Functional Recovery, Symptoms, and Quality of Life 1 to 5 Years After Traumatic Brain Injury

Key Points Question What is the course of functional, symptom, and quality of life outcomes 1 to 5 years after mild traumatic brain injury (mTBI) and moderate-severe traumatic brain injury (msTBI)? Findings In this cohort study of 1196 level I trauma center patients (859 mTBI, 188 msTBI, 152 orthopedic trauma controls) followed 5 years postinjury, msTBI was associated with increased mortality, yet msTBI survivors displayed improved independence from 1 to 5 years. mTBI survivors had poorer outcomes than controls. Meaning These results suggest that better understanding of TBI recovery, as well as increased clinical and community support for mTBI and msTBI, are warranted to address long-term risks associated with TBI.


Introduction
2][3][4][5] Most studies have terminated follow-up by 6 months, resulting in limited data on TBI's long-term natural history. 6Smaller, retrospective studies and investigations of restricted subgroups (eg, patients admitted to inpatient rehabilitation) indicate that TBI recovery is dynamic, continuing for years.3][14] Large-scale prospective studies characterizing the longterm outcomes of TBI may inform initiatives to develop systems of postacute care for TBI that better meet the long-term needs of this population, 14 while promoting clinical research aimed at developing disease-modifying therapies.
In recent years, data from large-scale prospective studies 5,8,15,16 have enriched understanding of the natural history of TBI.The multicenter Transforming Research and Clinical Knowledge in TBI (TRACK-TBI) study revealed that, contrary to popular belief that mild TBI (mTBI) (ie, Glasgow Coma Scale [GCS] score, 13-15) is a time-limited condition, the majority of level I trauma center patients with mTBI continue to report symptoms and injury-related problems with daily function at 1 year postinjury. 15,17Persons with moderate-severe TBI (msTBI) (GCS score, 3-12) in the TRACK-TBI sample have displayed varied outcomes, often achieving functional independence and returning to work but rarely making a complete recovery. 5Unlike functional outcomes, which show a clear doseresponse association with TBI severity, TBI-related symptoms show elevated rates of persistence in all level I trauma center patients with TBI, with weak and sometimes conflicting associations with TBI severity. 17,180][21] These findings highlight the value of capturing differing outcome measures over time.
Our objective was to examine the degree to which TBI across severities is associated with functional recovery, TBI-related symptoms, and quality of life from 1 to 5 years postinjury.We hypothesized that outcomes would continue to be poorest through the follow-up period post-msTBI, followed by mTBI, followed by orthopedic trauma control (OTC) groups.

Participants and Study Design
This manuscript follows Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for reporting observational studies.TRACK-TBI is a prospective, multicenter study recruiting participants from 18 level I trauma centers.The study was approved by the institutional review board of each enrolling institution.Participants or their legally authorized representatives completed written informed consent.The initial study enrolled participants from February 26, 2014, to July 27, 2018, within 24 of hours of injury with follow-up until 12 months postinjury.The clinical outcomes of the sample at these time points have been reported elsewhere. 5,15,22In 2019, the study began TRACK-LONG, which aimed to conduct up to 3 annual phone calls for living participants at least 2 years postinjury.The eligible sample for the present study of longer-term clinical outcomes were TBI or OTC participants who were at least 17 years old, had a known admission GCS score, and were not known to have died by 1 year postinjury (eTable 1 in Supplement 1).Of these, primary analyses focused on the participants who were not known to have died through 5 years postinjury and who completed at least 1 primary outcome from 2 to 5 years postinjury.Secondary analyses of mortality included another 20 participants who died between 1 and 5 years postinjury.

