Sheila Hancock, an actress who has given pleasure to thousands for decades, is a veteran motorist who has driven for 60 years, has had no driving accidents for which she was to blame and so has made no insurance claims. The Admiral Insurance Company celebrated her 82nd birthday by raising the premium for her 3-year old mini-Cooper from £873 to £ 2,246, i.e. £ 1373 in a single year. She shared her outrage with the Guardian Newspaper who researched actuarial statistics showing that, on average, claims by drivers aged from 18 to 20 average 67%, from 46 to 50 66% and from 76 to 80 65% of the premiums that they pay. Accident rates are highest in the teens and twenties, lowest in middle age rise only very slightly during the 70’s. So that these equalities mean that, from our 2nd to 7th decades insurers vary our premiums according to our projected accident costs. After age 80 these reassuring equalities break down. Surprisingly for those still driving in their 90’s insurers’ accident-cost/premium ratio drops to 52% so that insurers make 13% to 15% profit from nonagenarians than from teenagers. The Guardian quotes no accident cost/premium ratio for 80 year olds but, irrespective of car-type or claims history, it seems that an average 82 year old pays £392 while an average 62 year old pays £286 and this difference is increasing. During the past year average premiums have risen by £35 for 80 year olds but dropped by £50 for teenage drivers. Do sharp declines in driving competence in the late 70’s and 80’s justify this? Accident statistics show only a very slight rise in reported accidents between 70 and 90. This is not completely helpful when comparing claims costs that may reflect different kinds and numbers of accidents which may be more to less expensive for insurers.
Another difficulty is that as populations age so decade-group averages become increasingly poor indices of ability. In spite of clear everyday evidence we tend to ignore the blatant fact that variability in competence between people increases with group age so that differences between the least and most capable markedly widens. The most able 70 year olds in the UK are still running industries, steering politics, writing excellent novels or scientific papers or, like Ms Hancock, giving delightful performances but between 30% and 45% percent of their co-evals need sheltered accommodation or full-time care. In our Manchester longitudinal study a dwindling number of volunteers showed little or no measurable changes in mental ability as they aged from 65 to 85 but increasing numbers of their less fortunate peers suffered illnesses such as cardiovascular problems and diabetes that caused rapid losses of competence and earlier deaths [1,2]. We must expect that as motorists age a diminishing few will remain competent through their 60’s, 70’s and even in their late 80’s or 90’s, but an ever increasing majority will become at risk. Are there more convenient and fairer measures than accident statistics to discover motorists who are becoming risky so that they can give up driving before their increasingly probable accidents happen ?
My personal research on some problems of older drivers showed me that trying to answer this question is not only very hard but can get one into highly emotional discussions. Pragmatic traffic policemen saw me as a deranged scientist recklessly keen to unleash hordes of grey slayers. Older motorists saw me as a callous age-traitor, keen to snatch away their small joys and freedoms. How to recognize unsafe elderly before they cause damage? My favorite anecdote, to hint that this might not be easy, was that during most summers newspapers print stories of elderly drivers who bravely set out from their homes and are discovered, a day or more later exhausted and dehydrated in some distant part of the country. Obviously they were at grave risk while they were lost and bewildered, and we might well advise them to give up driving. But…….. We must nevertheless recognize that they have driven for very many hours and miles, often in a state of terminal fatigue, and without causing an accident. One explanation is that perceptual and motor “driving skills” sometimes seem to survive gross confusions of higher order thinking. Another is that the vigilance and competence of other road users makes the driving environment more forgiving than we usually suppose. (Current easy availability of GPS and other in-car navigation systems is probably not a good solution for such unfortunate geographical derangements because, as motorists grow older they have greater difficulty in attending to two things at once. This not only makes them uncertain in heavy urban traffic and at busy intersections  but particularly confuses them if they try to simultaneously drive and follow a GPS ). Road traffic policemen and other experts were generally unsympathetic and countered my tales of accident– free driving by deeply confused elderly with accounts of roadside checks that pick up oldies who, even when wearing their spectacles, are so visually disabled as to be almost blind. My ripostes with data from elderly motorists in Marin County California who responded to visual difficulties in a timely way in spite of great personal inconvenience seldom cut any ice  If people cannot recognize that being nearly blind puts them at risk how can we trust them to become aware of their more subtle problems?
