TRAUMATIC BRAIN INJURY (TBI) is well recognized as a major public health concern. The Centers for Disease Control and Prevention (CDC) estimate that annually in the United States, TBIs result in 53 000 deaths, 300 000 hospitalizations, 2.1 million emergency department (ED) visits, 1.1 million physician office visits, and 84 000 other outpatient department visits1; at least 3.2 million people in the United States currently live with long-term disability resulting from TBI.2 The CDC's disability prevalence estimate was derived solely from people hospitalized with TBI, which fails to capture the full extent and impact of TBI due to the much larger, less documented incidence of mostly mild TBI. Milder injuries are not typically treated in hospital inpatient settings but can result in negative outcomes.3–5 In fact, only 1 in 10 people who sustain a TBI of any severity in the United States annually are hospitalized. Approximately one-third of these individuals will have long-term disability,6 but it is unknown what percentage of people who sustain a nonhospitalized mild TBI also experience persisting disability. To better understand the prevalence of all severities of TBI and associated proportions of people with long-term disability, it is necessary to study the full spectrum of care pathways including hospitalizations, ED, and clinic and physician office visits for people with TBI and even those who do not receive medical care.
One method recommended by the CDC for studying TBI is to capture nonhospitalized and hospitalized TBIs through general population surveys.5 National surveys have asked about head injury in the last 12 months7 and state surveys have inquired about TBI in an individual's lifetime,8,9 but all were restricted to TBI with loss of consciousness (LOC). Limiting TBI to cases with LOC excludes the many individuals experiencing TBI with gaps in memory and other alterations of consciousness, but without LOC, thereby resulting in an underestimation of the full range of TBI severity.
Considerable research has been devoted to increasing the accuracy of self-reported lifetime histories of TBI. Corrigan and Bogner10,11 developed a systematic strategy to cue more accurate memories of TBI. The validity of the Ohio State University TBI Identification Method (OSU TBI-ID) is not based on elicitation of a veridical accounting of a person's lifetime history of TBI. Instead, the OSU TBI-ID provides data for calculating summary indices reflecting the likelihood that consequences have resulted from lifetime exposure to TBI. Initial validation research has supported the psychometric qualities of these summary indices. Reliability has been demonstrated by both interrater reliability and test-retest reliability.10,11 Predictive validity has been shown by the relations between indices of lifetime history and measures of cognitive performance, affective status, interpersonal functioning, and aggression.10,11
Paralleling this research are efforts that have focused on determining the prevalence of disability after TBI. Neurologic problems can arise after TBI, affecting cognition, movement, and sensation. Individuals can experience difficulty with activities of daily living as well as problems with productivity and socialization, which can result in reduced quality of life.12,13 The majority of research describing outcomes after TBI follows individuals treated in specific inpatient rehabilitation facilities14–17 and is therefore biased toward more serious injuries. Two statewide population-based follow-up outcome studies of TBI have been conducted,6,18 but they, too, have been limited to case patients whose acute care included at least 1 night of hospitalization.
Therefore, it was the purpose of this study to fill that gap in knowledge by using the validated OSU TBI-ID in a general population survey to investigate the prevalence and associated negative outcomes from all severities of TBI, treated in all settings, or not treated at all.
MATERIALS AND METHODS
We used a random digit–dialed survey to determine the population-based prevalence of a lifetime history of TBI of any severity and the prevalence of activity limitations and low satisfaction with life associated with TBI. A computer-assisted telephone interview (CATI) was programmed into 2 parts. The first part of the survey was an adaptation of the OSU TBI-ID for use by experienced interviewers without clinical backgrounds to collect a lifetime history of injury, and TBI in particular. The OSU TBI-ID first asks survey respondents to recall all lifetime injuries by inquiring about the number of injuries caused by specific etiologies (motor vehicle, sports, falls, etc). Next, questioning focuses on injuries to the head or neck before further narrowing the data collection to injuries involving loss or alteration of consciousness in order to capture all severities of TBI. For the CATI protocol, an initial series of questions asked the respondent to recall, in their lifetime, if they had ever been injured (1) in a vehicle crash, (2) by a fall, (3) playing sports, (4) by a falling or flying object, (5) by an explosion or blast, (6) because of violence, or (7) by some other cause. Follow-up questions to positive responses about each etiology included how many times they were injured by that etiology and the number of times an injury of that etiology had (1) included an injury to their head or neck, (2) resulted in an LOC or being knocked out, (3) resulted in a memory gap, and (4) resulted in feeling dazed or confused. Additional probes were used to identify their most serious TBI (or other injury if no TBI was reported) and where that injury was treated. Information gathered from this part of the survey was used to categorize each respondent into 1 of 6 injury groups based on the most severe injury incurred (no injury, injury without TBI, mild TBI without LOC, mild TBI with LOC <30 minutes, moderate TBI with LOC between 30 minutes and 24 hours, or severe TBI with LOC >24 hours), and 1 of 4 places of medical care for their most severe injury (hospital, ED, physician office/clinic, or no treatment received).
