Cognitive prerequisites for fitness to drive: Norm values for the TMT, UFOV and NorSDSA tests

Abstract

Background: Fitness-to-drive assessment is a growing area for occupational therapists. There are few off-road tests specially developed to assess fitness to drive, and several cognitive tests have no age-specific norms.

Aims/objectives: The aim was to identify and describe age-related norm values for the Trail Making Test, Nordic Stroke Driver Screening Assessment and Useful Field of View test, and to study inter-correlation between test results.

Materials and methods: The sample included 410 volunteers; 149 men and 261 women, mean age 52 ± 16.8 years. Commonly used off-road tests were used: TMT A and B, UFOV and NorSDSA.

Results: Normative data for the specific subtests and total score for NorSDSA and UFOV are provided and presented in four age groups. Age correlated with the results for most of the subtests.

Conclusions: Off-road cognitive test scores are necessary and valuable for occupational therapists in their contribution to the final decision on continued driving. In clinical practice, it can be difficult to interpret cognitive test results when working with driving assessments. Age-based norm values are suggested to be a way to provide clinicians with a benchmark against which scores can be compared.

Significance: Age-based norms can guide occupational therapists working with fitness to drive.

Keywords:

• Cognition
• driving assessment
• off-road test
• screening

Background

Community mobility and driving are important and growing areas for occupational therapists (OTs), especially fitness-to-drive assessments [1]. Some medical conditions, for example dementia and cardio-/cerebrovascular diseases, are likely to affect safe driving as a result of cognitive, perceptual and/or physical impairments [2]. Driving is a complex activity; it requires cognitive functions as well as appropriate behavior, based on the ability to make well-considered decisions in each traffic situation [1].

Cognitive tests are considered to be an important part of clinical decision making on continued driving or not in patients with cognitive impairments due to brain injury or disease [3]. There are a variety of screening tools, i.e. off-road tests, used for fitness-to-drive assessments, and these vary among countries [4,5]. However, executive functions as well as attention has been found to have large explanatory factors to on-road driving results and thus should be included in a cognitive assessment focusing on identifying persons who may have an increased risk for unsafe driving due to cognitive dysfunction [6–8]. In addition, an off-road assessment can help to identify strengths and weaknesses that may have an impact on fitness to drive and thus identify persons who may have an increased risk for unsafe driving [6,7].

Research has often focused on relationships between different cognitive tests and performance from an on-road test or a driving simulator [9–11]. In addition, research results are often presented for different diagnostic groups such as dementia [12] or stroke [13]. Accordingly, for clinical practice where the diagnoses, ages and disabilities vary, there are no clear-cut results, and no single test has been identified that can discriminate safe and unsafe drivers [1,3,8,14]. There are few tests specially developed to assess fitness to drive [4,15] and available tests vary from country to country. The Stroke Driver Screening Assessment (SDSA) is an example of a test designed as a screening measure to identify cognitive impairments of importance for safe driving. The SDSA has previously been evaluated for predictive validity [16], concurrent validity [17] and reliability [18]. A Nordic version has been further developed from the SDSA (the NorSDSA), and has been validated for stroke patients [19]. However, in a follow-up study, the total score for the Nordic version was shown to have low predictive value, i.e. to predict an on-road test result, for both stroke patients and patients with cognitive deficits/dementia [14]. It was recommended that NorSDSA be used together with other cognitive tests. Other cognitive screening tools have been developed and used by OTs outside the Nordic countries, such as the Drive Home Maze test [20], DriveSafe DriveAware [21] and OT-Dora [22]. However, based on our knowledge, OTs working with driving assessments in the Nordic countries often use the Trail Making Test (TMT), the NorSDSA and the Useful Field of View Test (UFOV). Those tests are available to OTs; they have been translated and are thus the most commonly used for driving assessments [23]. Even though the total score for the NorSDSA has been shown to be less predictive to identify unsafe drivers with different diagnoses, it would be of interest to look at each subtest to identify the most valid subtest for screening in clinical practice [14,24]. An on-road assessment allows the assessor to observe driving behavior, but it is time consuming and might not be possible for all OTs [1]. Moreover, even though an on-road assessment may be seen as the gold standard, it represents the driving behavior on one single occasion, and the outcome may not be reliable without completing the assessment with results from standardized cognitive tests [25]. Thus, off-road cognitive test scores are considered valuable and useful to screen patients and to guide clinical practice when determining fitness to drive [26].

