Relationships among postconcussional-type symptoms, depression, and anxiety in neurologically normal young adults and victims of mild brain injury $

Transcription

1 Archives of Clinical Neuropsychology ) 435±445 Relationships among postconcussional-type symptoms, depression, and anxiety in neurologically normal young adults and victims of mild brain injury $ Abstract Donald E. Trahan*, Christina E. Ross, Shirley L. Trahan Center for Behavioral Studies, 3560 Delaware, Suite 105, Beaumont, TX 77706, USA Accepted 30 December 1999 This study investigated the relationship between self-reports of postconcussional symptoms, depression, and anxiety in neurologically normal young adults and recovered victims of mild head injuries MHI). The participants were 496 young adults with no history of MHI or depression, 56 neurologically normal individuals with clinical depression, and 40 people with history of MHI. All completed the Beaumont Postconcussional Index BPCI), Beck Depression Inventory-II BDI-II), and the Beck Anxiety Inventory BAI). Groups were compared on frequency and severity of postconcussional symptoms, as well as general symptoms. Analysis revealed high correlations between scores on the Postconcussional Index PCI) and the BDI-II r = 0.68) as well as between PCI and BAI r = 0.64). Correlations between BDI-II, BAI, and the General Symptom Index GSI) were modest, but significant r = 0.44 and 0.48, respectively). MHI participants reported minimally higher scores on the PCI than the normative group. However, depressed individuals exhibited substantially higher endorsement of PCI symptoms and modestly higher endorsement of GSI symptoms than either the normative or MHI groups. The potential role that depression can have in producing, exacerbating, and maintaining PCS-like symptoms must be considered when evaluating and treating victims of MHI. D 2001 National Academy of Neuropsychology. Published by Elsevier Science Ltd. Keywords: Postconcussional symptoms; Depression; Anxiety; Mild head injuries; Postconcussional syndrome Postconcussional syndrome PCS) is a fairly common occurrence in victims of head injuries that are sufficient to produce a loss of consciousness exceeding 5 min or post- $ Portions of this article were presented at the 26th Annual Meeting of the International Neuropsychological Society, Honolulu, HI, February * Corresponding author. Tel.: ; fax: /01/$ ± see front matter D 2001 National Academy of Neuropsychology. PII: S )

2 436 D.E. Trahan et al. / Archives of Clinical Neuropsychology ) 435±445 traumatic amnesia exceeding 24 h Levin et al., 1987; Rutherford, 1989; Rizzo & Tranel, 1996). During the early stages of recovery, victims of head trauma may report a variety of symptoms including headaches, dizziness, nausea, vomiting, drowsiness, blurred vision, diplopia, irritability, reduced tolerance for frustration, anxiety, moodiness or depression, poor concentration, forgetfulness, insomnia, fatigue, poor hearing, hypersensitivity to noise, appetite changes, decreased coordination, and difficulties thinking clearly. In some cases, subjective symptoms may persist for years. This fact has stimulated considerable debate as to whether these prolonged symptoms, or the report of such, are due to subtle neurological effects of the injury or whether they are related more to factors such as litigation issues, including secondary gain, and emotional stress associated with life changes, including the development of depression. One problem is that many of these symptoms are also reported by adults with no history of head trauma or other neurological disease. This seems particularly true for individuals who are engaged in litigation or suffering from depression. Lees-Haley and Brown 1993) examined base rates of reporting of 27 common PCSsymptoms in samples of 170 individuals seeking claims for various types of job-related emotional stress and 50 medical control patients without history of head trauma. These authors found that despite the absence of concussion histories, claimants nevertheless reported significantly higher rates of endorsement than control patients for 19 of 27 PCSsymptoms. In a follow-up study, Fox et al. 1995) examined base rates of PCSsymptoms in 400 individuals seeking psychotherapy, samples of patients from departments of neurology, family practice, and internal medicine, and 296 people undergoing initial screening upon entry to a health maintenance organization. The latter subjects were considered to be a control group since they had no immediate health problems. Results revealed that individuals seeking psychotherapy reported significantly more PCSsymptoms than controls and internal medicine patients. The rate of endorsement by individuals seeking psychotherapy was no different from that of patients in the neurological group who were being evaluated or treated for actual