11-SchepersPotter.qxd Road deaths in the RSA increased tenfold from 1950 to 1998. During this period at least 300 000 people died on our roads. More than 100 000 were killed during the past ten years. In 1998 alone there were 9 068 fatalities and 36 246 serious injuries1. From this it should be apparent that the extent of road carnage, in our country, is rapidly reaching epidemic proportions. The direct cost associated with this is astronomical: The total cost of road accidents for 1998 was estimated at R12 696,7 million and this only included costs such as loss of output of the persons killed, damage to vehicles, medical costs, legal costs, etcetera2. According to the National Rapid Response Programme, conducted by the Council for Scientific and Industrial Research (Mynhardt & De Jager, 1993), man is responsible for 81,6% of road-collisions, the vehicle for 5,6% and the road and environment for 12,8%. These sources of variation are not independent of one another, but interact with one another. Driving a motor vehicle is a highly complex task and involves the driver as a total human being: His/her sensory abilities, motor abilities, cognitive abilities (including percept ual abilities, attention, information processing and alertness), affective disposition, emotional maturity, risk-taking behaviour, values, and attitudes towards road safety and authority. These are but a few of the psychological attributes that are of critical importance in driving a motor vehicle safely. Any substance (alcohol, drug or medicine), which affects these functions adversely, can aggravate the situation. During 2000 the Medical Research Council examined 1 968 transport-related deaths. (These represented 42,2% of the total of 4 658 transport-related deaths.) It was found that 54% of the cases had elevated blood alcohol concentrations (BACs). Of these, 89,2% had levels greater than 0,04% (0,04g/100ml.). Furthermore, over 50% of the drivers had elevated blood alcohol levels (mean BAC of 0,16%)3. The debilitating effects of elevated BACs on driving performance is generally acknowledged, but the negative effects of certain medicines, notably first-generation antihistamines, are less well- known. Accordingly, the sedating effects of certain antihistamines were examined. Weiler, Bloomfield, Woodworth, Grant, Layton, Brown, McKenzie, Baker and Watson (2000) examined the effects of fexofenadine, diphenhydramine, alcohol and a placebo on simulated driving performance. Their sample consisted of 40 licensed drivers with seasonal allergic rhinitis. Their ages varied from 25 to 44 years. In assessing driving performance they used the Iowa Driving Simulator. All the participants had to drive 45 miles on a two- lane rural highway. The experimental drive consisted of two phases. In the first phase the participants had to follow another vehicle for 13,5 miles. Phase 2 began when the lead vehicle turned off. The participants could then drive as they ‘normally would’. Towards the end of the drive the participants encountered a vehicle that unexpectedly turned in front of them. A truck with trailer occupied the oncoming lane. The intervention consisted of the following substances, administered at weekly intervals: � One 60 mg dose of fexofenadine � One 50 mg dose of diphenhydramine � Alcohol until the BAC reached 0,1% � A placebo For each participant the following variables were measured: � The constancy of the following distance between the two vehicles (i.e. the correlation between the participant’s vehicle velocity and the velocity of the lead vehicle). JOHANN M SCHEPERS Department of Human Resource Management Rand Afrikaans University PAUL POTTER University of Cape Town Lung Institute University of Cape Town CHRISTO VAN NIEKERK Innovara ABSTRACT The principal objective of the study was to examine the effect of Telfast® 180, an antihistamine, on driver behaviour, decision-making, and reaction time. A sample of 255 volunteers participated in the study. The sample was randomly divided into two groups, viz. a placebo and a Telfast group. One of the groups received Telfast and the other a placebo. Initially all the participants were given a standardised driving test, as well as a psychomotor test. Immediately thereafter the placebo group received their placebo tablets and the Telfast group their Telfast tablets. Exactly 2,6 hours later every participant was tested again, using the same tests as before. No statistically significant differences were found between the Placebo group and the Telfast group. No sedative effects due to Telfast® 180 were thus evident. OPSOMMING Die hoofdoelstelling van die studie was om die effek van Telfast® 180, ’n antihistamien, op bestuursgedrag, besluitneming en reaksietyd, te ondersoek. ’n Steekproef van 255 vrywilligers het aan die studie deelgeneem. Die steekproef is lukraak in twee groepe verdeel, te wete ‘n plasebogroep en ’n Telfastgroep. Om mee te begin is al die deelnemers aan ’n gestandaardiseerde bestuurstoets, asook ’n psigomotoriese toets, onderwerp. Onmiddellik daarna het die plasebogroep hul plasebotablette en die Telfastgroep hul Telfasttablette ontvang. Presies 2,6 uur later is al die deelnemers weer getoets met dieselfde toetse as voorheen. Geen statisties beduidende verskille is tussen die plasebogroep en die Telfastgroep gevind nie. Daar was dus geen sedatiewe effek as gevolg van Telfast® 180 nie. THE EFFECT OF TELFAST® 180 ON DRIVER BEHAVIOUR, DECISION-MAKING AND REACTION TIME Requests for copies should be addressed to: JM Schepers, Department of Human Resource Management, RAU University, PO Box 524, Auckland Park, 2006 76 SA Journal of Industrial Psychology, 2003, 29 (3), 76-86 SA Tydskrif vir Bedryfsielkunde, 2003, 29 (3), 76-86 � Steering instability (i.e. the root-mean-square deviation of the participant’s car from his/her preferred position in the lane). � The participant’s reaction time to the blocking vehicle. With regards to the participants’ ability to maintain a constant following distance, it was found that (1) after taking diphenhydramine their performance was significantly poorer than after taking alcohol, fexofenadine, or a placebo and (2) after taking alcohol their following distances were significantly shorter than after taking fexofenadine. Regarding