Risk and outcomes for traumatic brain injury in patients with mental disorders
- 1Department of Anaesthesiology, Taipei Medical University Hospital, Taipei, Taiwan
- 2Health Policy Research Centre, Taipei Medical University Hospital, Taipei, Taiwan
- 3School of Medicine, Taipei Medical University, Taipei, Taiwan
- 4Graduate Institute of Injury Prevention and Control, Taipei Medical University, Taipei, Taiwan
- 5Trauma and Emergency Centre, China Medical University Hospital, Taichung, Taiwan
- Correspondence to Professor T-L Chen, Department of Anaesthesiology, Taipei Medical University Hospital, affiliated with School of Medicine, College of Medicine, Taipei Medical University, 250 Wuxing St, Taipei 11031, Taiwan;
- Received 22 January 2012
- Revised 28 May 2012
- Accepted 13 June 2012
- Published Online First 8 July 2012
Objective To investigate the risk of traumatic brain injury (TBI) and post-injury mortality in patients with mental disorders.
Background Patients with mental disorders are at higher risk of injuries. However, the association between mental disorders and TBI is still not understood. We conducted case-control studies to investigate whether people with pre-existing mental disorders are at higher risk of TBI and post-injury mortality.
Methods Using reimbursement claims, we analysed 16 635 patients with TBI and 66 540 controls with adjustment of covariates to study the association of mental disorders and TBI. A nested case-control study was also conducted to analyse contributory factors for post-injury mortality.
Results People with mental disorders were at increased risk of TBI (odds ratio (OR) 1.94, 95% confidence interval (CI) 1.86 to 2.02). Men, older age, living in highly urbanised areas and patients on low income had a higher risk of TBI and post-injury mortality. Psychiatric medication intensity and frequency of psychiatric visits was significantly correlated with TBI in a severity dependent relationship (p for trend <0.0001). Patients receiving advanced psychiatric healthcare had an increased risk of TBI (OR 2.98, 95% CI 2.67 to 3.33) and post-injury mortality (OR 1.92, 95% CI 1.34 to 2.77). A history of receiving psychiatric related outpatient care (OR 1.77, 95% CI 1.70 to 1.85), hospitalisation (OR 3.21, 95% CI 2.79 to 3.70) or emergency visits (OR 3.53, 95% CI 3.15 to 3.94) were highly associated with subsequent TBI.
Conclusions Patients with mental disorders have an increased risk of TBI. Intensity of psychiatric medication is associated with increased post-injury mortality. Special attention to prevent TBI among this disabled population is mandatory.
Traumatic brain injury (TBI) is a common trauma worldwide. It causes disability and death, with an estimated 1.1 million emergency visits, 235 000 hospitalisations and 50 000 deaths in the USA every year.1–3 The socioeconomic impacts of disability following TBI are potentially long term or lifelong.3 The epidemiology, natural history, risk factors and treatment of TBI have been established.4 ,5 However, the high death rate from TBI is still a serious problem with global concern.4
Mental disorders also present a substantial global disease burden, with estimations of prevalence ranging from 4.3% to 26.4% across nations.6 ,7 Epidemiological studies show that mental disorders are common in individuals who suffer from traumatic injuries, including falls and hip fractures.8–14 More morbidities, longer length of stay, intensive care unit stay, higher medical expenditure, more postoperative complications and mortality were found in trauma victims with underlying mental disorders.11 ,15–18 Although the relationship between psychiatric illness and TBI has been studied, the impact of mental disorders on TBI has still not been globally validated.19–23 Furthermore, the association of severity of mental disorders with prevalence of TBI has not been determined. This study used the population based reimbursement claims to evaluate the association between underlying mental disorders and TBI in a case-control design. Furthermore, we conducted a nested case-control study to identify whether mental disorders play an important role in post-injury mortality after TBI.
