. We used the scale to measure subjective QOL. Two ratings were obtained for each item: the degree of importance of a given need area (1, not important; 5, most important) and how well the need in a given area was being met (1, not at all met well; 4, completely met). To study perceptions of unmet needs in each area, we developed the following formula. 36 UIN i I i (MA AA), where UIN i is the degree of unmet need in any area i taking into account the degree of importance in area i to the individual; I i is the importance rating for any area i (0 to 4); MA is the maximum attainment rating (MA 4, a constant); and AA i is the actual attainment rating in area i ( How well are needs met? 1to4). With this formula, areas of life in which needs are seen as fully met and are viewed as not important are set equal to zero. UIN i scores range from 0 to 12. Smaller scores signify less perceived need. When summing across the 16 areas to determine the total score, UIN i solely reflects unmet needs within areas that are important to individuals (ie, not rated 0 in terms of importance). The minimum UIN i for each area is 0 and the maximum is 12. These values generate a total score with a minimum of 0 and a maximum of 192. Thus, a total score ( UIN i ) of 0 indicates that all need areas have been rated as not important and/or as completely met, whereas a total score of 192 indicates that all 16 areas are judged most important by the respondent and their needs in all areas are not at all met. The measure was administered at T1 and T2 to assess change in self-reported unmet needs over time. Impact of TBI on Roles and Responsibilities Scale. The Impact of TBI on Roles and Responsibilities Scale (Impact Scale) consists of 9 items from the LLATBI that ask about the

3 DEPRESSION AND PSYCHOSOCIAL FUNCTIONING AFTER TBI, Hibbard S45 degree of perceived impact of TBI on an individual s social roles and responsibilities. Each item was rated as to degree of change on a 3-point scale (1, a positive change; 0, no change or mixed change; 1, a negative change). Items include who the person lived with, relationship with spouse, relationship with parents, responsibilities as a parent, responsibilities to others in the home, responsibilities in the household, degree to which the person saw others socially, being employed or a student, and the individual s income and income of others in the household. Responses were summed to create a composite measure of the total negative impact of TBI on social role and responsibilities, with high negative score indicative of greater negative impact of TBI on social roles and responsibilities. Reliability coefficients were computed and the overall was.77. The impact measure was administered only at T1. Life-3. The Life-3 45 is a single item that assesses a participant s current level of subjective QOL. The participant rates his/her overall QOL during the past month on a 7-point scale (0, terrible; 6, delighted). The scale is administered twice (time separated) within a given interview. Higher scores on this measure indicate higher levels of subjective QOL. The Life-3 was administered at T1 and T2. Brain Injury Screening Questionnaire. The Brain Injury Screening Questionnaire 46 (BISQ) consists of 100 functional symptoms commonly reported after brain injury. It was adapted from other TBI symptom checklists Participants are asked to identify symptoms that interfered with their ability to function on most days during the prior month (0, no; 1, yes). The BISQ was administered at T1 and T2 to assess change in TBI-related symptoms over time. The questionnaire is divided into 3 subscales: a physical scale of 19 symptoms (eg, sleep difficulties, sensory changes, headaches, clumsiness); a cognitive scale of 48 symptoms (eg, forgetting names, forgetting to take medication, problems with concentration, difficulty learning, problem solving); and a behavioral scale of 33 symptoms (eg, feeling moody, hitting or pushing others, feeling angry, heedless to danger). For this study, symptoms reflective of depression were removed from the latter scale, leaving 23 items in the revised behavioral subscale. Overall rating of health. Health rating is a single item, taken from the LLATBI, that assesses perceived health status. Participants rate overall health on a 5-point scale (0, excellent health; 4, poor health). The perceived health measure was administered at T1 and T2. Overall rating of pain. Overall pain rating is a single item, taken from the LLATBI, that assesses self-reported perceptions of pain. The experience of pain on most days during the prior month is rated on a 5-point scale (0, no pain; 4, very severe pain). The pain measure was also administered at T1 and T2. Data Analysis Creation of post-tbi depression groups. The presence of a DSM-IV major depression at T1 and T2 was used to determine patterns of post-tbi depression. Four groups were created: 91 individuals (48%) did not meet DSM-IV criteria for a major mood disorder at either T1 or T2, and were assigned to the no-depression group; 55 individuals (29%) were diagnosed with major depression at T1 that was resolved by T2, and they were assigned to the resolved depression group; 19 individuals (10%) were not depressed at T1 but met the criteria for DSM-IV diagnoses of a major depression at T2, and they were assigned to the late-onset depression group; and 27 individuals (14%) met the DSM-IV criteria for major depression at both T1 and T2, and they were assigned to the chronic-depression group. All