Mobility Impairments in Crash-Involved Older Drivers

Abstract

We examined associations of functional impairments with vehicle crashes using telephone interviews of 244 elderly at-fault crash involved Mobile County, Alabama, drivers and 475 crash free controls, frequency matched on age and gender. Police-investigated crash reports filed in 1996 were obtained from the Alabama Department of Public Safety. After controlling for potential confounding, reported difficulty walking ≥ 1/2 mile (OR 2.0; 95% CI 1.2, 3.6), moving outdoors (OR 2.7; 95% CI 1.1, 7.0), and increasing numbers of activity limitations (p for trend=0.04) were associated with crash involvement. A 50% increased odds of crashing was observed for subjects reporting prior falls.

Impairments in functional status increase with age and are correlated with high rates of hospital admission, institutionalization and death [Judge et al 1997, Hogan et al 1999, Reuben et al 1992]. Mobility, defined as the ability to move purposefully through one’s environment, typically involves such diverse daily activities as turning in bed, transferring from lying to sitting or standing, walking, climbing stairs, getting to places out of walking distance and driving a motor vehicle. Limited mobility impacts greatly on the independence of older adults and their ability to manage both the basic self care and instrumental activities of daily living (e.g., shopping and transportation) necessary for community residence [Judge et al 1997, Verbrugge and Jette 1994, Nagi 1976]. The spectrum of complications due to impaired mobility can range from pressure sores and pneumonia in frail bed-bound subjects to falling and motor vehicle crashes in older community dwelling adults.

Our work [Sims et al 1998] and that of Koepsell and associates (1994) show that falls and vehicle crashes are associated in some older drivers. Susceptible older adults consistently demonstrate multiple impairments due to the effects of medications, chronic medical conditions, affective disorders, and visual and cognitive dysfunction [Tinetti et al 1988, Nevitt et al 1989, Sims et al 1998, Owsley et al 1998, Fitten et al 1995, Foley et al 1995, Johansson et al 1996; Koepsell et al 1994, Marottoli et al 1994].

Our hypothesis that elderly with prior vehicle crashes would demonstrate associations with other mobility limitations, including falls, was tested in a case-control study of Mobile County, Alabama, drivers, aged 65 years and older.

PATIENTS AND METHODS

The current study represents a secondary analysis of data originally organized to assess the association of diabetes and its complications with crash involvement in older Mobile County, Alabama, drivers [McGwin et al 1999]. Details of this study are described in detail elsewhere, but are recapitulated here: All 39,687 residents of Mobile County, aged 65 years and older and possessing a valid driver’s license, were identified from data tapes supplied by the Alabama Department of Public Safety (DPS). Case subjects sustained at least one motor vehicle crash between January 1, 1996 and December 31, 1996, according to the DPS data file. Because telephone interviews were conducted, telephone numbers were sought for the 1,906 subjects known to have been crash involved in 1996. These were identified for 1,507 (80%) persons. Because of limited resources, a sub-sample of 560 randomly selected case subjects were chosen to be interviewed, and 447 participated, 84 declined, 24 were unable to complete the interview due to an impairment and 6 were deceased. Police-investigated accident reports, also obtained from the DPS, permitted a determination that 249 persons were at least partially responsible for crashing, and 244 subjects completed the interviews.

A random sample of 1,900 crash free drivers was selected from the DPS file. Of 1,520 potential controls with telephone numbers, 657 were randomly selected for interviews and 475 (74.1%) completed them. Although not individually matched, subjects were accrued such that cases and controls were balanced with regard to gender and age. Trained telephone interviewers, masked to case control status, collected self-report data on demographic variables (age, gender, race, education, marital status), medical diagnoses (cataracts, arthritis, cancer, detached retina, memory problems, hearing impairment, heart disease, epilepsy, glaucoma, hypertension, kidney disease, Parkinsonism, stroke, diabetes), medications, vision and functional status, using January 1, 1996 as a reference date. Information on driving habits included self-reported annual mileage, quality of driving, level of comfort with various driving situations and types of vehicles most often driven. All subjects were screened for possible cognitive impairment with a telephone-validated version of the Short Portable Mental Status Questionnaire (SPMSQ).

STATISTICAL ANALYSES

Frequency distributions were computed for demographic, driving, medical diagnostic and medication-related variables. In order to adjust for the simultaneous effects of multiple variables on the probability of suffering an at-fault crash, logistic regression was used to generate odds ratios and 95% confidence intervals [Kleinbaum, 1994]. Crude odds ratios were adjusted for the effects of control variables (demographic factors, cognitive status, crash involvement prior to 1996 and numbers of medical conditions and medications) using separate logistic regression models.