The selection of the 6 injury groups described earlier was based on an investigation of the CATI reliability, where a stratified random subsample of 200 individuals was called back an average of 11 months after their first interview to reassess their lifetime history of injury and TBI.19 The resulting test-retest reliability for the 6 injury groups used in this study (intraclass correlation coefficient = 0.78) was acceptable. However, count-based incidences of TBI were shown not to be reliable and were not used in this study.
The second part of the CATI survey gathered demographics and asked standardized questions about activity limitations and quality of life. The activity limitations question, from the CDC Behavioral Risk Factor Surveillance System (BRFSS), asked “Are you limited in any way in any activities because of physical, mental, or emotional problems?”20 Subjective quality of life was assessed using the 5-item Diener Satisfaction With Life Scale.21 A positive response to the activity limitation question, or a score less than the average rating of 20 on the Satisfaction With Life Scale,22 was considered negative outcomes.23
The methodology used for implementing the BRFSS was replicated in this study.24 Random telephone numbers from the state of Colorado were selected and dialed with the goal of gathering a representative sample of the general population. Cell phone numbers were not included, since the BRFSS did not include them until 2011, after the data collection was complete. Multiple attempts were made during daytime, evening, and weekend hours to contact the selected random households. Because of the low risk involved, the institutional review board granted approval to obtain verbal consent. Once contact was made with a household, family members 18 years and older were enumerated and the CATI selected a random family member to be interviewed. After completion of all interviews, the demographic characteristics of the study sample were compared with Colorado census data, and the most current BRFSS method of “raking and weight trimming”25 was used to adjust the sample to have the demographic characteristics of all 3803 587 Coloradoans 18 years and older reported in the 2010 census.26
Outcome analyses focused on the proportions of people reporting negative outcomes in subgroups of study participants defined by injury category and place of care. SAS 9.3 (SAS Institute Inc, Cary, North Carolina.) was used for raking and trimming, and SPSS 19.1 (released 2010. IBM SPSS Statistics for Windows, Version 19.0; IBM Corp, Armonk, New York) was used for all other analyses. The study design allowed the proportion of people with negative outcomes among various severities of TBI to be compared with the proportion reporting negative outcomes in people with no injuries and people with injuries without TBI, controlling for place of care.
A total of 2701 random digit–dialed surveys of 2594 English-speaking and 107 Spanish-speaking Coloradoans were completed using the 2-part CATI from 2008 to 2010. After removing nonworking and nonresidential numbers, ineligible households (out-of-state, no one 18 years or older, etc), and numbers with no answer after 15 attempts, the percentage of complete or partial interviews (57%) was comparable with the 2009 Colorado BRFSS (55%).27 Partial interviews were rare (<1% of the valid contacts), and they were not entered into the analysis database with the 2701 completed interviews. The time to complete the lifetime history of TBI portion of interview averaged 2 minutes for people with no injury, 5 minutes for people with non-TBI injury, and 6 minutes for people with TBI. Biases in the study sample (65% female; 11% aged 18-34 years, 26% aged 35-49 years, 35% aged 50-64 years, and 28% aged 65 years and older) were similar to biases in the 2009 Colorado BRFSS sample,27 and they were adjusted through the raking and trimming processes, resulting in a weighted study sample with a demographic distribution like the Colorado population aged 18 years and older (n = 3803 587; 50% female, 32% aged 18-34 years, 28% aged 35-49 years, 26% aged 50-64 years, and 14% aged 65 years and older). Both unweighted and weighted data included 84% white race.