further analyses used data from these 4 groups. To examine categoric variables, frequencies were computed for each group at T1 and T2 by using chi-square analyses to determine significant differences between groups at each time point. For continuous variables, means and standard deviations (SDs) were computed for each group at T1 and T2, and 1-way analyses of variance (ANOVAs) with post hoc analyses were used to determine significant differences among groups at each time point. To assess changes in response patterns between T1 and T2 within groups, repeated-measures ANOVAs for continuous variables were used to determine time by group interactions included in the models. Repeated-measure multivariate analyses of variance (MANOVAs) were used to analyze embedded BISQ subscales to determine time by group interactions included in the models. RESULTS Demographics Demographic data for the 4 groups are summarized in table 1. The mean age across groups was slightly more than 40 years, with the 4 groups comparable in terms of age at time of initial interview. Of the total sample, 23% of participants were under the age of 30, 28% were between 30 and 39, 19% were between 40 and 49, and 30% were age 50 or older. There were more men in the no depression group and more women in the chronic depression group, but this difference was not statistically significant. The groups were equivalent in academic achievement in that most had achieved high school or higher levels of education. Level of income was comparable for all 4 groups. The majority of participants were single, with the 4 groups being comparable in terms of marital status. Although most participants were white, ethnic diversity across the groups reflected national norms. 52 Participants presented with a full range of TBI severity, with the majority having more moderate to severe injuries. 53 Of the 188 individuals, 10% reported a period of being dazed and/or confused post-tbi, 18% reported a loss of consciousness (LOC) of less than 20 minutes, 10% report LOC between 21 and 60 minutes, 13% reported LOC of greater than 1 day to 1 week, 23% reported LOC for between 1 and 4 weeks, and 17% reported LOC greater than 4 weeks. Time since injury was equivalent across the groups, with participants being on average 2.5 years after TBI at the initial interview. Sixteen percent of the total sample had their initial interview within the first year after injury, 23% within the first full year, 18% within the second year, 29% within their third year, and 15% within their fourth year. Frequency of Pre- and Post-TBI Co-Occurring Psychiatric Disorders Psychiatric diagnoses before TBI were common across all groups in that 56% of the total sample (n 105/188) met criteria for a DSM-IV diagnosis before onset of TBI. The 4 groups were comparable in terms of frequency of pre DSM-IV diagnoses (table 2). Psychiatric diagnoses cut across a variety of DSM-IV classifications: mood disorders (24%) and substance abuse disorders (22%) were the most frequent, with frequencies of anxiety disorders (9%) and eating disorder (1%) being less common before onset of TBI. To determine if individuals with a premorbid mood disorder were at increased risk of post-tbi major depression, we explored patterns of pre-tbi mood disorders and post-tbi major depression (either T1 or T2). Of the total sample, 17% met criteria for a pre-tbi mood disorder with development of post-tbi major depression at either T1 or T2. Clearly, these individuals were at increased risk of post-tbi affective dis-

4 S46 DEPRESSION AND PSYCHOSOCIAL FUNCTIONING AFTER TBI, Hibbard Variable Table 1: Demographic Characteristics Across 4 Patterns of Depression After TBI No Depression (n 91) Resolved Depression (n 52) Late-Onset Depression (n 19) Chronic Depression (n 27) Mean age (y) Gender Men 57 (63) 24 (46) 9 (47) 10 (37) Women 34 (37) 28 (54) 10 (53) 17 (63) Education HS 23 (25) 13 (25) 7 (37) 7 (26) HS, some college 37 (41) 15 (29) 9 (48) 11 (41) College 31 (34) 23 (44) 3 (16) 9 (33) Other 0 1 (2) 0 0 Marital status Single 37 (42) 24 (46) 10 (56) 7 (26) Married/cohabit 27 (31) 15 (29) 5 (28) 9 (33) Separate/divorced/widow 23 (27) 13 (25) 3 (17) 11 (40) Ethnicity White 63 (70) 40 (78) 14 (74) 17 (63) Black 12 (13) 3 (6) 3 (16) 5 (19) Hispanic 7 (8) 4 (8) 2 (11) 4 (15) Other 8 (9) 4 (8) 0 1 (4) Income $0-$20,999/y 27 (31) 19 (38) 7 (41) 9 (35) $21,000 60,000/y 29 (34) 19 (38) 8 (47) 15 (58) $60,999 above/y 30 (35) 12 (24) 2 (12) 2 (8) Mean years post-tbi NOTE. Values are mean SD or n (%). Abbreviation: HS, high school. tress. Surprisingly, 21% of those with a pre-tbi mood disorder did not meet criteria for major depression at T1 or T2. Furthermore, 35% of individuals without pre-tbi mood disorders developed a major depression after TBI, with an additional 29% presenting with neither a pre- nor a post-tbi mood disorder. These data provide minimal support for a theory that pre-tbi mood disorders are solid predictors of post-tbi depression. Co-occurring psychiatric diagnoses (other than depression) were common across all 4 groups after TBI, with no significant differences between groups being identified at T1 or T2. As shown in table 2, co-occurring psychiatric diagnoses at initial assessment were most frequent in individuals with late-onset depression (74%), and least frequent in those with chronic depression (26%). At reassessment, such diagnoses increased in frequency in the 3 groups who experienced depression at any time post-tbi (ie, 73% of resolved depression, 84% of lateonset