steering instability it was found that (1) after taking fexofenadine the participants had significantly less steering instability than after taking diphenhydramine or alcohol (but not a placebo) (2) after taking alcohol the participants had the same or less steering instabilit y than after taking diphenhydramine (3) after taking diphenhydramine the participants crossed the centre line significantly more often than after taking fexofenadine or a placebo, and (4) after taking alcohol they crossed the centre line significantly more often than after taking fexofenadine or a placebo. With regards to the blocking vehicle, participants reacted significantly more slowly after consuming alcohol than after taking fexofenadine (2,21 seconds as against 1,95 seconds). From the foregoing it is clear that diphenhydramine impairs driving performance almost to the same extent as alcohol (for BACs of approximately 0,1%). The magnitude of the problem becomes clear if one considers that approximately 47% of Americans with allergies take first- generation-antihistamines, such as diphenhydramine (Weiler et al., 2000, p.8). It is estimated that approximately 20% of South Africans suffer from allergic rhinitis and take some form of antihistamine (Green, Potter, Plitt, Friedman, Hockman & Davis, 1998). If one considers that 866 536 drivers were involved in motor vehicle accidents during 1998 then the magnitude of the problem is clear4. The question now arises: What kind of antihistamines are safe for drivers, and how conclusive is the evidence? Vermeeren and O’Hanlon (1998) examined the effects of fexofenadine on driving performance and three psychomotor tests, namely critical tracking, choice reaction time, and sustained attention. Their sample consisted of 24 healthy volunteers. They found no evidence of driving impairment. However, there was a significant impairment of the critical tracking task for doses of 120 mg and 240 mg. The findings of Vermeeren and O’Hanlon (1998) are most important but need to be replicated on a larger sample. A 6-way cross-over design with 24 subjects extending over five days, cannot adequately control for all relevant extraneous variables. For example, the effect of heat stress on driver vigilance should be controlled for (Wyon, Wyon & Norin, 1996). Lack of sleep is also of critical importance (Mills, Spruill, Kanne, Parkman & Zhang, 2001). Hindmarch, Shamsi, Stanley and Fairweather (1999) studied the effects of fexofenadine, loratadine and promethazine on cognitive and psychomotor function. Their test battery included critical flicker fusion, choice reaction time and assessments of subjective sedation. Their sample consisted of 24 healthy volunteers, aged 19-58 years. Promethazine caused a significant reduction in critical flicker fusion threshold, a significant increase in recognition reaction time and a significant increase in ‘sleep-like’ activity. Neither fexofenadine nor loratadine caused a significant impairment of cognitive and psychomotor function. As this study was also based on a very small sample, it should be replicated on a larger one. Mann, Pearce, Dunn and Shakir (2000), who compared the sedating effects of loratadine, cetirizine, fexofenadine and acrivastine in a post-marketing surveillance study on 43 363 patients, found that the sedating effects of fexofenadine and loratadine were the lowest. In the light of the foregoing evidence it appears to be safe for drivers, pilots and operators, who work with dangerous equipment, to take fexofenadine or loratadine when suffering from allergic conditions, such as rhinitis. A brief pharmacological profile of fexofenadine (Telfast) will now be given, thereafter a South African study will be described wherein Telfast® 180 was used. Fexofenadine HCl, the active ingredient of TELFAST®, is a non sedating antihistamine with highly selective peripheral histamine H1-receptor antagonist activity indicated for the relief of symptoms associated with seasonal allergic rhinitis and for the relief of symptoms associated with chronic idiopathic urticaria (CIU) (Medicine Control Council, 2000). The H-1-antagonists are now commonly subdivided into two broad groups – the first generation or classical antihistamines and the second generation antihistamines with higher selectivity for H-1-receptors and lacking the undesirable central nervous system (CNS) actions. The second generation antihistamines are non-sedating and with little antagonist activity at other neurotransmitter receptors at therapeutic concentrations. These compounds are devoid of sedating effects due to poor CNS penetration, and possibly lowered affinities for central histaminic, cholenergic and adrenergic receptors (DeRuiter, (2001). Fexofenadine HCl is rapidly absorbed following oral administration of tablets, with the mean tmax occurring at 2,6 hours post-dose (Markham & Wagstaff, 1998). The single and multiple dose pharmacokinetics of fexofenadine are linear between 40 mg and 240 mg taken daily. It exhibits an antihistaminic effect within one hour, achieves maximum effect at six hours, and lasts 24 hours. No sedative or other CNS effects have been reported for fexofenadine and studies indicate that it does not cross the blood-brain barrier (Hindmarch et al., 1999; and Markham & Wagstaff, 1998). The most frequent adverse symptom reported is headache, with the incidence of fatigue and drowsiness comparable to those found in patients receiving placebo (Markham & Wagstaff, 1998). Objectives of the study The principal objective of the present study was to examine the effects of Telfast® 180, an antihistamine, on driver behaviour, decision making, and reaction time. The following secondary objectives were also set: � To determine the reliabilities of the various measures used. � To examine the effects of repeated testing (test-retest) on the performance of the two treatment groups (Telfast and placebo), the two gender groups, and the three age groups. � To ascertain whether there are gender differences in respect of the various measures used. � To determine whether there are age differences in respect of the various measures used. METHOD Sample A sample of 259 participants was randomly drawn from 600 volunteers – mostly doctors and pharmacists. However, four members