The Influence of Hearing Impairment on Driving Avoidance Among a Large Cohort of Older Drivers

Abstract

As people age, some of the commonly experienced psychomotor, visual, and cognitive declines can interfere with the ability to safely drive, often leading to situational avoidance of challenging driving situations. The effect of hearing impairment on these avoidance behaviors has not been comprehensively studied. Data from the American Automobile Association (AAA) Longitudinal Research on Aging Drivers (LongROAD) study were used to assess the effect of hearing impairment on driving avoidance, using three measures of hearing. Results indicated that hearing loss plays a complex role in driving avoidance, and that an objective hearing measure was a stronger predictor than hearing aid use and self-rated hearing. Greater hearing impairment was related to less nighttime and freeway driving, more trips farther than 15 mi from home, and lower odds of avoiding peak driving times. The moderating influence of hearing on both vision and cognition is also discussed, along with study implications and future research.

Keywords

driving hearing vision perceptual decline sensory problems driving exposure

Introduction

Prior to driving cessation, most older adults engage in driving reduction behaviors, including less driving overall and avoiding challenging situations such as driving at night, on freeways, during peak traffic times, in poor weather, in unfamiliar areas, and long distances (Braitman & Williams, 2011; Dickerson et al., 2019; McGuckin & Fucci, 2018; Ryvicker et al., 2020). Such behaviors may allow older adults to continue providing for their own mobility for a longer time.

Factors associated with driving reduction and avoidance include the presence of health conditions, increased frailty, cognitive impairment, concern about crashing, lack of driving comfort, lifestyle changes, and perceptual (especially visual) declines (Crowe et al., 2020; Devlin & McGillivray, 2014; Dickerson et al., 2007; Kandasamy et al., 2018; Molnar et al., 2013; Ragland et al., 2004). The influence of vision problems on driving reduction/avoidance has been studied extensively. For example, better self-rated vision is significantly associated with distinguishing drivers from nondrivers (Anstey et al., 2017) and less situational driving avoidance (e.g., intersections and nighttime driving; Charlton et al., 2006). In a U.S.-based longitudinal study, objectively measured visual acuity was related to driving cessation, whereas glaucoma and macular degeneration were not significant (Edwards et al., 2008). In that same study, however, visual acuity became nonsignificant after controlling for health and physical/cognitive performance. Another longitudinal study assessed the relationship between driving cessation and several different vision problems. Worse visual acuity, contrast sensitivity, and losses in the lower peripheral visual field were significantly related to driving cessation (Freeman et al., 2005). In other work, near visual acuity, contrast sensitivity, and useful field of view were all predictive of driving cessation, but not moving violations or crashes (Emerson et al., 2012). Although some of these findings are mixed, vision has been widely studied and findings generally suggest that it plays a role in driving outcomes. However, the type of visual problem (e.g., loss of acuity, contrast sensitivity, and visual field) seems to matter.

The evidence related to hearing impairment and driving outcomes is less clear, but critical to understand. Almost 25% of Americans aged 65 to 74 years have disabling hearing loss, as do 50% of those aged 75 years or older (National Institutes of Health, 2016). The potential effect of hearing loss on various driving outcomes have been assessed, using a variety of study designs and hearing measures. The findings are quite mixed and differences in how hearing has been measured may be a potential reason. For example, few studies have used objective measures of hearing to assess its relationship to driving outcomes, some studies have asked respondents about hearing aid use, and others have used self-reported (S-R) hearing. Among those using an objective assessment of hearing, two used an audioscope, but found no significant relationship with crash risk (Sims et al., 1998, 2000). Another study used a variety of hearing measures and found a significant interaction between hearing impairment (using audiometry) and distractors on driving performance (Hickson et al., 2010). Specifically, older adults with moderate or severe hearing impairment performed significantly worse in the presence of a distractor (a moderating effect). This suggests that hearing loss may increase demand on attentional capacity, thus increasing the negative effects of distraction while driving.

Results related to hearing aid use are also mixed. Use of a hearing aid was not significantly related to crash risk in some studies (McGwin et al., 2005; Sims et al., 2000), whereas at least one study found that it may increase risk for crash-related injuries (McCloskey et al., 1994). In a study where at-fault crashes were compared with controls, hearing aid use was nonsignificant (McGwin et al., 2005). Use of a hearing device was also found to be unrelated to driving status in a different study (MacLeod et al., 2014).

Using S-R measures to assess the effect of hearing on driving outcomes is most common in the literature. In terms of crash risk, a significant relationship was observed with S-R hearing loss in the right ear in an Australian study (where drivers sit on the right side of the vehicle and drive on the left side of the road; Ivers et al., 1999). In a French cohort study, no difference in S-R hearing difficulty was observed between crash-involved and non-crash-involved older adults (Lafont et al., 2008). Gallo and colleagues (1999) found a significant relationship between adverse driving events and S-R hearing impairment. The moderating effect of S-R hearing on the relationship between poor vision and crash rates has also been assessed. Findings suggested that hearing loss exacerbates the negative effect of vision problems (acuity and contrast sensitivity) on crashes (Green et al., 2013).

The relationship between S-R hearing impairment and driving reduction/cessation has also been assessed. Several studies compared hearing ability between older drivers and nondrivers. S-R hearing was worse among nondrivers than drivers (Kostyniuk et al., 2000) and among former drivers compared with current drivers (Forrest et al., 1997). Others have found no difference in S-R hearing between drivers and nondrivers (Anstey et al., 2017) or when comparing drivers with former drivers (Lafont et al., 2008). Gilhotra et al. (2001) reported that S-R severe hearing difficulty was significantly associated with driving cessation. S-R hearing impairment may also be related to driving fewer miles within the preceding 5 years (Forrest et al., 1997).