Inclusion Criteria
Inclusion criteria for TBI participants in the parent study were presentation within 24 hours of injury to a participating level I trauma center with clinical suspicion of TBI (ie, head computed tomography [CT] scan ordered) and either showing objective evidence of brain injury (eg, positive results from head CT) or displaying or reporting altered consciousness (American Congress of Rehabilitation Medicine TBI definition). 23OTCs presented within 24 hours of injury with orthopedic injuries and were excluded if they had physical signs of head trauma or met any criteria for TBI.Exclusions for all participants were being in custody of law enforcement, pregnancy, having nonsurvivable physical trauma, history of debilitating mental health disorders or neurological disease, and being non-English speaking (except for sites that enrolled Spanish-speaking individuals).

Outcome Measures Glasgow Outcome Scale-Extended
The Glasgow Outcome Scale-Extended (GOSE) estimates how injuries have affected daily function (eg, home independence and return to work, social, and other areas of preinjury daily functioning).
Respondents (participants or informants) report on new or worsened difficulties since injury 24,25 ; responses considered were given a score on an 8-point scale (with 1 indicating death; 2, vegetative state; 3, lower severe disability; 4, upper severe disability; 5, lower moderate disability; 6, upper moderate disability; 7, lower good recovery; and 8, upper good recovery).Deaths were identified through research assistants' contact with secondary contacts when pursuing follow-up visits.For the present study, GOSE scores reflected disability due to TBI and accompanying peripheral injuries, when present.Based on nonlinearity of the scale and differing traditional cut points for mTBI and msTBI, we dichotomized the GOSE to reflect both functional independence (GOSE score 5 or above) and complete functional recovery (GOSE score of 8).

Rivermead Post Concussion Symptoms Questionnaire
The Rivermead Post Concussion Symptoms Questionnaire (RPQ) assesses the presence and severity of 16 symptoms that are new or worsened since the injury. 26These diverse physical, cognitive, and emotional symptoms occur in all populations, 27 but are especially common after TBI. 28Injury-related symptoms are rated by participants as 2 (mild problem relative to preinjury), 3 (moderate), or 4 (severe); summed ratings yield a symptom severity score ranging from 0 to 64.Following a previously established threshold to project persistent postconcussion symptoms, we defined better (ie, lower) TBI symptom burden as RPQ of 15 or lower. 29,30

Quality of Life After Brain Injury-Overall Scale
The 6-item Quality of Life After Brain Injury-Overall Scale (QOLIBRI-OS) assesses health-related quality of life (QoL) in domains of cognitive, emotional, and physical function 31 on a 5-point scale (from not at all satisfied to very satisfied).Summed ratings (self-reported by participants) are transformed to produce a total score (100-point scale, with higher equaling better QoL).Scores were dichotomized as 52 or higher (better QoL) vs below 52 (worse QoL), a recommended cut score to indicate probable emotional distress and impaired health-related quality of life. 32,332) outcomes for a subset of individuals who were followed from year 1 through year 4 or 5.

Statistical Analysis
Finally, we computed cumulative mortality (ie, not known to have died at 1 year) for participants who died between 1 and 5 years postinjury, and compared groups on cumulative mortality using the log rank test.Tests were 2-tailed with α = .05;no adjustments were made for multiple comparisons.
Inverse probability weights were derived using the TWANG Shiny App software package for Windows developed by RAND Corporation (downloaded in December 2019). 34

Sample Characteristics
The

Prevalence and Group Differences in Clinical Outcomes from 1 to 5 Years Postinjury
Figure 1 illustrates the unweighted percentage (95% CI) of each primary outcome by group (eTable 3 for accompanying numbers, eTable 4 for primary analyses using propensity weighting, and eTable 5 for sensitivity analyses verifying results in a subset of individuals who were followed from year 1 to years 4 and 5 in Supplement 1).4).

