Self-Regulatory Behaviours of Older Drivers

Abstract

This study involved a survey of self-regulatory driving practices of 656 drivers aged 55 years and older. Types and prevalence of self-regulatory behaviours were described and several key characteristics of self-regulators were identified. Those who drove less than 100 kilometres per week were more likely to be female, 75+ years, retired, with arthritis, good/fair/poor ratings for decision-making, not the principal driver and not married. Those who avoided specific driving situations were also more likely to be female, 75+ years and not the principal driver. Avoidance was also associated with vision problems, lower confidence in driving situations and crash-involvement. Implications for promotion of safe driving practices are discussed.

It is generally assumed that older drivers ‘self-regulate’ their driving behaviour to minimise their risk of crashing. Self-regulation implies that drivers make adjustments in their driving behaviour that adequately match changing cognitive, sensory and motor capacities. Examples of such behaviours include reduction in driving distance and avoidance of busy traffic and night driving. While there is evidence that older drivers generally travel shorter distances on average than younger drivers [LTSA, 2000], evidence for widespread adoption of other types of self-regulatory practices that enhance safety is less definitive.

Researchers have proposed that most older drivers are able to regulate their driving adequately and compensate for age-related declines by reducing their annual driving and regulating when and where they drive [Eberhard, 1996]. While it is likely that many older drivers adjust their driving adequately to accommodate these changes, it is also possible that some fail to self-regulate appropriately and as a consequence, may be at a higher risk of crash involvement.

To date, studies reporting self-regulatory practices of older drivers have been predominantly based in Europe and North America [e.g. Lyman, McGwin and Sims, 2001; Holland and Rabbit, 1992]. Comparatively little is known about driving patterns of older Australian drivers and the effectiveness of these practices in reducing crash risk [Fildes 1997].

The processes involved in self-regulation are complex and the factors that influence the adoption of self-regulatory behaviours are multi-faceted. For the most part, the literature refers to these behavioural changes as compensatory, implying that older drivers change their behaviour in response to a loss of function. However, it is possible that drivers adopt particular driving practices for other reasons including lifestyle choices.

It is suggested that older individuals adopt considerable caution when driving [Eberhard, 1996]. Several studies have shown that most older drivers recognise that good vision is one of the most important elements for safe driving and often cite poor vision as a major determinant for reducing driving at night or in poor weather [Persson, 1993; Marottoli et al., 1993; Kostyniuk and Shope, 1998]. Other evidence suggests that older drivers tend to avoid complex traffic manoeuvres that require high cognitive demands [Hakamies-Blomqvist and Wahlström, 1998; Ball et al., 1998]. This suggests that at least some older adults are able to compensate well for limitations in their abilities.

In contrast, Rothman, Klein and Weinstein (1996) argued that people of all ages are poor at recognising the relationship between their own actions and potential risks. This may lead to optimism about one’s invulnerability, underestimation of risk and overestimation of one’s driving ability [Matthews, 1986]. In support of this notion, several studies have demonstrated that some older drivers do not adequately compensate for age-related changes in vision and cognitive abilities when driving [e.g. Stutts, 1998; Holland and Rabbit, 1992]. Interestingly, once drivers are made aware of their declining abilities, many make appropriate adjustments to their driving [Holland and Rabbit, 1992]. This suggests an important role for education.

The aim of this study was to gain an understanding of older drivers’ self-regulation practices. This paper focuses on the extent and type of self-regulatory practices and the characteristics of self-regulators in a group of Australian drivers.

METHOD

This study surveyed 656 drivers, recruited through seniors’ newspapers, auto club magazines, seniors’ clubs, retirement villages and local government aged care services. Eligible participants were drivers aged 55 years and older living in the State of Victoria, Australia. (Former drivers were also interviewed but these findings are not reported in this paper.)

The interview questionnaire was designed to gather information on the driving patterns of older drivers, transportation needs and decisions about driving cessation. A pilot study was conducted and subsequently, questions were refined with assistance from an expert panel of professionals in road safety (Project Advisory Committee). Interview questions were related to current driving patterns; recent changes in driving; confidence, difficulty and avoidance of driving situations; self-assessment of driving ability and functional abilities for safe driving. Drivers were also asked about demographic characteristics, health status and medical conditions, crash involvement and infringement history. Five experienced telephone interviewers conducted the interviews.

