The Long-Term Sequelae of Traumatic Brain Injury over 10 Years of Follow-Up

Abstract

Background

The acute effects of traumatic brain injury (TBI) are well documented, but there is no systematic quantification of its long-term sequelae in German-language literature. The purpose of this article is to compare the frequency of conditions linked to prior TBI with their frequency in the non-brain-injured population.

Methods

A matched cohort study was carried out on the basis of routine data from the BARMER statutory health insurance carrier. The exposure group consisted of patients treated over the period 2006–2009 for TBI at a variety of treatment intensities, including persons with multiple organ trauma. The control group consisted of BARMER insurees without prior TBI who were matched with the patients in the exposure group for age, sex, and pre-existing diseases. Late sequelae were sought in the routine data for a period of ten years after the injury. The outcome rates of the exposure and control groups were compared with Kaplan-Meier estimators and Poisson regression.

Results

114 296 persons with TBI in the period 2006–2009 were included in the study. The mortality within ten years of TBI was 305 per 1000 individuals. The relative mortality in the exposure group was higher than that in control individuals of the same age and sex, with an incidence rate ratio (IRR) of 1.67 (95% confidence interval, [1.60; 1.74]). Immobility, dementia, epilepsy, endocrine disorders, functional disorders, depression, anxiety, cognitive deficits, headache, and sleep disorders were also more common in the exposure group. Persons with TBI requiring high-intensity treatment displayed the highest relative incidence rates of the conditions studied over 10 years of follow-up. Persons who had been admitted to the hospital because of TBI had higher relative incidence rates for epilepsy and dementia than those who had been cared for on an outpatient basis.

Conclusion

Adverse sequelae of TBI can still be seen ten years after the exposure. These patients die earlier than persons without TBI and suffer earlier and more frequently from associated conditions.

Traumatic brain injury (TBI) is one of the major medical and socioeconomic challenges of our time (1). In Germany, 225 183 patients were hospitalized because of a TBI of any severity (ICD-10 S06: intracranial injury) in 2019 (2), and TBI ranked fifth among all inpatient diagnoses (2). Extensive study has already been devoted to certain sequelae of TBI, including dementia (3), epilepsy (4), and psychoorganic disorders (5), but a systematic study of all of its late consequences has been lacking to date.

In a project of the CNS Hannelore Kohl Foundation in collaboration with Barmer a statutory health insurance carrier, we performed post-TBI monitoring on the basis of routine insurance data. The aim of the study was to determine typical features of the long-term state of health of persons who have sustained a TBI.

Methods

This study was designed as a matched-cohort study according to “good practice in secondary data analysis” (6). Because all of the necessary data were already available in anonymized and standardized form, the approval of an ethics committee was not required. For the period under study (2005–2019), there were data on 7.7 million individuals from at least 1 year before their index time (TBI in case of) and for 10 years thereafter. These data documented the patient’s demographic features (age, sex, date of death) as well as full information on utilization of the German health care system, including the following:

• diagnoses registered in each quarter-year (as per the International Statistical Classification of Diseases and Related Health Problems, German Modification, ICD-10 GM)

• procedures according to the German prodecure code (Operationen- und Prozedurenschlüssel, OPS)

• outpatient medical activities according to the German standardized evaluation scale (Einheitlicher Bewertungsmaßstab, EBM)

• drug dispensings according to pharmaceutical central number (Pharmazentralnummer, PZN).

The subjects consisted of an exposure group and a control group. The exposure group comprised persons who sustained a TBI (ICD: S06) between 01/01/2006 and 12/31/2009 (one-year wash-out period). Because the data communicated by the hospitals to the insurer on the basis of the relevant German law (§ 301 SGB V) did not precisely specify the severity of injury or the duration of unconsciousness, the following indicators of a treatment-intensive TBI were defined:

• death on the day of the TBI and/or

• unconsciousness for longer than 30 minutes (a main or secondary diagnosis during hospitalization: S06.71-S06.73)

• and/or artificial ventilation and/or

• at least one of the following procedures: craniotomy, incision in the brain, hematoma evacuation, cerebrospinal fluid drainage, ICP measurement, intensive care (OPS: 5–012.0, 5–012.2, 5–013.1, 5–013.4f, 5–013.8, 5–013.x, 5–022.00f, 5–029.1, 5–029.10, 8–980).