Discussion
In this prospective, longitudinal study of former level I trauma center patients at 1 to 5 years post-TBI, we found both potential for continued functional recovery and persistently elevated rates of clinical impairment up to 5 years across all TBI severities.Survivors of msTBI demonstrated, on average, increasing odds of recovery of functional independence over time.As defined in this study, independence (ie, GOSE scores of 5 or above) reflected sufficient independence in home activities to be safely left alone for at least 24 hours, as well as the ability to conduct simple shopping tasks and arrange or provide one's own transportation.This important milestone indicates that persons with TBI, as well as individuals providing support to them, reclaim some autonomy in daily life.
Yet persons who reach this level of function may experience myriad symptoms, cognitive impairments, and physical limitations that disrupt life function and perceived life quality.For example, 83% of msTBI and 53% of mTBI participants reported incomplete functional recovery (ie, GOSE score below 8) at 5 years postinjury, reflecting stable odds of incomplete recovery from 1 to 5 years (ie, no interaction with year) that remained higher than the OTC group (38% at 5 years). 5,15milarly, the proportion of persons with lesser TBI-related symptom burden and better healthrelated quality of life was stable for all groups over time.The TBI groups showed comparable symptom and quality of life outcomes.Yet, both TBI groups experienced less favorable symptom and quality of life outcomes than the OTC group.These findings illustrate the differing relevance of injury variables across clinical outcomes and highlight the need to investigate the causal factors that drive long-term impairments captured by diverse outcomes.
Our findings align with those of the TBI Model Systems (TBI-MS) study, finding higher rates of functional improvement than decline from 1 to 5 years post-msTBI. 35However, the TBI-MS study also illustrated the variable recovery trajectories across individuals, with most demonstrating static

JAMA Network Open | Neurology
Functional Recovery, Symptoms, and Quality of Life 1 to 5 Years After Traumatic Brain Injury functional outcome and a minority declining.Like the TBI-MS study, we also found that despite potential for clinical improvements from 1 to 5 years, persons with msTBI have increased mortality rates relative to mTBI or OTC participants. 36,37It is unclear what factors contributed to the outcomes of the msTBI group in our sample.Other studies point to depression, physical comorbidities, cognitive functioning, and age as prognostic of long-term functional and quality of life outcomes after TBI. 8,11,35,38,39In light of the relative dearth of rehabilitation services and community supports beyond the first several months postinjury, 12 our findings justify intensifying efforts to provide longer-term rehabilitation, mental health treatment, and community services for persons with TBI.
The elevated rates of incomplete recovery among mTBI participants demonstrates the need for more comprehensive systems of care for level I trauma center patients irrespective of TBI severity and time since injury.
This study also revealed several social determinants of long-term TBI outcomes.Independent of injury group and time since injury, odds of less favorable outcomes manifested in persons who were older, female, and who had less baseline education, Medicaid or no insurance at the time of injury, and histories of prior TBI.Although cause of injury was not associated with outcomes, other research points to assaultive injury as prognostic of some outcomes. 40(Our sample had insufficient cases of assault to treat this category separately.)Taken together, our findings support the consensus that TBI recovery is influenced by multiple factors.Within individuals with TBI and GCS scores between 13 and 15, abnormal results on a head CT at hospital admission was not associated with clinical outcomes.This is consistent with research finding small or conflicting associations between TBI severity and self-report outcomes of symptoms and quality of life, 17,18,22 while contributing to knowledge on long-term functional outcome.These findings are likely to reflect the relatively severe grade of CT-GCS 13 to 15 TBI in level I trauma centers (eg, 27% of this subgroup of the TRACK-TBI sample have acute intracranial findings on magnetic resonance imaging around 2 weeks postinjury). 51,52

Limitations
This study had several limitations.Our group-level comparisons could not illustrate the varied trajectories of long-term recovery that are known to occur after TBI.Because TRACK-TBI only enrolled level I trauma center patients with TBI who underwent a clinical head CT scan, the findings cannot be generalized beyond these characteristics.Additionally, the study design precluded annual follow-up assessment for every patient during the entire 1-to-5-year follow-up period (eTable 1 in Supplement 1).However, our comparison of weighted and unweighted analysis and sensitivity analyses in a subset of persons followed from 1 to 5 years supports the robustness of our findings.
Additionally, because outcomes were gathered from research contact rather than public death databases, mortality may be underestimated.

