The Association Between Depression and Traumatic Brain Injury in Older Adults: A Nested Matched Case Control Study

Abstract

Objective: Determine association between depression and traumatic brain injury (TBI) in the older adult home care population of Ontario, Canada (2003-2013). Method: A nested matched case control study was used, and data were retrieved for service users 65 years or older who received home care between 2003 and 2013. Incident TBI cases were matched to four controls by sex, age, and assessment date. Odds ratios and multivariable conditional logistic regression were completed. Results: 554,313 service users were included of which 5,215 (0.9%) had a TBI and 39,048 (7.0%) had depression. Fall history was an effect modifier. The association between depression and TBI was 1.10 (95% confidence interval [CI] = [0.93, 1.31]) for those without a history of falling, while for those with a history of falling it was 1.24 (95% CI = [1.03, 1.48]) after adjusting for education and Alzheimer’s disease. Discussion: Depression is associated with sustaining a TBI in those with a history of falling.

Keywords

traumatic brain injury older adults depression and epidemiology

Introduction

Traumatic brain injury (TBI) is one of the most common causes of injury in youth and older adults (Ghajar, 2000; Langlois, Rutland-Brown, & Thomas, 2004; Langlois, Rutland-Brown, & Wald, 2006). Sustaining a TBI can have a number of long-term consequences that can affect an individual’s emotional (e.g., depression; see Ghajar, 2000; Langlois et al., 2006), physical (e.g., epilepsy and Alzheimer disease; see Holsinger et al., 2002; Horner et al., 2005; Plassman et al., 2000), and behavioral well-being (e.g., memory problems; see Ghajar, 2000; Langlois et al., 2006). The incidence of TBI in older adults aged 65 and above has been estimated at 155.9 per 100,000 population (Coronado, Thomas, Sattin, & Johnson, 2005) and costs more than CA$20 billion yearly (Coronado et al., 2005; Thompson, McCormick, & Kegan, 2008). Falls and motor vehicle collisions are the main cause of TBI in older adults (Ghajar, 2000); however, little is known regarding the risk factors associated with incident TBI cases. Research has shown that older adults are 40% more likely to have a medical condition prior to sustaining a TBI compared with younger adults (Kennedy et al., 2002; Stevens, Corso, Finkelstein, & Miller, 2006; Thompson et al., 2008). Any illness or chronic conditions associated with cognitive impairment or dementia (e.g., Alzheimer disease or other etiologies) are hypothesized risk factors for TBI (Plassman et al., 2000; Starkstein & Jorge, 2005).

Depression is a prevalent medical condition, especially in the older adult population. The accumulation of a long life of depressive events can put older adults at a particularly high risk of depression (Anderson, 2002). An estimated 8.9% of older adults have depression (Blazer, Hughes, & George, 1987) and depression is associated with a 16% mortality rate in older adults (Meats, Timol, & Jolley, 1991). As the North American population ages, it is hypothesized that the prevalence of depression is going to increase (Menzel, 2008; Swinkels, Neuman, & Allain, 2009).

Many studies have found that the occurrence of a TBI increases the likelihood of developing depression (Blazer et al., 1987; Meats et al., 1991; Menzel, 2008); however, depression has not been investigated as a risk factor for sustaining a TBI.

Medications used to treat depression have side effects that may increase the likelihood of falling and consequently sustaining a TBI (Wiese, 2011). Selective serotonin reuptake inhibitors (SSRIs) and tricyclic antidepressants are commonly used to treat geriatric depression. Approximately 10% of patients taking SSRIs develop hyponatremia, which can lead to an increased risk of falling (odds ratio [OR]: 1.84; 95% confidence interval [CI] = [0.91, 3.69]; Arfken, Wilson, & Aronson, 2001; Ruthazer & Lipsitz, 1993). Systematic reviews have found that tricyclic antidepressant use is associated with an increased risk of falling (OR: 1.48; 95% CI = [1.23, 1.77]; Ensrud et al., 2002; Leipzig, Cummings, & Tinetti, 1999).