Regardless of which tools are used for assessment, OTs need better support for interpretation of test results, especially when the age span among patients with cognitive impairments is large. OTs needs to rely on test results being valid, reliable and having age-based norm values for comparison [20]. Some of the commonly used tests have no or limited norm values and unknown reliability and validity. Norm values are available for TMT [27] and UFOV (for the elderly population) [28], however, the results are quite old and based on populations outside the Nordic countries, which might make them less reliable. No norm values are available for NorSDSA [14,19].

Moreover, one way to study the validity of the different tests might be to look at the inter-correlation of the test results and how the results correlate with age. Age-specific normative data based on a current Nordic sample on commonly used tests might guide therapists and physicians in analyzing and making recommendations and perhaps contribute to more equal assessments of fitness to drive. The aim of this study was to identify and describe age-related norm values for TMT, UFOV and NorSDSA.

Materials and methods

Participants

Participants were recruited from four different geographic areas from the north to the south of Sweden. The inclusion criteria were ≥18 years and having a driving license. The exclusion criteria were a medical condition affecting cognitive function, such as stroke or head injury, which was determined by the participants’ statement and confirmed at the time of assessment. All participants were volunteers and recruited through the Swedish driving license register, advertisements, local senior organizations and by word of mouth. The sample included 410 participants, 149 men and 261 women between 20 and 89 years of age; mean age, 52 years; SD, 16.8 years. The mean age was 56 ± 18 years for men and 50 ± 16 years for women.

Measurements

The TMT A and B, NorSDSA and UFOV tests were selected because they are commonly used, standardized off-road tests with specific methods for administration and scoring.

The trail making test

The TMT is a well-known cognitive test that aims to measure visual search and sequencing, information processing speed, divided attention and flexibility [29]. The test normally consists of the two subtests, A and B, completed in the shortest possible time and scored in seconds. In previous studies, performance on the TMT has been shown to be sensitive to age (TMT A and B) and education (TMT B), and reference data are available for comparison [27]. The test has poor face validity for driving but is commonly used both clinically and in research [30].

The nordic stroke driver screening assessment

The SDSA is a set of cognitive tests developed to evaluate fitness to drive with clients after a stroke [31]. The NorSDSA is the Nordic version of the SDSA [19] but no normative data for NorSDSA or SDSA have previously been reported. The NorSDSA comprises four subtests:

1. The Dot Cancelation Test comprises a sheet of paper with rows of groups of three, four or five dots. All groups of four should be marked. Time taken (seconds), number of misses and number of incorrect/false positives are noted.

2. The Direction Test comprises a 4 × 4 squared matrix. Large arrows in four different directions are placed along the left-hand side. Another four directions but with small arrows are placed across the top. A set of 16 cards depicting trucks and cars (two on each card) are traveling in four different directions. Each card should be placed on the matrix so that each truck is traveling in the direction of a large arrow and each car is traveling in the direction of a small arrow. Maximum score is 32 in 300 s.

3. The Compass Test comprises a 4 × 4 squared matrix. Four directions are aligned for each row and another four directions for each column along the top of the matrix. Cards have a picture of a roundabout joining eight roads on which two cars are traveling on two roads. The participant has to place the cards in the square at the intersection between (a) the row corresponding to the direction of travel of one car with (b) the column corresponding to the direction of travel of the second car. One point is given for each correctly placed car. The maximum score is 32 in 300 s.