were later excluded from the sample because their test records were incomplete. The final sample therefore consisted of 255 volunteers. All the participants had to be 18 years or older, EFFECT OF TELFAST® 180 77 and in possession of a valid South African drivers license. There were 195 men and 60 women in the sample. Eighty participants were younger than 30 years, 94 were between 30 and 40 years, and 79 were older than 40 years. Two participants were of unknown age. A double-blind design was used in the investigation: The sample was randomly divided into two groups. One group received Telfast (N = 128), and the other group a placebo (N = 127). None of the experimenters knew who received the placebo and who received Telfast. All the participants were treated exactly the same way. Measuring instruments All the participants were given a standardised driving test as well as a psychomotor test. The Driving Test The route that the participants had to drive was carefully delineated by cones. While driving along the route the participants had to execute seven carefully planned manoeuvres. An error was committed if the wheels of the car touched any of the cones. If a wrong turn was made, a time penalty of 0,25 sec was given. The following manoeuvres were specified: � 180° left turn � 180° right turn � 360° turn � Parking F: The participant was required to enter a parking spot with the front of the car first and by turning 90° to the right. � Parking R: The participant was required to enter a parking spot with the rear end of the car first and by turning 45° to the right. � Parking P: The participant was required to enter a parallel parking spot with the rear end of the car first. � Slalom: The participant was required to drive his/her car along a zigzag course, marked by cones. Penalties were incurred if the participant grated the gears of his/her car. All the errors made and penalties incurred were added together to obtain a total score. Similarly, the time taken to complete the route, yielded a time score. The Psychomotor Test The psychomotor test, which is known as the Decision Reaction Test (DR2) and which forms part of the Vienna Test System developed by Schuhfried (2001), was used. The Decision Reaction Test is computer based, and the test panel consists of: � A yellow light � A red light � A large decision timer button � A small reaction timer button � An audible warning signal (hooter) During the test the participants were presented with various combinations of lights and audible warnings. However, they were required to react only when a yellow light illuminated with a simultaneous audible warning. Any one of the following stimuli could be presented: � A yellow light � A yellow light with a simultaneous audible warning � A red light � A red light with a simultaneous audible warning � An audible warning. The test consisted of 48 stimuli of which 16 were the correct combination. During the test participants were to keep their index finger on the large decision timer button, and if the correct stimulus was presented they were to move their finger as fast as possible to the small reaction timer button and back again to the large decision timer button. The maximum time allowed between the presentation of the stimuli was 1,5 seconds. The test yielded four different scores, namely: � The participants’ mean decision time i.e. the time from the moment the stimulus was presented until they lifted their finger from the large button. � The participants’ mean reaction time i.e. the time from when they lifted their finger from the large button until they touched the small button. � The number of decision errors the participant committed. A decision error was committed if participants reacted to an incorrect stimulus, but did not carry it through as a reaction error. � The number of reaction errors participants committed. A reaction error occurred if participants reacted to an incorrect stimulus and touched the smaller button without correcting their incorrect decision. Strictly speaking, the four scores are not experimentally independent, but are nevertheless very useful. Procedure During the first phase of the testing programme all the participants were given the driving test as well as the psychomotor test. Immediately after the first series of tests the placebo group received their placebo tablets and the Telfast group their Telfast tablets. Exactly 2,6 hours later every participant was tested again, using the same tests as during the first phase. Research has shown that Telfast® 180 reaches a peak concentration in the blood after 2,6 hours (Markham & Wagstaff, 1998). Statistical analysis As a preliminary to the main study, the reliabilities of the various measures were determined. Before analysing the data for the main study the sample was divided into two gender groups, three age groups, and two treatment groups (placebo and Telfast). To ascertain whether there were any growth effects due to learning, the vectors of means of the pre-test and post-test scores were compared for the two treatment groups, the two gender groups, and the three age groups. Hotelling’s T2 for dependent samples (Morrison, 1976), followed by a series of t- tests (for dependent samples), were used. To determine the effect of the treatment variable on the test scores, the vectors of the pre-test and post-test means were compared for the placebo and Telfast groups. Hotelling’s T2 for independent samples, followed by a series of one-way analyses of variance, were used for this purpose. It merits mentioning that for two groups F = t2, and p(F) = p(t). A similar procedure was followed when comparing the vectors of means of the men and women. To compare the vectors of means of the three age groups, MANOVA, followed by a series of ANOVAs and Scheffé’s post hoc multiple comparisons, were used. To compare the factor structure of the research battery for the two treatment groups, inter-group factor analysis was used (Meridith, 1964a and 1964b). RESULTS Reliability of measuring instruments To begin with the reliabilities of the various measures used, were determined: The reliability of the driving test was estimated by means of Cronbach’s coefficient alpha (Schepers, 1992, pp. 42-44), using the variance-covariance matrix