The patients with TBI who had no evidence of intensive treatment according to the above criteria were divided into two groups depending on whether they were discharged into an inpatient setting (e.g., curative inpatient hospital stay) or an exclusively ambulatory setting. The date of TBI determined the index quarter.

The control group included only persons without TBI who matched the patients with TBI in terms of age, sex, and pre-existing diseases within the four quarters before the index quarter. The observed pre-existing diseases (classified according to Charlson 1987) (7) were:

• myocardial infarction

• heart failure

• peripheral vascular disease

• cerebrovascular disease

• dementia

• chronic pulmonary disease

• rheumatic disease

• ulcer disease

• mild liver disease

• diabetes with or without complications

• hemiplegia or paraplegia

• kidney disease

• solid tumors

• other cancer

• severe liver disease

• AIDS

For the documentation of a pre-existing disease, an inpatient principal diagnosis, outpatient diagnosis, or inpatient secondary diagnosis from two quarters was required (M2Q criterion). Moreover, the matching procedure took the use or non-use of anticoagulant drugs into account. 1 : 5 propensity score matching without deferral according to Rosenbaum and Robin (1983) (8) was used.

The primary outcome parameter examined was the death of the patient. The secondary outcome parameters were the nine conditions identified by Stocchetti and Zanier (9) as typical late sequelae of TBI:

• immobility

• dementia

• epilepsy

• endocrine disorder

• speech and visual impairment

• cognitive deficits (mental disorders due to brain damage or dysfunction)

• headache

• sleep disorders

• depression

The ICD-10 codes of the pre-existing conditions and late sequelae are listed in eTable 1.

The differences between the exposure group and the control group were analyzed statistically with Kaplan-Meier estimators and Poisson regressions. Kaplan-Meier curves were used to model the time to the onset of the late sequelae. 1-year survival rates [1ySR] and hazard ratios (HR) were determined. As the precise time of onset of the late sequelae could not be determined to the day, conditional (fixed effects) Poisson regressions were also used to compute relative risks (incidence ratios, IR) with 95% confidence intervals (10). The absolute difference in incidence rates between the exposure group and control group (ΔIR) is stated as well. The time at risk (10 years, truncated by death) served here as an offset (fixed log variable in Poisson regression for rate modeling). The control variables were the dichotomously coded pre-existing diseases and fixed effects for the exposed subjects and their five control subjects.

Results

Characteristics of the sample

114 296 patients who sustained a TBI during the period of observation were included in the study. A more detailed description of the study population can be found in eTable 2 und eFigure 1. The distribution was as follows:

eTable 2. Sample characteristics.

Category	TBI, outpatient treatment		TBI, inpatient treatment		TBI, intensive treatment
Overall	42 510	100 %	64 806	100 %	6  980	100 %
Age (years)
0–9	8 930	21 %	12 520	19 %	86	1 %
10–19	6 089	14 %	8 054	12 %	270	4 %
20–29	6 122	14 %	5 564	9 %	411	6 %
30–39	3 399	8 %	2 861	4 %	232	3 %
40–49	4 321	10 %	3 967	6 %	480	7 %
50–59	3 948	9 %	4 479	7 %	633	9 %
60–69	3 513	8 %	6 353	10 %	1 172	17 %
70–79	2 965	7 %	7 652	12 %	1 622	23 %
80–89	2 614	6 %	10 273	16 %	1 768	25 %
90+	609	1 %	3 083	5 %	306	4 %
Sex
female	24 292	57 %	35 613	55 %	3 257	47 %
male	18 218	43 %	29 193	45 %	3 723	53 %
Pre-existing diseases
myocardial infarction	911	2 %	3 041	5 %	691	10 %
heart failure	3 050	7 %	11 225	17 %	2 052	29 %
peripheral vascular disease	2 248	5 %	6 461	10 %	1 323	19 %
cerebrovascular disease	3 726	9 %	11 801	18 %	2 996	43 %
dementia	1 975	5 %	8 678	13 %	1 090	16 %
chronic pulmonary disease	9 117	21 %	13 266	20 %	1 547	22 %
rheumatic disease	831	2 %	1 792	3 %	278	4 %
ulcer disease	672	2 %	1 773	3 %	317	5 %
mild liver disease	2 450	6 %	5 201	8 %	1 035	15 %
diabetes without complication	3 213	8 %	9 194	14 %	1 804	26 %
diabetes with complication	1 276	3 %	3 952	6 %	712	10 %
hemiplegia or paraplegia	1 370	3 %	3 576	6 %	1 386	20 %
kidney disease	1 791	4 %	6 458	10 %	1 312	19 %
cancer	2 080	5 %	4 931	8 %	1 020	15 %
severe liver disease	117	0 %	396	1 %	135	2 %
solid tumors	443	1 %	1 233	2 %	247	4 %
AIDS	27	0 %	32	0 %	7	0 %
anticoagulation	4 358	10 %	11 514	18 %	2 214	32 %

TBI, traumatic brain injury

eFigure 1.