This paper focuses on a sub-set of issues addressed in the questionnaire including the driving patterns, confidence and avoidance of driving situations, medical status, functional abilities, crash involvement and infringement history of older drivers and the identification of variables associated with selected self-regulatory practices.

Descriptive statistics were used to describe the demographic characteristics, driving patterns and frequency of different types of self-regulatory practices amongst the sample. Chi-square analyses were conducted to explore the relationship between confidence and avoidance of specific driving situations. In addition, multivariate logistic regression modelling was used to describe the characteristics of those who adopted self-regulatory practice. In these analyses, the dependent measures were defined in terms of (i) weekly driving distance (self-regulators defined as those who drove 100 kilometres or less and non-self-regulators those who drove more than 100 kilometres) and (ii) avoidance of any driving situations (self-regulators defined as those who avoided any of the specified driving situations and non-self-regulators those who did not avoid any situations).

RESULTS

All participants (n=656) described themselves as ‘current drivers’ and all held a valid driver’s licence. Twenty-six percent of participants were aged between 55 and 64 years, 38 percent were aged between 65 and 74 years old and 36 percent were aged 75 years and older. Sixty-five percent of drivers were males. The majority of drivers (61%) indicated that they were not working while approximately one-third worked part time (including voluntary and paid work) and 10 percent worked full time. Approximately two-thirds (66%) of participants were married (including de facto or common law relationships), 20 percent widowed, 10 percent separated or divorced and 4 percent never married.

MEDICAL CONDITIONS AND SELF-RATED FUNCTIONAL ABILITIES FOR SAFE DRIVING

The three most commonly reported conditions were vision problems (77%) [these were predominantly presbyopia (33%) and myopia (27%)], arthritis (41%) and heart problems (23%). Drivers’ self-ratings of their functional abilities for safe driving were also considered (see Table 1). Participants were more likely to rate their overall health for safe driving as good (53%) rather than excellent (45%). Approximately half of the drivers rated their vision for safe driving during the daytime as excellent (49%) while about half rated their vision good (49%). Similar ratings were observed for decision-making, strength (upper and lower body) and head and neck mobility for safe driving, with the majority rating their abilities in these areas as either good (57–59%) or excellent (30–39%). Interestingly, for night driving, fewer participants (24%) rated their vision as excellent than they did for all other driving-related abilities (30–49%).

Table 1.

Self-Ratings of Functional Abilities for Safe Driving

Self- Rating	Overall health %	Vis day %	Vis night %	Dec spd %	Upp body strnth %	Low body strnth %	Head neck mob %
Exc	45	49	24	38	39	38	30
Good	53	49	52	58	57	58	59
Fair	1	2	19	4	4	4	11
Poor	<1	<1	5	<1	0	0	<1
Total	100	100	100	100	100	100	100

DRIVING PATTERNS

Overall, 86 percent of drivers reported that they were the principal driver in the household. Males were more likely than females to be the principal driver (90% vs. 78%) and drivers 75 years and older (90%) were more likely than those aged 55–64 years (85%) and 65–74 years (83%) to be the principal driver. Table 2 summarises the driving patterns of participants and reported changes in driving over the past five years.

Table 2.

Driving Patterns of Participants

Driving characteristic	Percent	N
Weekly distance driven (km)
≤ 20	4	24
21–50	9	58
51–100	20	128
101–200	21	138
> 200	46	306
Total		654
Change in driving frequency
More	14	92
Same	45	292
Less	41	271
Total		655
Change in driving speed
Faster	2	14
Same	58	381
Slower	40	261
Total		656
Change in driving quality
Better	11	73
Same	80	524
Not as good	9	58
Total		655

Approximately two-thirds of participants drove more than 100 kilometres weekly (67%). Seventy-eight percent of drivers indicated that they were driving about as much as they would like to. The majority of drivers (86%) indicated that they drove the same amount (45%) or less (41%), while comparatively few (14%) indicated that they drove more, compared to five years ago. Reasons for driving less were primarily related to changes in employment status (34%) or changes in lifestyle such as moving house (38%). Relatively few drivers (17%) identified health or age-related issues and a lack of confidence (6%) with specific driving situations, suggestive of appropriate self-regulation.