of the nine error scores. The variance- covariance matrix of the first driving test is given in Table 1. SCHEPERS, POTTER, VAN NIEKERK78 A striking feature of the variance-covariance matrix is the fact that all the variances of the error scores are unusually small, suggesting a very low reliability for the total error score. This was confirmed by Cronbach’s coefficient alpha (� = 0,284). The reliability of the second driving test is comparably low (� = 0,345). Both these coefficients are very low, suggesting that the error scores of the driving test are of doubtful value in the present study. Test-retest reliabilities were computed for all the measures used in the study. Separate analyses were done for the placebo group and for the Telfast group. The results are given in Table 2. As seen in Table 2 it is clear that the test-retest reliabilities of the time measures are considerably higher than for the error scores. The test-retest reliabilities of the time measures range from 0,765 to 0,843 for the placebo group, and from 0,669 to 0,825 for the Telfast group. The test-retest reliabilities of the error scores range from 0,237 to 0,496 for the Telfast group and from 0,376 to 0,465 for the placebo group. More weight should therefore be given to the time measures in the main study, than to the error scores. Effects of repeated testing The treatment groups To compare the pre-test and post-test means of the treatment groups, the means, variances and coefficients of skewness of the various measures were computed for the placebo group and the Telfast group. The results are given in Table 3. It is clear from Table 3 that there was a reduction in time from the pre-test to the post-test in respect of the Road Test (Time). This was true for both the placebo and the Telfast group. The reduction in time probably signifies a learning effect. However, as far as Decision Time and Reaction Time are concerned, there was an increase in time from the pre-test to the post-test. This was true for both the placebo and the Telfast group. It is probable that the participants were more cautious during the second tests. As far as the error scores are concerned, there was a consistent increase in the number of errors committed with respect to the Road Test, Decision Time and Reaction Time. This was true for both the placebo and the Telfast groups. A comparison of the differences in means of the pre-test and post-test scores of the treatment groups are given in Table 4. From Table 4 it is clear that all the differences in means (pre-test mean minus post-test mean) of the placebo group are statistically significant. However, for the Telfast group the differences in means of Decision Time and Reaction Time are not statistically significant. The gender groups To compare the means of the pre-test and post-test scores of the gender groups, the means, variances and coefficients of skewness of the various measures were computed for the men and women. The results are given in Table 5. EFFECT OF TELFAST® 180 79 TABLE 2 TEST-RETEST CORRELATIONS OF THE VARIOUS MEASURES IN RESPECT OF THE PLACEBO GROUP AND THE TELFAST GROUP Placebo Group Telfast Group Variables N Test-retest reliability Test-retest correlation N 1. Road test: Time 127 0,843 0,825 128 2. Decision time 127 0,791 0,765 128 3. Reaction time 127 0,765 0,669 128 4. Road test: Errors 127 0,376 0,382 128 5. Decision time: Errors 127 0,465 0,496 128 6. Reaction time: Errors 127 0,389 0,237 128 TABLE 1 VARIANCE – COVARIANCE MATRIX OF THE SUBTESTS OF THE FIRST ROAD TEST Variables N 1 2 3 4 5 6 7 8 9 1. 1800 Left 255 0,2476 0,0665 0,0052 0,0568 -0,0141 0,0079 0,0131 -0,0042 -0,0094 2. 1800 Right 255 0,0665 0,2565 0,0049 0,0283 0,0032 0,0051 -0,0055 -0,0009 0,0208 3. 3600 Turn 255 0,0052 0,0049 0,0115 0,0068 0,0026 0,0028 0,0066 -0,0002 0,0021 4. Parking F 255 0,0568 0,0283 0,0068 0,4378 0,0444 -0,0100 -0,0071 -0,0042 0,0145 5. Parking B 255 -0,0141 0,0032 0,0026 0,0444 0,4567 -0,0005 0,0448 -0,0047 -0,0006 6. Parking P 255 0,0079 0,0051 0,0028 -0,0100 -0,0005 0,0890 0,0066 -0,0017 0,0060 7. Slalom 255 0,0131 -0,0055 0,0066 -0,0071 0,0448 0,0066 0,1139 -0,0019 0,0081 8. Use of gears 255 -0,0042 -0,0009 -0,0002 -0,0042 -0,0047 -0,0017 -0,0019 0,0268 -0,0029 9. Route 255 -0,0094 0,0208 0,0021 0,0145 -0,0006 0,0060 0,0081 -0,0029 0,1904 Note: 1 1 1 1 2 1 2 2 2 2 1 1 1 0,6191 ( ) 0,6191; 0,0086 ( 1) 72 1,8302 _ 81 0,0086 2,4493; 2,4493 0,284 _ k k gh g hk k g h ghgh g h g h k g g k k k gh g gh g h g g h g g r s s r s s for g h K K s K C s r s s s s C = = = = = = = = ≠ = = = = − = × = + = α = = = ∑ ∑ ∑ ∑ ∑ ∑ ∑ ∑ From Table 5 it is clear that the pattern of means of the men and women are very similar to that of the placebo and Telfast groups, and probably for the same reasons. A comparison of the differences in means of the pre-test and post-test scores of the men and women is given in Table 6. From Table 6 it is clear that, for men, all the differences in means are statistically significant. However, for the women, the differences for Decision Time, Reaction Time, and Road Test (Errors), are not statistically significant. It is interesting to note that the pattern of differences in means, for men, is identical to that of the placebo group. The pattern for the women, on the other hand, is very similar to that of the Telfast group. The age groups To compare the means of the pre-test and post-test scores, of the various age groups, the means, variances and coefficients of skewness of the scores were computed for the participants younger than 30 years, for those between 30 and 40 years, and for those older than 40 years. The results are given in Table 7. From Table 7 it is evident that the pattern of means of the two younger age groups (less than 30 years, and between 30 and 40 years) is very similar to that of the placebo group, whilst the pattern of the older group (40 years and more) is very similar to that of the Telfast group. This raises the issue of representativeness of the placebo and Telfast groups in terms of gender and age. The gender and age distributions of the placebo and Telfast groups were therefore examined. There was no difference in the gender distributions of the placebo and Telfast groups [�2 (1) = 2,078; p = 0,149]. Similarly, no difference was