Collectively, these findings suggest that the relationship between hearing impairment and driving outcomes among older drivers is not well understood. It may also be a complex relationship, given that several studies identified hearing impairment as a potential moderator, amplifying the relationship between distractors on driving performance (Hickson et al., 2010) and poor vision on crashes (Green et al., 2013). Research that specifically focused on how hearing impairment may affect driving reduction/avoidance is particularly scarce. The goal of the current study was to clarify the role of hearing impairment on driving reduction/avoidance, using a large cohort of older drivers and several different measures of hearing (which could explain some mixed results in previous research). Specific hypotheses follow:

Hypothesis 1 (H1): Greater hearing impairment will be associated with more driving reduction and avoidance.

Hypothesis 2 (H2): The objective measure of hearing (Whisper Test [WT]) will have a stronger association with driving reduction/avoidance than other hearing measures.

Hypothesis 3 (H3): Greater hearing impairment will moderate the relationships between visual acuity and cognition with driving reduction/avoidance, resulting in an increase in those behaviors.

Method

Data Source

This study used data from the American Automobile Association (AAA) Longitudinal Research on Aging Drivers (LongROAD) study, a large multisite prospective cohort study with five locations in the United States (Ann Arbor, MI; Baltimore, MD; Cooperstown, NY; Denver, CO; and San Diego, CA). LongROAD includes respondents from a variety of geographic locations and backgrounds, and includes a wealth of driving and sociodemographic information (see Li et al., 2017 for full detail). Data were collected using questionnaires, in person clinical assessments, and GPS/dataloggers installed in participants’ vehicles. Data from 2,030 individuals were analyzed in this study. Participants from the New York site were not included because of differences in their administration of the visual acuity test. Baseline questionnaires/assessments were used for each participant (enrollment occurred between July, 2015 and March, 2017). Objective driving data were collected continuously, and the first 12 months of driving data were used to eliminate any seasonal differences. Study protocols were reviewed and approved by the University of Michigan institutional review board (HUM00094031).

Measures

Dependent variables

Objectively measured dependent variables were derived from the dataloggers installed in participants’ vehicles. The dataloggers automatically recorded location, time of day, vehicle speed, heading, and whether or not the actual participant was driving the vehicle. Percent of trips at night represented the percentage of all trips where at least 80% of a trip occurred during nighttime (defined as civil twilight or a solar angle greater than 96°). Percent of trips during peak times was the percentage of trips that occurred between 7:00 a.m. and 9:00 a.m. and between 4:00 p.m. and 6:00 p.m. on weekdays. Percent of trips on freeways was operationalized as the percentage of trips where at least 20% of the distance traveled was at a speed of at least 60 mph. Percent of trips in unfamiliar areas represented the percentage of trips at least 15 mi from home. Each of these was reverse-coded, such that a higher value represented more avoidance of these situations. These same behaviors were measured subjectively, using the question stem, “Do you try to avoid driving . . .” followed by the specific behavior, “. . . at night?” “. . . in rush hour traffic?” “. . . on the freeway?” and “. . . in unfamiliar areas?” Response choices included yes and no for each item.

Independent variables of interest

Hearing was operationalized using three methods. Hearing aid use was assessed by asking, “Do you ever wear a hearing aid?” with yes and no response options. Self-rated hearing was assessed with the question, “Is your hearing excellent, very good, good, fair, or poor?” coded such that a higher score represented worse hearing. For those who used hearing aids, this item did not specify whether they should assess their corrected or uncorrected hearing, so that was open to interpretation for those respondents. The WT (Pirozzo et al., 2003) was performed in both ears (with hearing aids in place for those who used them) and participants could pass or fail in each ear. This test has been compared with audiometric assessment, with results suggesting that it has strong performance indicators (sensitivity: 79%, specificity: 91%; Strawbridge & Wallhagen, 2017).

Visual acuity was measured using the Tumbling E chart and then converted to Logarithm of the Minimum Angle of Resolution (LogMAR) scores. In the LogMAR scale, zero is average (equivalent to 20/20) and higher values represent greater impaired acuity. The total range of this variable in our sample was approximately 0.8, so it was standardized when included in the regressions to allow for a meaningful interpretation of a 1-unit increase (representing a standard deviation increase). Cognition was operationalized using the Trail Making Test Part B. Higher values indicate longer time to complete the test, representing poorer cognitive function (see Tombaugh, 2004, for norms based on age and education).

Control variables

Control variables included age, gender, race, relationship status, household income, education, health, and depressive symptoms. Age was measured as a continuous variable. Gender was recorded as male or female by the interviewer. S-R race and ethnicity were collapsed to White, Non-Hispanic; Black, Non-Hispanic; and Other due to low numbers of people in some groups. Relationship status categories were collapsed to married/partnered, separated/divorced, widowed, and never married due to low cell counts and for ease of interpretation. Household income was measured in five groups, including less than US$20,000; US$20,000 to US$49,999; US$50,000 to US$79,999; US$80,000 to US$99,999; and US$100,000 or more. Education included high school or less, some college, bachelor’s degree, and graduate degree. Health was measured by asking, “How satisfied are you with your health?” Responses ranged from 1 (not at all satisfied) to 5 (completely satisfied) on a 5-point scale. Depressive symptoms were represented by the summed responses to four items, which all began with the stem, “In the past 7 days, I felt . . .” followed by worthless, helpless, depressed, and hopeless. Response choices ranged from never to always. More than 76% of the sample reported never for all four items, so this variable was dichotomized to represent the never group and those who reported anything else.