The majority of drivers reported driving about the same (58%) or slower (40%) now compared with five years ago. Of those who indicated that they drove slower, most attributed this to safety-related issues (38%), including speed restrictions (34%) and enforcement (16%).

Most drivers (80%) thought that their quality of driving was about the same as it was five years ago. Only a relatively small proportion (11%) of those who indicated that their quality of driving had changed said they believed their driving was ‘better’ while 9 percent rated their driving as ‘not as good’ as it was previously.

Drivers were asked to rate their level of confidence and to indicate whether they intentionally avoided specific driving situations in the previous six months (Table 3). In general, participants were very confident in the majority of situations. This was particularly evident for right hand turns with signals (i.e. fully controlled turning phase) (93%). In contrast, fewer drivers (55%) were very confident when driving at night and only 44 percent of drivers indicated that they were very confident driving at night when wet.

Table 3.

Confidence Ratings and Avoidance Patterns for Different Driving Situations

Driving situation	Confidence level (%)
	Very	Mod	Not at all	Avoid %)
Rain	61	38	1	14
Merging	68	30	2	6
Busy traffic	71	28	2	22
Night	55	36	9	25
Night when wet	44	44	12	26
Changing lanes	74	25	<1	15
Intersectn: no light	70	28	2	-
RH turn: no light	75	23	2	-
RH turn: light/no arrow	77	23	<1	-
RH turn: light/arrow	93	7	0	-
Roundabouts	82	17	1	-
Any intersection	-	-	-	10

Overall, the majority of participants indicated that they did not intentionally avoid various driving situations. The most commonly avoided driving situations were driving at night, particularly when wet, and driving in busy traffic.

No specific interview questions were asked about avoidance of particular types of intersections. However, responses relating to avoidance of ‘any intersections’ were followed up for types of intersections avoided. Of the 10 percent of drivers who indicated that they avoided intersections, the majority indicated that they avoided intersections without traffic lights (77%) and intersections without fully controlled right turn arrows (30%). The most frequently reported reasons for avoiding intersections were concerns for safety and crash avoidance. Reasons for avoiding other driving situations were:

Rain:	Safety factors (66%)
Merging:	Personal preference/comfort factors (28%)
Busy roads:	Personal preference/comfort factors (40%)
Night/Wet night:	Vision problems (53%/54%, respectively)

Those driving situations with the highest reported levels of avoidance behaviours were examined in more detail. Of interest was whether drivers with lower confidence ratings were more likely to avoid driving situations than those with higher ratings. A relationship was found between confidence ratings and avoidance of driving at night (χ²(1)=189.1, p<.0001), driving at night when wet (χ²(1)=148.71, p<0.0001), and driving in busy traffic (χ²(1)=74.74, p<0.0001). As shown in Table 4, those drivers who were very confident were much less likely to avoid each of these driving situations than those who were moderately confident or not at all confident. However, only about half of those drivers (44–50%) who said they were moderately confident or not at all confident, reported avoiding these driving situation as a general rule, while the remainder (50–56%) of this less confident group did not intentionally avoiding these situations.

Table 4.

Percentage of Drivers Who Avoid Driving in Certain Conditions by Confidence

	Percentage Who Avoid Driving:
	At Night			Wet Night			Busy Traffic
Confidence	Yes	No	N	Yes	No	N	Yes	No	N
Very	4	96	364	2	98	284	13	87	462
Mod/Not	50	50	292	45	55	367	44	56	193
Total	25	75	656	26	74	651	22	78	655

CRASH AND INFRINGEMENT HISTORY

Fifteen percent of drivers reported that they had been involved in a crash in the past two years. Most of these occurred during the day (83%). Seventy-two percent of all crashes occurred on the road, 7 percent occurred on private property and 21 percent in carparks. Fifteen percent of drivers had incurred an infringement, other than a parking fine, in the past two years. Males (18%) were more likely to incur an infringement than females (11%).