Patient selection flowchart

• 6% were treatment-intensive cases;

• 57% underwent inpatient treatment;

• 37% underwent outpatient treatment only.

The age distribution and the prevalences of pre-existing diseases are given in eTable 2. A comparison of the matching of the exposure group and the control group is shown in eTable 3.

eTable 3. Comparison of matched groups.

Category	TBI, outpatient treatment				TBI, inpatient treatment				TBI, intensive treatment
	Exposure group		Comparison group		Exposure group		Comparison group		Exposure group		Comparison group
	N	% of total	N	% of total	N	% of total	N	% of total	N	% of total	N	% of total
Overall	42 510	100 %	211 825	100 %	64 806	100 %	323 459	100 %	6 980	100 %	35 966	100 %
Age (years)
0–9	8 930	21 %	44 184	21 %	12 520	19 %	61 591	19 %	86	1 %	423	1 %
10–19	6 089	14 %	30 416	14 %	8 054	12 %	40 230	12 %	270	4 %	1 316	4 %
20–29	6 122	14 %	30 595	14 %	5 564	9 %	27 751	9 %	411	6 %	2 021	6 %
30–39	3 399	8 %	16 974	8 %	2 861	4 %	14 227	4 %	232	3 %	1 119	3 %
40–49	4 321	10 %	21 551	10 %	3 967	6 %	19 805	6 %	480	7 %	2 337	6 %
50–59	3 948	9 %	19 677	9 %	4 479	7 %	22 390	7 %	633	9 %	3 128	9 %
60–69	3 513	8 %	17 563	8 %	6 353	10 %	31 621	10 %	1 172	17 %	5 799	16 %
70–79	2 965	7 %	14 865	7 %	7 652	12 %	38 267	12 %	1 622	23 %	8 596	24 %
80–89	2 614	6 %	12 964	6 %	10 273	16 %	51 984	16 %	1 768	25 %	9 621	27 %
90+	609	1 %	3 036	1 %	3 083	5 %	15 593	5 %	306	4 %	1 606	4 %
Sex
female	24 292	57 %	120 945	57 %	35 613	55 %	177 709	55 %	3 257	47 %	16 752	47 %
male	18 218	43 %	90 880	43 %	29 193	45 %	145 750	45 %	3 723	53 %	19 214	53 %
Pre-existing diseases
Myocardial infarction	911	2 %	4 197	2 %	3 041	5 %	13 139	4 %	691	10 %	2 840	8 %
Heart failure	3 050	7 %	14 278	7 %	11 225	17 %	48 878	15 %	2 052	29 %	9 078	25 %
Peripheral vascular disease	2 248	5 %	10 191	5 %	6 461	10 %	29 898	9 %	1 323	19 %	6 605	18 %
Cerebrovascular disease	3 726	9 %	14 462	7 %	11 801	18 %	46 122	14 %	2 996	43 %	8 565	24 %
Dementia	1 975	5 %	7 583	4 %	8 678	13 %	32 408	10 %	1 090	16 %	4 246	12 %
Chronic pulmonary disease	9 117	21 %	40 385	19 %	13 266	20 %	62 435	19 %	1 547	22 %	7 715	21 %
Rheumatic disease	831	2 %	3 729	2 %	1 792	3 %	8 378	3 %	278	4 %	1 395	4 %
Ulcer disease	672	2 %	26 10	1 %	1 773	3 %	7 109	2 %	317	5 %	1 419	4 %
Mild liver disease	2 450	6 %	11 094	5 %	5 201	8 %	23 793	7 %	1 035	15 %	4 762	13 %
Diabetes without complication	3 213	8 %	16 055	8 %	9 194	14 %	44 696	14 %	1 804	26 %	8 656	24 %
Diabetes with complication	1 276	3 %	6 001	3 %	3 952	6 %	17 995	6 %	712	10 %	3 685	10 %
Hemiplegia or paraplegia	1 370	3 %	46 70	2 %	3 576	6 %	13 032	4 %	1 386	20 %	2 172	6 %
Kidney disease	1 791	4 %	7 950	4 %	6 458	10 %	26 616	8 %	1 312	19 %	5 365	15 %
Cancer	2 080	5 %	9 694	5 %	4 931	8 %	23 518	7 %	1 020	15 %	5 041	14 %
Severe liver disease	117	0 %	518	0 %	396	1 %	1 321	0 %	135	2 %	345	1 %
Solid tumors	443	1 %	1 985	1 %	1 233	2 %	5 371	2 %	247	4 %	1 101	3 %
AIDS	27	0 %	79	0 %	32	0 %	124	0 %	7	0 %	19	0 %
Anticoagulation	4 358	10 %	19 234	9 %	11 514	18 %	49 837	15 %	2 214	32 %	10 263	29 %