The Department of Health of Taiwan integrated 13 insurance systems into a universal coverage healthcare programme in 1995; this covered more than 99% of the 23 million population of Taiwan by the end of 2008. We conducted a population based case-control study using reimbursement claims from 1 000 000 insured people selected randomly from the 23 million beneficiaries covered by Taiwan's National Health Insurance Programme. The claims data included patient personal characteristics and records for all medical services, both outpatient and inpatient, from 1996 to 2008. The National Health Research Institute, Taiwan, collects and provides the above database available for public access. Patient information, including date of birth, sex, residency and income (inferred from insurance fees), was retrieved and the validity of the database has been verified.24 This study was conducted in accordance with the Helsinki Declaration. To protect personal privacy, the electronic database was decoded, with patient identifications scrambled for further public access for research. According to the National Health Research Institute regulations, informed consent is not required because patient identification numbers were decoded and scrambled. However, this study was evaluated and approved by Taiwan's National Health Research Institutes.25
Selection of study subjects
We identified 16 635 patients diagnosed with TBI in 2005–2008 as study cases. For each TBI patient, four controls without a history of TBI were randomly selected during the same index period. Sociodemographic factors with significant covariates were analysed for their potential association with TBI. To investigate factors associated with post-injury mortality, a nested case-control study was conducted. Among 16 635 patients with TBI in 2005–2008, cases with inhospital death were defined as post-injury mortality and the survivors of TBI were considered as controls.
According to the International Classification of Diseases, 9th revision, clinical modification (ICD-9-CM), we defined TBI as ICD-9-CM 800-805 and 850-854, and mental disorders as ICD-9-CM 290-319. Sociodemographic variables included sex, age, urbanisation level and income (low income or not). The age of the study case was defined when the TBI event occurred; the age of the control was coincident with the timing of the TBI cases selected. The population density (persons/km2) was defined by dividing the 359 townships and city districts' populations by the area (km2) of each administrative unit and categorising them into four quartiles: low, moderate, high and very high urbanisation.26 Low income was defined as patients qualifying for waived medical co-payment, as certified by the Bureau of National Health Insurance. To validate the pre-TBI medical service utilisation for mental disorders as a potential impact on disease severity associated with TBI, we categorised the frequency of psychiatric outpatient visits into quartiles: 1–2, 3–4, 5–16 and ≥17 within the pre-TBI 24 month period. To understand the potential relationship between medication for mental disorders and TBI, days of having prescriptions for antipsychotic drugs, antidepressants and anxiolytics within the pre-TBI 24 month period were divided by 730 days to calculate the psychiatric medication index.27 This psychiatric medication index was divided into quartiles to analyse the association between level of psychiatric medication (indicating whether they were heavily medicated or not) and risk of TBI. Whether patients with mental disorders who received specific treatments, including electroconvulsive therapy, occupational therapy, re-educative or intensive psychotherapy and other biofeedback therapies in the pre-TBI 24 months, were at a higher risk of TBI and subsequent mortality was evaluated and compared with people without mental disorders. Furthermore, we examined whether patients with a history of outpatient care, emergency care or hospitalisation due to mental disorders as the primary diagnosis within the follow-up period were at higher risk of TBI.
We compared sociodemographic factors, coexisting medical conditions and mental disorders' status between cases with TBI and controls using χ2 tests. Types of psychosis were specified individually among cases and controls. Multivariate logistic regression analysis was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) of TBI associated with mental disorders after controlling for sociodemographic variables, including sex, age, urban residence, low income status and coexisting medical conditions. The OR associated with each mental disorder was assessed using multivariate logistic regression analysis after adjusting for sociodemographic variables and coexisting medical conditions. Nested case-control analysis was used to compare various mental disorder scenarios among TBI patients with or without post-injury mortality. Multivariate logistic regression analysis was applied to estimate the OR of post-injury mortality in association with mental disorders when controlling for sociodemographic variables and coexisting medical conditions. We calculated the risk of TBI for patients with various frequencies of outpatient visits for psychiatric care (pre-TBI within 24 months), different levels of psychiatric medication, receiving specific psychiatric interventions, and history of receiving emergency care or hospitalisation for psychiatric care during the 24 month pre-TBI period. All analyses were performed with SAS for Windows (V.9.1, SAS Institute Inc.). A p value <0.05 was taken to represent statistical significance. All tests were two tailed.