neurological complaints or from those in the family practice group who were being treated for various medical illnesses. Gouvier et al. 1992) examined base rates of reporting of PCSsymptoms in college students with and without history of head trauma. Results from this study revealed no significant differences in the rates of PCSsymptoms between students with no history of head trauma and those who had suffered head injuries that had resolved to the point to allow a return to school. Results did reveal that the frequency, intensity, and duration of symptoms were correlated positively with a measure of daily stress. Sawchyn et al. 1998) also examined rates of PCSsymptom endorsement in college students with and without history of head trauma sufficient to cause loss of consciousness. Results of this study revealed no differences between students without history of trauma and those reporting head injuries. However, the rate of PCSsymptom reporting did correlate significantly r = 0.55) with scores on the Beck Depression Inventory Beck & Steer, 1987). These studies have certainly contributed to our knowledge of base rates of PCSsymptomatology in neurologically normal adults and the relationship between those symptoms and the presence of other psychological and medical conditions. Nevertheless, some problems remain. First of all, a number of studies have used checklists that simply ask respondents to indicate ``yes'' or ``no'' with regard to whether they have experienced certain symptoms Oddy et al.,

3 D.E. Trahan et al. / Archives of Clinical Neuropsychology ) 435± ). Checklists of this type do not distinguish between respondents experiencing only rare symptom occurrence e.g., less than once per month) and those having a high rate of symptoms e.g., daily or more). Secondly, such checklists do not request information on duration or severity of symptoms, which can vary considerably from one person to another. Gouvier et al. 1992) attempted to address this problem by developing the Postconcussion Syndrome Checklist. This checklist has respondents rate the frequency, intensity, and duration of nine common symptoms associated with PCS, including headaches, dizziness, irritability, memory problems, difficulty concentrating, visual disturbance, aggravation by noise, judgment problems, and anxiety. While this measure represents a substantial improvement over previous checklists, it does not include a number of other symptoms commonly associated with PCS, including drowsiness, insomnia, fatigue, moodiness/depression, poor tolerance for frustration, nausea/vomiting, appetite disturbance, and decreased coordination. Trahan et al. 1997) addressed these limitations by constructing the Beaumont Postconcussional Index BPCI), which examines the reported frequency, intensity or severity, and duration of a wide range of PCSsymptoms including some not surveyed by previous measures. Normative studies using the BPCI have revealed that neurologically normal young adults generally report a low incidence of PCSsymptoms, with 59% reporting frequencies of ``never'' or ``almost never'' for 17 of the symptoms surveyed. Another 31% reported frequencies of only once per month to once per week. Average ratings of severity also were quite low Trahan & Ross, 1998). The question remains, however, as to the relationship between PCSsymptomatology, as measured by the BPCI, and the presence of psychological conditions such as depression and anxiety. The answer to this question is crucial clinically in formulating effective treatment plans. However, there also are important implications for forensic neuropsychology. Victims of even minor head trauma commonly report a wide array of physical, cognitive, and emotional symptoms, which they frequently attribute to effects of ``brain injury''. However, these individuals may be clinically depressed as well. Consequently, there is the critical question regarding the etiology of these reported symptoms. This study was designed to examine the relationship between PCSsymptomatology, depression, and anxiety in neurologically normal young adults and recovered victims of mild brain injuries. Based on results from previous studies, we hypothesized that there would be a modest relationship between BPCI scores and measures of depression and anxiety even among nondepressed neurologically normal subjects. We also predicted that BPCI scores of patients with clinical depression would be significantly higher than those of normal adults and at least as high as those of recovered victims of mild brain injuries. 1. Method 1.1. Participants The normative sample included 496 participants 143 male, 353 female) with no history of neurological conditions such as cerebrovascular accident, seizures, brain neoplasms, infectious disease involving the central nervous system, or head trauma with loss of consciousness.