TBI, traumatic brain injury

Persons with TBI who underwent treatment of different intensities differed markedly with respect to their age and pre-existing diseases. Treatment-intensive TBI tended to be found in older patients with higher rates of pre-existing disease.

Survival-time analysis

Survival rates after TBI were lower than the survival rates of the control groups (Figure 1, eFigure 2) for all of the following situations:

Figure 1.

Kaplan-Meier estimators for survival after treatment-intensive traumatic brain injury (TBI) for the exposure group and the control group

• TBI treated in an outpatient setting 1YSR = 97.9%; HR = 1.19 [1.11; 1.18])

• TBI treated in an inpatient setting (1YSR = 92.7%; HR = 1.27 [1.25; 1.29]

• treatment-intensive TBI (1YSR = 61.7%; HR = 2.34 [2.27; 2.42]).

HRs were not constant over time after TBI. After TBI with outpatient care, the HR was 0.97 [0.90; 1.04] in the first year and 1.09 [1.06; 1.13] for survivors in the following nine years. After inpatient-treated TBI, the HR was 1.13 [1.10; 1.17] in the first year and 1.14 [1.12; 1.16] for survivors in subsequent years. After treatment-intensive TBI, the HR was 2.35 [2.21; 2.50] in the first year and 1.19 [1.14; 1.25] for survivors in subsequent years. Thus, the HR was significantly above 1 for survivors in all situations, even one year after TBI (figure 1).

Late sequelae of TBI

Patients with TBI of any treatment intensity had a higher incidence of all late sequelae than the corresponding control subjects (table 1). The highest absolute differences in incidence within 10 years per 1 000 insured persons, other than for death (ΔIR = 67), were for depression (ΔIR = 70), cognitive impairment (ΔIR = 30), sleep disturbance (ΔIR = 29), and speech and visual impairment (ΔIR = 24). The highest relative difference was for headache (IRR = 4.89 [9.27; 10.86]) (table 1).

Table 1. Incidence of TBI-associated outcomes per 1 000 insurees up to 10 years after a TBI.

Outcome	IR	95% CI	Inc TBI (per 1000)	Inc Control (per 1000)
Death	1.67	[1.60; 1.74]	305	238
Immobility	1.39	[1.33; 1.44]	14	12
Dementia	1.66	[1.57; 1.75]	36	31
Epilepsy	1.99	[1.93; 2.04]	51	28
Endocrine disorders	1.71	[1.62; 1.80]	5	3
Speech and visual impairments	1.12	[1.10; 1.14]	528	504
Depression, anxiety	1.34	[1.33; 1.36]	458	388
Cognitive deficits	1.91	[1.85; 1.96]	101	72
Headaches	4.89	[4.50; 5.32]	3	1
Sleep disorders	1.34	[1.32; 1.35]	179	151

CI, confidence interval; Inc, incidence; IR, incidence ratio; TBI, traumatic brain injury

Figure 2 shows the relative incidence rates of TBI-associated sequelae as a function of the intensity of treatment for TBI. All late sequelae had an IRR greater than 1. For some late sequelae, the higher the treatment intensity, the higher the IRR. This was particularly true among persons who underwent intensive treatment for TBI, with respect to death (IRR = 3.20 [2.91; 3.53]), epilepsy (IRR = 4.81 [4.4; 5.27]), endocrine disorders (IRR = 5.54 [4.58; 6.69]), and headache (IRR = 13.46 [9.81; 18.47]). On the other hand, no dependence of IRR on treatment intensity was found for dementia, language and visual impairment, depression, or sleep disorders (figure 2).