Research shows that antidepressants are associated with an increased risk of falling; however, the relationship between depression and sustaining a TBI has not been fully investigated. The objective of this study is to determine the association between depression and TBI in older adults using home care in Ontario from 2003 to 2013. This study will also assess the modifying effects of having a history of falling. It is hypothesized that TBI will be positively associated with the presence of depression in the older adult population. It is further hypothesized that the magnitude of effect will be greater for those with a history of falling than for those without.

Method

Data Source

The data source used for this study was the Ontario Association of Community Care Access Centers (OACCAC) home care database. The OACCAC database contains longitudinal assessments of service users expected to be in Ontario home care for 60 or more days (InterRAI, 2009). Community Care Access Centers (CCAC) administer home care in Ontario and determine the eligibility for long-term care (Ontario Ministry of Health and Long-Term Care, 2006). Service users are assessed approximately every 6 months for the duration of their service, or when there is a significant change in status (which includes sustaining a TBI; Ontario Ministry of Health and Long-Term Care, 2006).

Assessment Instrument

Analyses are based on the Resident Assessment Instrument–Home Care (RAI-HC) assessment system, which is used regularly in home-based health care settings. Items in the RAI-HC have been validated by a variety of international studies (Carpenter et al., 2004; Chi Chou, Kwan, Lam, & Lam, 2006; InterRAI, 2014; Landi et al., 2000; Morris et al., 1997). The RAI-HC assessment is administered by trained assessors (nurses or social workers), who use all sources of information available to complete all sections of the instrument. The RAI-HC has a variety of items, many of which are organized into scales and clinical assessment protocols (e.g., Depression Rating Scale) to help interpret the meaning of the items.

Study Design

A nested matched case control study was used to determine the association between depression and sustaining a TBI in older adults using home care in Ontario. Individuals were eligible for inclusion in this study if they were aged 65 years or older and were administered home care between 2003 and 2013. Cases and controls were selected from the cohort where follow-up began at the time of initial assessment and ended at the time of first TBI (for cases), or discharge from home care, relocation outside of Ontario, death, or the end of the assessment period in 2013 for controls. All persons included in the study had an initial/admission assessment and at least one reassessment. Each case was randomly matched to four controls (individuals without a TBI at the time of the case’s TBI) by date of cohort entry (±3 months of admission to home care), age (±1 year), and sex to control for potential confounding factors, including time at risk. Matching was used because age and sex are known to be associated with sustaining a TBI (Ghajar, 2000; Langlois et al., 2004; Langlois et al., 2006) and controlling for time at risk is always prudent when conducting a matched case control study (Goldstein & Zhang, 2009). Goldstein and Zhang (2009) have shown that using four controls per case minimizes the amount of efficiency lost when performing a nested case control study compared with a full cohort study analysis (Goldstein & Zhang, 2009). Ethical review was not required to conduct this study due to the use of anonymized data used for secondary analysis (Canadian Institute of Health Research, National Sciences and Engineering Research Council of Canada, & Social Sciences and Humanities Research Council of Canada, 2014).