Data Management and Analysis

SAS Version 9.4 (SAS Institute Inc., 2016) was used for all analyses. Descriptive statistics were generated to assess variables and check for errors. Bivariate statistical tests were performed to assess initial relationships and inform regression model building. Linear regression techniques were used for the objective outcomes and binary logistic regression was used for the subjective outcomes, with separate models for each. Self-rated hearing (measured using an ordinal-level scale) was assessed separately as a categorical versus a continuous variable (given its reasonably normal distribution). That assessment suggested that the continuous treatment of this variable produced more statistically significant relationships, so it was included as a continuous predictor in all models. Variance inflation factors (VIFs) were calculated to evaluate multicollinearity for all predictors and no problems were identified. VIFs also determined whether the three hearing variables could be included in the same models. Models with only one hearing variable included were also compared with models with all three hearing variables to see the extent to which parameter estimates changed when controlling for other hearing measures. We noted very little difference between models and no multicollinearity issues, thus models included all hearing variables together. Finally, two-way statistical interactions were assessed between each hearing variable with visual acuity and cognition for each outcome. When statistically significant interactions were observed, they were retained in the final model; when the interactions were not significant, they were removed and only main effects were reported. Our model-building process included the fitment of eight separate models with the same predictors. We began with control variables only, and then added vision, followed by cognition. We next added the hearing variables and assessed potential interactions. This process allowed us to examine changes in parameters and model statistics at each stage.

Results

Descriptive Findings

Results of descriptive statistics for control variables are shown in Table 1. Participants ranged in age from 65 to 79 years, with mean of 70.9 years. Women comprised 51.9% of the sample. The majority identified their race as White (83.5%) and were married/partnered (66.0%). Respondents reported relatively high incomes and educational levels. The mean health score was 3.77, which was close to very satisfied, and about three quarters of the sample reported never experiencing any depressive symptoms during the past 7 days.

Table 1.

Descriptive Statistics for Control Variables.

Characteristic	M (SD) or N (%)
Age	70.89 (4.08)
Gender
Male	978 (48.18%)
Female	1,052 (51.82%)
Race
White	1,695 (83.50%)
Black	170 (8.37%)
Other	165 (8.13%)
Relationship status
Married/partnered	1,339 (65.96%)
Divorced/separated	358 (17.64%)
Widowed	233 (11.48%)
Never married	100 (4.93%)
Income
<US$20,000	66 (3.25%)
US$20,000–US$49,999	397 (19.56%)
US$50,000–US$79,999	496 (24.43%)
US$80,000–US$99,999	312 (15.37%)
>=US$100,000	759 (37.39%)
Education
<High school	166 (8.18%)
Some college	458 (22.56%)
Bachelor’s degree	491 (24.19%)
Graduate degree	915 (45.07%)
Health	3.77 (0.82)
Depressive symptoms
Never (all symptoms)	1,547 (76.21%)
>Never	483 (23.79%)

Descriptive findings for key predictors of interest and outcome variables are shown in Table 2. The mean for visual acuity was 0.093 on the LogMAR scale, indicating slightly worse than average acuity (~20/25). The average time to complete the Trail Making Test Part B test was 91 s (within age-based norms; Tombaugh, 2004). The mean for self-rated hearing was 2.32 (between good and very good hearing). Only 16.7% reported wearing a hearing aid; 11.3% failed the WT in both ears, 10.6% failed in only one ear, and 78.1% passed both ears.

Table 2.

Descriptive Statistics for Predictors of Interest and Outcomes.

Characteristic	M (SD) or N (%)
Predictors of interest
Visual acuity	0.09 (0.12)
Cognition	91.13 (40.57)
Hearing aid
No	1,690 (83.25%)
Yes	340 (16.75%)
Self-rated hearing	2.32 (0.95)
Whisper Test
Pass both ears	1,585 (78.08%)
Fail one ear	215 (10.59%)
Fail both ears	230 (11.33%)
Outcomes
Objective
% night	7.38 (5.55)
% day	92.62 (5.55)
% freeway	14.09 (11.36)
% non-freeway	85.91 (11.36)
% >15 mi	32.35 (21.35)
% ≤ 15 mi	67.65 (21.35)
% peak time	16.76 (5.54)
% nonpeak	83.24 (5.54)
Subjective
Avoid night	693 (34.14%)
No night avoid	1,337 (65.86%)
Avoid freeway	212 (10.44%)
No freeway avoid	1,818 (89.56%)
Avoid unfamiliar	327 (16.11%)
No unfamiliar avoid	1,703 (83.89%)
Avoid peak time	1,237 (60.94%)
No peak avoid	793 (39.06%)

For the objective measures of driving avoidance, most trips occurred during the day (92.6%), on non-freeways (85.9%), within 15 mi of home (67.6%), and during nonpeak time (83.2%). For the subjective measures, 34.2% reported that they avoid night driving, 10.5% avoid freeways, 16.1% avoid unfamiliar areas, and 61.0% avoid peak time driving. Three online supplemental tables (available in the online appendix) show descriptive statistics for each predictor and outcome as a function of each hearing variable. The Supplemental Table A1 shows results for hearing aid use, Supplemental Table A2 for self-rated hearing, and Supplemental Table A3 for the WT.

Multiple Regression Findings

Linear regression results from the objectively measured driving outcomes are shown in Table 3. Those outcomes represent the percentage of trips in less challenging situations (e.g., daytime), so a positive beta value can be interpreted as less travel in challenging situations, thus more driving avoidance. Note that some beta values are small because of the large range of the variable, and that a 1-unit increase represents only a modest change. These can be manually adjusted if assessing a larger change is desired. For example, to adjust age from a 1-year to a 5-year increase, the beta can simply be multiplied by 5.

Table 3.

Regression Results for Objective Driving Avoidance Variables.