Conclusions
This study was unique in its long-term follow-up of persons with mTBI, msTBI, and orthopedic injury.
That msTBI survivors at 1 year postinjury experienced both greater increases in functional independence and elevated rates of mortality relative to mTBI and OTC groups demonstrates the need to advance understanding of factors that contribute to long-term TBI outcomes.That mTBI was also associated with persistently elevated rates of adverse clinical outcomes from 1 to 5 years postinjury demonstrates that all level I trauma center patients with TBI should be more closely monitored and treated for injury sequelae, from the acute through chronic recovery period.Finally, the urgency of providing better clinical and community support is paramount.

JAMA Network Open | Neurology
Mixed-effects regression modeling was conducted using SAS statistical software version 9.4 (SAS Institute), and other statistical analyses were conducted in SPSS Statistics for Windows version 26 (IBM Corp).

Figure .
Figure.Unweighted Mean Percentages of Patients With Better Outcomes From 1 to 5 Years Postinjury by Group

To facilitate drawing JAMA Network Open | Neurology
Characteristics of subsamples with vs without outcome data were compared descriptively (percentages, means) and with Mann-Whitney U tests and Fisher exact tests.

Downloaded From: https://jamanetwork.com/ on 09/24/2023 inferences
for the full enrolled sample, all primary analyses incorporated inverse probability weighting.This approach more heavily weights data from groups who disproportionately did not have follow-up data to estimate parameters for the fully enrolled sample (eMethods in Supplement 1).Analyses were also performed without weighting, with a trivial effect on results.Preliminary analyses also compared the mTBI, msTBI, and OTC groups on demographic and injury variables using Kruskal-Wallis H tests and Fisher exact tests to inform selection of covariates in multivariable regression models.For the primary analyses, we first computed the percentage of persons within each group who met each outcome at 1, 2, 3, 4, and 5 years.Second, we used mixed effects logistic regression models to evaluate the degree to which group, year since injury (treated linearly), and group × year estimated odds of each outcome, controlling for relevant sociodemographic and injury covariates (age, sex, race, ethnicity, insurance type, education, prior TBI, cause of injury).Categories for race included Alaskan Native or Inuit, Asian, Black, Native Hawaiian or Pacific Islander, Indian, White, and unknown; the source from which examiners collected racial and ethnic data was not documented.Groups with insufficient numbers of individuals who were functionally dependent (OTC group in primary analyses, and additionally the mTBI group without acute intracranial findings in supplemental analyses) were excluded from logistic regression analysis of functional independence (GOSE score 5 or above).Contrasts of within-group changes in outcome measures over time were calculated to interpret group × year interactions.Nonsignificant interaction terms were dropped to facilitate interpretability of main effects.In secondary analyses, we computed percentages for each GOSE domain to further describe the groups' impairment or function.Supplemental analyses examined (1) the outcomes of the mTBI group stratifying the group by the presence or absence of acute intracranial findings on clinical head CT scans performed upon hospital admission and (

Table 2
Table1) or were of substantive interest as potential social determinants of TBI outcomes.Unadjusted, unweighted regression models revealed quite similar group and year effects as the more complex multivariable models (eTable 6 in Supplement 1).
provides results of propensity-weighted mixed effects logistic regression models evaluating the effects of group, year, and group × year on the odds of each outcome, controlling for several sociodemographic and injury variables.Covariates were selected JAMA Network Open | Neurology Functional Recovery, Symptoms, and Quality of Life 1 to 5 Years After Traumatic Brain Injury JAMA Network Open.2023;6(3):e233660.doi:10.1001/jamanetworkopen.2023.3660(Reprinted) March 20, 2023 4/15 Downloaded From: https://jamanetwork.com/ on 09/24/2023 because they demonstrated different distributions across groups (