RESULTS

Complete data was collected on 244 case and 475 control subjects. The majority of cases experienced one vehicle crash (81.0%) in 1996, with the remainder sustaining up to 3. Analyses were carried out with and without information contributed by subjects known to be cognitively impaired, as indicated by 3 or more errors on the SPMSQ, but no differences in outcomes were observed.

Table 1 illustrates the lack of statistically significant differences between crash involved and crash free subjects with respect to age and gender. African-American elderly were marginally more likely to experience a crash relative to older whites, as were subjects who drove every day. Having experienced a crash prior to 1996 and increasing numbers of miles driven annually were significantly associated with crash involvement.

Table 1.

Demographic & driving characteristics of crash involved cases & non-crash involved controls residing in Mobile County, Alabama.

	Cases (n=244)	Controls (n=475)
	%	%	OR (95% CI)
Age (years)
65–68	21.3	25.7	1.0 (referent)
69–72	25.4	24.4	1.3 (0.8–2.0)
73–77	25.8	25.7	1.2 (0.8–1.9)
78–93	27.5	24.2	1.4 (0.9–2.1)
P for trend				.21
Gender
Female	50.4	51.0	1.0 (referent)
Male	49.6	49.0	1.0 (0.8–1.4)
Race
White	74.6	80.0	1.0 (referent)
Black	23.0	16.8	1.5 (1.0–2.2)
Other	2.4	3.2	0.8 (0.3–2.0)
Quality of driving
Excellent/good	82.7	86.8	1.0 (referent)
Average/fair/poor	17.3	13.2	1.4 (0.9–2.1)
Prior crash	36.1	21.1	2.1 (1.5–3.0)
Drives everyday	9.8	10.5	1.3 (1.0–1.6)
Annual miles driven
≤4,000	26.3	35.1	1.0 (referent)
>4,000–≤8,000	26.2	21.5	1.7 (1.1–2.6)
>8,000–≤13,000	21.3	22.1	1.3 (0.9–2.0)
>13,000	26.6	21.3	1.7 (1.1–2.6)
P for trend				.03

As indicated in Table 2, the numbers of medical diagnoses and medications also did not impact upon the odds of crashing. Table 3 presents self-reported functional impairments and their relationship to vehicle crashes. Older drivers describing difficulty moving out of doors and walking one-half or more miles were crash prone, as were those with increasing numbers of functional impairments.

Table 2.

Self-reported medical conditions and medicines before January 1, 1996 among at-fault crash involved drivers & crash free Mobile County, Alabama drivers.

	Cases (n=244)	Controls (n=475)
	%	%	OR (95% CI)
No. of medical conditions
0	16.3	20.4	1.0 (referent)
1–2	43.9	44.2	1.2 (0.8–1.9)
>=3	39.8	35.4	1.4 (0.9–2.2)
P for trend				.16
No. of medicines
0	42.3	42.1	1.0 (referent)
1–2	43.4	45.5	1.0 (0.7–1.3)
>=3	14.3	12.4	1.2 (0.7–1.9)
P for trend				.65

Less than 2% of cases reported epilepsy, detached retina, diabetic retinopathy, diabetic neuropathy, Parkinson’s disease, amputations, or memory impairment.

Table 3.

Self-reported functional impairments among crashers and non-crashers in Mobile County, Alabama.¹

Impairment	Cases (n=244)	Controls (n=475)	OR crude (95% CI)		OR adjusted (95% CI)
	%	%
Carrying a heavy object 100 yards	22.3	17.5	1.4 (0.9–2.0)		1.4 (0.9–2.1)
Walking >= ½ mile	11.6	6.3	2.0 (1.1–3.3)		2.0 (1.2–3.6)
Using stairs	7.9	7.4	1.1 (0.6–1.9)		0.9 (0.5–1.7)
Moving outdoors	5.0	1.7	3.1 (1.2–7.6)		2.7 (1.1–7.0)
Walking between rooms	2.5	1.7	1.5 (0.5–4.3)		1.4 (0.5–4.3)
Getting in & out of bed	2.1	1.0	3.3 (0.8–14.0)		3.2 (0.7–13.9)
Dressing & undressing	2.5	1.1	2.4 (0.7–7.9)		2.4 (0.7–8.4)
Difficulty with:2
1 activity	18.0	14.3	1.4 (0.9–2.1)		1.4 (0.9–2.1)
2 activities	5.7	4.4	1.4 (0.7–2.9)		1.4 (0.7–3.8)
3+ activities	5.7	3.6	1.8 (0.9–3.7)		1.8 (0.8–3.8)
P for trend				.03		.04
≥1 fall in 1995	14.0	9.5	1.6 (0.9–2.6)		1.5 (0.9–2.6)
Falls requiring medical attention	3.7	3.2	1.2 (0.5–2.7)		1.2 (0.5–2.9)

¹ =Crude odds ratios adjusted for the effects of age, race, gender, miles driven annually, previous vehicle crash, cognitive status and numbers of diagnoses and medications, using logistic regression. One percent or fewer cases reported difficulty with feeding, using the lavatory, or washing and bathing.