Figure 2.

Relative incidence of traumatic brain injury (TBI)-associated late sequelae up to 10 years after TBI, by severity of injury

As for the time of onset of TBI-associated late sequelae, there are those that arise shortly after the injury and those that only arise much later. Table 2 lists the mean interval, in years, from the index date to the onset of each late sequela. The higher the treatment intensity for TBI, the earlier the late sequelae were detected. In particular, after intensive treatment for TBI, death, headache, and epilepsy tended to manifest themselves in the short term, and immobility and dementia over the long term. Patients who were hospitalized for TBI tended to manifest early cognitive deficits (table 2).

Table 2. Mean intervals (years) from index date for traumatic brain injury (TBI) and onset of TBI-associated outcomes.

Outcome	TBI, outpatient treatment		TBI, inpatient treatment		TBI, intensive treatment
	TBI	Control	TBI	Control	TBI	Control
Death	5.05	5.72	4.01	5.12	2.27	5.33
Immobility	5.85	6.22	5.46	5.92	4.39	5.99
Dementia	5.07	5.4	4.09	4.78	4.49	4.94
Epilepsy	4.55	5.31	4.06	4.91	2.49	5.21
Endocrine disorders	5	5.58	5.09	5.43	2.86	5.09
Language and visual impairments	4.31	4.42	4.06	4.15	3.53	4.03
Depression, anxiety	4.88	5.21	4.81	5.06	3.54	4.52
Cognitive deficits	4.91	5.49	4.22	4.92	2.91	5.06
Headache	3.71	5.54	3.76	4.89	2.88	6.1
Sleep disorders	5.28	5.62	4.88	5.21	4.41	4.94

Discussion

Our study of the long-term sequelae of traumatic brain injury in 114 296 cases thus revealed that persons who sustained a TBI went on to suffer significantly more short- and long-term sequelaie in terms of mortality and morbidity than non-brain-injured control subjects who were matched for age and pre-existing diseases. The clinical severity of TBI was difficult to classify, as routine health insurance data include no information on the degree of impairment of consciousness (e.g., the Glasgow Coma Scale) or on its duration. The entity designated here as “treatment-intensive TBI” corresponds only roughly to severe TBI as it is usually defined. Treatment-intensive TBI was found in 9.7% of hospitalized individuals and in 6.1% of all patients with TBI. Wide variations are seen in data drawn from the literature for comparison. Cuthbert et al. (11) and Langlois (12) reported 20% severe TBI in the inpatient setting in the USA; an even higher figure (26%) was reported by Lawrence et al. (13) for the United Kingdom. On the other hand, Styrke et al. concluded that severe TBI makes up only 2% of TBI in Scandinavia (14). In the present study, we found that patients who sustained TBI of different treatment intensities hads different characteristics: TBI requiring intensive treatment is more likely to be seen in older patients with higher pre-existing morbidity. Injuries in young children lead more commonly to hospitalization but are less likely to require intensive treatment (15). This study could not address potential differences between TBI of different causes, because the health insurance data do not include the requisite information on the cause of injury. Nor were we able to determine the severity of TBI as it is usually defined, which is not equivalent to that which is designated here as treatment intensity.

As for mortality, data from the German Federal Statistical Office for the year 2019 yielded a rate of 2.9% per year (225 183 inpatient TBI cases, 6 467 deaths) (2). This figure describes the association between TBI and death only in close temporal relation to trauma. In the present data set, the overall mortality within 10 years was 3.1%. Patients who received intensive treatment for TBI mainly had higher mortality than the control group right after the trauma. Moreover, both patients who received inpatient treatment for TBI and those who were treated exclusively as outpatients had a higher mortality than the control subjects over the ensuing 10 years.

As for the late sequelae of TBI, it was found that patients with TBI are more likely to develop additional diseases, even in comparison to persons without TBI who were matched for the same pattern of pre-existing diseases. Even by the end of the 10-year follow-up period of this study, the rate of developing new diseases among persons who had sustained a TBI did not subside toward the rate prevailing in non-brain-injured controls. Similar results are found in the literature.