Study Measures

TBI

The primary outcome for the study was incident cases of TBI. There are two ways to capture TBI in the RAI-HC data: the “Head Trauma” item and text entries (InterRAI, 2009). The definition of Head Trauma as stated in the RAI-HC user manual is “damage to the brain as a result of physical injury to the head” (InterRAI, 2009). TBI is indicated as present on a service users’ RAI-HC if the doctor has indicated it affects the service users’ status, requires treatment, or symptom management. TBI is also indicated as present if the disease is monitored by a home care professional or led to a hospitalization in the 90 days prior to the RAI-HC (or since last assessment if less than 90 days; InterRAI, 2009). Cases were considered incident because nurses or social workers completing the RAI-HC were instructed not to include past TBI, which ensures that only new cases were captured. If a service user had more than one TBI (either during a 1-year period or during the whole 10-year period from 2003 to 2013), then only the first TBI was counted as a case and subsequently matched to four controls for the analysis. On the RAI-HC, TBI is coded as either not present (leaving the check box blank), (1) present—not subject to focused treatment or monitoring by home care professional, or (2) present—monitored or treated by home care professional (InterRAI, 2009). For this study, we recoded TBI as present if either (1) or (2) were indicated on the RAI-HC. In addition to capturing TBI using the “Head trauma” item, text entries were used to capture TBI. Any text diagnosis that referred to head trauma, concussion, closed head injury, head injury, or acquired brain injury was also used to capture cases of TBI. This method of capturing TBI from RAI-HC data has been validated by Foebel et al. (2013). They compared the incidence of TBI from RAI-HC data with TBI incidence from the Canadian Institute for Health Information (CIHI), and the National Ambulatory Care Reporting System (NACRS), which captures inpatient hospital and emergency department records, to determine reliability and validity. They found the RAI-HC assessment to have a sensitivity of 0.23, a specificity of 0.99, a positive predictive value of 0.22, and a kappa statistic of 0.22 when comparing the RAI-HC measure of TBI with CIHI and NACRS data (Foebel et al., 2013).

Depression

Depression was measured from the RAI-HC assessment conducted at the time of TBI for cases, and from the assessment nearest to the index date (time of matched case’s TBI) for the controls. Depression was assessed using seven items from the mood, anxiety, and behaviors section of the RAI-HC. Items are coded similar to TBI, and a service user was considered depressed if they had a score greater than or equal to 3 out of 14 on the seven-item scale (Burrows, Morris, Simon, Hirdes, & Phillips, 2000). The scale has been validated using the Cornell Scale for Depression and the Hamilton Depression Rating Scale (Burrows et al., 2000). Specifically, a cut-point score of 3 or greater on the Depression Rating Scale maximized sensitivity (0.78 for Cornell and 0.94 for Hamilton) with minimum loss of specificity (0.77 for Cornell and 0.72 for Hamilton) when tested against cutoffs for mild to moderate depression (Burrows et al., 2000).

Antidepressants

Antidepressant use was measured as a potential confounder and effect modifier using the RAI-HC (InterRAI, 2009). Antidepressant use was captured as a categorical variable, under the receipt of psychotropic medication heading. Antidepressant use is considered present if antidepressants were taken in the last 7 days (or since last assessment). For cases, antidepressant use was captured at the time of TBI and for controls from the assessment date nearest to the index data (time of matched case’s TBI).

Demographics

Demographic characteristics were obtained from the RAI-HC and included sex, age, aboriginal origin, and highest level of education completed (InterRAI, 2009). For cases, demographics were measured at the time of TBI and for controls from the assessment nearest to the index date (time of matched case’s TBI).

History of falling

History of falling was measured as a potential confounder and effect modifier using the RAI-HC. History of falling was captured as a categorical variable, under the heading of fall frequency (InterRAI, 2009). Fall frequency is reported as the number of times fallen in the last 90 days (or since the last assessment if less than 90 days). On the RAI-HC, fall frequency is coded as “0” for none or “9” if more than nine falls (InterRAI, 2009). For this study, history of falling was recoded as one or more falls or no falls. For cases, history of falling was measured at the time of TBI and for controls from the assessment nearest to the index date (time of matched case’s TBI).

Neurological diseases

Neurological diseases were assessed as potential effect modifiers because their presence may influence the association between depression and TBI. Neurological disease was reported using the list of disease diagnoses and included: Alzheimer disease, dementia other than Alzheimer, and Parkinsonism (InterRAI, 2009). Neurological diseases were ascertained as present similar to depression and the measures of neurological disease have been validated (Foebel et al., 2013). For cases, neurological diseases were measured at the time of TBI and for controls from the assessment nearest to the index date.

Data Analysis

Univariate descriptive analyses were conducted on all variables to check for errors and outliers, examine distributions, and to examine responses for each of the variables. Counts and proportions were determined for all variables and means and standard deviations were reported for continuous variables.