Table 1 .
Sample Characteristics and Group Comparisons : OR, 0.81; 95% CI, 0.64-1.03;P=.08)(Table2).Beyond group and year, odds of functional independence were lower in persons who were older (OR per age 10 years, 0.67; 95% CI, 0.55-0.82)and were higher in those with non-Medicaid insurance (vs Medicaid or no insurance: OR, 2.38; 95% CI, 1.26-4.50).When comparing participants in each group who were dependent vs independent within each GOSE domain (eg, home independence, travel, work, social functioning) across the 5 years of follow-up, domains of independence in the home and independence outside the home (eg, a Unweighted P values reported from Mann-Whitney and Fisher exact tests.bThe source of race and ethnicity (eg, medical records vs participant report) was not collected.cOther or unknown race categories included Alaska Native or Inuit, Asian, American Indian, mixed race, Native Hawaiian or Pacific Islander, and unknown.dOther insurance categories included: insurance e Injury cause categories included act of mass violence, other, other nonintentional injury, other road traffic incident, suicide attempt, and unknown.fWitnessed and suspected categories collapsed.year

Table 4 .
Cumulative Mortality From 1 to 5 Years Postinjury Among Persons Believed to Be Living at 1 Year a a Cells report cumulative mortality and number of deaths (cumulative).P value from log rank test.
14,[41][42][43][44][45][46][47][48][49]Nuanced interrogation of the biological mechanisms by which age and sex affect TBI recovery will fuel the design of more effective treatments.Additionally, efforts to provide follow-up care for TBI must account for difficulties accessing health care that result from insufficient health insurance coverage for many Americans with TBI.50 Functional Recovery, Symptoms, and Quality of Life 1 to 5 Years After Traumatic Brain Injury 49.Miller T, Kallenbach MD, Huber DL, Brett BL, Nelson LD.Relationship between neighborhood disadvantage and mild traumatic brain injury symptoms.J Head Trauma Rehabil.Published online October 14, 2022.doi:10.1097/HTR.000000000000080950.Asemota AO, George BP, Cumpsty-Fowler CJ, Haider AH, Schneider EB.Race and insurance disparities in discharge to rehabilitation for patients with traumatic brain injury.J Neurotrauma.2013;30(24):2057-2065.doi: 10.1089/neu.2013.309151.Yuh EL, Mukherjee P, Lingsma HF, et al; TRACK-TBI Investigators.Magnetic resonance imaging improves 3-month outcome prediction in mild traumatic brain injury.Ann Neurol.2013;73(2):224-235.doi:10.1002/ana.23783 52.Yue JK, Yuh EL, Korley FK, et al; TRACK-TBI Investigators.Association between plasma GFAP concentrations and MRI abnormalities in patients with CT-negative traumatic brain injury in the TRACK-TBI cohort: a prospective multicentre study.Lancet Neurol.2019;18(10):953-961. doi:10.1016/S1474-4422(19)30282-0Summary of Participants From the Initial TRACK-TBI Study Considered for the Present Study Who Were Eligible for TRACK-LONG Follow-up Appointments by Year Post-injury and Number Followed eTable 2. Overview of Sample of Interest and Predictors of Being Followed 1-5 Years eTable 3. Percentage of Individuals With Favorable Recovery From 1-5 Years Post-injury (Unweighted) eTable 4. Percentage of Individuals With Favorable Outcome From 1-5 Years Post-injury (Weighted) eTable 5. Percentage of Individuals Who Completed Both a Year 1 and Year 4 or 5 Outcome Assessment With Favorable Outcome at Year 1 and Year 4/5 Post-injury (Weighted) eTable 6. Mixed Effects Logistic Model Depicting Group, Year, and Group × Year Effects on Favorable Outcome (Without Propensity Weighting or Additional Covariates) eTable 7. Multivariable Mixed Effects Logistic Model Depicting Effects of Group, Year, and Other Variables on Odds of Favorable Outcome From 1-5 Years Post-injury (With Propensity Weighting) eTable 8. Frequencies of Each Glasgow Outcome Scale-Extended Domain in the Control Group