The weighted study data indicate that 19.8% of adult Coloradoans reported no injury in their lifetime, 37.7% reported injury but no TBI, and a total of 42.5% reported having a lifetime history of TBI. Mild TBI was predominant, with 18.1% reporting their most severe injury as mild TBI without LOC and 18.4% reporting mild TBI with LOC; moderate TBI accounted for only 3.6%, and severe TBI was reported by only 2.4% of Coloradoans. Of those reporting a TBI, 23.1% were hospitalized for their most severe TBI, 38.5% were treated in an ED, 9.8% were seen in a physician's office, and 27.5% did not seek medical care. Table 1 shows the distribution of injury categories by the place of treatment. It indicates that 89% of those with severe TBI were treated in an inpatient hospital setting. Among Coloradoans reporting moderate TBI, 49% were hospitalized, 37% received treatment in the ED, and 9% sought no medical care. Among people with mild TBI (both mild TBI groups with and without LOC), about 40% were seen in the ED whereas about 30% did not seek any care. The places of care for those reporting injury but no TBI were distributed similarly to those of the mild TBI groups, except there were more people seen in physician offices and fewer did not seek medical care.
Table 2, based on weighted data, presents both the estimated counts of the Colorado population aged 18 years and older and the proportion of the people who report activity limitations and low satisfaction with life among injury groups and then among injury groups by place of care. For both activity limitations and low satisfaction with life, the highest rate of poor outcomes occurred in people with severe TBI, with drops in prevalence as the severity of TBI decreased. All 4 injury groups with TBI had higher prevalence of poor outcomes than the group with injury but no TBI; the group of participants reporting no injury in their lifetime had the lowest prevalence. When the prevalence of people with both negative outcomes was examined within each place of care column, the same general pattern was seen in most cases. Interestingly, for moderate and severe injuries, as treatment setting became less intense, the percentage reporting a negative outcome increased.
Confidence intervals for the total prevalence are reported in the next-to-last column of Table 2. They indicate that some injury groups are significantly different from each of the other 5 injury groups (eg, for both outcomes, this is true of the no injury group, as the confidence interval of the no injury group does not overlap with the confidence interval of any other group). When each group is not significantly different from all other groups, the exceptions are groups immediately adjacent in the severity continuum. The final column of Table 2 compares the prevalence of an injury group with that for the no injury group, expressing the difference as a percentage of the no injury group. The increase in prevalence of activity limitations in people with TBI compared with no injury ranges from 79% to 307%, rising as the severity of the TBI increases. For low satisfaction with life, the increase in prevalence ranges from 78% to 407% as the severity of TBI increases, compared with people with no injury.
The prevalence in Table 2 was used to calculate estimates of the actual number of adults in Colorado reporting poor outcomes in the severity and place of care groups. It is estimated that 986 068 Coloradoans report an activity limitation because of physical, mental, or emotional problems (26%) and 576 728 report low satisfaction with life (15%). While a large number of Coloradoans with poor outcomes report a lifetime history of any TBI with hospitalization (151 919 with activity limitations and 92 839 with low life satisfaction), substantially more people with nonhospitalized TBI report these poor outcomes (402 303 with activity limitations and 225 065 with low life satisfaction). The greater number of people with poor outcomes and nonhospitalized TBI resulted from the fact that the nonhospitalized TBI group was more than 3 times larger than the hospitalized TBI group, although the rates of poor outcomes were higher in those with hospitalized TBI (41% with activity limitations and 25% with low satisfaction with life) than in those with nonhospitalized TBI (33% with activity limitations and 18% with low satisfaction with life).
The scope and seriousness of TBI as a public health problem are highlighted when disability is examined by a general population survey that includes all severities of TBI treated in all settings or not treated at all. High percentages of adult Coloradoans (42.5%) reported a lifetime history of TBI, and nearly a quarter (24.4%) of the population had a TBI with LOC. The severity category of worst TBI reported in a person's lifetime was associated with activity limitations and low satisfaction with life. A clear gradient of these negative outcomes was seen, with the highest proportion of people with activity limitations and low satisfaction with life reporting moderate to severe TBI; however, even those reporting mild TBI had more negative outcomes than those reporting injuries but no TBI and those not reporting any injuries in their lifetime. This pattern generally held no matter where (or if) treatment occurred.