4 438 D.E. Trahan et al. / Archives of Clinical Neuropsychology ) 435±445 Individuals also were excluded from this sample if they reported alcohol use exceeding two drinks per day, any current use of recreational drugs, clinical depression, or any other major psychiatric illness. Participants ranged in age from 18 to 39 years M = 22.01, SD = 4.74). The racial composition of the sample was 74% Caucasian, 15% African American, and 11% other primarily Hispanic and Asian). The educational level of participants was as follows: 1) 6±11 years of education 1%), 2) 12±15 years of education 90%), and 3) 16 or more years of education 9%). These participants were drawn primarily from two settings: undergraduate classes at a mid-sized public university in Southeast Texas and the general population of a mid-sized urban community in Southeast Texas. All participants were recruited by personal contacts the authors had with local agencies. A second sample included 56 participants 4 male, 52 female) with clinical depression. Individuals were included in the sample if they reported having a history of clinical depression and had BDI scores of 20 or greater. This dual criterion was used to exclude individuals with histories of depression who had recovered to the point that they were no longer reporting significant depressive symptomatology. These participants also were screened in the same manner as those in the normative sample. Those with histories of neurological disease, head trauma with loss of consciousness, alcohol abuse, or recreational drug use were excluded from the study. These participants ranged in age from 18 to 39 years M = 21.64, SD = 4.26). The racial composition of the sample was 65% Caucasian, 33% African American, and 2% other. The educational level of these participants was as follows: 1) 12±15 years of education 95%) and 2) 16 years or more of education 5%). These participants were drawn from the same settings as the normative sample. A third sample mild head injuries, MHI) included 40 participants 20 male, 20 female) with histories of MHI. Participants in this sample reported head trauma resulting in loss of consciousness of less than 30 min, with no neurological complications. All received some initial medical attention, but none required prolonged hospitalization due to the MHI. Individuals with acute injuries i.e., occurring within the previous 12 months) were excluded from the study. All participants had recovered sufficiently to allow a return to full-time student status in college or to regular employment. Information regarding the exact time interval between the injury and this study was not available. Subjects with histories of depression, neurological disease, alcohol abuse, or recreational drug use were not included in this sample. These participants ranged in age from 18 to 40 years M = 21.98, SD = 5.31). The racial composition of the sample was 78% Caucasian, 11% African American, and 11% other. The educational level of these participants was as follows: 1) 6±11 years of education 10%), 2) 12±15 years of education 85%), and 3) 16 or more years of education 5%). Analysis of variance ANOVA) was used to determine whether there were group differences in either age or education. Analysis of data revealed no differences for either age [F 2,589) = 0.15, p = 0.86, or education, F 2,589) = 0.02, p = 0.98] Materials The BPCI is a 55-item survey containing 33 questions pertaining to traditional postconcussional symptoms. The items are designed to measure frequency, intensity, or duration of these symptoms. Most of the items allow one to rate frequency or intensity on a 5-point

5 D.E. Trahan et al. / Archives of Clinical Neuropsychology ) 435± scale. The 33 items together form a composite score referred to as the Postconcussional Index PCI). The potential range of scores on the PCI is 27±157. The BPCI also contains 10 questions relating to frequency of other physical and cognitive symptoms that may be seen in association with more serious neurological injuries and other medical conditions, but only rarely with mild head trauma alone e.g., shortness of breath, chest pains, problems getting lost, problems with oral and reading comprehension, difficulties writing, etc.). These items together form a composite score referred to as the General Symptom Index GSI). Items on the GSI are rated on a 5-point scale ranging from 1 never or almost never) to 5 daily or more). Scores on the GSI may range from 10 to 50. Finally, the BPCI includes an 8-item Validity Index VI). The VI is composed of items that are endorsed in a given direction only rarely i.e., less than 15% of the time) by neurologically normal adults. The items of the VI assess ideas and opinions unrelated to head trauma e.g., ``I have developed a strong dislike for children and do not see why people think they are so special''). The purpose for the VI was to help determine if respondents were generally endorsing symptoms or ideas, even uncommon ones, unrelated to the physical and cognitive effects of head trauma. Each item of the VI is scored 0 or 1. VI scores may range from 0 to Procedure All participants were treated in accordance with ethical standards established by the American Psychological Association. Participants were given a brief description of the general purpose of the study. All were asked to read and sign a consent form. Only three subjects declined to participate after reading the informed consent. Participating respondents then completed a brief personal history questionnaire requesting information about age, gender, race, education, and medical history. Participants then completed the BPCI, Beck Depression Inventory-II BDI-II: Beck et al., 1996b) and Beck Anxiety Inventory BAI: Beck et al., 1996a). 2. Results Analysis of the relationships between gender, race, and BPCI variables was conducted using a 2 3 ANOVA, with gender being the first factor and race being the second factor. Separate analyses were conducted for the normative, head trauma, and depressed groups. For the normative group, gender had a significant effect on the PCI F = 17.19, p < 0.01). Females scored 3.5 points higher on the composite PCI. However, this difference is of questionable significance clinically and represents an average difference of only 0.11 point for each of the 33 items composing the index. Gender had no effect on either the GSI F = 0.35, p = 0.51) or VI F = 0.18, p = 0.59). Race had no effect on the PCI F = 0.22, p = 0.77), GSI F = 0.82, p = 0.51), or VI F = 2.53, p = 0.08). A second analysis was conducted using level of education as the second variable. These analyses revealed that education had no effect on the PCI F = 0.49, p = 0.81), GSI F = 1.60, p = 0.15), or VI F = 0.68, p = 0.67). It is important to note, however, that 90% of the participants fell within the same educational level i.e., 12±15 years). The lack of correlation may be due in part to restriction of range.