There is a huge literature on when and why people give up driving. A 1995 U.S. study of the driving expectancy of 4699 motorists aged 70 and over  found that those aged from 70 to 74 could expect another 11 years of driving, often terminated only by their deaths but also by voluntary withdrawal. In a typically thorough 1992 Finnish study  all license holders born in 1922 were asked how and why they had decided to stop or continue. The main reason for stopping was poor health, though most also reported increased driving stress. All but 6.9% of those who had given up said that they had done so without advice from family, friends or medics. Like all other surveys this study also found that withdrawal from driving is not an abrupt decision but a gradual process. Older drivers gradually drive less, year on year, and become increasingly selective about the times of day and the routes on which they drive until they discover that even reducing difficulties in this way is not enough, and give up completely. A key point is that this self-modification of driving behaviour does, in general, seem to work very well. The numbers of accidents for older drivers who maintain a high mileage are equal to, or less than those of their peers who now drive much less . Clearly self-monitoring is effective, and people know when they are still safe and when they are beginning to be at risk. Recognition of driving stresses alters their behaviour until they finally stop driving, in spite of the strong temptation to give themselves the benefit of the doubt. Giving up driving is hard, particularly for men, who tend to feel that a car is essential for more aspects of their lives than women do and people living in rural areas may have very little access to other forms of transport to manage their daily lives. US studies have found that having to give up driving greatly attenuates the lives of both men and women by shrinking the already sparse range of out-of-home activities that remain possible for them . This difficult decision can be eased by provision of check lists such as the “Driving decisions Workbook”  or the Manchester Driver Behaviour Questionnaire  which, by probing their memories of difficulties they have encountered in various driving situations can help them to become more clearly aware of their problems – or, indeed, to relieve them of anxieties and more cheerfully plan their remaining driving careers.
So, rather than simply stopping people from driving at some determined age or waiting for them to disqualify themselves can we find “gold standard” tests that will identify older drivers who are becoming risky ? A study of 1910 elderly drivers by Karlene Ball and colleagues  found that among individuals with adequate vision, being older, being a man, having a history of falls, having poorer scores on a behavioral test of frontal lobe function (the “trails test”) and on a complex test of attention (Ball’s useful field of view test) all predicted future risk of self-at-fault crashes. Most other studies roughly agree with at least some of these findings, but there is a growing consensus that health problems and the biological changes accompanying old age are even more powerful predictors. Reading about Ms Hancock’s confrontation wither the Admiral Insurance Company made me re-visit data gathered and published many years ago on a sample of 555 gallant Manchester volunteers who not only took a battery of more than 50 different physiological and mental tests but agreed to be evaluated, while driving their own cars, by experienced driving instructors (see, for instance,  and other papers by the same authors). The conclusions, given below are from an exceptionally thorough re-analysis of this data-set by Mei Foong Low . As in other analyses of this data set Mei Foong Low found that quite simple biological indicators such as strength of hand-grip or the maximum force of leg-thrusts and a measure of lung capacity did significantly predict both whether or not individuals has experienced any accidents during the last 5 years and how many accidents they had experienced. The new twist was that if drivers’ self -reports of all illnesses with which they had been had been diagnosed were also taken into consideration predictions from grip strength, leg-thrust and lung capacity were greatly weakened or disappeared. This suggests that these rather arbitrary and peripheral measures of well-being mainly pick up differences in general health status and wellness that are more basic determinants of driving skills. Among other significant predictors were peripheral vision, hearing, balance, joint flexibility and of decision speed and attention, including the tests of attention and of frontal lobe function included in the study led by Karlene Ball. Though, in our study, neither Ball’s test of frontal function (the Trails test) nor her test of attention (The Useful Field of View test) significantly predicted frequency of accidents, two of our other tests of attention and decision speed did. However the catch was the same as has been found in her study and in all of the others that I have read. While statistical tests confirm that associations between peoples’ scores on some tests test scores and their accident records are significant, in the sense that they are much stronger than we would expect to happen by chance, no test, or measure, or combination of measures has accounted for more than a very small amount of the huge variability between individuals. In fact even our” best” tests accounted for no more than 2% of the total variation in accident records between people. In short, within a very large group of people those who had poor test scores were indeed more likely to have had accidents the power of this prediction was so low that it would not be sensible or fair to make decisions about whether particular individuals were at risk only on the basis of their test results. Among many possible reasons why they are such weak predictors is that the vast experience of driving that individuals gain during their lifetimes compensates for other declines in their mental and so emerged in Mei Foong Low’s exceptionally thorough analyses as a significant predictor of their accident records.