4. The Road Sign Recognition test consists of 12 cards with pictures of different traffic situations, e.g. railway crossing. Another 19 cards with pictures of traffic signs are given to the participants to match appropriate signs with the traffic situations. Total time allowed is 3 min and 1 point is given for each correct answer. In addition, the possibility to extend the response time to 5 min as suggested by Selander et al. [14] was used. The maximum score is 12.

Higher scores on Directions, Compass and Road Sign Recognition are considered better than lower scores. Six scores are derived from these subtests, but only four are entered into an equation derived from discriminant function analysis. Based on the results for Dot Cancelation (time and errors), Compass (score) and Road Sign Recognition (score in 3 min), the test provides a weighted overall total score (the higher the score the better).

The useful field of view

The UFOV test is a computer-based visual and cognitive test comprising three subtests. The first subtest aims to measure processing speed, the second measures processing speed for a divided attention task and the third measures processing speed for a selective attention task [32].

In the first task, the person identifies an object (a car or a truck) presented in the center of the screen, for changing lengths of time. In the second part, the person identifies an object (car or truck) as before but also localizes a simultaneously presented target (another car) on the periphery of the screen, which is alternately positioned at eight different points around the screen. The third task is similar to part two but the objects in the periphery are embedded in distracters (triangles) to make the task more difficult. The score represents the display speed in milliseconds (14–500 ms) at which the participant can perform the tasks correctly; lower scores indicate better (faster) performance [28].

The results from the UFOV test estimate risk by quantifying the visual field over which a driver can process rapidly presented visual information and therefore supposedly drive safely. The test has repeatedly been shown to be correlated to driving performance and the incidence of car crashes [33], and has sufficient validity and reliability [28,32]. However, norm values for a Nordic population and for a younger age group are lacking.

Statistics

All statistical analyses were performed using the Statistical Package for the Social Science (SPSS 25.0). Based on the type of variable and the distribution of the results, non-parametric, as well as parametric tests were used. A correlation analysis based on the entire material and all included tests was performed to analyze inter-correlations between different test results as well as between age and test results, using the Spearman correlation test. Interpretation of the correlation coefficient as described by Mukaka [33] was used. A correlation coefficient r_s > 0.5 was considered moderate and r_s > 0.7 was considered high. Norm values were calculated and analyzed for each subtest in UFOV as well as, NorSDSA. In addition, norm values were calculated for the sum of UFOV scores and NorSDSA total score. The participants were categorized into four groups based on age. The categorization was based on thorough evaluations of data from scatterplots as well as, statistical comparisons between groups. The Kruskal-Wallis H test was used for all analyses of test results according to age groups. p < 0.05 was considered statistically significant. To decide the sample size for the study, a sample size calculation was performed with a confidence level of 95% and an excepted margin of error of 5%, resulting in a minimum recommended sample size of 377.

Ethical considerations

No medical or other risks related to participation in the study could be identified. All participants were fully informed before giving written consent. All data forms were coded and no individual could be identified by the researchers. All procedures were in accordance with ethical standards of the Declaration of Helsinki. Ethical approval was obtained from the Ethics Regional Board in Linköping (Dnr 2016/181-31; 2016/271-32) and Stockholm (2006/1524-31). The participants were all volunteers and were not given any reimbursement.

Results

A correlation analysis of all the test results showed that age had an r value ≥0.5 for all variables except UFOV subtest 1 (Table 1). Because age was found to be strongly associated with the results for all tests, norm values are presented for different age groups. Four different age groups were identified: (1) 20–39 years, 29% men; (2) 40–59 years, 30% men; (3) 60–69 years, 36% men; and (4) 70–89 years, 59% men. Education only had an effect on some of the test results for the oldest age group and was thus not further described. Gender did not have any homogeneous effect on the test results for any age group; women aged 40–49 years and 60–69 years had somewhat lower values on UFOV III and UFOV Total (p values from 0.01 to 0.04).

Table 1. Results from spearman’s correlation test.

