Hearing Impairment Affects Older People's Ability to Drive in the Presence of Distracters

Louise Hickson, PhD; Joanne Wood, PhD; Alex Chaparro, PhD; Philippe Lacherez, PhD; Ralph Marszalek, BPsychSci(Hons)


J Am Geriatr Soc. 2010;58(6):1097–1103 

In This Article



One hundred seven older drivers were recruited to participate from staff and students at Queensland University of Technology, the University of Queensland 50+ Research Register, and the wider community. Participants included 59 men and 48 women with a mean age of 73.5 ± 5.6 (range 62–88). There were no specific selection criteria other than that participants should be current drivers and legally eligible to drive. This was to enable selection of a representative sample of current older drivers. All participants had binocular visual acuity that was better than driver licensing standards in Australia of 20/40. Better-eye visual acuity had a mean of 0.01 ± 0.10 logMAR (equivalent to 20/20) (range –0.2–0.3). The Mini-Mental State Examination was used to give an indication of cognitive function, and all but two of the participants scored above the criterion level of 23[17] (mean 28.6 ± 1.7, range 21–30). The majority of participants had completed secondary (38.3%) or tertiary education (41.1%), with only 20.6% finishing school at the primary level.

All participants lived in the community and reported a range of driving experience. On average, they drove 5 ± 1.7 days of each week (range 1–7). The majority (72%) had more than 50 years of driving experience; 21.5% had 41 to 50 years of experience, and 6.5% had 20 to 40 years of experience. Seventy percent drove more than 60 km per week. Participants reported an average of one crash per 25 driver years. Fifteen percent of participants reported that they wore a hearing aid in one or both ears when driving. Details about hearing aid use at other times were not recorded.


Participants attended two testing sessions. The first was a laboratory session where demographic information was collected, along with assessments of vision, cognition, and hearing. The second session was an assessment of driving performance on a closed-road circuit. The study was conducted in accordance with the requirements of the Queensland University of Technology Human Research Ethics Committee. All participants were given a full explanation of the experimental procedures, and written informed consent was obtained, with the option to withdraw from the study at any time.

Hearing There were three measures of hearing: pure-tone audiometry, speech perception testing, and self-reported hearing disability. These audiometric tests were conducted without hearing aids to define participants' hearing impairment and status, because this is the benchmark for describing the degree of hearing impairment. The first two tests were conducted in the quietest available room in the research wing of the School of Optometry at Queensland University of Technology. Pure-tone audiometry and speech perception testing were conducted using a Madsen Itera II by GN Otometrics (Taastrup, Denmark), with circumaural ME70 noise excluding TDH39 and a B71 bone conductor and headband. Pure-tone air conduction thresholds were obtained in both ears at 250, 500, 1,000, 2,000, and 4,000 Hz, and bone conduction was obtained at 500, 1,000, 2,000, and 4,000 Hz.

Speech perception in quiet and in noise was assessed using sentence lists originally developed for the Hearing in Noise Test.[18] The recorded sentences were presented bilaterally at each participant's most comfortable listening level in quiet (range 60– 105 dB, sound pressure level mean 74 ± 9 dB). Participants heard four lists of 10 sentences in different listening conditions in the following order of increasing difficulty: in quiet, at 10 dB signal to noise ratio (SNR), at 5 dB SNR, and at 0 dB SNR. The recording used was of a female Australian speaker, and the noise was a speech-shaped broadband noise. Participants receive a percentage correct word score for each list. An example sentence is: "Big dogs can be dangerous." Participants received a percentage correct score for each test condition based on the total number of words they repeated correctly for each of the 10 sentences in a list. All but one of the participants completed this assessment.

The 25-item Hearing Handicap Inventory for the Elderly (HHIE)[19] was used as a measure of hearing difficulties in everyday life. The response options and scoring for each item on the questionnaire are yes (4), sometimes (2), and no (0). A total score and a score for two subscales (social and emotional) are calculated. The social subscale has 12 items, for example, "Does a hearing problem cause you to avoid groups of people?" and the emotional subscale has 13 items, for example, "Does a hearing problem cause you to be nervous?" Total scores can range from 0 to 100, with higher scores indicative of greater self-reported hearing disability.