found with respect to the age distributions of the two groups [�2 (2) = 0,299; p = 0,861]. A comparison of the differences in means of the pre-test and post-test scores of the three age groups is given in Table 8. From Table 8 it is clear that all the differences in means of the 30 to 40 year old group are statistically significant. For the youngest age group the difference in means of the Road Test (Errors) is statistically not significant, and for the oldest group the differences in means as regards Decision Time and Reaction Time are statistically not significant. From the foregoing it is apparent that the measures used are sensitive enough to detect gender and age differences pertaining to the pre-test and post-test scores. SCHEPERS, POTTER, VAN NIEKERK80 TABLE 3 MEANS, VARIANCES AND COEFFICIENTS OF SKEWNESS OF THE PRE-TEST AND POST-TEST SCORES IN RESPECT OF THE TREATMENT GROUPS Placebo Group Telfast Group N Means Variances Skewness N Means Variances Skewness PRE-TEST 1. Road test 1: Time 127 135,66 945,63 1,493 128 141,90 2352,35 2,503 2. Decision time A 127 354,79 4558,14 0,747 128 346,59 4316,39 0,213 3. Reaction time A 127 466,75 6700,35 1,129 128 467,16 5621,29 0,539 4. Road test 1: Errors 127 1,57 2,60 1,519 128 1,87 2,18 0,874 5. Decision time A: Errors 127 1,54 1,44 1,132 128 1,85 2,74 1,245 6. Reaction time A: Errors 127 0,53 0,60 1,675 128 0,54 0,60 1,637 POST-TEST 1. Road test 2: Time 127 112,30 453,83 1,844 128 117,59 653,95 1,251 2. Decision time B 127 365,46 4869,01 0,703 128 354,05 3191,82 0,050 3. Reaction time B 127 481,27 5645,61 0,757 128 471,65 4225,79 0,194 4. Road test 2: Errors 127 2,11 3,32 1,065 128 2,20 3,12 1,051 5. Decision time B: Errors 127 2,54 3,78 1,140 128 2,93 3,74 1,191 6. Reaction time B: Errors 127 1,17 1,37 1,019 128 1,24 1,59 2,245 TABLE 4 DEPENDENT T-TESTS: COMPARISON OF MEANS OF PRE-TEST AND POST-TEST SCORES IN RESPECT OF THE TREATMENT GROUPS Placebo Group Telfast Group Variables X D t-value df p X D t-value df p 1. Road test: Time 23,36 15,338 126 < 0,001* 24,32 8,884 127 < 0,001* 2. Decision time -10,68 -2,707 126 0,008* -7,46 -1,973 127 0,051 3. Reaction time -14,52 -3,022 126 0,003* -4,48 -0,880 127 0,381 4. Road test: Errors -0,54 -3,152 126 0,002* -0,34 -2,064 127 0,041* 5. Decision time: Errors -1,00 -6,455 126 < 0,001* -1,08 -6,714 127 < 0,001* 6. Reaction time: Errors -0,65 -6,476 126 < 0,001* -0,70 -6,062 127 < 0,001* Note: XD = Difference in means: pre-test mean minus post-test mean * Statistically significant Placebo group Hotelling T2 = 316,630 F(6, 121) = 50,678; p < 0,001 Telfast group Hotelling T2 = 130,320 F(6, 122) = 20,865; p < 0,001 EFFECT OF TELFAST® 180 81 TABLE 5 MEANS, VARIANCES AND COEFFICIENTS OF SKEWNESS OF THE PRE-TEST AND POST-TEST SCORES IN RESPECT OF THE GENDER GROUPS Men Women N Means Variances Skewness N Means Variances Skewness Pre-test 1. Road test 1: Time 195 129,92 1397,23 3,967 60 167,64 1424,03 0,529 2. Decision time A 195 348,56 4098,52 0,391 60 357,52 5559,41 0,634 3. Reaction time A 195 458,92 5732,03 0,842 60 493,07 6655,18 0,969 4. Road test 1: Errors 195 1,66 2,48 1,252 60 1,92 2,11 1,075 5. Decision time A: Errors 195 1,69 2,09 1,399 60 1,70 2,21 1,146 6. Reaction time A: Errors 195 0,55 0,60 1,504 60 0,47 0,59 2,193 Post-test 1. Road test 2: Time 195 109,62 423,47 2,143 60 132,28 615,39 0,674 2. Decision time B 195 358,31 3872,52 0,633 60 364,37 4647,69 0,248 3. Reaction time B 195 469,09 4550,80 0,378 60 500,32 5531,07 0,942 4. Road test 2: Errors 195 2,15 3,45 1,108 60 2,18 2,49 0,730 5. Decision time B: Errors 195 2,74 3,31 0,928 60 2,70 5,40 1,468 6. Reaction time B: Errors 195 1,28 1,45 1,515 60 0,97 1,49 2,494 TABLE 6 DEPENDENT T-TESTS: COMPARISON OF MEANS OF PRE-TEST AND POST-TEST SCORES IN RESPECT OF THE GENDER GROUPS Men Women Variables X D t-value df p X D t-value df p 1. Road test: Time 20,30 11,444 194 < 0,001* 35,36 12,237 59 < 0,001* 2. Decision time -9,74 -3,140 194 0,002* -6,85 -1,190 59 0,239 3. Reaction time -10,17 -2,595 194 0,010* -7,25 -0,926 59 0,358 4. Road test: Errors -0,48 -3,408 194 0,001* -0,27 -1,462 59 0,149 5. Decision time: Errors -1,05 -8,645 194 < 0,001* -1,00 -3,809 59 < 0,001* 6. Reaction time: Errors -0,73 -8,455 194 < 0,001* -0,50 -3,052 59 0,003* Note: XD = Difference in means: pre-test mean minus post-test mean * Statistically significant Men Hotelling T2 = 210,795 F(6, 189) = 34,227; p < 0,001 Women Hotelling T2 = 191,670 F(6, 54) = 29,238; p < 0,001 TABLE 7 MEANS, VARIANCES AND COEFFICIENTS OF SKEWNESS OF THE PRE-TEST AND POST-TEST SCORES IN RESPECT OF THE VARIOUS AGE GROUPS <30 years 30-40 years 40+ years Variables Means Variance Skewness Means Variance Skewness Means Variance Skewness PRE-TEST 1. Road test 1: Time 146,24 2153,26 2,735 133,53 1529,66 2,783 136,12 1194,76 1,653 2. Decision time A 333,46 4020,18 0,709 344,81 3819,32 0,365 376,63 4682,83 0,351 3. Reaction time A 444,44 5594,45 0,623 458,79 4924,30 0,601 500,30 6672,11 1,288 4. Road test 1: Errors 1,89 2,84 0,997 1,55 1,69 1,245 1,76 2,83 1,221 5. Decision time A: Errors 1,61 2,06 1,188 1,64 2,02 1,541 1,85 2,34 1,258 6. Reaction time A: Errors 0,50 0,63 1,934 0,54 0,62 1,831 0,57 0,56 1,088 POST-TEST 1. Road test 2: Time 117,91 574,57 1,507 110,78 530,80 1,785 115,64 477,87 1,194 2. Decision time B 346,14 3962,47 0,609 354,61 3072,24 -0,030 381,71 4603,03 0,724 3. Reaction time B 455,86 5041,41 1,022 476,54 4412,36 0,350 498,72 4673,00 0,492 4. Road test 2: Errors 2,20 3,02 0,988 2,10 3,27 1,515 2,18 3,33 0,654 5. Decision time B: Errors 2,50 4,00 0,818 2,46 2,77 1,043 3,20 3,70 1,156 6. Reaction time B: Errors 1,13 1,78 2,124 1,01 1,09 1,726 1,49 1,46 1,292 Sample size 80 94 79 Differences between treatment groups In the light of the main objective of the study, the question arose as to whether there were any statistically significant differences between the pre-test and post-test means of the placebo group and the Telfast group. To compare the means of the placebo group with that of the Telfast group, the data shown in Table 3 were used. First, the vectors of the means of the placebo group and the Telfast group were compared by means of Hotelling’s T2-test. This was done separately for the pre-test and post-test scores. Second, the means of each of the measures were compared