6 440 D.E. Trahan et al. / Archives of Clinical Neuropsychology ) 435±445 Analysis of data for the depressed group was conducted in the same manner. These analyses revealed that gender had no effect on the PCI F = 0.75, p = 0.53), GSI F = 0.63, p = 0.51), or VI F = 1.26, p = 0.27). However, these findings must be viewed with caution because there were only four males in this group. Similarly, race had no effect on the PCI F = 1.54, p = 0.23), GSI F = 0.46, p = 0.55), or VI F = 0.65, p = 0.50). ANOVA was not conducted for education because there was essentially no variability, with 95% of these individuals falling in the educational category 4 i.e., 12±15 years). Analysis of data for the head trauma group was conducted in the same manner. These results indicated that gender had a significant effect on the PCI F = 4.67, p = 0.04). Females scored about 10 points higher on the average than males. However, gender had no effect on either the GSI F = 2.66, p = 0.11), or VI F = 0.07, p = 0.74). Race had no effect on either the PCI F = 0.23, p = 0.76), GSI F = 0.12, p = 0.92), or VI F = 0.73, p = 0.50). A second analysis revealed that education had no affect on PCI F = 0.10, p = 0.98), GSI F = 0.47, p = 0.76), or VI F = 0.46, p = 0.76). The next analyses were conducted on the combined sample of 592 individuals composing the three groups. The purpose of combining these groups was to examine the relationship between PCI, BDI-II, and BAI scores in all adults, without regard to psychological or neurological status. Pearson product-moment correlations were calculated between the three BPCI indexes and both Beck scores. These correlations are presented in Table 1. All of these correlations were significant at the level. However, the percentage of variance accounted for by these correlations ranged from only 6% for the relationship between the General Symptom and Validity Indexes to 46% for the relationship between the PCI and the Beck Depression Inventory Score. Analyses of differences between the normative, MHI, and depressed groups on the BPCI were conducted using one-way ANOVAs. These analyses are presented in Table 2. As can be seen, significant group differences were observed for all variables. Post hoc analyses were conducted using the Newman±Keuls procedure alpha level = 0.05). These analyses revealed that MHI participants reported significantly higher BDI-II scores than normal subjects, although the mean difference was only 3.5 points. On the other hand, depressed subjects had substantially higher BDI-II scores than either of the other groups. This was expected, since an elevated BDI score was one of the criteria for inclusion in the depression group. Analysis of data for BAI revealed no difference between the normative and MHI groups. However, depressed individuals again reported significantly more symptoms than either of the other groups. Table 1 Correlations between BPCI, BDI-II, and BAI scores PCI GSI Validity BDI-II BAI PCI ± GSI ± ± Validity ± ± ± BDI-II ± ± ± ± 0.62 All correlations were significant at the level. BPCI = Beaumont Postconcussional Index; PCI = Postconcussional Index; GSI = General Symptom Index; BDI-II = Beck Depression Inventory-II; BAI = Beck Anxiety Inventory.