depression, 93% of chronic depression). Co-occurring psychiatric disorder was also found in 53% of the no-depression group. The predominant co-occurring psychiatric disorders at both T1 and T2 were anxiety disorders (19%, 16%, respectively), with other depressive disorders (10%, 5%, respectively), substance abuse disorders (4%, 6%, respectively), and somatoform disorders (0%, 1%, respectively) being less frequently diagnosed. Depressive Symptoms Across Patterns of Post-TBI Depression The 4 groups differed significantly in depression severity, as measured by the BDI, at T1 (F 1, , P.001) and at T2 (F 1, , P.001) (table 2). The no-depression group reported significantly fewer depressive symptoms than all other Table 2: Coexisting Psychiatric Disorders and Severity of Depressive Symptoms Across 4 Patterns of Depression Post-TBI Variable No Depression (n 91) Resolved Depression (n 52) Late-Onset Depression (n 19) Chronic Depression (n 27) Pre-psychiatric disorder* 44 (48%) 33 (64%) 9 (47%) 19 (70%) Co-occurring psychiatric disorder Time 1 47 (52%) 23 (44%) 14 (74%) 7 (26%) Time 2 48 (53%) 38 (73%) 16 (84%) 25 (93%) BDI Time 1 (n 88) (n 49) (n 17) (n 25) Mean SD F 1, Time 2 (n 87) (n 48) (n 17) (n 25) Mean SD F 1, *Includes any DSM-IV diagnosis, including major depression, before TBI. Queried at T1 only. Excludes DSM-IV diagnosis of major depression post-tbi. P.001. ANOVA

5 DEPRESSION AND PSYCHOSOCIAL FUNCTIONING AFTER TBI, Hibbard S47 Fig 1. Self-report of depressive symptoms by the 4 patterns of depression over time. groups at T1 (all contrasts, P.01) and at T2 (all contrasts, P.01). The resolved-depression group reported significantly fewer depressive symptoms than the chronic-depression group at both T1 (P.001) and at T2 (P.01). Repeated-measures ANOVAs revealed a time by group interaction (F 1,3 5.09, P.01), with a significant decline in BDI scores in the resolved-depression group at T2 and a significant increase in BDI scores for the late-onset depression group at T2 (fig 1). These differences were in the expected direction, providing empirical support for the existence of the subgroups. Psychosocial Functioning Across Patterns of Depression Post-TBI Psychosocial functioning was broadly defined to include perceived QOL, extent of unmet important needs, effect of TBI on social roles and responsibilities, and extent to which cognitive and behavioral challenges affected functioning after TBI. Findings are summarized in table 3 and described below. Quality of life. The groups differed significantly in reports of QOL, as measured by the Life-3, at both T1 (F 1,3 8.24, P.001) and at T2 (F 1, , P.001). At T1, the nodepression group had significantly higher QOL ratings than the other groups (all contrasts, P.01). At T2, QOL ratings in the no-depression group remained significantly greater than in either the chronic (P.001) or the late-onset depression (P.001) groups, with equivalent QOL ratings for the resolved and no-depression groups. Repeated-measures ANOVAs revealed a significant time by group interaction (F 1,3 2.90, P.05), with a significant decline in QOL reported by the chronic-depression group at T2 (fig 2). Unmet important needs. The groups differed significantly in regard to perceived important unmet needs, as measured by the UIN scale, at both T1 (F 1,3 8.49, P.001) and T2 (F 1,3 4.73, P.01). At T1, the no-depression group reported significantly fewer unmet needs (better QOL) than either the chronic-depression (P.001) or the late-onset depression (P.01) groups. The no-depression group continued to report significantly fewer unmet needs at T2 whereas the level of unmet needs was now equivalent (P.05) for the resolved and no-depression groups. Interaction effects on the repeated-measures ANOVAs were not significant, which suggests that overall ratings of unmet important needs within groups were stable over time. Impact of TBI on roles and responsibilities. The groups differed significantly (F 1,3 4.11, P.01) in the degree of impact TBI had on social roles and responsibilities, as measured by the Impact Scale. Post hoc analysis revealed that participants in the no-depression group reported significantly less negative impact than did the chronic-depression or the resolved-depression groups (P.05) at T1. Cognitive symptoms. We computed MANOVAs for the cognitive scale of the BISQ. The groups differed significantly in the extent to which cognitive symptoms affected functioning at T1 (F 1, , P.001) and T2 (F 1, , P.001). Post hoc analyses revealed that the no depression group reported significantly fewer TBI cognitive symptoms than all other groups at T1 (all contrasts, P.001) and at T2 (all contrasts, P.001). We computed repeated-measures MANOVAs to examine the consistency of cognitive symptoms reported across groups over time. We found 1 time by group interaction (F 1,3 4.58, P.01), with significant declines in cognitive symptoms reported at T2 in the resolved-depression group. Behavioral symptoms. We computed MANOVAs for the behavioral scale of the BISQ. The groups differed significantly in their reports of the extent to which behavioral symptoms interfered with functioning at T1 (F 1, , P.001) and at T2 (F 1, , P.001). Post hoc analyses revealed that the no-depression group reported significantly fewer behavioral symptoms than all other groups at T1 (all contrasts, P.001) and at T2 (all contrasts, P.001). Repeated-measures MANOVAs were not significant; therefore, patterns of reported behavioral symptoms within groups were stable over time.