Of the 16 635 TBI cases and 66 540 controls without TBI, mental disorders were more prevalent in the TBI group than in controls (OR 1.94, 95% CI 1.86 to 2.02) (table 1). Men were at higher risk of TBI than women (OR 1.34, 95% CI 1.29 to 1.38). Compared with people aged 30–39 years, older people aged ≥70 years had the highest risk of TBI (OR 1.87, 95% CI 1.75 to 2.01), followed by the younger group aged 20–29 years (OR 1.75, 95% CI 1.65 to 1.85), those aged 60–69 years (OR 1.41, 95% CI 1.31 to 1.52) and those aged 50–59 years (OR 1.16, 95% CI 1.09 to 1.23). Increased risk of TBI was also found in people living in less urbanised areas (OR 1.38, 95% CI 1.31 to 1.45) or with a low income (OR 1.92, 95% CI 1.75 to 2.11). Hypertension (OR 1.09, 95% CI 1.03 to 1.15), diabetes (OR 1.32, 95% CI 1.24 to 1.41), stroke (OR 1.95, 95% CI 1.78 to 2.15), epilepsy (OR 3.20, 95% CI 2.74 to 3.75) and renal dialysis (OR 1.49, 95% CI 1.20 to 1.87) were coexisting diseases associated with TBI risk. Based on age stratification, the adjusted OR of TBI associated with mental disorders for people aged 20–29, 30–39, 40–49, 50–59, 60–69 and ≥70 years were 2.17 (95% CI 1.98 to 2.39), 2.47 (95% CI 2.24 to 2.73), 2.07 (95% CI 1.89 to 2.27), 1.58 (95% CI 1.43 to 1.74), 1.52 (95% CI 1.35 to 1.70) and 1.74 (95% CI 1.59 to 1.91), respectively (data not shown in tables).
The associations between specific types of mental disorder and TBI are shown in table 2. Compared with people without TBI, patients with specified non-psychotic mental disorders due to brain damage had the highest risk of TBI (OR 8.32, 95% CI 6.17 to 11.2), followed by patients with alcohol or drug induced mental disorders (OR 4.40, 95% CI 3.08 to 6.30) and emotional disturbance specific to childhood and adolescence (OR 3.85, 95% CI 1.34 to 11.1).
Among all TBI victims, 495 patients (2.97%) suffered from inhospital mortality after TBI (table 3). The risk of post-injury TBI mortality was higher in men than in women (OR 2.02, 95% CI 1.65 to 2.47). Compared with patients aged 20–29 years as the reference group, older patients (≥70 years) had the highest risk of TBI (OR 13.3, 95% CI 8.75 to 20.2), followed by those aged 60–69 years (OR 4.37, 95% CI 2.74 to 6.99), 50–59 years (OR 3.71, 95% CI 2.37 to 5.80) and 40–49 years (OR 2.41, 95% CI 1.51 to 3.84). TBI patients living in highly urbanised areas or having a low income also were found to have an increased risk of post-injury mortality, with an OR of 1.96 (95% CI 1.51 to 2.54) and OR of 2.01 (95% CI 1.46 to 2.75), respectively. Stroke (OR 1.39, 95% CI 1.06 to 1.82), renal dialysis (OR 5.62, 95% CI 3.55 to 8.90) and diabetes (OR 1.31, 95% CI 1.04 to 1.66) were coexisting diseases associated with post-injury mortality.
Compared with people without mental disorders, psychiatric patient groups with the maximal medication index had the highest risk of TBI (OR 2.56, 95% CI 2.34 to 2.81), with a dose-dependent relationship between psychiatric medication index and risk of TBI (p for trend <0.0001) (table 4). Patients with mental disorders receiving specific psychiatric treatments were also at higher risk of TBI compared with people without mental disorders (OR 2.98, 95% CI 2.67 to 3.33). Number of psychiatric outpatient visits during the 24 month pre-TBI period was significantly correlated with incidence of TBI; patients with ≥17 visits for psychiatric outpatient services were at the highest risk (OR 3.36, 95% CI 2.76 to 4.08). Compared with individuals without mental disorders, the ORs of TBI for patients who had a history of psychiatric related emergency care or hospitalisation during the pre-TBI period were 3.53 (95% CI 3.15 to 3.94) and 3.21 (95% CI 2.79 to 3.70), respectively.