7 D.E. Trahan et al. / Archives of Clinical Neuropsychology ) 435± Table 2 Comparisons between BPCI, BDI-II, and BAI scores for normative, MHI, and depressed groups Normative group MHI group Depressed group Test index M SD RNG M SD RNG M SD RNG F p BPCI PCI ± ± ± <0.001 GSI ± ± ± <0.001 VI ± ± ± <0.001 BDI-II ± ± ± <0.001 BAI ± ± ± <0.001 RNG = Range 10th and 90th percentiles); BPCI = Beaumont Postconcussional Index; BDI-II = Beck Depression Inventory-II; BAI = Beck Anxiety Inventory; MHI = mild head injury; PCI = Postconcussional Index; GSI = General Symptom Index; VI = Validity Index. Analysis of data for the PCI revealed no differences between the normative and MHI groups, with mean scores varying by less than 3 points for the entire Index. On the other hand, depressed participants reported significantly higher scores, with the mean PCI score being almost 50% higher than that reported for either of the other groups. Analysis of data for the GSI revealed statistically significant differences between all these groups. However, mean differences were small i.e., for the normative vs for depressed). All groups actually reported a relatively low frequency of symptoms on this index. Analysis of data for the VI revealed a similar pattern, with depressed participants reporting significantly higher scores than either the control or MHI groups, which did not differ from one another. Mean scores for all these groups were low overall. However, 18% of depressed subjects reported scores of 3 or higher on the VI, where only 1% of the normative sample reported similarly elevated scores. In the next analysis, we compared the frequency of endorsement of individual items from the PCI for the three groups. Table 3 lists the percentage of respondents endorsing frequencies of twice per week or more for each symptom. A c 2 analysis was conducted to determine if the observed differences were statistically significant. Results revealed significant differences in reported frequencies for 12 of 17 symptoms. The c 2 statistic could not be meaningfully calculated for four symptoms i.e., vomiting, blurred vision, confusion, and incoordination) because of cell frequencies below 5 for more than one cell. A further analysis of the frequency tables reveals that MHI respondents reported higher than expected frequencies for three of these symptoms: dizziness, forgetfulness, and fatigue. On the other hand, depressed respondents reported higher than expected frequencies on all 12 of these symptoms. Table 4 presents a similar analysis for the nine items that are rated for severity. The numbers in the table reflect the percentage of respondents reporting a severity of 3 or higher on a scale of 1 to 5. Inspection of Table 4 reveals a very similar pattern of reporting for all three groups, with depressed respondents reporting the highest percentages for all nine symptoms surveyed. Unfortunately, low expected cell frequencies precluded statistical analysis for five items. However, for the analyses that were completed, depressed respondents reported higher than expected rates of endorsement for headaches, insomnia, irritability, and

8 442 D.E. Trahan et al. / Archives of Clinical Neuropsychology ) 435±445 Table 3 Percentage of subjects reporting symptoms twice per week or more Symptom Normative group MHI group Depressed group c 2 p Headache ns Dizziness <0.01 Nausea <0.01 Vomiting nc Drowsiness <0.01 Blurred vision nc Insomnia <0.01 Irritability <0.01 Anxiety <0.01 Depression <0.01 Frustration <0.01 Forgetfulness <0.01 Poor concentration <0.01 Fatigue <0.01 Appetite change <0.01 Confusion nc Incoordination nc ns = not significant; nc = not calculated. MHI = mild head injury. fatigue. Patterns of endorsement for the MHI respondents were similar to the normative control group for these nine items. Similar comparisons were made for normative and depressed groups for individual items on the GSI and VI. For the GSI items, very few normative subjects reported frequencies of twice per week or greater, with the average rate of endorsement being only 3.8% per item. Respondents in the MHI group reported only slightly higher rates of endorsement for GSI items, with the average rate of endorsement per item being 7.8%. Even more individuals in the depressed group reported frequencies of twice per week or Table 4 Percentage of subjects reporting moderate-to-severe symptomatology Symptom Normative group % MHI group % Depressed group % c 2 p Headache <0.01 Dizziness nc Blurred vision nc Insomnia <0.01 Irritability <0.01 a Forgetfulness nc Fatigue <0.01 Confusion nc Incoordination nc nc = not calculated; MHI = mild head injury. a Distribution of scores contains one cell with expected frequency of 4.