CHARACTERISTICS OF SELF-REGULATORS AND NON-SELF-REGULATORS

Logistic regression was used to model the self-regulatory behaviour of older drivers. The purpose of the modelling was to identify key characteristics that were indicative of self-regulatory behaviour in the study group of older drivers. It was determined that an appropriate procedure would be to classify drivers using a dichotomous outcome: those who exhibited self-regulatory behaviour and those who did not. Exploratory analyses considered two dependent variables considered exemplary of self-regulatory behaviour:

• Weekly driving distance (≤100km = self regulator);

• Avoidance of one or more driving situations (Yes = self-regulator).

Although other self-regulatory practices were examined in this study, it was not practical or statistically desirable to explore all of these options using regression modelling. Weekly driving distance provided an indication of level of reduced exposure as a self-regulatory strategy. In addition, the avoidance variable provided some useful insights into characteristics of drivers who engaged in avoidance of selected driving situations identified in the literature as potentially difficult for older drivers.

The initial selection of predictor variables was made on the basis of a priori knowledge (previous literature) and included various demographic variables as well as medical conditions, functional abilities for safe driving and crash history. The statistical significance of these potential predictor variables was first confirmed using chi-square analyses.

The logistic regression analyses measured the proportionate change in the odds ratio when moving from the reference category to the category of interest (e.g. females, compared with the reference group of males). These are reported in the summary tables below as relative odds ratios.

Summary figures for multivariate logistic regression analyses based on self-regulation of weekly driving distance (≤ 100 km) are shown in Table 5. The analyses identified characteristics that may be considered typical of a self-regulating driver, and conversely, a non self-regulating driver. Odds ratios (and 95% confidence intervals) are reported for each variable, adjusted for all other variables in the model.

Table 5.

Summary of Multivariate Model Statistics for Prediction of ‘Weekly Driving Distance ≤ 100km’

				95% C.I. for Odds Ratio
Variable	Reference	Sig.	Relative Odds Ratio	Lower	Upper
Gender	Male	0.001	2.10	1.37	3.22
55-64 yrs	75yrs+	0.001	0.39	0.23	0.67
65-74 yrs	75yrs+	0.001	0.50	0.33	0.76
Arthritis	No Arth	0.068	1.41	0.98	2.04
Decisions	Good/fair/poor	<0.001	0.50	0.33	0.74
Princ Driver	No	0.005	0.47	0.27	0.80
Marit Stat	Not married	0.062	0.66	0.43	1.02
PT Empl	Retired	0.002	0.51	0.34	0.77
FT Empl	Retired	0.005	0.26	0.10	0.66
Constant		0.059	2.07

Females were more likely to be self-regulators with respect to weekly driving distance than males. The adjusted odds ratio (see Table 5) for gender (with males as the reference group) was 2.10. Age appeared to be monotonically associated with self-regulatory behaviour. The odds of a 55–64 year old driver being a self-regulator was only 39% that of a driver aged 75 years or older, while the odds of a 65–74 year old driver being a self-regulator was 50% that of a driver aged 75 years or older. Thus, as age increased the odds of being a self-regulator increased. Employment was also monotonically associated with self-regulation of driving distance. The odds of a driver who is part-time employed being a self-regulator was 51% that of a retiree, while the odds of a driver who is employed full-time being a non self-regulator was 26% that of a retiree. Drivers with arthritis were 1.4 times as likely to be self-regulators than those without arthritis. The odds of self-regulating for those who rated their speed of decision-making as excellent was only 50% that of drivers who rated their decision-making as good, fair or poor. Those who were the principal driver were less likely to be a self-regulator as those who considered themselves not to be the main driver in the household (OR: 0.47). In addition, the odds of drivers who were not married (nor in a de facto/common law relationship) being a self-regulator was only 66% that of a married driver.

In sum, those who drove shorter distances (100 kilometres or less per week) were more likely to be

• Female;

• 75 years and older;

• Retired;

• With arthritis;

• Good/fair/poor decision-making for safe driving

• Not the principal driver in the household;

• Not married.

Summary figures for logistic regression analyses based on avoidance of driving situations are shown in Table 6. Three of the same variables were found to be predictive of self-regulation as were identified for weekly distance travelled: gender, age and principal driver status.

Table 6.