	UFOV 1	UFOV 2	UFOV 3	UFOV total	TMT A	TMT B	NorSDSA	Age
UFOV 1	1.00
UFOV 2	0.613	1.00
UFOV 3	0.377	0.586	1.00
UFOV, total	0.434	0.668	0.986	1.00
TMT A, time	0.365	0.436	0.571	0.579	1.00
TMT B, time	0.325	0.433	0.484	0.484	0.659	1.00
NorSDSA	−0.208	−0.358	−0.528	−0.530	−0.467	−0.512	1.00
Age	0.372	0.543	0.754	0.760	0.545	0.504	−0.502	1.00

Bold values >0.5 are accepted as moderate.

Analyses presented for the different age groups on UFOV subtests 1–3 and UFOV total score, TMT A and B and the NorSDSA total score showed a difference between groups for all tests and especially for results for UFOV 3 and UFOV total score (Table 2).

Table 2. Results from UFOV subtests and total score, TMT a and TMT B by age groups (mean, SD, median, percentiles).

	Age groups				P	H
	20–39 years (n = 107)	40–59 years (n = 139)	60–69 years (n = 89)	≥70 years (n = 75)
UFOV 1					0.001	14.4
Mean	15.69	15.88	15.88	25.27
SD	2.73	2.8	2.02	24.73
Median	14.8	14.8	15.7	16
10%–90%	14–18.2	14–19	14.8–17	14.8–50
UFOV 2					<0.001	37.6
Mean	17.45	20.72	35.62	95.06
SD	9.24	20.28	46.37	95.48
Median	14.9	15	18.3	61.5
10%–90%	14–23	14–27	14.8–66	16–248
UFOV 3					<0.001	119.4
Mean	40.25	64.16	124.67	238
SD	30.04	35.56	78.17	120.15
Median	31.8	55	103	220
10%–90%	15–63	24–112	51–220	96–399
UFOV, total					<0.001	120.7
Mean	73.35	100.77	177.39	358.33
SD	35.08	48.63	104.84	205.04
Median	64.5	92.4	138.4	305
10%–90%	44.5–107.2	56–151	80–299	140.7–701
TMT A, time					<0.001	37.3
Mean	23	26.4	32.3	39.6
SD	6.5	8.6	11.4	12.8
Median	22	25	29	37
10%–90%	15.4–32	17–37	21.6–50	26–60
TMT B, time					<0.001	32.2
Mean	55.2	57	75.1	97.9
SD	21.6	21.1	29.5	41.6
Median	52	55	70	90
10%–90%	35–82	36–85	45.5–118	53–141
NorSDSA, total score					<0.001	39.1
Mean	3.11	3.09	2.51	1.59
SD	0.74	0.62	0.76	1.09
Median	3.26	3.15	2.65	1.56
10%–90%	2.2–3.8	2.25–3.77	1.53–3.46	0.32–2.81

Norm values are presented as means and standard deviations as well as medians and percentiles for each age group (Table 3). The median values for the Direction and Compass scores in NorSDSA for the first three age groups were all the maximum score. This indicates a ceiling effect for these subtests. When analyzing differences in test results for the different subtests in NorSDSA, the results showed that there was a difference between age groups for all tests except Dot Cancelation misses. When further analyzing the results between two age groups at a time, most results were age sensitive for the two oldest age groups. When comparing the two youngest groups, the differences are less obvious in some of the tests, because these groups are performing close to maximum (Table 3).

Table 3. Results from NorSDSA subtests by age groups (mean, SD, median, percentiles).