6 S48 DEPRESSION AND PSYCHOSOCIAL FUNCTIONING AFTER TBI, Hibbard Table 3: QOL Indicators Across 4 Patterns of Depression After TBI Variable No Depression Resolved Depression Late-Onset Depression Chronic Depression ANOVA Life-3 Time 1 (n 91) (n 50) (n 19) (n 27) Mean SD F 1,3 8.24* Time 2 (n 89) (n 47) (n 18) (n 24) Mean SD F 1, * UIN scale Time 1 (n 90) (n 51) (n 19) (n 27) Mean SD F 1,3 8.49* Time 2 (n 89) (n 47) (n 18) (n 24) Mean SD F 1,3 4.73* Impact scale Time 1 (n 91) (n 51) (n 19) (n 27) Mean SD F 1, Cognitive scale Time 1 (n 90) (n 51) (n 19) (n 27) Mean SD F 1, * Time 2 (n 89) (n 47) (n 18) (n 24) Mean SD F 1, * Behavioral scale Time 1 (n 90) (n 51) (n 19) (n 27) Mean SD F 1, * Time 2 (n 89) (n 47) (n 18) (n 24) Mean SD F 1, * *P.001. P.01. Health Indicators Across 4 Patterns of Depression After TBI We assessed health status with a subjective rating of overall pain, general health, and extent to which physical symptoms post-tbi affected functioning within the community. Findings are summarized in table 4 and described below. Ratings of pain. The groups reported comparable levels of perceived pain at T1; at T2, they differed significantly (F 1,3 4.37, P.01). Post hoc comparisons revealed that the no-depression and resolved-depression (P.01) groups reported significantly less pain than did the chronic-depression group. Interaction effects on the repeated-measures ANOVAs were not significant; therefore, the overall rating of pain within groups was stable over time. Ratings of health. The groups differed significantly in health ratings both at T1 (F 1,3 4.89, P.01) and at T2 (F 1,3 3.80, P.01). At T1, the no-depression and the resolveddepression groups rated their health as being significantly better than the chronic-depression group (P.01). At T2, the no-depression group continued to rate their health as significantly better than the chronic-depression group (P.05). Interaction effects on the repeated-measures ANOVAs were not significant; therefore, overall rating of health within groups was stable over time. Physical symptoms. We computed MANOVAs for the physical scale of the BISQ. The groups differed significantly in their report of the extent to which physical symptoms affected functioning at T1 (F 1, , P.001) and at T2 (F 1, , P.001). Post hoc analyses revealed that the no-depression group reported significantly fewer TBI physical symptoms than did the other groups at T1 (all contrasts, P.001) and at T2 (all contrasts, P.001). Interaction effects on the repeated-measures MANOVAs were not significant, indicating that patterns of reporting physical symptoms within groups were stable over time. The Relationship of Pre-TBI Psychiatric Disorders to Patterns of Depression and Psychosocial Functioning Post-TBI Given the unexpected elevated rates of psychiatric disorders before TBI across the 4 groups (table 2), we did additional analyses to examine the relationship of preexisting psychiatric diagnoses to demographic, injury-related, and psychosocial variables. Individuals with pre-tbi psychiatric disorders were equivalent to those without such disorders in terms of age, ethnicity, income, education, marital status, years post-tbi, and severity of TBI. Similarly, individuals with psychiatric disorders before TBI reported levels of psychosocial functioning (ie, QOL, unmet important needs, cognitive and emotional symptoms) and health status (ie, overall health, pain, physical symptoms) that were equivalent to those of individuals without a premorbid history. Therefore, pre-tbi psychiatric disorders were not predictive of either depression or the level of psychosocial functioning after TBI. The Relationship of Post-TBI Co-Occurring Psychiatric Disorders to Patterns of Depression and Psychosocial Functioning High rates of post-tbi co-occurring psychiatric disorders were noted in all 4 groups at T1 and T2 (table 2). Further analyses were undertaken to examine the relationship of these co-occurring psychiatric diagnoses post-tbi on psychosocial functioning by combining the relatively small subsamples of individuals in the chronic-depressed and late-onset depression groups. Chi-square tests, recomputed for categorical variables, revealed no significant differences within groups in terms of demographic variables (ie, marital status, gender, education, ethnicity, income). Recomputation of t tests for continuous variables revealed significant within-group differences at T1 for age (t 4.65, P.001), years postinjury (t 2.26, P.05),