A crude bivariate association between depression and TBI was determined using a contingency table and OR with 95% confidence interval. A sensitivity analysis of the crude unmatched association between depression and TBI was conducted using the validity results published by Foebel et al. (2013) assuming nondifferential misclassification in the TBI outcome measure. A sensitivity of 0.23 from Foebel and colleagues (Ontario Ministry of Health and Long-Term Care, 2006) indicates that an adjustment for TBI will require 77% more cases. Thus, 77% more cases were added to both the exposed and unexposed case group and the adjusted measure of association was then calculated and compared with our unadjusted measure. Results were stratified by history of falling and effects were estimated with the Cochran-Mantel-Haenszel OR. Multivariable analysis was conducted using conditional logistic regression modeling to estimate the OR and 95% confidence interval for the association between TBI and depression. A three-step modeling process was used (Kleinbaum & Klein, 2010). First, the modifying effects of history of falling, antidepressant use, Alzheimer disease, dementia and Parkinsonism on the association between TBI and depression were tested. Effect modifiers were considered statistically significant if the p value of the interaction term was less than or equal to .05. Potential confounders as identified from the survey of the literature were included in the gold standard (GS) model as being potential confounders (Kleinbaum & Klein, 2010): sex, aboriginal origin, age, marital status, falls frequency, antidepressant use, Alzheimer disease, dementia, and Parkinsonism. Second, the potential confounding factors were evaluated by methodically determining different subsets of potential confounders that gave comparable estimates of effect as the GS model (i.e., within 10% of the GS model; Kleinbaum & Klein, 2010). The final model was selected based on which subset model was closest to the GS model (Kleinbaum & Klein, 2010). If more than two models gave similar estimates compared with the GS model, then the most parsimonious subset with acceptable precision was selected as the final, adjusted model. All data were analyzed using SAS software, version 9.4 (SAS Institute, 2015).

Results

The total sample size for this study was 554,313 service users, of which 5,215 (0.9%) had sustained a TBI (Table 1). Of the 5,215 cases of TBI, 4,188 were captured using the “head trauma” item and 1,027 were captured using the text entries section of the RAI-HC. The selected controls were found to be representative of service users without TBI, except for on the matched variables, which had a distribution identical to the cases (Table 1).

Table 1.

Univariate Descriptive Analyses of the Characteristics of Service Users Without TBI, Matched Controls (Matched on Age (±1 Year), Sex, and Date of Admission (±3 Months) to Home Care) and TBI Cases in the Older Adult Home Care Population of Ontario from 2003 to 2013.