Variable	Daytime driving(avoiding night)		Lower speed(avoiding freeway)		Near home(avoiding far trips)		Nonpeak(avoiding peak)
Variable	β	SE	β	SE	β	SE	β	SE
Age	0.160***	0.032	0.138*	0.065	0.519***	0.121	0.096**	0.033
Gender (ref = male)
Female	1.803***	0.259	2.337***	0.527	5.381***	0.982	0.229	0.264
Race (ref = White)
Black	−2.544***	0.464	2.550**	0.945	9.049***	1.762	0.360	0.476
Other	−0.732^†	0.444	−1.991*	0.903	3.334*	1.684	−0.646	0.453
Relationship status (ref = married/partnered)
Divorced/separated	−1.663***	0.362	−0.272	0.737	−1.095	1.375	−0.362	0.370
Widowed	−1.372***	0.413	0.328	0.840	−1.843	1.566	0.138	0.421
Never married	−2.085***	0.582	−0.300	1.183	4.129^†	2.207	0.640	0.594
Income (ref >= US$100,000)
<US$20,000	−0.058	0.752	4.373**	1.530	9.070**	2.853	1.053	0.768
US$20,000–US$49,999	0.113	0.393	3.959***	0.800	4.294**	1.492	1.463***	0.402
US$50,000–US$79,999	0.212	0.329	3.155***	0.700	−0.265	1.249	0.728*	0.336
US$80,000–US$99,999	0.242	0.364	0.884	0.740	−2.043	1.381	0.552	0.371
Education (ref ≤ HS)
Some college	−0.480	0.494	−1.772^†	1.005	−2.409	1.874	−0.530	0.504
Bachelor’s degree	−0.179	0.503	−0.352	1.024	−1.617	1.909	−0.531	0.514
Graduate degree	−0.751	0.488	−0.859	0.994	−2.237	1.853	−0.834^†	0.498
Health	−0.071	0.154	−0.099	0.314	−0.907	0.585	−0.236	0.157
Depressive symptoms (ref = never)
>Never	−0.257	0.293	0.325	0.595	2.672*	1.110	0.051	0.299
Visual acuity (z)	−0.002	0.124	0.300	0.253	2.219***	0.471	0.138	0.127
Cognition	0.008*	0.003	0.034***	0.007	0.013	0.012	0.003	0.004
Hearing aid (ref = no)	0.066	0.341	−0.256	0.693	1.574	1.293	0.086	0.348
Self-rated hearing	−0.222^†	0.135	−0.125	0.274	−0.513	0.511	−0.241^†	0.138
Whisper Test (ref = pass both)
Fail one ear	−0.065	0.403	1.939*	0.820	−2.577^†	1.529	0.901	0.944
Fail both ears	0.930*	0.414	1.575^†	0.842	−4.782**	1.570	−1.887^†	1.019
Cognition*WT1	ns	—	ns	—	ns	—	−0.005	0.009
Cognition*WT2	ns	—	ns	—	ns	—	0.020*	0.009

Note. ref = reference category; z = standardized; WT1, WT2 = Whisper Test fail in one ear, two ears; ns = not significant.

†

p < .10. *p < .05. **p < .01. ***p < .001.

As expected, many of the control variables were related to avoiding specific driving situations (see Table 3). Among variables of interest, worse visual acuity was only associated with more travel close to home. Worse cognition was significantly related to less nighttime and freeway travel. The three hearing variables differed in their relationships to objective driving behaviors (the assessment of H1 and H2). Neither hearing aid use nor S-R hearing was significantly related to any of the objective behaviors. However, failing the WT was related to all objective behaviors. Compared with passing the test in both ears, failing in both ears was related to less nighttime driving. Failing the test in one ear compared with passing in both was associated with less freeway driving (p < .05), and failing in both ears approached significance (p = .06). Counterintuitively, more trips over 15 mi were taken among those who failed the test in both ears (p < .01) and failing in one ear approached significance (p = .09). A significant interaction between cognition and the WT was observed for driving during peak times (an assessment of H3). Although the main effects were not significant, the significant interaction suggested that lower cognition increased avoidance of peak time driving only for those who failed the WT in both ears (p < .05). All other two-way interactions between each hearing variable and the visual acuity and cognition variables were separately assessed, but were not statistically significant.

Binary logistic regression results from the subjective outcomes are shown in Table 4. Note that odds ratios (ORs) are included for all variables, but should be interpreted with caution for those with a significant interaction, as that makes them conditional (conditional ORs are reported in text with an explanation). Some betas/ORs are small because they represent only a modest change in the predictor. These can be manually adjusted by multiplying the beta by the factor of interest (e.g., by 5 for a 5-year increase in age), and then exponentiating the resulting value to calculate the OR.

Table 4.

Regression Results for Subjective Driving Avoidance Variables.