	Age groups				P	H
	20–39 years (n = 107)	40–59 years (n = 139)	60–69 years (n = 89)	≥70 (n = 75)
Dots, time^a					<0.001	39.3
Mean	321.5	341.2	387.1	414
SD	63.8	62	83.3	92.3
Median	311	335	381	395
10%–90%	256–404	274–429	294–481	310–549
Dots, misses^a					0.748	0.6
Mean	6.1	5.7	6.4	9.8
SD	5.5	5.4	6.4	8.7
Median	4	4	5	8
10%–90%	1–14	1–14	1–13	1–25
Direction, score					0.001	9.1
Mean	31.9	31.6	31.7	29.6
SD	0.5	1.1	1.2	4.8
Median	32	32	32	32
10%–90%	32–32	30–32	32–32	24–32
Direction, time					<0.001	23.2
Mean	111.8	119	151.2	207.3
SD	50.1	51.4	67.3	68.1
Median	97	103	134	200
10%–90%	67–183	65–189	81–274	118–300
Compass, score^a					<0.001	31.6
Mean	31.2	30.8	28.7	23.9
SD	3	2.7	4.5	6.2
Median	32	32	32	24
10%–90%	30–32	88–32	22–32	15–32
Compass, time					<0.001	65.5
Mean	209.5	245	270.3	291.1
SD	52.4	47.8	41.9	18.14
Median	199	248	300	300
10%–90%	145–300	180–300	208–300	258–300
Road signs, 3 min^a					<0.001	43.9
Mean	9.7	9.7	7.8	6.2
SD	2.1	2.7	2.2	2.3
Median	10	10	8	6
10%–90%	6–12	7–12	5–11	4–10
Road signs, 5 min					<0.001	27.7
Mean	10.5	10.8	9.7	7.7
SD	1.8	1.8	1.8	2.2
Median	11	11	10	7
10%–90%	9–12	8–12	7–12	5–11

Kruskal-Wallis H test.

^aVariables included in the formula for the NorSDSA total score.

Correlation analyses based on a comparison with NorSDSA subtests showed that correlations between NorSDSA and UFOV TMT A and B were highest for those tests that were based on speed (Dot Cancelation had a correlation coefficient >0.5 with UFOV 3, UFOV Total and TMT A and B; Direction and UFOV 3 and UFOV Total; Compass and UFOV 3 and UFOV Total, as well as TMT A and B; (Table 4).

Table 4. Correlation analysis for NorSDSA subtests, UFOV subtests and TMT a and B.

	UFOV 1	UFOV 2	UFOV 3	UFOV total	TMT A	TMT B
Dot cancelation, time^a	0.131	0.383	0.637	0.640	0.598	0.537
Dot cancelation, misses^a	0.014	0.176	0.141	0.168	0.181	0.226
Direction, score	0.074	−0.166	−0.274	−0.268	−0.150	−0.125
Direction, time	0.121	0.335	0.617	0.614	0.441	0.258
Compass, score^a	−0.089	−0.277	−0.471	−0.460	−0.363	−0.447
Compass, time	0.205	0.436	0.724	0.729	0.508	0.522
Road marks, 3 min^a	−0.039	−0.258	−0.462	−0.467	−0.399	−0.391
Road marks, 5 min	0.077	−0.109	−0.290	−0.295	−0.453	−0.434

Accepted level for moderate correlation, r_s > 0.5.

^aVariables included in the formula for the NorSDSA total score.

Discussion

The results from this study offer age-based norms for cognitive tests that are used by OTs in some countries when making fitness-to-drive recommendations. The results identified that age has an effect on the results for all but one of the subtests and thus should be considered in clinical practice when interpreting test results (Tables 2 and 3). Previous studies have also shown that increasing age affects the results for the UFOV and TMT A and B [27,28]. In addition, the results showed a moderate correlation between several of the tests, which confirms that the different tests are inter-related but not to 100%. If the correlations had been even higher, one might recommend OTs to exclude one or more of the tests because in that case, the test might not add any new information. There are changes associated with ageing and health status, e.g. cognition and visual functioning, that may have a negative impact on safe driving [34]. However, there are also other factors that may have a positive impact on driving performance, such as driving experience, knowledge, driving habits and insight [6]. In addition, older drivers may limit their exposure according to traffic intensity [35], which can be a way of adapting to the existing driving environment.