The confidence intervals of the stair-step prevalence of poor outcomes across the 6 injury groups indicated that people reporting any severity of TBI had significantly higher percentages of activity limitations and significantly higher percentages of low satisfaction with life. Compared with the no injury group, people reporting mild TBI without LOC had more than a 75% increase in the likelihood of both poor outcomes; mild TBI with LOC increased the likelihood of both poor outcomes by more than 100%; moderate TBI increased the likelihood of both poor outcomes by more 200%; and severe TBI increased the likelihood of activity limitations by more than 300% and the likelihood of low satisfaction with life by more than 400%.
Most prior research has focused on annual incidence of TBI rather than lifetime prevalence. The supplement to the 1991 National Health Interview Survey7 asked about head injury with LOC in the past 12 months, with follow-up questions asking where (or if) medical care was sought. The annual incidence of nonfatal TBI with LOC was 618 per 100 000 population, with 75% seeking medical care (14% in clinics or offices, 35% in EDs, and 25% in an hospital setting).28 Other research has estimated the incidence of TBI seen in the ED but not admitted at 392 per 100 000 population,29 and the incidence of mild TBI at 503 per 100 000 population.30 The present study adds support to previous findings that TBIs treated in an hospital setting comprise the minority of TBI, exceeded by TBI seen in EDs and by mild TBI.
Previous research asking a single question about lifetime history of TBI in statewide BRFSS surveys identified 10% to 37% of the population as reporting a TBI,8,9 but different wording of the questions—including requiring the respondent to know the definition of TBI—may partially explain the wide range in percentages. The current findings extend the high end of that range, but this is likely due to the far more comprehensive OSU TBI-ID methodology.
This research expanded our understanding of outcomes associated with a lifetime history of TBI by including all severities of TBI treated in all settings (or not treated at all). The first statewide population-based follow-up outcome study of people with TBI was also conducted in Colorado, but it was limited to adults hospitalized with TBI.6 Thirty-seven percent of the sample had functional limitations, and 28% reported only fair to poor quality of life 1 year postinjury. A population-based study of hospitalized TBI in South Carolina documented similar findings.18 These proportions with negative outcomes are similar to adult Coloradoans hospitalized for their TBI in the present research. However, the current study also identified substantial proportions of people with activity limitations and low life satisfaction who had less severe TBI not resulting in hospitalization that were generally higher proportions than those occurring among people without a history of TBI.
While the proportion of people with disability after TBI was higher in those with more severe injuries, the absolute number of people reporting disability after TBI was higher in those with less severe injuries. Among the estimated 554 222 Coloradoans with current activity limitations who reported a lifetime history of TBI, 80% reported only mild TBI and 73% were not hospitalized. The fact that most people with activity limitations were not hospitalized raises the question of whether (despite differences in the period of observation) the current estimate of 3.2 million people living in the United States with long-term disability after TBI is low, since it was based solely on those hospitalized with TBI.2 If these Colorado results were replicated nationally, the estimated number of people living in the United States with long-term disability who have experienced TBI might be closer to 10 million. While the current methodology does not allow inference about the extent to which the disability was caused by the TBI, the co-occurrence of TBI with disability nonetheless has public health implications.
The importance of TBI as a public health concern is further underscored by the concentration of poor outcomes with TBI. Of all people reporting activity limitations in Colorado, 56% also report TBI; 55% of people with low life satisfaction also report a TBI. Again, whether or not the poor outcomes found in this study were caused by the lifetime history of TBI, the association of these 2 factors defines a clear population in need of services and interventions. Furthermore, higher percentages of negative outcomes were observed for similar severity of injuries treated in less intense settings. This relation may also suggest that people with injuries receiving inadequate medical attention or follow-along services have greater consequences.