I would love to encourage the scintillating Ms. Hancock in her confrontation with her insurers by giving her numbers to persuade them that because we are all so very different from each other, and because these differences increase as we age, it is grossly unjust to lump all of us into very rough categories so that the most able and harmless of us can subsidize the costs of the more dangerous. I am sure that this is the case, and that more effective and equitable means must and can be found. I am very sorry that this is not yet, and certainly not by me.
1. Rabbitt, P., Lunn, M., & Wong, D. (2006). Understanding terminal decline in cognition and risk of death. European Psychologist, 11(3), 164-171.
2. Rabbitt, P., Lunn, M., Pendleton, N., & Yardefagar, G. (2011). Terminal pathologies affect rates of decline to different extents and age accelerates the effects of terminal pathology on cognitive decline. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 66(3), 325-334.
3. Ponds, R. W., Brouwer, W. H., & Van Wolffelaar, P. C. (1988). Age differences in divided attention in a simulated driving task. Journal of Gerontology, 43(6), P151-P156
4. Dingus, T. A., Hulse, M. C., Mollenhauer, M. A., Fleischman, R. N., McGehee, D. V., & Manakkal, N. (1997). Effects of age, system experience, and navigation technique on driving with an advanced traveler information system. Human Factors: The Journal of the Human Factors and Ergonomics Society, 39(2), 177-199.
5. Brabyn, J. A., Schneck, M. E., Lott, L. A., & HaegerstrÖM-Portnoy, G. (2005). Night driving self-restriction: vision function and gender differences. Optometry & Vision Science, 82(8), 755-764.
6. Daniel J. Foley, Harley K. Heimovitz, Jack M. Guralnik, and Dwight B. Brock. Driving Life Expectancy of Persons Aged 70 Years and Older in the United States. American Journal of Public Health: August 2002, Vol. 92, No. 8, pp. 1284-1289
7. Hakamies-Blomqvist, L., & Wahlström, B. (1998). Why do older drivers give up driving?. Accident Analysis & Prevention, 30(3), 305-312.
8. Langford, J., Methorst, R., & Hakamies-Blomqvist, L. (2006). Older drivers do not have a high crash risk—A replication of low mileage bias. Accident Analysis & Prevention, 38(3), 574-578. Marottoli, R. A., de Leon, C. F. M., Glass, T. A., Williams, C. S., Cooney, L. M., & Berkman, L. F. (2000). Consequences of driving cessation decreased out-of-home activity levels. The Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 55(6), S334-S340.
9. Eby, D. W., Molnar, L. J., Shope, J. T., Vivoda, J. M., & Fordyce, T. A. (2003). Improving older driver knowledge and self-awareness through self-assessment: The driving decisions workbook. Journal of safety research, 34(4), 371-381.
10. Parker, D., McDonald, L., Rabbitt, P., & Sutcliffe, P. (2000). Elderly drivers and their accidents: the Aging Driver Questionnaire. Accident Analysis & Prevention,32(6), 751-759..
11. Ball, K. K., Roenker, D. L., Wadley, V. G., Edwards, J. D., Roth, D. L., McGwin, G., … & Dube, T. (2006). Can High‐Risk Older Drivers Be Identified Through Performance‐Based Measures in a Department of Motor Vehicles Setting?.Journal of the American Geriatrics Society, 54(1), 77-84.
12. Mei Foong Low (2003) Assessing the Association Between Cognitive and Physiological Measures and Car Accidents among Older drivers. MSc thesis , Department of Statistics, University of Oxford.