7 DEPRESSION AND PSYCHOSOCIAL FUNCTIONING AFTER TBI, Hibbard S49 Fig 2. QOL across 4 patterns of depression after TBI. pain (t 3.36, P.001), and QOL (t 2.99, P.01), and at T2 for unmet important needs (t 2.01, P.05) and QOL (t 2.90, P.05). One-way MANOVAs revealed significant within-group differences for TBI-related physical symptoms at T1 (t 4.64 P.001) and T2 (t 3.39, P.001), cognitive symptoms at T1 (t 3.65, P.001) and T2 (t 3.35, P.001), and behavioral symptoms at T1 (t 4.8, P.001) and T2 (t 4.10, P.001). Post hoc analyses revealed significant differences within the no-depression group for pain at T1, and for QOL, physical, cognitive, and emotional symptoms at T1 and T2, and the resolved depression group for age and cognitive and behavioral symptoms at T1 and T2 (all contrasts, P.01). Because of small sample sizes in the resolved-depression group, further analyses of these differences were not done. Given the high frequency of co-occurring psychiatric disorders post-tbi within the no-depression group (table 2), we split the group into 2 subgroups: no psychiatric disorder post-tbi (n 43, 23% of total sample) and no depression but another psychiatric disorder post-tbi (n 48, 26% of total sample). Data were reanalyzed comparing the 5 groups: no psychiatric Table 4: Health Indicators Across 4 Patterns of Depression After TBI Variable No Depression Resolved Depression Late-Onset Depression Chronic Depression ANOVA Overall pain ratings Time 1 (n 91) (n 51) (n 19) (n 27) Mean SD Time 2 (n 90) (n 49) (n 19) (n 26) Mean SD F 1, Health ratings Time 1 (n 91) (n 51) (n 19) (n 27) Mean SD F 1, Time 2 (n 90) (n 49) (n 19) (n 26) Mean SD F 1, Physical symptoms Time 1 (n 90) (n 51) (n 19) (n 27) Mean SD F 1, * Time 2 (n 89) (n 47) (n 18) (n 24) Mean SD F 1, * *P.001. P.01.

8 S50 DEPRESSION AND PSYCHOSOCIAL FUNCTIONING AFTER TBI, Hibbard Variable Table 5: QOL Indicators Across 5 Patterns of Psychiatric Disorders After TBI No Disorder Another Psychiatric Disorder Post-TBI Resolve Depression Late-Onset Depression Chronic Depression ANOVA Physical scale Time 1 (n 42) (n 48) (n 51) (n 19) (n 27) Mean SD F 1, * Time 2 (n 42) (n 47) (n 47) (n 18) (n 24) Mean SD F 1, * Cognitive scale Time 1 (n 42) (n 48) (n 51) (n 19) (n 27) Mean SD F 1, * Time 2 (n 42) (n 47) (n 47) (n 18) (n 24) Mean SD F 1, * Behavioral scale Time 1 (n 42) (n 48) (n 51) (n 19) (n 27) Mean SD F 1, * Time 2 (n 42) (n 47) (n 47) (n 18) (n 24) Mean SD F 1, * Overall pain rating T1 (n 43) (n 48) (n 51) (n 19) (n 27) Mean SD F 1, T2: (n 42) (n 48) (n 49) (n 19) (n 26) Mean SD Life-3 T1 (n 43) (n 48) (n 50) (n 19) (n 27) Mean SD F 1, * T2 (n 42) (n 47) (n 47) (n 18) (n 24) Mean SD F 1,4 8.11* *P.001. P.01. disorder, no depression but another psychiatric disorder, resolved depression, late-onset depression, and chronic depression. Means and SDs for significant variables in the post hoc analysis were contrasted for the 5 groups (table 5). One-way ANOVAs revealed significant differences for pain at T1 and for physical, cognitive, and behavioral symptoms and QOL, at both T1 and T2 (all contrasts, P.01). Post hoc comparisons revealed that the no psychiatric disorder group reported significantly less pain at T1 and less cognitive, physical, and behavioral symptoms and better QOL at both T1 and T2, when compared with the group with another psychiatric disorder post-tbi (all contrasts, P.01). DISCUSSION Our primary goals in this longitudinal study were to determine if there were patterns of emergence, resolution, and chronicity of depression in individuals up to 5 years post-tbi and to examine the relationship of identified patterns to perceived psychosocial functioning over time. Our findings suggest that differing patterns of depression post-tbi exist and that psychosocial functioning varies in a logical fashion with these patterns. Forty-eight percent of our sample did not meet DSM-IV criteria for a major depression at either that initial assessment or the repeat assessment 1 year later. These individuals reported significantly better psychosocial functioning than did those individuals who developed depression after TBI. Subsequent analyses were made within this group because of the high frequency of psychiatric disorders (other than depression) noted at both assessment intervals. Individuals without psychiatric diagnoses post-tbi reported better psychosocial functioning and health than did those who had a psychiatric disorder other than depression post-tbi, thus suggesting the toll caused by any psychiatric diagnoses on daily functioning after TBI. Among persons who met DSM-IV criteria for major depression after TBI, 29% had resolution of their depression, a decline in self-reported depressive symptoms, and improvements in psychosocial functioning at reassessment. An additional 10% had a late onset of depression (at reassessment) accompanied by an increase in depressive symptoms and a decline in psychosocial functioning. The remaining 29% were depressed at both initial and repeat assessments, with stable but significantly lower levels of psychosocial functioning at reassessment and a further decline in QOL over time. The patterns of post-tbi depression we identified and the direction of perceived changes in psychosocial functioning that accompanied these patterns argue for the validity of self-report in individuals with TBI and the sensitivity of the SCID in categorizing psychiatric challenges after TBI. Our findings provide a context in which to understand the often confusing data from earlier research, that is, declines in depression during the initial year post-tbi, with increasing prevalence in the ensuing years ,16-21 The early declines may well reflect the experience of individuals whose depression has resolved, whereas the later increases in depression may well reflect the experiences of individuals who have either late-onset depression or who remain chronically depressed after TBI. Our findings highlight the changing nature of depression post-tbi and the need for timely interventions to avert late-onset depressions while addressing the psychologic needs of those who remain chronically depressed. In this study, neither demographic variables (ie, age, gender, education, race, income) nor TBI factors (ie, injury severity, years postinjury) were predictive of the onset or the nature of