Variable	Service users without TBI	Matched controls	TBI cases
N	549,098	20,823	5,215
Sex
Male	200,094 (36.4)	9,741 (46.8)	2,442 (46.8)
Female	348, 972 (63.6)	11,082 (53.2)	2,773 (53.2)
Aboriginal origin
Yes	3,900 (0.7)	134 (0.6)	73 (1.4)
No	545,167 (99.3)	20,689 (99.4)	5,141 (98.6)
Missing	31		1
Age
65-69	43,896 (8.0)	2,505 (12.0)	635 (12.2)
70-74	65,189 (11.9)	3,084 (14.8)	775 (14.9)
75-79	101,609 (18.5)	4,146 (19.9)	1,060 (20.3)
80-84	137,949 (25.1)	5,175 (24.9)	1,252 (24.0)
85-89	121,557 (22.1)	3,900 (18.7)	998 (19.1)
>90	78,898 (14.4)	2,013 (9.7)	495 (9.5)
M (SD)	81.9 (7.6)	80.2 (7.7)	80.2 (7.7)
Education level
8th grade or less	132,442 (24.1)	5,466 (26.3)	1,259 (24.1)
9th-12th grade	175,281 (31.9)	7,624 (36.6)	1,657 (31.8)
Post-secondary	105,488 (19.2)	3,854 (18.5)	1,248 (23.9)
Graduate degree	11,945 (2.2)	468 (2.3)	162 (3.1)
Unknown	123,895 (22.6)	3,409 (16.4)	888 (17.1)
Missing	47	2	1
Depression
Yes	38,497 (7.0)	1,435 (6.9)	551 (10.6)
No	510,601 (93.0)	19,388 (93.1)	4,664 (89.4)
Antidepressant use
Yes	111,401 (20.3)	5,091 (24.5)	1,435 (27.5)
No	437,697 (79.7)	15,732 (75.6)	3,780 (72.5)
Falls frequency
No falls	356,871 (65.0)	14,115 (67.8)	2,323 (44.5)
One or more falls	192,183 (35.0)	6,708 (32.2)	2,892 (55.5)
Missing	44	3
M (SD)	0.7 (1.4)	0.7 (1.4)	1.34 (1.97)
Alzheimer disease
Yes	45,401 (8.3)	1,950 (9.4)	439 (8.4)
No	503,697 (91.7)	18,873 (90.6)	4,776 (91.6)
Dementia
Yes	82,257 (15.0)	2,914 (14.0)	1,174 (22.5)
No	466,841 (85.0)	17,909 (86.0)	4,041 (77.5)
Parkinsonism
Yes	22,202 (4.0)	962 (4.6)	254 (4.9)
No	526,896 (96.0)	19,861 (95.4)	4,961 (95.1)

Note. TBI = traumatic brain injury.

A total of 5,209 of the original 5,215 cases could be matched to four controls; the remaining six cases were matched to three controls. All 5,215 cases were included in the regression analyses. Some controls were matched to more than one case and the total matched case control sample size was 26,038. Table 2 shows the number of matched controls with depression for cases with and without depression. The OR of the crude association between TBI and depression was 1.54 (95% CI = [1.43, 1.64]).

Table 2.

Exposure Among Matched Sets (Matched on Age, Sex, and Date of Admission to Home Care) Showing Number of Cases and Matched Controls With a History of Depression in the Older Adult Home Care Population of Ontario From 2003 to 2013 (N = 5,215 Matched Pairs).

Cases with depression	Number of controls with depression					Total number of sets
Cases with depression	0	1	2	3	4	Total number of sets
Yes	377	149	25	0	0	551
No	3,432	1,172	52	3	5	4,664

A sensitivity analysis was conducted on the outcome measure of TBI using the sensitivity of 0.23 as determined by Foebel and colleagues (30). An OR of 1.77 (95% CI = [1.66, 1.85]) was determined using adjusted, nonmatched data. This suggests our matched results are a conservative estimate.

Multivariable analysis was conducted using conditional logistic regression modeling to estimate adjusted ORs and 95% confidence intervals for the association between depression and TBI. Analyses suggested that education level (OR: 1.98, 95% CI = [1.78, 2.11]), history of falling (OR: 2.61, 95% CI = [2.46, 2.77]), and Alzheimer disease (OR: 0.89, 95% CI = [0.80, 0.99]) were significantly associated with sustaining a TBI. Furthermore, effect modification and confounding were assessed. History of falling was determined to be an effect modifier. The association between depression and TBI was 1.10 (95% CI = [0.93, 1.31]) for those without a history of falling, while for those with a history of falling it was 1.24 (95% CI = [1.03, 1.48]) after adjusting for level of education and Alzheimer disease.

Discussion

The crude analysis revealed a positive association between depression and TBI, and after controlling for confounders, history of falling was found to be an effect modifier in that association. True to the initial hypothesis, the association between depression and TBI was smaller in magnitude among those without a history of falling compared with those with a history of one or more falls, after adjusting for other factors (i.e., level of education, Alzheimer disease). Given the low sensitivity of the TBI outcome measure, all ORs are likely underestimates of the true association between depression and TBI.