by means of a series of one-way analyses of variance. The results are given in Table 9. From Table 9 it is clear that the vectors of the pre-test means of the two treatment groups, do not differ statistically significantly [F(6, 248) = 1,561; p = 0,159]. Similarly, the vectors of post-test means of the two treatment groups do not differ significantly [F(6, 248) = 1,346; p = 0,237]. None of the F-ratios, regarding the individual measures, are statistically significant. No sedative effects due to Telfast® 180 were thus evident. Differences between men and women To gain further evidence of the discrimination power of the measures used, the means of the men were compared with those of the women. These were computed separately for the pre-test and post-test scores. The data shown in Table 5 were used. First, the vectors of means of the men and women were compared by using Hotelling’s T2-test. This was done separately for the pre-test and post-test scores. Second, the means of each measure were compared by a series of one-way analyses of variance. The results are given in Table 10. SCHEPERS, POTTER, VAN NIEKERK82 TABLE 8 DEPENDENT T-TESTS: COMPARISON OF MEANS OF PRE-TEST AND POST-TEST SCORES IN RESPECT OF THE VARIOUS AGE GROUPS < 30 years 30-40 years > 40 years Variables X D t-value df p X D t-value df p X D t-value df p 1. Road test: Time 28,34 8,550 79 <0,001* 22,75 9,791 93 <0,001 * 20,48 8,092 78 <0,001* 2. Decision time -12,68 -2,787 79 0,007* -9,80 -2,199 93 0,030 * -5,08 -0,955 78 0,343 3. Reaction time -11,43 -2,115 79 0,038* -17,76 -3,000 93 0,003 * 1,58 0,231 78 0,818 4. Road test: Errors -0,31 -1,386 79 0,170 -0,54 -2,762 93 0,007 * -0,42 -2,111 78 0,038* 5. Decision time: Errors -0,89 -4,499 79 <0,001* -0,82 -4,845 93 <0,001 * -1,35 -7,009 78 <0,001* 6. Reaction time: Errors -0,63 -4,249 79 <0,001* -0,47 -4,184 93 <0,001 * -0,92 -6,868 78 <0,001* Note: XD = Difference in means: pre-test mean minus post-test mean * Statistically significant < 30 years Hotelling T2 = 93,456 F(6, 74) = 14,590; p < 0,001 30-40 years Hotelling T2 = 145,358 F(6, 88) = 22,924; p < 0,001 > 40 years Hotelling T2 = 140,121 F(6, 73) = 21,894; p < 0,001 TABLE 9 ANOVA: COMPARISON OF THE MEANS OF THE TREATMENT GROUPS IN RESPECT OF THE PRE-TEST AND POST-TEST MEASURES Source Dependent variables Type III sum of squares DF Mean Square F p(F) TREATMENT GROUPS: PRE-TESTS Placebo vs. Telfast 1. Road test 1: Time 2487,413 1,253 2487,413 1,506 0,221 2. Decision time A 4288,015 1,253 4288,015 0,966 0,327 3. Reaction time A 11,034 1,253 11,034 0,002 0,966 4. Road test 1: Errors 5,747 1,253 5,747 2,408 0,122 5. Decision time A: Errors 6,371 1,253 6,371 3,043 0,082 6. Reaction time A: Errors 0,008 1,253 0,008 0,014 0,906 TREATMENT GROUPS: POST-TESTS Placebo vs. Telfast 1. Road test 2: Time 1785,459 1,253 1785,459 3,221 0,074 2. Decision time B 8310,557 1,253 8310,557 2,064 0,152 3. Reaction time B 5898,731 1,253 5898,731 1,196 0,275 4. Road test 2: Errors 0,550 1,253 0,550 0,171 0,680 5. Decision time B: Errors 9,909 1,253 9,909 2,639 0,106 6. Reaction time B: Errors 0,303 1,253 0,303 0,205 0,651 Note: Pre-tests Hotelling’s Trace = 0,038 F(6, 248) = 1,561; p = 0,159 Post-test Hotelling’s Trace = 0,033 F(6, 248) = 1,346; p = 0,237 From Table 10 it is clear that the vectors of pre-test means of the men and women differ significantly [F(6, 248) = 10,998; p<0,001]. Similarly, the vectors of post-test means of the men and women differ significantly [F(6, 248) = 13,053; p<0,001]. Furthermore, as far as the individual measures are concerned, the means of the Road Test (time) and Reaction Time, differ significantly. This is true for both the pre-test and post- test scores. Differences between age groups Lastly, the effect of age on the test scores was explored. The data in Table 7 were used for this purpose. As a first step, MANOVA was used to compare the vectors of means of the three age groups. This was done separately for the pre-test and post-test scores. Thereafter, a series of one-way analyses of variance were used to compare the means of each measure. The results are given in Table 11. From Table 11 it is clear that the vectors of pre-test means of the three age groups differ significantly [F(12, 490) = 3,009; p<0,001]. Similarly, the vectors of post-test means of the three age groups differ significantly [F(12, 490) = 3,537; p<0,001]. As far as the individual measures are concerned, the three age groups differed with respect to Decision Time and Reaction EFFECT OF TELFAST® 180 83 TABLE 10 ANOVA: COMPARISON OF THE MEANS OF THE GENDER GROUPS IN RESPECT OF THE PRE-TEST AND POST-TEST MEASURES Source Dependent variables Type III sum of squares DF Mean Square F p(F) GENDER GROUPS: PRE-TESTS Men vs. women 1. Road test 1: Time 65305,70 1,253 65305,70 46,531 <0,001 2. Decision time A 3677,40 1,253 3677,40 0,828 0,364 3. Reaction time A 53488,95 1,253 53488,95 8,994 0,003 * 4. Road test 1: Errors 3,11 1,253 3,11 1,296 0,256 5. Decision time A: Errors 0,00 1,253 0,00 0,001 0,971 6. Reaction time A: Errors 0,35 1,253 0,35 0,584 0,446 GENDER GROUPS: POST-TESTS Men vs. women 1. Road test 2: Time 23557,90 1,253 23557,90 50,313 <0,001 * 2. Decision time B 1684,40 1,253 1684,40 0,416 0,520 3. Reaction time B 44733,49 1,253 44733,49 9,360 0,002 * 4. Road test 2: Errors 0,05 1,253 0,05 0,017 0,896 5. Decision time B: Errors 0,09 1,253 0,09 0,023 0,880 6. Reaction time B: Errors 4,56 1,253 4,56 3,126 0,078 Note: Pre-tests Hotelling’s Trace = 0,266 F(6, 248) = 10,998; p < 0,001 *Statistically significant Post-test Hotelling’s Trace = 0,316 F(6, 248) = 13,053; p < 0,001 TABLE 11 ANOVA: COMPARISON OF THE MEANS OF THE AGE GROUPS IN RESPECT OF THE PRE-TEST AND POST-TEST MEASURES Source Dependent variables Type III sum of squares DF Mean Square F p(F) AGE GROUPS: PRE-TESTS < 30 years vs. 30-40 years vs. 1. Road test 1: Time 7559,20 2,250 3779,60 2,330 0,099 > 40 years 2. Decision time A 80108,88 2,250 40054,40 9,647 <0,001* 3. Reaction time A 134676,30 2,250 67338,20 11,852 <0,001* 4. Road test 1: Errors 4,98 2,250 2,49 1,035 0,357 5. Decision time A: Errors 2,70 2,250 1,35 0,633 0,532 6. Reaction time A: Errors 0,20 2,250 0,10 0,163 0,850 AGE GROUPS: POST-TESTS < 30 years vs. 30-40 years vs. 1. Road test 2: Time 2332,13 2,250 1166,07 