9 DEPRESSION AND PSYCHOSOCIAL FUNCTIONING AFTER TBI, Hibbard S51 post-tbi depression. Our findings add to those of earlier studies 11,12,14 and highlight the need to routinely screen individuals post-tbi for onset of depression. Although pre-tbi psychiatric diagnoses have been reported as predictive of post-tbi depression, 11,16,21,24 these factors were not predictive in our study. Indeed, although 17% of our total sample presented with a pre-tbi mood disorder and subsequently developed post-tbi major depression, 21% with pre-tbi depression did not experience major depression after TBI. Post-TBI co-occurring psychiatric diagnoses are common after TBI. 11,12,21 These diagnoses were common in our study, with the predominant co-occurring diagnoses post TBI being anxiety disorders. Individuals who experienced depression after TBI had the greatest frequency of co-occurring psychiatric disorders. At reassessment, a further increase in such diagnoses was noted, in the majority of such individuals. Furthermore, other psychiatric disorders post-tbi in a small cohort of individuals who did not become depressed were also related to declines in psychosocial functioning. These findings point to the need for a broad-based assessment of psychiatric challenges post-tbi, including screening for anxiety and substance abuse disorders. Our study provides insights into the markedly different life experiences of individuals up to 5 years post-tbi, in terms of their mood and their overall psychosocial functioning in the community. Our findings show that depression can increase or decrease over time. The results validate previous findings that documented the negative toll that post-tbi depression takes on perceptions of health, 10 reports of depressive symptoms, 17,22 decline in social roles and responsibilities, 22,26,35 and QOL. 36,37 We found that patterns of depression were related to reported levels of pain after TBI. Although pain and suffering have been associated with chronic depression in other rehabilitation patient groups, 54,55 they have not been a traditional area of focus in the study of depression after TBI. Based on current findings, further investigation of post-tbi pain is indicated given that pain may serve as either a stimulus or maintainer of depressed mood post-tbi. The high incidence of depression found in this sample makes it equally important to recognize that a small cohort of individuals experienced neither depression nor other psychiatric disabilities after TBI. It is not surprising that they reported having significantly fewer depressive symptoms and better psychosocial functioning than their peers and appeared to be better adjusted after their injury. Future exploration of factors that buffer these individuals from developing depression or other psychiatric illness is clearly indicated. Research directed at exploring within person variables, such as coping mechanisms, 56 degree of discrepancy between pre- and post-tbi selves, 57 as well as contextual factors (ie, community, family and environmental supports), that may help to enhance the resiliency 58 seen in these individuals are suggested. Such research can lead to valuable insights that can be used to minimize the impact of psychiatric challenges for many people after TBI. Major depressive episodes are often time limited. 13 In our sample, 29% of individuals experienced a resolution of their depression, which suggests that for many people who are post-tbi, major depression is an episodic event, and, perhaps, a normal phase in the overall adjustment process after TBI. Resolution was accompanied by improvement in psychosocial functioning and perceived enhancement of one s health. In this study, we did not explore contextual and interpersonal factors or interventions that might facilitate resolution during this early phase of adjustment. These issues clearly deserve study. Our study highlights the risk of late-onset depression post- TBI. Individuals who had late-onset depression presented with increased depressive symptoms at the initial assessment but did not meet DSM-IV criteria for major depression at that time. It can be hypothesized that these individuals presented with a subclinical depression. Additionally, this group had a high prevalence of post-tbi co-occurring psychiatric disorders at both initial (74% of group) and at follow-up (84% of group) evaluations. Combined, these findings suggest a profile for an individual at heightened risk of delayed-onset depression after TBI. Such individuals require frequent monitoring of their mood and proactive psychologic interventions to prevent development of major depression. A less sanguine finding was the frequency (14%) of chronic depression in the sample, a frequency clearly greater than the estimated lifetime prevalence rate of 6% reported in community-based samples. 