Depression following TBI has been investigated extensively and found to occur more frequently in persons who suffered a TBI (Fedoroff et al., 1992; Jorge et al., 2004; Kreutzer, Seel, & Gourley, 2001); however, depression has not been specifically investigated as a potential risk factor for sustaining a TBI using validated measures. A number of factors could explain why the observed association was not larger. For example, service users who are depressed may not be receiving treatment, which would decrease their likelihood of falling as various studies have indicated that depression medications increase the likelihood of falling (Arfken et al., 2001; Ensrud et al., 2002; Leipzig et al., 1999; Ruthazer & Lipsitz, 1993). Depression could also be undetected or undiagnosed. Depression also leads to a more sedentary lifestyle in older adults and this may decrease the likelihood of falling and sustaining a TBI (Hart, 2003; Sosin, Sniezek, & Thurman, 1996).

Research has shown that falls and fall-induced injuries in the older adult population are increasing in frequency and that approximately 35% of older adults aged 65 and above living in the community will fall every year, of which about half fall recurrently (Kannus, Sievanen, Palvanen, Jarvinen, & Parkkari, 2005). Our study is the first to formerly assess the association between depression and TBI and consequently the first study to investigate the modifying effects of falls on this relationship. However, the literature has shown that depression is associated with an increased risk of falling (Hausdorff, Rios, & Edelberg, 2001) and that there are common risk factors that can predict both an increased risk of falling and depression (poor self-rated health, poor cognitive status, impaired activities of daily living, two or more primary care clinic visits in the past month, and slow walking speed; Biderman, Cwikel, Fried, & Galinsky, 2002). Based on this information, it is plausible that the association between depression and TBI would be modified by a history of falling, and our results confirmed our initially stated hypothesis.

The results of the sensitivity analysis indicate that the study findings are likely an underestimation of the true association between depression and TBI. This was expected due to the low sensitivity in the TBI outcome measure. An accurate case definition and operational measure of TBI need to be developed to improve the accuracy of etiologic associations between various conditions and TBI.

The limitations of this study were the use of a low sensitivity instrument to capture TBI, the amalgamation of all severities of TBI under one diagnosis, the inability to examine the cause of TBI and history of TBI as potential confounders, the inability to examine recurrent TBIs (using Poisson regression) due to the small numbers of multiple TBIs, and the inability to establish causality. The sensitivity was poor for the TBI outcome measure; however, because the misclassification is considered nondifferential in nature, the estimate is biased toward the null. This indicates that the study estimate is conservative.

The impact of TBI is widespread and is a major public health concern among older adults. This etiologic study has investigated a major mental health condition among older adults and its association with sustaining a TBI. The results suggest that depression is associated with TBI; however, the results cannot be interpreted as causal due to the study design. Longitudinal studies should be conducted in the future to ensure the temporal relationship between depression and TBI.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Dr. Kristman is supported by the Canadian Institutes of Health Research through a New Investigator Award in Community-Based Primary Health Care.

References

Anderson

A. J.

(2002). Treatment of depression in older adults. International Journal of Psychosocial Rehabilitation, 6, 69-78.

Arfken

C. L.

Wilson

J. G.

Aronson

S. M.

(2001). Retrospective review of selective serotonin reuptake inhibitors and falling in older nursing home residents. International Psychogeriatrics, 13, 85-91.

Biderman

Cwikel

Fried

A. V.

Galinsky

(2002). Depression and falls among community dwelling elderly people: A search for common risk factors. Journal of Epidemiology and Community Health, 56, 631-636.

Blazer

Hughes

George

(1987). The epidemiology of depression in an elderly community population. The Gerontologist, 27, 281-287.

Burrows

A. B.

Morris

J. N.

Simon

S. E.

Hirdes

J. P.

Phillips

(2000). Development of a minimum data set-based depression rating scale for use in nursing homes. Age and Ageing, 29, 165-172.