The findings also support the need to establish routine screening for lifetime history of TBI in agencies and organizations that serve persons with activity limitations due to physical or mental disability. Unidentified TBI can potentially decrease the accessibility and effectiveness of services and treatment. Cognitive and executive functioning deficits can impede treatment because such difficulties may directly interfere with a person's ability to fully participate. The misinterpretation of executive functioning deficits by providers unaware of TBI may lead to incorrect attributions of poor motivation or lack of cooperation, and this can undermine the therapeutic relationship. Treatment providers who are aware that an individual has a history of TBI can learn to accommodate their treatment and approach to facilitate better engagement and retention in treatment.
This study demonstrated the feasibility of using a general population survey to examine the prevalence of all severities of TBI. A reliable self-report of the most severe TBI occurring in a person's lifetime was collected, along with information on associated outcomes using a random digit–dialed CATI survey. This study offers a methodology that could be replicated in other states or nationally. Future research might include an evaluation of the relative risk of negative outcomes within various TBI severity levels and injury groups after controlling for demographics and place of care. As has been shown in previous research, the OSU TBI-ID proved useful in eliciting memories of TBI and other injuries. The basic strategy of using specific etiologies to cue memories of a wide range of injuries before narrowing the focus of questions to TBI seems sound.
A major weakness of the survey methodology used in this study is that it relies on self-report, although good test-retest reliability was found nearly a year later. A second weakness is the potential lack of representativeness resulting from the random digit–dialing methodology and the exclusion of cell phones. While the study followed BRFSS protocols and used the best available procedures for weighting the data to census parameters, this does not ensure that systematic differences do not exist between responders and nonresponders (or between land line– and cell phone–only users), and the methodology does not allow comparisons between these groups. This study's methodology also restricts us to commenting on associations between TBI and activity limitations and low life satisfaction (rather than attributing causation). Finally, this study was limited to adults. Future research to address these limitations and to develop interventions that might reduce activity limitations is warranted. In addition, future studies should control for other factors found to be associated with satisfaction with life and activities of daily living, although another study using this data set suggested that the relations hold when basic demographics are controlled.19
1. Coronado V, McGuire L, Faul M, Sugerman D, Pearson W. The epidemiology and prevention of TBI. In: Zasler N, Katz D, Zafonte R, eds. Brain Injury Medicine. New York, NY: Demos; 2012;45–56.
2. Zaloshnja E, Miller T, Langlois J, Selassie A. Prevalence of long-term disability from traumatic brain injury
in the civilian population of the United States, 2005. J Head Trauma Rehabil. 2008;23(6):394–400.
3. Alexander M. Mild traumatic brain injury
: pathophysiology, natural history, and clinical management. Neurology. 1995;45(7):1253–1260.
4. Kushner D. Mild traumatic brain injury
: toward understanding manifestations and treatment. Arch Intern Med. 1998;158(15):1617–1624.
5. National Center for Injury Prevention and Control. Report to Congress on Mild Traumatic Brain Injury
in the United States: Steps to Prevent a Serious Public Health Problem. Atlanta, GA: Centers for Disease Control and Prevention; 2003.
6. Whiteneck G, Brooks C, Mellick D, Harrison-Felix C, Sendroy-Terrill M, Noble K. Population-based estimates of outcome after hospitalization for traumatic brain injury
in Colorado. Arch Phys Med Rehabil. 2004;85(suppl 2):S76–S81.
7. Adams P, Benson V. Current estimates from the National Health Interview Survey. Vital Health Stat. 1991;10(184):1–232.
8. Bureau of Health Statistics and Data Management. New Hampshire Behavioral Risk Factor Surveillance System Questionnaire. Concord, NH: New Hampshire Department of Health and Human Services; 1998.
9. Health Statistics Section. Colorado Behavioral Risk Factor Surveillance System Questionnaire. Denver, CO: Colorado Department of Public Health and Environment; 2001.
10. Bogner J, Corrigan J. Reliability and predictive validity of the Ohio State University TBI identification method with prisoners. J Head Trauma Rehabil. 2009;24(4):279–291.
11. Corrigan J, Bogner J. Initial reliability and validity of the Ohio State University TBI Identification Method. J Head Trauma Rehabil. 2007;22(6):318–329.
12. Rosenthal M, Griffith E, Kreutzer J, Pentland B, eds. Rehabilitation of the Adult and Child With Traumatic Brain Injury
. 3rd ed. Philadelphia, PA: FA Davis; 1999.