Certain cognitive domains such as attention, executive functions and memory have been found to be important and, according to the driver licensing authorities, they need to be addressed in an assessment of ability to continue car driving [36,37]. No single tool exists that includes all the functions and skills needed for safe driving [15]. Thus, a combination of cognitive tests is needed, as confirmed in several reviews [1,38,39]. The lack of strong correlation between different test results in the present study further confirms that the different tests measure different cognitive abilities (Table 1). Only some of the available cognitive tests used to assess fitness to drive are presented in this study, based on the tests that are used by OTs in the Nordic countries.

The NorSDSA aims to measure several cognitive aspects, such as focused and sustained attention, mental speed and the ability to attend to two visual dimensions at the same time [31]. One of the strengths of the NorSDSA is that it has face validity, which is important in clinical practice. However, the total score, an equation derived from a discriminant function analysis, comes from one single study based on the performance of 97 stroke patients [19]. It is commonly used in the Nordic countries but as a single tool, it is inadequate to predict on-road outcome, especially for neurological conditions other than stroke; it is recommended that it is used only in combination with other tests [14]. Suggested norm values in Table 3 could be used as a support in decision making within clinical practice.

The UFOV test has repeatedly been shown to be correlated to driving performance [28,32], because lack of attention to relevant traffic objects might increase crash risk. The different subtests in the UFOV have been described to measure different cognitive areas. In the present study, subtest 3 (selective attention) was found to be the most sensitive part of the UFOV in separating the different age groups (Table 2). In addition, UFOV subtest 3 has been found to be the most valuable subtest in selecting those who need an on-road assessment [40]. In the study by Edwards et al. [28], the authors presented the data for the total group of participants. Only norms for subtest 2 were included but were they were presented in different age groups than in this study. However, overall, the older participants in this study seemed to perform slightly better than the participants in the study by Edwards et al. [28].

TMT A and B was not developed for driving assessment, however it is often used in clinical practice. It has been found to be useful to forecast unsafe driving in older persons [41] and is sensitive to age (which is confirmed in the present study; Table 2) and education [27]. In comparison with previously published norms [27], the younger participants had similar test results but participants >60 years old performed somewhat better.

Study limitations

Results from the study should be interpreted with caution due to study limitations such as the relatively small sample, gender skewness, cognitive dysfunction might preexist because no medical examination or medical record review was undertaken. According to gender skewness, there was a difference in the results for UFOV III and UFOV Total and in two of the age groups only; no other differences related to gender were found. Based on analyses of the test results by gender, we consider the gender skewness to be of less interest for the study aim and conclusions. No information on former crash involvement was available for the study population, a factor that might have affected the willingness to participate. Looking at median values compared with mean values (Tables 2 and 3), the mean values were increased for some individuals (not the same individuals for all tests). However, we believe this strengthens the description of a norm population. Because there was skewness in the results for the entire study group and all variables, we chose to describe the results as medians as well as mean values (Tables 2 and 3). When comparing the norm values with results from other studies [27,28,42], we can confirm that the oldest age group in the present study had higher scores on the UFOV test and the TMT.

Cognitive tests might not be sufficient for every assessment because aspects such as driving behavior and driving experience might be of significant relevance. A simulated driving test, some other cognitively demanding activity or an on-road assessment might serve as a complement when relevant.

Clinical implications

Return to driving after a stroke or traumatic brain injury is an important goal for many clients and therefore it needs to be included in a rehabilitation program [41]. OTs are often asked to make fitness-to-drive assessments based on a client’s cognitive and physical functions [43]. This can be a complex task because the guidelines and tools are not well defined [36]. No single test can predict driving performance [44] and in addition to cognitive tests, an assessment should also include some kind of behavioral observations when necessary. Presented norm values should be seen as a guide to provide clinicians with a benchmark against which scores can be compared. However, other factors of importance need to be considered, such as driving behavior and driving experience [45].

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Disclosure statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Additional information

Funding

This study was funded by the Swedish Transport Agency, Vinnova, County Council Östergötland and Medical Research Council of Southeast Sweden.