	Nighttime avoidance			Freeway avoidance			Unfamiliar avoidance			Peak avoidance
Variable	β	OR	SE	β	OR	SE	β	OR	SE	β	OR	SE
Age	−0.003	1.003	0.013	−0.036^†	0.964	0.020	0.003	1.003	0.017	0.011	1.011	0.012
Gender (ref = male)
Female	0.867***	2.380	0.109	0.894***	2.444	0.177	0.654***	1.924	0.141	−0.165^†	0.848	0.100
Race (ref = White)
Black	0.404*	1.498	0.185	0.433^†	1.542	0.239	0.419*	1.521	0.212	−0.207	0.813	0.178
Other	0.382*	1.480	0.179	−0.013	0.987	0.279	0.512*	1.669	0.208	−0.169	0.844	0.171
Relationship status (ref = married/partnered)
Divorced/separated	−0.068	0.934	0.147	−0.495*	0.610	0.221	−0.400*	0.671	0.190	−0.012	0.988	0.140
Widowed	0.114	1.120	0.164	0.043	1.044	0.225	−0.068	0.934	0.200	0.113	1.119	0.160
Never married	−0.172	0.842	0.234	−1.359**	0.257	0.485	−0.025	0.976	0.279	0.576*	1.778	0.244
Income (ref >= US$100,000)
<US$20,000	0.786**	2.194	0.302	1.155**	3.174	0.403	0.712*	2.038	0.347	0.448	1.565	0.302
US$20,000–US$49,999	0.510**	1.665	0.162	1.041***	2.832	0.250	0.529**	1.698	0.204	0.283^†	1.327	0.154
US$50,000–US$79,999	0.328*	1.388	0.138	0.755***	2.127	0.224	0.314^†	1.369	0.180	0.026	1.027	0.127
US$80,000–US$99,999	0.502***	1.652	0.152	0.497^†	1.643	0.260	0.281	1.325	0.203	0.119	1.126	0.141
Education (ref =< high school)
Some college	−0.145	0.865	0.199	−0.475^†	0.622	0.257	−0.257	0.774	0.227	0.076	1.079	0.190
Bachelor’s degree	−0.042	0.959	0.203	−0.220	0.803	0.266	−0.360	0.698	0.239	0.253	1.288	0.194
Graduate degree	−0.093	0.911	0.198	−0.309	0.734	0.262	−0.294	0.745	0.230	0.187	1.206	0.187
Health	−0.370***	0.691	0.065	−0.097	0.908	0.094	−0.219**	0.803	0.079	−0.196**	0.822	0.061
Depressive symptoms (ref = never)
>Never	0.181	1.199	0.118	0.315	1.371	0.171	0.274^†	1.315	0.145	0.374**	1.454	0.117
Visual acuity (z)	0.190***	1.209	0.051	0.359^†	1.432	0.203	0.098	1.103	0.064	−0.057	0.945	0.048
Cognition	0.003*	1.003	0.001	0.006**	1.006	0.002	0.008***	1.008	0.002	−0.000	1.000	0.001
Hearing aid (ref = no)	−0.192	0.825	0.147	−0.542*	0.581	0.260	−0.268	0.765	0.197	−0.064	0.938	0.131
Self-rated hearing	0.078	1.081	0.056	−0.001	0.999	0.087	0.031	1.031	0.071	0.088^†	1.091	0.053
Whisper Test (ref = pass both)
Fail one ear	0.045	1.045	0.167	0.010	1.010	0.269	1.108*	3.029	0.496	−0.259^†	0.772	0.155
Fail both ears	0.000	1.000	0.175	0.523*	1.688	0.257	0.461	1.585	0.546	−0.413**	0.662	0.159
Visual acuity (z)*SR hearing	ns	—	—	−0.163*	0.850	0.081	ns	—	—	ns	—	—
Cognition*WT1	ns	—	—	ns	—	—	−0.011*	0.989	0.005	ns	—	—
Cognition*WT2	ns	—	—	ns	—	—	−0.005	0.995	0.005	ns	—	—

Note. OR = odds ratio; ref = reference category; z = standardized; SR = self-rated; WT1, WT2 = Whisper Test fail in one ear, two ears; ns = not significant.

†

p < .10. *p < .05. **p < .01. ***p < .001.

Many of the control variables were related to subjective driving avoidance (see Table 4). In terms of the variables of interest, worse visual acuity was significantly associated with higher odds of subjective nighttime driving avoidance and approached significance for freeways (p = .08 for the main effect). Lower cognition was related to increased odds of all avoidance behaviors except peak time. Hearing aid use was significantly related to lower odds of freeway avoidance but no other behaviors (H1 and H2). Worse S-R hearing approached significance for peak time avoidance (p = .09, H1 and H2). For freeway avoidance, the interaction between S-R hearing and visual acuity was significant (H3). Conditional ORs were calculated for hearing at various levels of vision. The relationship was only significant for those with low vision (at least 0.495 LogMAR [about 20/63 on the Snellen scale]). For those with the worst vision in the sample (LogMAR = .699), worse self-rated hearing was associated with lower odds of reporting freeway avoidance (OR = 0.428, p < .05).

The WT was not significantly related to nighttime driving avoidance, but was related to the other three behaviors (H2). Compared with passing in both ears, failing in both was related to 69% higher odds of reporting freeway avoidance. There was a statistically significant interaction between cognition and the WT for avoiding unfamiliar areas, and conditional ORs were calculated for cognition at each level of hearing (H3). For respondents who passed in both ears only, every 10-s increase in time to complete the Trail Making Test was related to 8.3% higher odds of reporting unfamiliar area avoidance (OR = 1.008, p < .05). No other interactions between any hearing variables and vision/cognition on any of the subjective behaviors were statistically significant.

The WT was also significantly related to subjective peak time avoidance, where those who failed in both ears had significantly lower odds (33.8%) of reporting avoidance (H1 and H2). Results for those who failed in one ear were similar (22.8% lower), but only approached significance (p = .09). To assess whether hearing loss in a specific ear mattered (as suggested by Ivers et al., 1999), models were also explored that used WT results from each individual ear, rather than the combined measure. Individual ear measures were only significant for two of the eight outcomes assessed, with a different ear for each outcome.

Discussion

Three hypotheses were assessed in this study: H1—worse hearing would be associated with more driving reduction/avoidance, H2—the objective measure of hearing would have a stronger association with driving avoidance than subjective measures, and H3—hearing would moderate the relationships between visual acuity and cognition with driving avoidance, leading to greater avoidance. The second hypothesis was supported, whereas the other two were partially supported.

Regarding H1, worse hearing was related to a higher level of some aspects of driving reduction/avoidance, but there was no effect or the opposite effect in others. Worse hearing (assessed by the WT) was significantly related to less objectively measured nighttime and freeway driving, but more trips 15 mi or farther away from home. Worse hearing (WT) was also significantly related to higher odds of reporting freeway and unfamiliar area avoidance. However, this test was counterintuitively related to decreased odds of peak time avoidance. These mixed results suggest that the effect of hearing on driving avoidance is complex and differs depending upon the measurement approach to hearing and driving avoidance. Differences between objective and subjective driving may be due to S-R measures being assessed at baseline, while driving was measured continuously during participants’ first study year. Early modeling also addressed H1, using objective driving exposure outcomes (e.g., number of days driven). None of the hearing variables were associated with those outcomes, so results are not reported here, given the focus on hearing impairment.