2 Individuals with chronic depression had significant pre- and post-tbi co-occurring psychiatric disorders, severe depressive symptoms, significantly lower psychosocial functioning, and a worsening of their QOL at reassessment. Clearly, they require intensive treatment to counter the chronicity of their mood disorder, as well as their co-occurring psychiatric disorders. Given the reported declines in health in this group, clinicians must broaden their assessment and intervention approaches to include exploration of health concerns because these coexisting health challenges may be contributing to the chronicity of the depression. Limitations We examined the relationship between patterns of depression and self-reported psychosocial functioning. No attempts were made to equate causality between the patterns observed and self-reported levels of psychosocial functioning. It could be argued that sole reliance on self-report of psychosocial functioning, especially in depressed individuals, might better reflect their perceptions, rather than the reality, of their psychosocial functioning. Our study lacked external validation of self-reported mood or psychosocial functioning; consequently, the lack of external validation limits interpretation of our findings. Similarly, the study could be criticized for its sole reliance on self-report of the severity of TBI. In communitybased research, external documentation (eg, computed tomography scans, magnetic resonance imaging, medical records) is typically not available. Despite the lack of such documentation, our participants were able to report estimated durations of their altered mental state after TBI. Therefore, our sample, while not representative of a population-based study, provides a broad picture of psychosocial adjustment in individuals with a full continuum of TBI severity. It could also be argued that the high prevalence of depression we found is reflective of a selection bias or convenience sample effect. Indeed, participants were recruited from a larger study that focused on QOL and community integration. However, the finding that 48% of our participants were not depressed would argue against an obvious selection bias. Clearly, external validation of reported TBI severity, observed mood, and observed psychosocial functioning could only strengthen our findings and should be considered in future research. The study was limited to a 1-year follow-up. Although this time interval is viewed as of adequate duration over which to assess an episode of major depression, 13 the patterns of post- TBI depression we identified may well continue to shift over time. For example, individuals whose depression resolved at reassessment might develop a repeat episode of major depression if they were followed for a longer time. Conversely, individuals in the late-onset depression group may remit after

10 S52 DEPRESSION AND PSYCHOSOCIAL FUNCTIONING AFTER TBI, Hibbard a period of time and/or they may become chronically depressed. Given the study s limited follow-up, our findings must be viewed as preliminary. We will reassess the individuals in this study; this third assessment will provide a longer timespan over which to examine the stability of identified patterns. In this study, individuals were initially assessed at varying points post-tbi, but within a 4-year period, and then assessed 1 year later for stability, resolution, or chronicity of mood disorder. This methodologic design can be viewed as a potential limitation of the study because it could be argued that select time intervals for example, within the first year post-tbi might show the greatest within-person changes in mood. Segmenting groups by discrete time intervals post-tbi was beyond the scope of this study but should be considered when planning future research efforts. It could be argued that the presence of postconcussion symptoms (eg, difficulty with concentration, fatigue) may have inflated the prevalence rates of depression. Although we did not specifically explore this issue, reliance on DSM-IV criteria for diagnosis and the moderate-to severe levels of self-reported depression on the BDI identified in the 3 depression groups (chronic, late onset, resolution) argue that depression, not postconcussion symptoms, accounted for the majority of documented negative psychosocial impacts. Future research should consider teasing apart this TBI-relevant issue. A final limitation was our failure to track interventions provided to participants before or between assessments. These interventions could have included the use of antidepressants and psychotherapy, which can clearly affect mood and psychosocial outcomes. Other rehabilitation interventions for example, cognitive remediation, physical therapies, and pain management could have affected psychosocial outcomes as well. Conversely, barriers to community integration (eg, lack of available services financial or insurance limitations in procuring needed services, lack of transportation) were