The psychiatric medication index also was associated with post-injury mortality. Increased risk of post-injury mortality was also found in patients who had pre-TBI specific psychiatric treatment (OR 1.92, 95% CI 1.34 to 2.77); the number of visits for psychiatric outpatient care within the 24 month pre-TBI period was associated with post-injury mortality. Patients with mental disorders receiving pre-TBI psychiatric related hospitalisation had an OR of 2.75 (95% CI 1.83 to 4.15) for post-injury mortality.
In the further vehicle related stratification among inpatient TBI cases, mental disorders was much more associated with vehicle unrelated TBI (OR 2.04, 95% CI 1.90 to 2.20) than vehicle related TBI (OR 1.65, 95% CI 1.49 to 1.82) (data not shown in tables).
This population-based study demonstrated that patients with mental disorders were at a significantly higher risk of TBI in socioeconomic associated, disease specific and severity dependent patterns. The risk of TBI and post-injury mortality was highly associated with the intensity of pre-TBI psychiatric medication and specific psychotic related treatment. To the best of our knowledge, this is the first study to investigate the dose–response effect of psychiatric medication and TBI risk. In addition, this was also the first study which validated the association of pre-TBI severity of mental illnesses in relation to clinical aspects with the risk for TBI. TBI victims with mental disorders receiving psychiatric interventions, psychiatric related emergency service or hospitalisation before their injury were more susceptible to post-injury mortality after TBI.
Our population-based investigation demonstrated that TBI and post-injury mortality were related to sociodemographic differences. The elderly also showed a higher risk of TBI and post-injury mortality.28 ,29 We suggested that older people limited by movement disorders, reduced activity and polypharmacy might experience falls and subsequent TBI.27 Gender difference in TBI and post-injury mortality was also noted: men had a higher risk of TBI and post-injury mortality than women due to greater exposure to perilous jobs and dangerous activities.28 ,29 In contrast, better outcome in women after TBI might be due to female sex hormones and supplemental oestrogen therapy.30 A previous study of hospitalised TBI victims highlighted the significantly increased risk facing patients residing in rural areas.31 Compared with highly urbanised areas, less urbanised areas had more unemployed, blue collar, elderly and disabled residents facing a greater risk of TBI. In contrast, higher post-injury mortality was seen in patients in highly urbanised areas, who are more at risk of traffic accidents, violence or suicidal behaviours that can lead to a worse TBI outcome. Lower income is also a risk determinant for fatal and non-fatal injury.32–34 Our data showed a consistently increased risk of TBI and post-injury mortality among people with lower incomes who were more likely to be blue collar workers, living in hazardous environments and having less knowledge about injury prevention or recognition. These factors may all contribute to the potential delay in medical care and lead to a worse outcome after TBI.32–34
Although patients with dementia, depression or schizophrenia have been reported to have an increased risk of falls, hip fractures and other injuries,8–14 the association between specific mental disorders in full spectrum and the incidence of TBI afterwards has not been investigated.20 Previous studies were limited by focusing on specific patient groups, such as veterans, homeless people or prisoners, with relatively low ORs19–22 and without specifying the mental illness,21 without proper adjustment for sociodemographic factors20 and having small sample sizes.22 ,23 Among all of the subtypes analysed, we demonstrated that the risks for TBI among patients with different mental disorders were disease specific. Specified non-psychotic mental disorders due to brain damage, alcohol or drug induced mental disorders, and emotional disturbances were the top three mental disorders associated with TBI in this population-based study.19–23 Why patients with mental disorders are prone to a significantly higher risk of TBI is multifactorial. First, psychiatric patients' poor attention due to disease per se or psychiatric medication is the most important cause of vehicle collisions, occupational injuries or falls that might lead to TBI.8–14 Suicide or self-harm is another cause of TBI for patients with mental disorders.35 Psychiatric disorder is clearly the strongest risk factor for suicide; among 1887 suicides in Finland, 52% had a history of mental disorders.35 Third, fighting or violence is also one of the causes of TBI among patients with mental disorders.36 ,37 In addition, psychiatric patients who take psychiatric medications have reduced bone mineral density, which may be considered a risk factor for skull bone or related fractures.38 ,39
Our data showed that the post-injury mortality of TBI was severity dependent on patients' psychiatric related utilisation of medical resources before injury. Although the risk of injuries was known to be higher among patients receiving antipsychotics or antidepressants,9 ,14 we first demonstrated the dose dependent impact of psychiatric medication on the prevalence of TBI and post-injury mortality. This provided important information to surgical teams about both prediction of TBI risk among patients with mental disorders and their potential outcome after TBI for patients receiving different doses of psychiatric medications before injury. The risk for TBI and mortality was also significantly higher when psychiatric patients received specific psychiatric interventions or emergency service or hospitalisation for mental illness before TBI. As we know, inadequate insight into self-care for trauma, delay of medical treatment for TBI,15 low quality medical care17 and interactions between psychiatric and pain medications for trauma might all contribute to why psychiatric patients had a poorer outcome after TBI.16 Our data, the background healthcare information for this specific population in terms of their utility of medical resources before TBI, may be useful as clinical predictors of TBI and post-injury mortality.