H2 was supported in that the WT (the objective hearing measure) was statistically related to more factors than the other hearing measures. Results of the multicollinearity assessment also showed that the three hearing measures accounted for different aspects of driving avoidance. Although use of a hearing aid indicates declining auditory ability, it improves hearing and can compensate for declining ability. That may explain why hearing aid use was only related to one outcome and it suggested lower odds of freeway avoidance. The main effect of S-R hearing was unrelated to all outcomes (the significant interaction is discussed later). Overall, the measurement of hearing is important, which is consistent with previous research comparing how different hearing measures are related to a variety of functional outcomes (Choi et al., 2016). Future researchers should consider that there may be a stronger association between objective (compared with perceived) hearing measures and driving avoidance.

In terms of H3, hearing was extensively explored as a moderator in the current study and three significant interactions were observed. Lower cognition increased objective peak time avoidance for those who failed the WT in both ears, the predicted direction of influence. This is consistent with previous research where the effect of a cognitive distractor on driving ability was exacerbated for those with severe hearing loss (Hickson et al., 2010).

The interactions for subjective outcomes were unexpected. Worse cognition increased the odds of reporting unfamiliar area avoidance, but only for those with good hearing. An explanation for this inconsistency is unclear and attempts at replication with different samples should be assessed in future research. For participants with poor vision, worse S-R hearing was related to lower odds of reporting freeway avoidance. This is contrary to findings by Green and colleagues (2013), where hearing loss worsened the effect of poor vision on crash risk. Variable differences could explain this inconsistency: an objective measure of crash involvement over 5 years versus reported avoidance of freeways. Those with both poor vision and hearing may find freeway driving easier than the negotiation of intersections and traffic more common on surface streets. Different measures of vision (e.g., contrast sensitivity) could also produce different results. Although the current study assessed driving avoidance, and most previous research assessed different driving outcomes, the mixed hearing-related findings in the current study are consistent with previous mixed results (see Anstey et al., 2005; MacLeod et al., 2014 for reviews).

Results related to control variables are similar to previous research (see Dickerson et al., 2019, 2007; Eby et al., 2019, for reviews). Age, gender, race, relationship status, socioeconomic status, health, and depressive symptoms were all related to at least some outcomes. Future research might further explore whether such factors (e.g., gender) modify the relationship between hearing loss and driving avoidance, given known demographic differences in these outcomes. There was also general consistency in which control variables were related to objective versus subjective measures, but some factors were more strongly associated with one measure than another (e.g., age with objective measures; income with subjective, not objective nighttime avoidance). This suggests that driving avoidance behaviors are not all the same. These behaviors are predicted by somewhat different factors and are engaged in at different rates. The overall similar but slightly different findings between objective versus subjective outcomes is consistent with previous research (Molnar et al., 2018).

Limitations and Strengths

This study has some limitations. The sample has higher educated and higher earning participants than the general population, but study sites represent people from diverse socioeconomic backgrounds (Li et al., 2017). The study did not allow for causal assessment, only associations. To address our hypotheses, we conducted eight different regressions and assessed several interactions per model; that number of tests may have resulted in some statistically significant findings due only to chance. The underlying reasons for differences observed in the study remain unclear; some people avoid driving situations as a response to declining abilities, whereas others may have always avoided them, or experienced a lifestyle change (Molnar et al., 2013). Finally, different measures of cognition, vision, or hearing could yield different results. For example, the S-R hearing item did not specify whether respondents should consider their hearing with or without correction, which may have introduced unwanted variance. Likewise, the WT was administered by trained researchers, not clinicians.

The strengths of this study include the large cohort of older drivers and the inclusion of both objective and subjective measures of a variety of avoidance behaviors. Three hearing measures were also included. The moderating effect of hearing on visual acuity and cognition was also comprehensively examined.

Conclusion

Hearing impairment affects driving avoidance among older adults in a complex way. Worse hearing was related to less actual nighttime driving; less actual freeway driving, but more trips farther than 15 mi from home; and lower odds of reporting peak time avoidance. Hearing significantly interacted with visual acuity and cognition. Lower cognition was related to less peak time driving, but only for those with poor hearing. Worse S-R hearing was related to lower odds of freeway avoidance for those with poor visual acuity. Worse cognition was associated with higher odds of avoiding unfamiliar areas, but only for those with good hearing. Finally, the objective hearing measure was a better predictor of driving avoidance than hearing aid use or S-R hearing. These findings are important because hearing loss can sometimes be mitigated. As such, public health interventions designed to initiate hearing assessment and/or hearing aid use could point toward the possibility of continued driving as another potential benefit, particularly given other consequences (e.g., loneliness, social disengagement, and memory declines) of hearing loss (Huang et al., 2020; Moorman et al., 2020).

Supplemental Material

sj-pdf-1-jag-10.1177_0733464821999223 – Supplemental material for The Influence of Hearing Impairment on Driving Avoidance Among a Large Cohort of Older Drivers

Supplemental material, sj-pdf-1-jag-10.1177_0733464821999223 for The Influence of Hearing Impairment on Driving Avoidance Among a Large Cohort of Older Drivers by Jonathon M. Vivoda, Lisa J. Molnar, David W. Eby, Scott Bogard, Jennifer S. Zakrajsek, Lidia P. Kostyniuk, Renée M. St. Louis, Nicole Zanier, David LeBlanc, Jacqui Smith, Raymond Yung, Linda Nyquist, Carolyn DiGuiseppi, Guohua Li and David Strogatz in Journal of Applied Gerontology

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This article was funded by the AAA Foundation for Traffic Safety.

ORCID iDs

Jonathon M. Vivoda

Carolyn DiGuiseppi

Supplemental Material

Supplemental material for this article is available online.