13. Thurman D, Alverson C, Dunn K, Guerrero J, Sniezek J. Traumatic brain injury
in the United States: a public health perspective. J Head Trauma Rehabil. 1999;14(6):602–615.
14. Caplan B, ed. The TBI Model Systems of Care Project. Journal of Head Trauma Rehabil. 1996;(spec issue):1–99.
15. Cifu D, Kreutzer J, Marwitz J, et al. Etiology and incidence of rehospitalization after traumatic brain injury
: a multicenter analysis. Arch Phys Med Rehabil. 1999;80(1):85–90.
16. Harrison-Felix C, Newton C, Hall K, Kreutzer J. Descriptive findings from the Traumatic Brain Injury
Model Systems National Database. Arch Phys Med Rehabil. 1996;39(11). doi:10.1097/00001199-199610000-00002.
17. Rosenthal M, ed. The TBI Model Systems of Care Project. Journal of Head Trauma Rehabil. 1993;(spec issue):1–116.
18. Selassie A, Pickelsimer E, Gravelle W, Gu J, Broadhurst S. South Carolina Traumatic Brain Injury
Follow-up Registry. Charleston, SC: Medical University of South Carolina; 2003.
19. Cuthbert J, Whiteneck G, Corrigan J, Bogner J. The reliability of a computer assisted telephone interview version of Ohio State University TBI Identification Method [published online ahead of print June 11, 2014]. J Head Trauma Rehabil. doi:10.1097/HTR.0000000000000075.
20. Center for Disease Control and Prevention. 2008 Behavioral Risk Factor Surveillance System Questionnaire. Center for Disease Control and Prevention Web site. http://www.cdc.gov/brfss/annual_data/pdf-ques/2008brfss.pdf
. Published December 13, 2007. Accessed August 25, 2013.
21. Diener E, Emmons R, Larsen R, Griffin S. The Satisfaction with Life Scale. J Pers Assess. 1985;49(1):71–75.
22. Diener E. Understanding Scores on the Satisfaction With Life Scale. University of Illinois Web site. http://internal.psychology.illinois.edu/∼ediener/Documents/Understanding%20SWLS%20Scores.pdf
. Published February 13, 2006. Accessed November 2, 2014.
23. Corrigan J, Cuthbert J, Harrison-Felix C, et al. US population estimates of health and social outcomes 5 years after rehabilitation for traumatic brain injury
. J Head Trauma Rehabil. 2014;29(6):E1–E9. doi:10.1097/HTR.0000000000000020.
24. Center for Disease Control and Prevention. Overview: BRFSS 2008. Center for Disease Control and Prevention Web site. http://www.cdc.gov/brfss/annual_data/2008/overview_08.rtf
. Accessed August 25, 2013.
25. Battaglia M, Izrael D, Hoaglin D, Frankel M. Tips and tricks for raking survey data (a.k.a. Sample balancing. In: Proceedings of the Survey Research Methods Section of the American Statistical Association. Alexandria, VA: American Statistical Association; 2004:4740–4745.
26. US Census Bureau. 2010 Census of Population and Housing, Population and Housing Unit Counts. CPH-2-7, Colorado. Washington, DC: US Census Bureau; 2012.
27. Center for Disease Control and Prevention. Behavioral Risk Factor Surveillance System 2009 Summary Data Quality Report. Center for Disease Control and Prevention Web site. http://www.cdc.gov/brfss/annual_data/2009/pdf/2009_Summary_Data_Quality_Report.pdf
. Published February 18, 2011. Accessed March 17, 2015.
28. Sosin D, Sniezek J, Thurman D. Incidence of mild and moderate brain injury in the United States, 1991. Brain Inj. 1996;10(1):47–54.
29. Guerrero J, Thurman D, Sniezek J. Emergency department visits associated with traumatic brain injury
: United States. Brain Inj. 2000;14(2):181–186.
30. Bazarian J, MCcClung J, Shah M, Cheng Y, Flesher W, Kraus J. Mild traumatic brain injury
in the United States, 1998-2000. Brain Inj. 2005;19(2):85–91.