9 D.E. Trahan et al. / Archives of Clinical Neuropsychology ) 435± greater, with the average rate of endorsement being 15.2% per item. The percentage of subjects reporting frequent symptoms was numerically higher in the depressed group for all 10 items. However, because both groups reported relatively low rates of endorsement, c 2 analysis could not be meaningfully conducted for 8 of 10 symptoms due to expected all frequencies less than 5. The two analyses that were conducted revealed that depressed subjects reported higher than expected rates for shortness of breath p < 0.01) and trouble remembering names p < 0.01). Analysis of data for the VI revealed an average rate of endorsement of 9.2% per item for the normative group, 11.8% for the MHI group, and 25.5% for the depressed group. c 2 analyses revealed significant differences p < 0.05) for five of eight items, with depressed individuals reporting higher than expected frequencies on all five items and MHI subjects reporting higher frequencies on two items. c 2 could not be conducted for three items that had very low rates of endorsement for all groups resulting in expected cell frequencies less than Discussion This study examined the relationship between self-reports of PCSsymptomatology, depression, and anxiety in a sample of 496 normal adults, 56 neurologically normal depressed young adults, and 40 recovered victims of MHI. Results revealed moderate to high correlations between the PCI, which is a composite index of traditional PCSsymptoms, and Beck Depression and Anxiety Inventory scores. This finding is consistent with previous reports in the literature Sawchyn et al., 1998). Additionally, individuals with clinical depression scored significantly higher on the PCI than either normal adults or MHI victims, while differences between normal and recovered MHI participants were either minimal or nonsignificant altogether. A logical question, however, is whether this difference is because depression causes individuals to perceive themselves as having more symptoms, or whether the PCI score is higher in depressed patients simply because PCSand depression are characterized clinically by some of the same symptoms. Certainly, PCSand depression are characterized by a number of common symptoms including, among others, moodiness, fatigue, irritability, poor tolerance for frustration, insomnia, poor concentration, forgetfulness, and reduced cognitive efficiency. Consequently, it is possible that the positive correlations observed between PCI and depression scores may be due in part to these shared symptoms. However, this would not explain why depressed subjects who had no history of head trauma or any other major health problems would also report higher frequencies of endorsement for all 10 symptoms from the GSI, which are not associated with the syndrome of clinical depression. On the GSI, depressed subjects reported higher rates of endorsement for such symptoms as problems getting lost, reduced reading comprehension, difficulty comprehending what other people say, slurred speech, difficulty writing, chest pains, and shortness of breath. Furthermore, the fact that clinical depression and PCSmay share some symptoms does not explain theoretically why depressed individuals, with no history of head trauma or other neurological disease, would report a substantially higher rate of PCSsymptoms than victims of MHI who actually suffered brain injuries.

10 444 D.E. Trahan et al. / Archives of Clinical Neuropsychology ) 435±445 Depressed individuals also endorsed items on the VI at two to three times the rate observed in normative subjects and twice the rate of MHI subjects. A closer look at the VI items, however, reveals that although all items from the VI were endorsed at a very low rate by MHI and normative subjects, the five items endorsed more frequently by depressed individuals tended to have a common thread or theme that might be described as ``negative outlook.'' For example, depressed individuals more commonly reported negative attitudes toward their mothers and fathers, preoccupation with thoughts of death, and a general negativism toward authority and society. Thus, the higher scores of depressed individuals on the VI may have been due, in part, to general pessimism, which is common with depression, and to dissatisfaction with family than to intentional over-endorsement of these symptoms. Results from this study, along with those reported in previous investigations, have a number of implications for those working with patients in both clinical and legal settings. First of all, depression often coexists with other clinical conditions, including head trauma. Individuals who are depressed, even those with no established