Canadian Institute of Health Research, National Sciences and Engineering Research Council of Canada, & Social Sciences and Humanities Research Council of Canada. (2014). Tri-council policy statement—Ethical conduct for research involving humans. Retrieved from http://www.pre.ethics.gc.ca/pdf/eng/tcps2-2014/TCPS_2_FINAL_Web.pdf

Carpenter

Gambassi

Topinjova

Schroll

Finne-Soveri

Henrard

J. C.

. . . Bernabei

(2004). Community care in Europe. The aged in home care project (AdHOC). Aging Clinical and Experimental Research, 16, 259-269.

Chi

Chou

Kwan

C. W.

Lam

E. K.

Lam

T. P.

(2006). Use of the minimum data set—home care: A cluster randomized controlled trial among the Chinese older adults. Aging & Mental Health, 10, 33-39.

Coronado

V. G.

Thomas

K. E.

Sattin

R. W.

Johnson

R. L.

(2005). CDC traumatic brain injury surveillance system: Characteristics of persons aged 65 and older hospitalized with a TBI. Journal of Head Trauma Rehabilitation, 20, 215-228.

10.

Ensrud

K. E.

Blackwell

T. L.

Mangione

C. M.

Bowman

P. J.

Whooley

M. A.

Bauer

D. C.

. . . Nevitt

M. C.

(2002). Central nervous system-active medications and risk for falls in older women. Journal of the American Geriatrics Society, 50, 1629-1637.

11.

Fedoroff

P. J.

Starkstein

S. E.

Forrester

A. W.

Geisler

F. H.

Jorge

R. E.

Arndt

S. V.

Robinson

R. G.

(1992). Depression in patients with acute traumatic brain injury. The American Journal of Psychiatry, 149, 918-923.

12.

Foebel

A. D.

Hirdes

J. P.

Heckman

G. A.

Kergoat

M. J.

Patten

Marrie

R. A.

(2013). Diagnostic data for neurologic conditions in interRAI assessments in home care, nursing home and mental health care settings: A validity study. BMC Health Services Research, 13, 457-468.

13.

Ghajar

(2000). Traumatic brain injury. The Lancet, 356, 923-929.

14.

Goldstein

Zhang

(2009). Efficiency of the maximum partial likelihood estimator for nested case control sampling. Bernoulli, 15, 569-597.

15.

Hart

L. E.

(2003). Physical activity and depression in older adults. Clinical Journal of Sport Medicine, 13, 274-277.

16.

Hausdorff

J. M.

Rios

D. A.

Edelberg

H. K.

(2001). Gait variability and fall risk in community-living older adults: A 1-year prospective study. Archives of Physical Medicine and Rehabilitation, 82, 1050-1056.

17.

Holsinger

Steffens

D. C.

Phillips

Helms

M. J.

Havlik

R. J.

Breitner

J. C.

Plassman

B. L.

(2002). Head injury in early adulthood and the lifetime risk of depression. Archives of General Psychiatry, 59, 17-22.

18.

Horner

M. D.

Ferduson

P. L.

Selassie

A. W.

Labbate

L. A.

Kniele

Corrigan

J. D.

(2005). Patterns of alcohol use one year after traumatic brain injury: A population based, epidemiological study. Journal of the International Neuropsychological Society, 11, 322-330.

19.

InterRAI. (2009). Home care (HC) assessment form and user’s manual (Version 9.1). Retrieved from http://catalog.interrai.org/HC-home-care-manual

20.

InterRAI. (2014). Home care. Retrieved from http://www.interrai.org/home-care.html

21.

Jorge

Robinson

R. G. E.

Moser

Tateno

Crespo-Facorro

Arndt

(2004). Major depression following traumatic brain injury. Archives of General Psychiatry, 61, 42-50.

22.

Kannus

Sievanen

Palvanen

Jarvinen

Parkkari

(2005). Prevention of falls and consequent injuries in elderly people. The Lancet, 366, 1885-1893.

23.

Kennedy

R. L.

Henry

Chapman

A. J.

Nayar

Grant

Morris

A. D.

(2002). Accidents in patients with insulin-treated diabetes: Increased risk of low-impact falls but not motor vehicle crashes—A prospective register-based study. Journal of Trauma-Injury Infection & Critical Care, 52, 660-666.