² =The sum of self-reported functional impairments.

Statistically marginal associations of incident vehicle crashes with reports of one or more falls in the prior year and problems carrying a heavy object 100 yards were also noted. The odds of crashing were increased for subjects describing difficulty dressing and undressing, walking between rooms, getting out of bed and falls requiring medical attention, but these results were not statistically significant.

DISCUSSION

The findings from this study support the hypothesis that crash prone older drivers will describe other functional limitations, particularly those associated with mobility. While generally fit, elderly reporting difficulty walking a half mile or more and moving out of doors were more likely than unaffected controls to have had a crash. A potential association of crashing with prior falling is also noted.

While not directly addressing mobility issues in crash involved seniors, prior studies are supportive of a crash impaired mobility relation: Marottoli and associates observed a 2.3-fold relative risk of an automobile crash, a moving violation or being stopped by police in older subjects, who described fewer blocks walked (1994). In a case-control study of older health maintenance organization members who were injured in state-recorded motor vehicle crashes, Koepsell and colleagues noted a similarly marginal association of crashing with falls (OR 1.4; 95% CI 0.9–2.4) (1994). Although the fall crash association is weak, several studies document highly similar, if not identical, risk factors for falls and crashes in the elderly [Tinetti et al 1988, Nevitt et al 1989, Sims et al 1998, Owsley et al 1998, Fitten et al 1995, Foley et al 1995, Johansson et al 1996; Koepsell et al 1994, Marottoli et al 1994]. For example, sedative use, depression and cognitive impairment, among others, have been implicated in these events [Tinetti et al 1988 and 1995, Hemmelgarn et al 1997, Sims et al 1998, Owsley et al 1998, Fitten et al 1995].

Mobility impairments in older adults appear to predict more general disability and susceptibility to other geriatric syndromes. In a large prospective study of independent, community dwelling elderly, Guralnik and colleagues documented strong associations of incident disability in the activities of daily living and mobility-related impairments in subjects scoring in the lowest range of a physical performance battery (1995). A more recent longitudinal study reported high rates of subsequent functional disability, institutionalization and death after 4 years among subjects, who at baseline described difficulty or dependence in basic self-care activities [Gill et al 1998]. Similar results are described in other elderly populations [Hogan et al 1999, Ostir et al 1998, Ho et al 1997].

Tinetti and associates reported on the co-occurrence of falls, incontinence and functional disability in a prospective cohort study of more than 1000 community-living elderly (1995). Shared risk factors for incontinence, falling and functional dependence included slow timed chair stands, decreased arm strength, vision and hearing limitations and either anxiety or depression. The observed relation between vehicle crashes, functional impairments and possibly falls is in concert with these results.

Unlike other publications on older driver crashes, case status in the current study was based upon police-investigated crash reports, from which responsibility for the crashes could be ascertained. This advantage limited the effects of recall bias and misclassification in the current study. The availability of information on driving habits also permitted adjustment for driving exposure (i.e., annual miles driven), an important confounding variable.

Our results should be cautiously interpreted due to several limitations. The study from which the data for this report was derived depended on self-report for information on the independent variables [McGwin et al, 1999]. This is particular concern for medical diagnoses, which may have been undiagnosed and of which cases and controls could have been unaware. Haapanen and colleagues, however, noted the general reliability of self-report diagnoses, for which easily communicated criteria exist (1997). We also considered only the number of chronic diseases, rather than relying on specific diagnoses. For roughly 20% of potentially eligible cases and controls, telephone numbers could not be identified, but information on age and gender from the DPS data tapes revealed that there were no meaningful differences between subjects with and without telephone numbers. With regard to the quality of information provided by persons with possible cognitive impairment, defined by 3 or more errors on the SPMSQ, analyses were carried out with and without data from these individuals. No significant differences in study outcomes were observed. Thus, the results of the current study are unlikely to be changed by these limitations.

Crash involvement among older drivers has been linked to other mobility impairments, which according to the literature appear to predict further functional declines. This raises speculation that interventions aimed at one adverse event (e.g., vehicle crashes) could reduce the incidence of others (e.g., falls). Further research is warranted on the potential relationships among mobility impairments in elderly persons.