major health problems, often report more frequent and more severe symptoms than those who are not depressed. This appears true for a wide range of symptoms, including some that are not related in any logical fashion to the syndrome of depression as traditionally defined. Depressed individuals often perceive themselves as being more disabled physically, cognitively, and emotionally than their nondepressed counterparts. Thus, differential diagnosis and treatment of individuals with known health problems should take into consideration the potential role that depression can have in producing, exacerbating, and maintaining such symptoms. This is particularly true when treating victims of head trauma who at times report prolonged PCSsymptoms that seem out of proportion to what is known about the nature and severity of the injury. Failure to consider depression as a possible factor in the prolongation of symptoms may lead to erroneous conclusions regarding the etiology of such symptoms and to incomplete or ineffective treatment plans. A few additional words of caution are in order. First, future studies with the BPCI and other similar instruments need to investigate these same issues in groups of subjects with moderate and severe brain injuries. More severe injuries certainly have the potential to produce more and longer lasting symptoms. Consequently, results from this study may not generalize completely to these other groups. Second, both the normal and depressed groups in this study were composed predominantly of females, while most of the head trauma victims were male. This potentially could produce some problems with generalizability of our results. However, the absence of clinically significant gender differences on BPCI variables would likely minimize this problem. Finally, future studies with the BPCI or other similar instruments need to examine whether a specific group of symptoms may help to distinguish reliably between depressed subjects and individuals experiencing actual postconcussional symptoms secondary to head injury. If such a group can be found, it would facilitate our ability to differentially diagnose these conditions. References Beck, A. T., & Steer, R. A. 1987). Manual for the Revised Beck Depression Inventory. San Antonio, TX: The Psychological Corporation.

11 D.E. Trahan et al. / Archives of Clinical Neuropsychology ) 435± Beck, A. T., Steer, R. A., & Brown, G. 1996). Beck Anxiety Inventory: Manual. San Antonio, TX: The Psychological Corporation. Beck, A. T., Steer, R. A., & Brown, G. 1996). Beck Depression Inventory: Manual 2nd Edn.). San Antonio, TX: The Psychological Corporation. Fox, D. D., Lees-Haley, P. R., Earnest, K., & Dolezal-Wood, S. 1995). Base rates of postconcussive symptoms in health maintenance organization patients and controls. Neuropsychology 9, 606±611. Gouvier, W. D., Cubic, B., Jones, G., Brantley, P., & Cutlip, Q. 1992). Postconcussive symptoms and daily stress in normal and head-injured college populations. Arch Clin Neuropsychol 7, 193±211. Lees-Haley, P. R., & Brown, R. S. 1993). Neuropsychological complaint base rates of 170 personal injury claimants. Arch Clin Neuropsychol 8, 203±209. Levin, H. S., Gary, H. E., High, W. E., Mattis, S., Ruff, R., Eisenberg, H. M., Marshall, L. F., & Tabaddor, K. 1987). Minor head injury and the postconcussional syndrome: methodological issues in outcome studies. In H. S. Levin, J. Graham, & H. M. Eisenberg Eds.), Neurobehavioral Recovery from Head Injury pp. 262±275). New York: Oxford Univ. Press. Oddy, M., Humphrey, M., & Uttley, D. 1978). Subjective impairment and social recovery after closed head injury. J Neurol, Neurosurg, Psychiatry 41, 611±616. Rizzo, M., & Tranel, D. 1996). Overview of head injury and postconcussive syndrome. In M. Rizzo & D. Tranel Eds.), Head Injury and Postconcussive Syndrome pp. 1±18). New York: Churchill Livingstone. Rutherford, W. H. 1989). Postconcussion symptoms: relationship to acute neurological indices, individual differences, and circumstances of injury. In H. S. Levin, H. M. Eisenberg, & A. L. Benton Eds.), Mild Head Injury pp. 217±228). New York: Oxford Univ. Press. Sawchyn, J. M., Brulot, M. M., & Strauss, E. 1998). Postconcussion syndrome checklist in a university sample. Paper presented at the 26th Annual Meeting of the International Neuropsychological Society, Honolulu, HI, February. Trahan, D. E., & Ross, C. E. 1998). Base rates of postconcussional symptoms in neurologically normal young adults. Paper presented at the 26th Annual Meeting of the International Neuropyschological Society, Honolulu, HI, February. Trahan, D. E., Ross, C. E., & Trahan, S. L. 1997). Beaumont Postconcussional Index. Beaumont, TX: Center for Behavioral Studies.