not assessed. These barriers may have served to prolong depression in some individuals while precipitating it in others. Future research efforts should incorporate these broader contextual and environmental enhancers and barriers when assessing outcomes after TBI. CONCLUSIONS This study adds to the literature that documents increased prevalence of both major depression and other psychiatric disorders after TBI ,16-18,21-25 In addition, it is the first study to document distinct patterns of post-tbi depression and to relate these patterns to psychosocial functioning. Thus, it provides an initial look at the fabric of life within the first 5 years after TBI and the often devastating impact that depression has on a person s perceptions of his/her overall functioning after injury. A strength of the study is the large sample of individuals who participated. They presented with a full continuum of self-reported TBI severity and were recruited from throughout the United States. Patterns of emergence, resolution, and chronicity of depression identified in the study; the impact of mood disturbances on psychosocial functioning; and the lack of predictive value of a prior mood or psychiatric disorder in determining who will become depressed post-tbi, highlight the need for routine screening for mood disturbances many years after a TBI. Elevated rates of co-occurring psychiatric disorders after TBI argue for a broadening of screening to include assessment of potential anxiety disorders and substance abuse challenges. The findings point to the importance of early and intensive psychologic and rehabilitative interventions if health care professionals are to have a positive impact on the lives of people after TBI. Finally, future longitudinal research is indicated to replicate patterns of depression identified in this initial study and to determine whether these patterns remain stable over time. Acknowledgment: We thank Wayne A. Gordon, PhD, for his careful editing of this manuscript. References 1. Consensus conference. Rehabilitation of persons with traumatic brain injury. NIH Consensus Development Panel on Rehabilitation of Persons With Traumatic Brain Injury [see comments]. JAMA 1999;282: Bourdon KH, Rae DS, Locke BZ, Narrow WE, Regier DA. Estimating the prevalence of mental disorders in U.S. adults for the Epidemiologic Catchment Area Survey. Public Health Prev 1992;107: Brooks N, Campsie L, Symington C, Beattie A, McKinlay W. The effect of head injury on patients and relatives within seven years of injury. J Head Trauma Rehabil 1987;2(2): Brooks D, McKinlay W. Personality and behavioral change after severe blunt head injury: a relative s view. J Neurol Neurosurg Psychiatry 1983;46: Leach LR, Frank RG, Bouman DE, Farmer J. Family functioning, social support and depression after traumatic brain injury. Brain Inj 1994;9: Rutherford WH, Merrett JD, McDonald JR. Sequelae of concussion caused by minor head injuries. Lancet 1977;1: Schoenhuber R, Gentilini M. Anxiety and depression after mild head injury: a case control study. J Neurol Neurosurg Psychiatry 1988;51: Van Zomeren A, Van den Berg W. Residual complaints of patients two years after severe head injury. J Neurol Neurosurg Psychiatry 1985;48: Varney NR, Martzke JS, Roberts RJ. Major depression in patients with closed head injury. Neuropsychology 1987;1: Fann JR, Katon WJ, Uomoto JM, Esselman PC. Psychiatric disorders and functional disability in outpatients with traumatic brain injuries. Am J Psychiatry 1995;152: Van Reekum R, Bolago I, Finlayson MA, Garner S, Links PS. Psychiatric disorders after traumatic brain injury. Brain Inj 1996; 10: Hibbard MR, Uysal S, Kepler K, Bogdany J, Silver J. Axis I psychopathology in individuals with traumatic brain injury. J Head Trauma Rehabil 1998;13(4): First M, Spitzer R, Gibbon M, Williams J. User s guide for the Structured Clinical Interview for DSM-IV Axis I Disorders SCID. Washington (DC): Am Psychiatric Pr; Seel RT, Kreutzer JS, Rosenthal M, Hammond FM, Corrigan JD, Black K. Depression after traumatic brain injury: a National Institute on Disability and Rehabilitation Research Model System multicenter investigation. Arch Phys Med Rehabil 2003;84: Kreutzer JS, Seel RT, Gourley E. The prevalence and symptom rates of depression after traumatic brain injury: a comprehensive examination. Brain Inj 2001;15: Deb S, Lyons I, Koutzoukis C, Ali I, McCarthy G. Rate of psychiatric illness 1 year after traumatic brain injury. Am J Psychiatry 1999;156: Jorge RE, Robinson RG, Starkstein SE, Arndt SV. Influence of major depression on 1 year outcome in patients with traumatic brain injury. J Neurosurg 1994;81: Jorge RE, Robinson RG, Arndt SV, Starkstein SE, Forrester AW, Geisler F. Depression after traumatic brain injury: a one year longitudinal study. J Affect Disord 1993;27: Corrigan JD, Smith-Knapp K, Granger CV. Outcomes in the first 5 years after traumatic brain injury. Arch Phys Med Rehabil 1998;79: Kreutzer JS, Witol AD, Sander AM, Cifu DX, Marwitz JH, Delmonica R. A prospective longitudinal multicenter analysis of alcohol use patterns among persons with traumatic brain injury. J Head Trauma Rehabil 1996;11(5):58-69.