Summary of Multivariate Model Statistics for Prediction of ‘Avoidance of Any Driving Situations’

				95% C.I. for Odds Ratio
Variable	Reference	Sig.	Relative Odds Ratio	Lower	Upper
Gender	Male	0.001	1.8	1.26	2.57
55-64 yrs	75yrs+	0.076	0.68	0.45	1.04
65-74 yrs	75yrs+	0.215	0.79	0.54	1.15
Princ Driver	No	0.011	0.51	0.30	0.86
Vision	No	0.028	1.53	1.05	2.25
Accident	No	0.069	1.54	0.97	2.46
Constant		0.053	1.90

Females were more likely than males to be self-regulators (OR: 1.8). Age was monotonically associated with self-regulatory behaviour. The odds of a 55–64 year old driver being a self-regulator was 68% that of a 75+ year old driver, while the odds of a slightly older 65–74 year old driver being a self-regulator were 79% that of a 75+ year old driver. That is, as age increases the odds of being a self-regulator increased. The odds of avoidance self-regulatory behaviours amongst those who were the principal driver was 51% that of drivers who considered themselves to not be the main driver in the household. A self-reported vision problem was a significant predictor of avoidance of specific driving situations. Those with vision problems were approximately 1.5 times as likely to be self-regulators as those without vision problems. Interestingly, accident involvement was also a significant predictor of avoidance. Those drivers who reported being involved in an accident in the last two years were approximately 1.5 times as likely to be a self-regulator than those who had not been involved in an accident.

In sum, those who avoid any of the designated driving situations were more likely to be

• Female;

• 75 years and older;

• With a vision condition;

• Not the principal driver in the household;

• Involved in an accident (last 5 years)

Additional exploratory analyses were conducted to consider confidence rating as a potential predictor variable given that bivariate analyses revealed significant relationships between avoidance and confidence for three specific driving situations. For this model, confidence ratings were categorised dichotomously (very confident versus moderately or not at all confident) in any of the eight situations. This model showed that confidence was highly predictive of avoidance of any of the designated driving situations. Those who were very confident in all of the eight driving situations were 4 times as likely to be non self-regulators (OR: 4.1, 95% CI: 2.9 and 5.7). Interestingly, the inclusion of the ‘confidence’ variable in the regression model tended to reduce the relative importance of all the other predictor variables (gender, age, vision and accident involvement). This finding warrants further investigation.

DISCUSSION

The broad aim of the study was to describe the prevalence and types of self-regulatory practices adopted by older drivers and to identify characteristics of those who self-regulate and those who do not. Telephone interviews were conducted with 656 drivers aged 55 years and older from urban areas, country towns and rural areas in the state of Victoria, Australia.

Over two-thirds of all drivers reported driving more than 100 kilometres weekly and the majority were satisfied with their amount of driving. Although the general perception of drivers was that their driving quality remained unchanged, about 40% of drivers said that they were driving less and slower now than they were 5 years ago. Reasons for reductions in the amount of driving included general lifestyle changes, such as moving house and employment changes while fewer than 20% of drivers who reduced their amount of driving attributed this to health or general ageing issues. Reasons for driving slower predominantly focussed on safety issues and adherence to road rules.

Overall, the majority of drivers reported being very confident in the majority of driving situations. Two notable exceptions were for night driving and driving at night when wet, where ‘very confident’ ratings were made by only around half of the drivers. A relatively small proportion of drivers reported avoiding various driving situations. Highest avoidance levels were seen for busy traffic, night driving and driving at night when wet. About one-quarter of the participants reported avoiding these situations. More than half of drivers who avoided night driving or driving at night when wet reported doing so because of vision-related issues, especially adjusting to glare from lights. This was consistent with findings that approximately one-quarter of participants rated their vision for night driving as only fair or poor. The most common reason for avoiding busy traffic was not related to health or functional ability but rather to personal preference, with many reporting that busy traffic was not enjoyable and made them feel uncomfortable.

The findings also highlighted that not all drivers with lower confidence ratings were adopting appropriate self-regulatory practices. Indeed, only one-half of those participants who rated their confidence as good, fair or poor, who arguably should avoid these driving situations, reported self-regulating. It is also worth noting that, while few in number, some participants with high confidence ratings were also self-regulating when perhaps they may not need to do so.