References

Anstey

K. J.

Hosking

D. E.

Eramudugolla

(2017). The epidemiology of driving in later life: Sociodemographic, health and functional characteristics, predictors of incident cessation, and driving expectations. Accident Analysis & Prevention, 107, 110–116. https://doi.org/10.1016/j.aap.2017.07.026

Anstey

K. J.

Wood

Lord

Walker

J. G.

(2005). Cognitive, sensory and physical factors enabling driving safety in older adults. Clinical Psychology Review, 25(1), 45–65. https://doi.org/10.1016/j.cpr.2004.07.008

Braitman

K. A.

Williams

A. F.

(2011). Changes in self-regulatory driving among older drivers over time. Traffic Injury Prevention, 12(6), 568–575. https://doi.org/10.1080/15389588.2011.616249

Charlton

J. L.

Oxley

Fildes

Oxley

Newstead

Koppel

O’Hare

(2006). Characteristics of older drivers who adopt self-regulatory driving behaviours. Transportation Research Part F: Traffic Psychology and Behaviour, 9(5), 363–373. https://doi.org/10.1016/j.trf.2006.06.006

Choi

J. S.

Betz

Deal

Contrera

K. J.

Genther

D. J.

Chen

D. S.

Gispen

F. E.

Lin

F. R.

(2016). A comparison of self-report and audiometric measures of hearing and their associations with functional outcomes in older adults. Journal of Aging and Health, 28(5), 890–910. https://doi.org/10.1177/0898264315614006

Crowe

C. L.

Kannoth

Andrews

Strogatz

DiGuiseppi

Hill

Eby

D. W.

Molnar

L. J.

Mielenz

T. J.

(2020). Associations of frailty status with low-mileage driving and driving cessation in a cohort of older drivers. Geriatrics, 5(1), Article 19. https://doi.org/10.3390/geriatrics5010019

Devlin

McGillivray

J. A.

(2014). Self-regulation of older drivers with cognitive impairment: A systematic review. Australasian Journal on Ageing, 33(2), 74–80. https://doi.org/10.1111/ajag.12061

Dickerson

A. E.

Molnar

L. J.

Bédard

Eby

D. W.

Berg-Weger

Choi

Grigg

Horowitz

Meuser

Myers

O’Connor

Silverstein

N. M.

(2019). Transportation and aging: An updated research agenda to advance safe mobility among older adults transitioning from driving to non-driving. The Gerontologist, 59(2), 215–221. https://doi.org/10.1093/geront/gnx120

Dickerson

A. E.

Molnar

L. J.

Eby

D. W.

Adler

Bédard

Berg-Weger

Classen

Foley

Horowitz

Kerschner

Page

Silverstein

N. M.

Staplin

Trujillo

(2007). Transportation and aging: A research agenda for advancing safe mobility. The Gerontologist, 47(5), 578–590. https://doi.org/10.1093/geront/47.5.578

10.

Eby

D. W.

Molnar

L. J.

Louis

R. M. S.

(2019). Perspectives and strategies for promoting safe transportation among older adults. Elsevier.

11.

Edwards

J. D.

Ross

L. A.

Ackerman

M. L.

Small

B. J.

Ball

K. K.

Bradley

Dodson

J. E.

(2008). Longitudinal predictors of driving cessation among older adults from the ACTIVE clinical trial. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 63(1), P6–P12.

12.

Emerson

J. L.

Dawson

J. D.

E. Y.

Anderson

S. W.

Rizzo

(2012). Predictors of driving outcomes in advancing age. Psychology and Aging, 27(3), 550–559. https://doi.org/10.1037/a0026359

13.

Forrest

Y.-Z.

Bunker

C. H.

Songer

T. J.

Coben

J. H.

Cauley

J. A.

(1997). Driving patterns and medical conditions in older women. Journal of the American Geriatrics Society, 45(10), 1214–1218. https://doi.org/10.1111/j.1532-5415.1997.tb03772.x

14.

Freeman

Muñoz

Turano

West

(2005). Measures of visual function and time to driving cessation in older adults. Optometry and Vision Science, 83(8), 765–773.

15.

Gallo

J. J.

Rebok

G. W.

Lesikar

S. E.

(1999). The driving habits of adults aged 60 years and older. Journal of the American Geriatrics Society, 47(3), 335–341. https://doi.org/10.1111/j.1532-5415.1999.tb02998.x

16.

Gilhotra

J. S.

Mitchell

Ivers

Cumming

R. G.

(2001). Impaired vision and other factors associated with driving cessation in the elderly: The Blue Mountains Eye Study. Clinical & Experimental Ophthalmology, 29(3), 104–107. https://doi.org/10.1046/j.1442-9071.2001.00411.x

17.

Green

K. A.

McGwin

Owsley

(2013). Associations between visual, hearing, and dual sensory impairments and history of motor vehicle collision involvement of older drivers. Journal of the American Geriatrics Society, 61(2), 252–257. https://doi.org/10.1111/jgs.12091

18.

Hickson

Wood

Chaparro

Lacherez

Marszalek

(2010). Hearing impairment affects older people’s ability to drive in the presence of distracters. Journal of the American Geriatrics Society, 58(6), 1097–1103. https://doi.org/10.1111/j.1532-5415.2010.02880.x

19.

Huang

A. R.

Deal

J. A.

Rebok

G. W.

Pinto

J. M.

Waite

Lin

F. R.

(2020). Hearing impairment and loneliness in older adults in the United States. Journal of Applied Gerontology. Advance online publication. https://doi.org/10.1177/0733464820944082

20.

Ivers

R. Q.

Mitchell

Cumming

R. G.

(1999). Sensory impairment and driving: The Blue Mountains Eye Study. American Journal of Public Health, 89(1), 85–87. https://doi.org/10.2105/AJPH.89.1.85

21.

Kandasamy

Betz

M. E.

DiGuiseppi

Mielenz

T. J.

Eby

D. W.

Molnar

L. J.

Hill

Strogatz

(2018). Self-reported health conditions and related driving reduction in older drivers. Occupational Therapy in Health Care, 32(4), 363–379. https://doi.org/10.1080/07380577.2018.1522681

22.