According to our data, there are several clinical strategies to improve the healthcare quality of these specific populations. First, TBI patients with the top three risk group diagnoses (specified non-psychotic mental disorders due to brain damage, alcohol or drug induced mental disorders, and emotional disturbances) should be considered for special care for the prevention of TBI. Family or social support, avoiding using motor vehicles and other protective policies could minimise the potential TBI incidents. Second, a history detailing doses of psychiatric medication and other factors concerning the utilisation of medical resources in patients with mental disorders could be considered important outcome predictors when handling TBI by emergency or neurosurgical teams. Third, integrated care, including psychiatrists, trauma teams, neurosurgeons and intensivists, should be considered for better care of this specific population with TBI.
This study was limited by the data provided by insurance claims which might be underestimated because not all patients with a psychiatric illness seek medical treatment. Those who are treated are more likely to be patients with severe or less manageable psychiatric symptoms. Similarly, TBI events might be underestimated because patients who suffer from minor TBI might not seek medical services. A higher suicide rate can be expected in patients with mental disorders, but information on suicide was not included in Taiwan's National Health Insurance Research Database. Furthermore, the exact aetiologies of TBI were not available in our reimbursement data, so it was unclear when TBI resulted from violence, vehicle related collisions, falls, firearms, sports injuries, accidental or non-accidental causes.40 In this study, prescriptions do not equate with patient adherence because we had no information on whether the users were indeed taking antipsychotics. Our study successfully demonstrated that medication index, history of emergency care and hospitalisation for mental disorders were associated with TBI, but these indexes cannot totally represent the severity of disease. In addition, this was a study of association and not direct causation; it was retrospective, not prospective. Prospective studies are still needed to validate causative relationships between mental disorders and TBI and its outcomes. Finally, although we controlled for several confounders, residual confounding is always possible.14
Our case-control study provided population-based evidence that having a mental disorder is an important TBI risk factor with subtype specificity. The current studies demonstrated risk factor analysis and clinical suggestions in full spectrum, including prevention, risk assessment and outcome related information to TBI patients with mental disorders. Our data further demonstrated a dose-dependent relationship between psychiatric medication/treatment and TBI risk and outcome. We also noted that TBI victims with recent specific psychiatric treatment or inpatient services would face an increased risk of post-injury mortality. Strategies to prevent TBI and meticulous care to reduce post-injury mortality should be routinely considered for this specific population with mental disorders.
This study is based in part on data obtained from the National Health Insurance Research Database provided by the Bureau of National Health Insurance, Department of Health and managed by the National Health Research Institutes. The interpretation and conclusions contained herein do not represent those of the Bureau of National Health Insurance, Department of Health or the National Health Research Institutes.
H-CC and T-LC contributed equally to this study.
Competing interests None.
Ethics approval The National Health Research Institutes of Taiwan collects and provides the database available for public access. Patient information, including date of birth, sex, home location and income (inferred from insurance fees), was retrieved using scrambled identification and the validity of the database has been verified.24 This study was exempt from ethics approval due to patient identification numbers being decoded and scrambled to safeguard patient privacy and confidentiality. However, this study was evaluated and approved by Taiwan's National Health Research Institutes.
Provenance and peer review Not commissioned; externally peer reviewed.