Data on the incidence of spasticity after TBI are scarce in the literature; the reported figures range from 13% to 20% (16), and up to 75% after severe TBI (17). A meta-analysis by Gu et al. revealed an increased risk of developing dementia (pooled odds ratio 1.81 [1.53; 2.14]) at follow-up times from 4 to 40 years (18). Epilepsy can develop at any time after TBI. The occurrence of seizures more than 7 days after TBI has been reported in the literature in 1.9% to 30% of patients who have suffered a TBI (19); we were able to show, in addition, that seizures can also appear for the first time even in the late phase (ca. 5 years after TBI), whatever the severity of the trauma. The incidence of pituitary insufficiency after moderate and severe TBI has been estimated at 27.5% [22.8; 28.9] (20). This complication received little attention until recently, which may explain the lower rate that we found in the present study during 10 years of observation after injuries sustained from 2006 to 2009.

Visual disturbances of all types were reported among American veterans in 33 of 50 cases of blast-related TBI and 34 of 49 cases of non-blast-related TBI (21). The 10-year follow-up study by Chen et al. 2017 also revealed a markedly higher probability of visual impairment over time after TBI (22). As for complex language and speech disorders after TBI, Wilde et al. found aphasia in 26% and dysarthria in 46% of cases in one of the few publications dealing with this subject (23).

According to the English SHEFBIT study, depression was present in up to 56.3% [52.8; 59.8] of patients with TBI within 10 weeks and was still present in 41.2% [37.6; 44.9] at one year (24). In a review published in 2011, Guillamondegui et al. reported that 33% of patients with a more severe course after TBI suffer from depression at one year (25). We have shown that most of the development of depression in patients who have sustained a TBI actually occurs several years after the injury, yet this finding should be seen in perspective, considering the high prevalence of depression in the general population. In 2016, Li et al. reported that age-associated cognitive decline is more rapid in persons who have sustained a TBI. Headache is seven times more common in persons who have had a mild TBI than in persons without TBI (26). We found an overall incidence three times higher than in the non-TBI group, and this difference persisted throughout the observation period.

In 2011, Castriotta et al. reported that 46% of persons who sustained a TBI suffered from sleep disorders afterward (including 23% sleep apnea, 11% post-traumatic hypersomnia, and 6% narcolepsy) (27). Although sleep disorders were less common in our data set, there is still a clear difference in prevalence between the TBI patients and the controls over the entire period of observation.

Dementia, visual disturbances, depression, and sleep disturbances are correctly said to appear often after TBI, yet these entities are so common in the general population that further study is needed to determine the increase that can truly be attributed to TBI.

Limitations

Causal inferences cannot be drawn from this study because there was no randomized control; this is, of course, an inherent limitation of any study of the sequelae of traumatic brain injury in human beings. The evidence base for the association between TBI and late sequelae naturally becomes smaller over time.

The exposure group and the (non-randomized) control group may have differed in ways that affected the results. Any differences that may have been due to age, sex, and the pre-existing conditions that we considered were compensated for by the statistical procedure of propensity score matching. No data were available on possible differences in socioeconomic circumstances or in the taking of antiplatelet drugs, so these factors could not be analyzed. The persons included in the study may also have had additional characteristics, such as multiple organ system trauma or the presence of an advance directive, that could markedly affect the outcome.

We assume that physicians’ diagnoses reflect a true burden of disease. Detection bias can arise, however, if patients who have sustained a TBI are followed up more closely than the general population. This effect is probably small in view of the long time horizon of ten years.

The severity of TBI could not be ascertained in this study according to common measures such as the Glasgow Coma Score. Therefore, a comparison to other studies, clinical studies in particular, is only possible to a very limited extent. Our stratification by treatment intensity used is generally associated with TBI severity but cannot be taken as equivalent, because TBI of any degree of severity can receive treatment of any degree of intensity. In particular, the procedures mentioned (OPS codes) may also have been performed for moderate or mild TBI.

Our inability to consider the severity of TBI according to the usual measures may have affected our findings in case any influencing factors were at work that we did not assess and that were associated with TBI severity, but not with treatment intensity. One conceivable influencing factor is the sex of the patient; yet, in a sensitivity analysis, no effect modification by sex was found ( eFigure 2). This does not rule out the possibility of effect modification by other factors, e.g., socioeconomic status.

Overview

Evaluating routine health insurance data of patients with TBI compared to persons without TBI enables new insights into the major long-term consequences, but also puts some of the already recognized sequelae of TBI into a new perspective. This type of analysis meets with difficulties in taking the different degrees of severity of TBI into account.