2,208 0,112 > 40 years 2. Decision time B 55300,03 2,250 27650,00 7,217 0,001* 3. Reaction time B 73036,03 2,250 36518,00 7,782 0,001* 4. Road test 2: Errors 0,53 2,250 0,27 0,083 0,921 5. Decision time B: Errors 28,68 2,250 14,34 4,158 0,017* 6. Reaction time B: Errors 10,63 2,250 5,32 3,739 0,025* Note: Pre-tests Wilks’ Lambda = 0,867 F(12, 490) = 3,009; p < 0,001 *Statistically significant Post-test Wilks’ Lambda = 0,847 F(12, 490) = 3,537; p < 0,001 Time. This applied to both the pre-test and post-test scores. Furthermore, the age groups also differed with regard to Decision Time (Errors) and Reaction Time (Errors) as far as the post-test scores are concerned. To find out which of the age groups differed from one another in terms of the pre-test and post-test means, Scheffé’s post hoc multiple comparisons technique was used. The results are given in Table 12. From Table 12 it is clear that the youngest group’s (<30 years) pre-test and post-test scores differ from those of the oldest group (40+ years) in terms of Decision Time and Reaction Time. The 30-40 year old group differs from the 40+ group in terms of Decision Time and Reaction Time, as far as the pre- test scores are concerned. For the post-test scores there are statistically significant differences in respect of Decision Time, as well as Decision time (Errors) and Reaction time (Errors). The mean Decision time and mean Reaction time of the 40+ group are consistently longer than those of the two younger groups. They also made more Decision and Reaction Time errors. To examine the factor structure of the measures used for the placebo and Telfast groups, an inter-group factor analysis was done. For the purpose of this analysis the Road Test (Errors) was discarded because of its low reliability. The average factor matrix of the placebo and Telfast groups is given in Table 13. This matrix was obliquely rotated to simple structure by means of a Direct Quartimin rotation. From Table 13 it appears that the first factor is well determined with high loadings on Decision Time A and B and Reaction Time A and B. Factor 2 is a doublett with high loadings on Road Tests 1 and 2 (Time). Factor 3 is well determined with high loadings on Decision Time A and B (Errors), and Reaction Time A and B (Errors). Factor 1 can be identified as Decision Time/Reaction Time, Factor 2 relates to driving performance, and Factor 3 to Decision and Reaction errors. The communalities of the various measures are given in the last column of Table 13. The communalities of the time measures range from 0,705 to 0,874, and the communalities of the error scores range from 0,308 to 0,610. If the communality of a variable is taken as a lower bound for its reliability, then the reliabilities of the measures used in the present study are acceptable for research purposes (Guttman, 1957; and Mulaik, 1965, 1966). TABLE 13 AVERAGE FACTOR MATRIX IN RESPECT OF THE PLACEBO AND TELFAST GROUPS (DIRECT QUARTIMIN ROTATION) Variables Factor 1 Factor 2 Factor 3 h 2 j 1. Road test 1: Time -0,017 0,934 -0,043 0,874 2. Road test 2: Time 0,030 0,880 0,046 0,777 3. Decision time A 0,894 -0,055 -0,030 0,803 4. Decision time B 0,896 -0,007 0,016 0,803 5. Reaction time A 0,859 -0,006 -0,001 0,738 6. Reaction time B: 0,832 0,078 0,007 0,705 7. Decision time A: Errors -0,091 -0,068 0,602 0,375 8. Decision time B: Errors -0,054 0,100 0,773 0,610 9. Reaction time A: Errors 0,099 -0,084 0,540 0,308 10. Reaction time B: Errors 0,019 0,092 0,653 0,435 The factor variances of the placebo and Telfast groups are given in Table 14. TABLE 14 FACTOR VARIANCES IN RESPECT OF THE PLACEBO AND TELFAST GROUPS Groups Factor 1 Factor 2 Factor 3 Placebo group 1,14 0,69 0,98 Telfast group 0,86 1,31 1,02 Total group 2,00 2,00 2,00 From Table 14 it appears that the variance of Factor 1 is slightly higher for the placebo group than for the Telfast group. This implies that Factor 1 is slightly better developed for the placebo group than for the Telfast group. The opposite is true for Factor 2. Here the variance of the Telfast group is considerably larger than that of the placebo group. SCHEPERS, POTTER, VAN NIEKERK84 TABLE 12 SCHEFFÉ’S MULTIPLE COMPARISONS: COMPARISON OF THE MEANS OF THE VARIOUS AGE GROUPS IN RESPECT OF THE PRE-TEST AND POST-TEST MEASURES Means of age groups Pairs of age groups < 30 years 30-40 years 40+ years 1/2 1/3 2/3 PRE-TEST 1. Road test 1: Time 146,24 133,53 136,12 0,118 0,287 0,915 2. Decision time A 333,46 344,81 376,63 0,513 <0,001 * 0,006* 3. Reaction time A 444,44 458,79 500,30 0,458 <0,001 * 0,002* 4. Road test 1: Errors 1,89 1,55 1,76 0,368 0,874 0,685 5. Decision time A: Errors 1,61 1,64 1,85 0,993 0,597 0,642 6. Reaction time A: Errors 0,50 0,54 0,57 0,937 0,852 0,974 POST-TEST 1. Road test 2: Time 117,91 110,78 115,64 0,127 0,825 0,384 2. Decision time B 346,14 354,61 381,71 0,668 0,002 * 0,017* 3. Reaction time B 455,86 476,54 498,72 0,142 0,001 * 0,108 4. Road test 2: Errors 2,20 2,10 2,18 0,929 0,997 0,957 5. Decision time B: Errors 2,50 2,46 3,20 0,989 0,060 0,033* 6. Reaction time B: Errors 1,13 1,01 1,49 0,820 0,152 0,031* Note: * Statistically significant Factor 2 is therefore better developed for the Telfast group than the placebo group. The variances of Factor 3 are virtually the same for the two groups. The intercorrelations of the factors of the placebo and Telfast groups are given in Table 15. TABLE 15 INTERCORRELATIONS OF FACTORS IN RESPECT OF THE PLACEBO AND TELFAST GROUPS Placebo group Telfast group Factor 1 Factor 2 Factor 3 Factor 1 Factor 2 Factor 3 Factor 1 1,00 0,02 0,00 1,00 0,01 -0,06 Factor 2 0,02 1,00 0,00 0,01 1,00 0,04 Factor 3 0,00 0,00 1,00 -0,06 0,04 1,00 From Table 15 it is clear that the factor structures of the two groups are similar. Furthermore, they are strictly orthogonal, despite the fact that an oblique rotation was used. There was therefore no indication of a regression towards an undifferentiated state due to the ingestion of Telfast® 180. DISCUSSION OF RESULTS As far as the secondary objectives of the study are concerned it was found that the Driving Test as well as the Decision-making and Reaction Time Test yielded acceptable reliabilities, particularly as far as the time scores were concerned. All the measures were sensitive enough