Kostyniuk

L. P.

Shope

J. T.

Molnar

L. J.

(2000). Reduction and cessation of driving among older drivers in Michigan: Final report (UMTRI-2000-6). University of Michigan Transportation Research Institute. https://deepblue.lib.umich.edu/bitstream/handle/2027.42/1315/93411.0001.001.pdf?sequence=2

23.

Lafont

Laumon

Helmer

Dartigues

J.-F.

Fabrigoule

(2008). Driving cessation and self-reported car crashes in older drivers: The impact of cognitive impairment and dementia in a population-based study. Journal of Geriatric Psychiatry and Neurology, 21(3), 171–182. https://doi.org/10.1177/0891988708316861

24.

Eby

D. W.

Santos

Mielenz

T. J.

Molnar

L. J.

Strogatz

Betz

M. E.

DiGuiseppi

Ryan

L. H.

Jones

Pitts

S. I.

Hill

L. L.

DiMaggio

C. J.

LeBlanc

Andrews

H. F.

Bogard

Chihuri

Engler

A.-M.

Feng

, & the LongROAD Research Team. (2017). Longitudinal Research on Aging Drivers (LongROAD): Study design and methods. Injury Epidemiology, 4(1), Article 22. https://doi.org/10.1186/s40621-017-0121-z

25.

MacLeod

K. E.

Satariano

W. A.

Ragland

D. R.

(2014). The impact of health problems on driving status among older adults. Journal of Transport & Health, 1(2), 86–94. https://doi.org/10.1016/j.jth.2014.03.001

26.

McCloskey

L. W.

Koepsell

T. D.

Wolf

M. E.

Buchner

D. M.

(1994). Motor vehicle collision injuries and sensory impairments of older drivers. Age and Ageing, 23(4), 267–273. https://doi.org/10.1093/ageing/23.4.267

27.

McGuckin

Fucci

(2018). Summary of travel trends: 2017 National Household Travel Survey (FHWA-PL-18-019; pp. 1–101). Federal Highway Administration. https://nhts.ornl.gov/assets/2017_nhts_summary_travel_trends.pdf

28.

McGwin

Xie

Mays

Joiner

DeCarlo

D. K.

Hall

T. A.

Owsley

(2005). Visual field defects and the risk of motor vehicle collisions among patients with glaucoma. Investigative Ophthalmology & Visual Science, 46(12), 4437–4441. https://doi.org/10.1167/iovs.05-0750

29.

Molnar

L. J.

Eby

D. W.

Charlton

J. L.

Langford

Koppel

Marshall

Man-Son-Hing

(2013). Driving avoidance by older adults: Is it always self-regulation? Accident Analysis & Prevention, 57, 96–104. https://doi.org/10.1016/j.aap.2013.04.010

30.

Molnar

L. J.

Eby

D. W.

Vivoda

J. M.

Bogard

S. E.

Zakraksek

J. S.

St Louis

R. M.

Zanier

Ryan

L. H.

LeBlanc

Smith

Yung

Nyquist

DiGuiseppi

Mielenz

T. J.

Strogatz

(2018). The effects of demographics, functioning, and perceptions on the relationship between self-reported and objective measures of driving exposure and patterns among older adults. Transportation Research Part F: Traffic Psychology and Behaviour, 54, 367–377. https://doi.org/10.1016/j.trf.2018.02.026

31.

Moorman

S. M.

Greenfield

E. A.

Lee

C. S. H.

(2020). Perceived hearing loss, social disengagement, and declines in memory. Journal of Applied Gerontology. Advance online publication. https://doi.org/10.1177/0733464820909244

32.

National Institutes of Health. (2016). Quick statistics about hearing. https://www.nidcd.nih.gov/health/statistics/quick-statistics-hearing

33.

Pirozzo

Papinczak

Glasziou

(2003). Whispered voice test for screening for hearing impairment in adults and children: Systematic review. BMJ, 327(7421), Article 967. https://doi.org/10.1136/bmj.327.7421.967

34.

Ragland

D. R.

Satariano

W. A.

MacLeod

K. E.

(2004). Reasons given by older people for limitation or avoidance of driving. The Gerontologist, 44(2), 237–244. https://doi.org/10.1093/geront/44.2.237

35.

Ryvicker

Bollens-Lund

Ornstein

K. A.

(2020). Driving status and transportation disadvantage among Medicare beneficiaries. Journal of Applied Gerontology, 39(9), 935–943. https://doi.org/10.1177/0733464818806834

36.

SAS Institute Inc. (2016). Statistical Analysis Systems (9.4)[Computer software].

37.

Sims

R. V.

McGwin

Jr Allman

R. M.

Ball

Owsley

(2000). Exploratory study of incident vehicle crashes among older drivers. The Journals of Gerontology: Series A, 55(1), M22–M27. https://doi.org/10.1093/gerona/55.1.M22

38.

Sims

R. V.

Owsley

Allman

R. M.

Ball

Smoot

T. M.

(1998). A preliminary assessment of the medical and functional factors associated with vehicle crashes by older adults. Journal of the American Geriatrics Society, 46(5), 556–561. https://doi.org/10.1111/j.1532-5415.1998.tb01070.x

39.

Strawbridge

W. J.

Wallhagen

M. I.

(2017). Simple tests compare well with a hand-held audiometer for hearing loss screening in primary care. Journal of the American Geriatrics Society, 65(10), 2282–2284. https://doi.org/10.1111/jgs.15044

40.

Tombaugh

T. N.

(2004). Trail Making Test A and B: Normative data stratified by age and education. Archives of Clinical Neuropsychology, 19(2), 203–214. https://doi.org/10.1016/S0887-6177(03)00039-8

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.15 MB