Regression modelling identified key characteristics of self-regulators amongst older drivers. Not surprisingly, those who drove no more than 100 kilometres per week and those who avoided specific driving situations were more likely to be 75 years and older. Gender was also a significant predictor of self-regulatory practices with females more likely than males to drive fewer kilometres weekly and avoid potentially risky driving situations. Interestingly, the presence of another driver in the household was associated with both types of self-regulatory behaviours. Those who were not the principal driver in the household were more likely to be self-regulators than those who were not the principal driver. This is an intuitive result since the presence of another driver would provide an opportunity for sharing the amount of driving and possibly when and where trips were taken. Curiously, those who were not married (or in a defacto or common law relationship) were also more likely to be self-regulators in terms of limiting the weekly travel distance. It is possible that having a partner creates a greater need for travel and possibly leads to longer and/or more frequent trips. Of interest was the relationship between self-regulation and self-reported medical conditions and functional abilities. Previous studies have identified vision impairment and cognitive impairment as a significant predictor of both amount of driving and avoidance of driving in certain situations [Ball et al., 1998; Stutts, 1998; Lyman et al., 2001]. In this study, the presence of arthritis was associated with reduced driving distance while vision problems were associated with avoidance of potentially risky driving situations. Only one of the self-reported functional abilities assessed in this study, speed of decision-making for safe driving, was associated with self-regulation. Those who rated their decision-making as only good, fair or poor (rather than excellent) were more likely to drive fewer kilometres per week. However, this variable did not influence avoidance behaviours. One might expect, for example, that those who rated their decision-making as less than excellent, might also opt to avoid complex intersections or driving in busy traffic which arguably place a load on information processing capacities. A unique finding of this study was that driving confidence was associated with avoidance behaviours. Those who rated themselves as moderately or not at all confident in at least one of the eight driving situations were more likely to be self-regulators.

Interestingly, crashes were more prevalent amongst those who self-regulated by avoiding potentially risky driving situations. It is important to note that the survey did not provide information about the relative timing of crashes and adoption of self-regulatory behaviour. However, it is plausible that older drivers were likely to avoid potentially risky driving situations following their involvement in an accident. Future research is needed to better identify the relationship between self-regulatory driving practices and crash risk.

LIMITATIONS OF THE STUDY

Notwithstanding the use of multiple recruitment methods and sources, it is possible that the sample may not be entirely representative of Australian drivers aged 55 years and older. For example, it could be argued that the sample of volunteers may be more interested in driving and have a greater need to drive than the general population from which they are drawn.

Due to constraints on interview time, it was not possible to explore changes in self-regulatory practice using a longitudinal study design. Hence, it was difficult to determine whether those who avoided certain driving situations had done so only recently or whether they had avoided these situations all their driving life. In addition, functional abilities related to driving were assessed by self-report. Ideally, these should be measured using standardised tests of cognition and other abilities related to driving. However, one could also argue that if drivers perceive their abilities in these areas to be less than optimal for safe driving, then this indeed may be sufficient justification for self-regulation. While this might be the case for the majority of drivers, those with poor cognitive capacity and poor insight are likely to provide inaccurate information about their health status or driving patterns.

CONCLUSIONS AND RECOMMENDATIONS

The results of this study confirmed for a sample of Australian drivers many of the findings from previous research with drivers in other countries [e.g. Ball et al., 1998; Hakemies-Blomqvist and Wahlström, 1998; Lyman et al, 2001; Kostyniuk, and Shope, 1998; Marottoli et al., 1993; Persson, 1993]. In general, this study found evidence for age-related changes in reduced driving distances as well as avoidance of specific driving situations. In addition, a major contribution of this study has been to explore characteristics, other than age, that are associated with self-regulatory driving practices.

This study has provided a rich source of information about drivers’ self-regulatory practices. The findings highlight the need for strategies to promote through educational materials and programs, the adoption of self-regulatory practices consistent with declines in functional ability and presence of medical conditions known to be associated with crash risk. The study also highlighted the need for further research to explore the relationship between self-regulation and functional impairment, using standardised tests of cognition, attention, visual perception etc to assess functional abilities. In addition, case control studies with crash-involved and crash-free drivers might be useful in developing a better understanding of the effectiveness of self-regulation in reducing crash risk.