24.

Kleinbaum

D. G.

Klein

(2010). Modeling strategy for assessing interaction and confounding. In Kleinbaum

D. G.

Klein

(Eds.), Logistic regression: A self-learning text (pp. 203-240). New York, NY: Springer.

25.

Kreutzer

J. S.

Seel

R. T.

Gourley

(2001). The prevalence and symptom rates of depression after traumatic brain injury: A comprehensive examination. Brain Injury, 15, 563-576.

26.

Landi

Tua

Onder

Carrara

Sgadari

Rinaldi

. . . Bernabei

(2000). Minimum data set for home care: A valid instrument to assess frail older people living in the community. Medical Care, 38, 1184-1190.

27.

Langlois

J. A.

Rutland-Brown

Thomas

K. E.

(2004). Traumatic brain injury in the United States: Emergency department visits, hospitalizations, and deaths. Atlanta, GA: Centers for Disease Control and Prevention, National Center for Injury Prevention and Control.

28.

Langlois

J. A.

Rutland-Brown

Wald

M. W.

(2006). The epidemiology and impact of traumatic brain injury. The Journal of Head Trauma Rehabilitation, 21, 375-378.

29.

Leipzig

R. M.

Cummings

R. G.

Tinetti

M. E.

(1999). Drugs and falls in older people: A systematic review and meta-analysis. Journal of the American Geriatrics Society, 47, 30-39.

30.

Meats

Timol

Jolley

(1991). Prognosis of depression in the elderly. British Journal of Psychology, 159, 659-663.

31.

Menzel

J. C.

(2008). Depression in the elderly after traumatic brain injury: A systematic review. Brain Injury, 22, 375-380.

32.

Morris

Fries

B. E.

Steel

Ikegami

Bernabei

Carpenter

G. I.

. . . Topinková

(1997). Comprehensive clinical assessment in community setting: Applicability of the MDS-HC. Journal of the American Geriatrics Society, 45, 1017-1024.

33.

Ontario Ministry of Health and Long-Term Care. (2006). Eligibility criteria for CCAC services. Retrieved from http://www.health.gov.on.ca/english/providers/pub/manuals/ccac/ccac_3.pdf

34.

Plassman

B. L.

Havlik

R. J.

Steffens

D. C.

Helms

M. J.

Newman

T. N.

Drosdick

. . . Breitner

J. C.

(2000). Documented head injury in early adulthood and risk of Alzheimer’s disease and other dementias. Neurology, 55, 1158-1166.

35.

Ruthazer

Lipsitz

L. A.

(1993). Antidepressants and falls among elderly people in long- term care. American Journal of Public Health, 85, 746-749.

36.

SAS Institute. (2015). SAS statistical software (Version 9.4). Cary, NC: Author.

37.

Sosin

D. M.

Sniezek

J. E.

Thurman

D. J.

(1996). Incidence of mild and moderate brain injury in the United States, 1991. Brain Injury, 10, 47-54.

38.

Starkstein

S. E.

Jorge

(2005). Dementia after traumatic brain injury. International Psychogeriatrics, 17, S93-S107.

39.

Stevens

J. A.

Corso

P. S.

Finkelstein

E. A.

Miller

T. R.

(2006). The costs of fatal and non-fatal falls among older adults. Injury Prevention, 12, 290-295.

40.

Swinkels

Neuman

J. H.

Allain

T. J.

(2009). A prospective observational study of falling before and after knee replacement surgery. Age Ageing, 38, 175-181.

41.

Thompson

H. J.

McCormick

W. C.

Kegan

S. H.

(2008). Traumatic brain injury in older adults: Epidemiology, outcomes and future implications. Journal of the American Geriatrics Society, 54, 1590-1595.

42.

Wiese

B. S.

(2011). Geriatric depression: The use of antidepressants in the elderly. British Columbia Medical Journal, 53, 341-347.