to pick-up test-retest effects, and to differentiate between the gender and age groups. With regard to the major objectives of the study it was found that there were no differences on any of the measures between the Telfast and placebo groups. As regards the structure of the measures used, the factor variances of the two treatment groups are very similar: The first factor is slightly better developed for the placebo group than the Telfast group. Factor 2 is considerably better developed for the Telfast group than the placebo group, and Factor 3 is equally well developed for both groups. The factor intercorrelations are virtually the same for the two groups, and indicate that the factors are essentially orthogonal to one another. The factor structures of the two groups are therefore essentially the same. The fact that the two intercorrelation matrices are essentially orthogonal implies that there was no regression towards an undifferentiated state due to the ingestion of Telfast® 180. There is therefore no evidence that Telfast® had a sedating effect on the participants who completed the Driving Test and the Psychomotor Tests. The findings of the present study are in keeping with the pharmacological profile of Telfast: Fexofenadine HC1, the active ingredient of Telfast, is a non sedating antihistamine with highly selective peripheral histamine H1-receptor antagonist activity (Hindmarch et al., 1999; and Markham & Wagstaff, 1998). It is claimed to be free of the negative effects of first generation antihistamines. The results of the present study are supported by the findings of Weiler et al. (2000) who found that fexofenadine did not affect driving performance to the same extent as diphenhydramine or alcohol. The findings of the present study also support those of Vermeeren and O’Hanlon (1998). They found no evidence of driving performance impairment in respect of fexofenadine. Also Hindmarch et al. (1999) found that neither fexofenadine nor loratadine caused a significant impairment of cognitive and psychomotor function. Mann et al. (2000) found that the sedating effects of fexofenadine and loratadine were the lowest compared to cetirizine and acrivastine. The findings of the present study are very promising indeed. No indications of driving performance impairment, associated with the use of Telfast® 180, were found. However, further research should investigate whether the continuous use of Telfast® 180 has any negative effects. ACKNOWLEDGEMENTS The authors wish to thank the following persons for their contributions to the study: BMW Advanced Driving Gary Clackworthy – Training Manager Kevin Smith – Marketing Manager Statistical Consultation Service, RAU Janes du Toit Aventis Pharma (Pty) Ltd Dr Christo van Niekerk – Medical Director Dr Etienne Fourie – Medical Advisor Jenny Henderson – Clinical Development Manager Rhoda du Toit – Data Manager Telsa de Lange – Project Assistant Leandra Lombard – Project Coordinator Michelle Simpson – Research Nurse Iselle Viljoen – Research Nurse Secretarial Service Annetjie Boshoff REFERENCES DeRuiter, J. (2001). Histamine H2-antagonists, proton pump inhibitors and other drugs that alter gastric acidity. Available from: http://www.duc.auburn.edu/~deruija/hist_acid.pdf (Accessed 2nd April 2002). Green, R.J., Potter, P.C., Plitt, M., Friedman, R., Hockman, M. & Davis, G. (1998). The South African allergic rhinitis working group. South African Medical Journal, 88 (11), 1366-7. Guttman, L. (1957). Simple proofs of relations between the communalit y problem and multiple correlations. Psychometrika, 22, 147-157. Hindmarch, I., Shamsi, Z., Staneley, N. & Fairweather, D.B. (1999). A double-blind, placebo-controlled investigation of the effects of fexofenadine, loratadine and promethazine on cognitive and psychomotor function. British Journal of Clinical Pharmacology, 48, 200-206. Mann, R.D., Pearce, G.L., Dunn, N. & Shakir, S. (2000). Sedation with “non-sedating” antihistamines: four prescription-event monitoring studies in general practice. British Medical Journal, 320, 1184-1187. Markham, A. & Wagstaff, A.J. (1998). Fexofenadine. Drugs, 55 (2), 269-274. Medicine Control Council 32/5.7.1/0447 (2000). Meridith, W. (1964a). Notes on factorial invariance. Psychometrika, 29, 177-185. Meridith, W. (1964b). Rotation to achieve factorial invariance. Psychometrika, 29, 187-206. Mills, K.C., Spruill, S.E., Kanne, R.W., Parkman, K.M. & Zhang, Y. (2001). The influence of stimulants, sedatives, and fatigue on tunnel vision: Risk factors for driving and piloting. Human Factors, 43 (2), 310. Morrison, D.F. (1976). Multivariate statistical methods (2nd ed.). New York: Wiley. Mulaik, S.A. (1965). Reliability as the upper limit of a test’s communality. Perceptual and Motor Skills, 20, 646-648. EFFECT OF TELFAST® 180 85 Mulaik, S.A. (1966). Inferring the communality of a variable in a universe of variables. Psychological Bulletin, 66 (2), 119-124. Mynhardt, D.C. & De Jager, P.J.D. (1993). Nasionale Snelresponsprogram: Opsomming van bydraende faktore tot noodlottige padverkeersbotsings vanaf 1 Januarie 1993 tot 31 Desember 1993. Projekverslagnommer PR93/808, Pretoria. Schepers, J.M. (1992). Toetskonstruksie: Teorie en praktyk. Johannesburg: RAU-Press. Schuhfried, G. (2001). Computer-aided procedures for ability and personality diagnostics. Catalogue 09/2001. Mödling, Austria. Vermeeren, A. & O’Hanlon, J.F. (1998). Fexofenadine’s effects, alone and with alcohol, on actual driving and psychomotor performance. Journal of Allergy and Clinical Immunology, 101 (3), 306-311. Weiler, J.M., Bloomfield, J.R., Woodworth, G.G., Grant, A.R., Layton, T.A., Brown, T.L., McKenzie, D.R., Baker, T.W. & Watson, G.S. (2000). Effects of fexofenadine, diphenhydramine, and alcohol on driving performance. A randomised placebo-controlled trial in the Iowa driving simulator. Annals of Internal Medicine, 132 (5), 354-363. Wyon, D.P., Wyon, I. & Norin, F. (1996). Effects of moderate heat stress on driver vigilance in a moving vehicle. Ergonomics, 39 (1), 61-75. Road safety statistics: 1. Information supplied by Transportek. 2. Information supplied by Transportek. 3. Information supplied by the Medical Research Council. 4. Information supplied by Transportek. SCHEPERS, POTTER, VAN NIEKERK86