Driving Driving performance was assessed on a 5-km closed-road circuit representative of a rural road and free of other vehicles (except for one following car). Participants drove a right-hand drive sedan (1997 Nissan Maxima) with an automatic transmission and power steering. If participants normally wore glasses or hearing aids while driving, they wore these during the assessment. Participants were given a practice run during which they were able to familiarize themselves with the car, the road circuit, and the driving tasks. The practice run was identical to each test run except that it was driven in the opposite direction to the recorded run, so as to minimize familiarity effects. It included driving without distraction and then with the auditory and visual distracters added separately so that participants had the opportunity to practice all components of the assessment before the recorded run. Participants were told that they would be required to perform a number of concurrent tasks while driving at what they felt was a safe speed, to drive in their own lane except when avoiding hazards, and to drive as they normally would under the circumstances. Two research personnel, one seated in the passenger seat of the vehicle and the other in the rear seat, who scored different aspects of driving performance, assessed performance. A subsequent check on the reliability of the coding was performed to independently assess one of the driving measures for 20 random participants and revealed an interrater reliability (Pearson correlation coefficient) of 0.99 (P<.001).

Time to Complete the Road Course. An experimenter in the vehicle recorded the total time taken to complete the circuit.

Road Sign Recognition. The road sign recognition task required participants to report the information on any of 54 road signs located along the course (e.g., stop, yield) containing a total of 77 items of information. A participant's score represents the total number of correctly reported items of information.

Road Hazard Recognition and Avoidance. Participants were required to report and avoid hitting any of nine, large, low-contrast foam rubber road hazards centered across the driving lane. The road hazards were constructed from sheets of 180- × 80- × 5-cm gray–brown foam rubber, with a mean reflectance of 10%. Although the hazards could be felt when driven over, they had little effect on vehicle control. The position of the road hazards was randomized between each lap; during any given trial, nine of a total of 11 hazards were positioned on the course. Performance was measured as the number of road hazards hit.

Gap Perception. Nine pairs of traffic cones with variable lateral separations were also positioned throughout the course. Equal numbers of cones were set to be not wide enough, just wide enough, and obviously wide enough for the test vehicle to pass through. Participants were instructed to report whether the clearance between cones was sufficient for the vehicle to pass through and, if so, to attempt to do so. If the cone separation was judged to be too narrow, they were instructed to drive around the cones. The separation of the cones was varied between each lap. Performance was measured as the number of cone gaps judged correct.

Composite Driving Z Score. A composite score was also derived to compare the overall driving performance of the individual participants with that of the whole group and included road sign recognition, cone gap perception, course time, and number of road hazards hit, as per previous studies.[20,21] Z scores for each of these four component driving measures were determined, and the mean Z score for each participant was calculated to give an overall score. Equal weighting was assigned for all tasks.

Participants drove the track three times: without distraction, with auditory distraction, and with visual distraction. The order of conditions was randomized. The secondary distraction task required participants to verbally report the sums of numbers presented through a computer speaker (auditory distracter) or through a dashboard mounted liquid crystal display monitor (visual distracter) while driving. The monitor was positioned just to the left of the steering wheel on the dashboard, slightly below driver eye height. The visual task consisted of the simultaneous presentation of pairs of numbers (e.g., 1+5) subtending between 3.5° and 4.8° of visual angle at the viewing distance of participants. The auditory stimuli were presented at a comfortable listening level set by the participant using an adaptive procedure. Pairs of numbers were presented roughly every 3.5 seconds. Performance measures for the distracter tasks included the percentage of correct responses, incorrect responses, and nonresponses.

Data Analysis

First, the association between various measures of hearing function and the driving performance measures was examined. Pure-tone average (PTAvge) hearing levels were obtained by averaging 500, 1,000, 2,000, and 4,000 Hz. Hearing loss in the better ear was classified based on this average using the following criteria: 25 db or less hearing level (HL) was normal hearing; 25.1 to 40.0 dB HL was mild hearing loss; 40.1 to 60.0 dB HL was moderate hearing loss; greater than 60.0 dB HL was severe hearing loss.[22,23] Although 15 participants wore hearing aids for driving, degree of hearing impairment was defined in the unaided condition for all participants, because the sensorineural hearing impairment that occurs in older adults causes reduced audibility of sound as well as significant distortion of sound; wearing a hearing aid does not alleviate the latter problem of distortion.[24] Thus, classification of degree of hearing impairment is based on unaided results. Because age was also found to be highly correlated with driving performance in the sample, analyses were conducted using age as a covariate, so age-related decrements in performance (potentially related to other functional domains) were accounted for. Because some hearing measures showed highly skewed distributions, bivariate correlations were examined using nonparametric (Spearman) correlations adjusting for age. Mixed analyses of variance (ANOVAs) were conducted for each driving performance measure with the repeated-measures factor of distraction (none, auditory, and visual) and the between-subjects factor of hearing (normal to mild vs moderate to severe). Measures of performance were also calculated for the secondary tasks, and these were analyzed using the same mixed ANOVA model.

An a priori power analysis was conducted for the analyses, assuming sphericity and a reasonable reliability for the measures (0.7), and showed that the sample size of 107 allowed greater than 80% power to detect an effect of moderate size (Cohen's f=0.25) for all effects.