Self-administration and Reliability of Computerized Neurobehavioral Tests among Egyptian Pesticide Workers

Background: Behavioral Assessment and Research System (BARS) is a computerbased test system designed to assess neurobehavioral function and neurotoxicity in humans. This system is available in 5 languages including Arabic. Objectives: To assess the potential use of the Arabic version of computerized neurobehavioral system (BARS) and to compare performance of Egyptian pesticide workers to non-exposed populations using the same computerized battery. Participants and Methods: This study involved the administration of eight neurobehavioral tests from the Arabic computerized neurobehavioral test battery (BARS) to Egyptian workers occupationally exposed to pesticides (n= 25) as compared to non-exposed Egyptian workers (n= 25). One-week test-retest reliability of the computerized battery was measured among non-exposed participants. Results: Performance of pesticide non-exposed Egyptian workers did not show any significant differences between test and re-test (i.e., after one week). Performance of pesticide exposed workers was significantly lower in most of the administered computerized tests as compared to non-exposed Egyptian and US populations. Conclusions: The current findings demonstrate the potential utility of the Arabic computerized BARS in occupational epidemiological research especially in the short-term intervals. Stability of the administered Arabic BARS tests over the short-term interval makes it broadly applicable in assessing exposures at different workplaces and with different cultural and educational levels.


INTRODUCTION
Since the late 1960's neurobehavioral performance tests have been used to assess the effects of occupational exposure in adult workers (1)(2)(3) .The use of neurobehavioral tests to assess workplace exposure has continued to increase and neurobehavioral tests have become the most efficient methods (in terms of cost and time) to screen for adverse effects of neurotoxic exposures in adult workers (4)(5)(6) .

Behavioral Assessment and Research
System (BARS) is a computer-based test system designed to assess neurobehavioral function and neurotoxicity in humans (7,8) .It has been effectively applied among poorly educated populations and young children, with minimal support from a human examiner (9- 15) .Instructions were written in very direct and simple words.Each simply-stated step was presented on the screen, followed by practice on that step or concept.Feedback was provided for correct (smiling face) and incorrect (frowning face) performance at each step of the instructions/practice (9) .
To achieve the goal of testing noneducated participants, spoken instructions were implemented.Thus, BARS could be used with participants who had low (or no) reading skills and the instructions could be presented in any language by using the appropriate graphics (pictures) of the instructions and associated sound files.
One of the biggest challenges in using computers for testing in the workplace is that many workers with low-level of education are not at all familiar with computers and they are reluctant to even touch the keyboard.This led to the development of the 9BUTTON (formerly named DataSled) unit (9,16)  Although reliability of the original English BARS has been tested (16,17) and was applied to identify neurobehavioral deficits in several studies (10,12,13,18,19) ,  (20) .
Test-retest reliability involves administering the same test to a group of participants on two different occasions (21,23) .The interval of time may be as short as same day or it can be as long as several years (24,25) .Studies of neurotoxic exposure tend to rely on shorter intervals (26) .This study was conducted to assess the potential use of the Arabic version of computerized neurobehavioral system (BARS) and to compare the performance of a group of Egyptian workers occupationally exposed to pesticides to that of non-exposed populations using the same computerized battery.

Participants
This study involved administration of the    2).

Statistical Analysis
The data were statistically analyzed using tests did not show statistically significant differences from Time 1 to Time 2 (i.e., after one week).However, although difference scores of the US population on all measures were fairly small, significant differences were detected on Symbol Digit and Finger Tapping (preferred hand) (Table 2).Mean difference scores between T1 and T2 of both non-exposed Egyptian and US populations were compared using Mann-Whitney test, where no significant differences were reported in any of the administered tests (Table 2).
Pairwise comparisons were done to compare performance of the Egyptian pesticide workers to that of the nonexposed Egyptian and US populations at the first session (Time 1) (Table 3).
Egyptian non-exposed participants showed better performance than pesticide workers.3).

DISCUSSION
The Behavioral Assessment and Research System (BARS), a computerized neurobehavioral test system, was initially targeted for use with a broad range of working populations with different educational levels and cultural backgrounds (7,8) .Previous research has found the English version of BARS tests to be reliable across a one-week interval in a normative sample.
The correlation coefficients ranged from 0.44 to 0.92 between session 1 and session 2 on seven standard neurobehavioral tests (16,17) .
This neurobehavioral tests allow the use of alternative forms that can be modified before test administration (10,13) .The use of alternate forms may reduce the amount of practice effect, a finding consistent with the literature (27) .However, presence of practice effects should not be surprising as re-Bull High Inst Public Health Vol.37 No. 2 [2007]  administration of the same measures (28) or even alternate forms of the measures tends to result in improved performance especially at short time intervals (20) .Dikmen et al., (1999) (20) (16) .** p < 0.05 using paired t test.*** no statistical significance between mean difference (1) and mean difference (2) using Mann-Whitney test.

2 . 3 . 4 . 5 .
Simple Reaction Time (SRT; response speed): The subject is instructed to respond by pressing a 9BUTTON as quickly as possible after they see a stimulus presented on the screen or when a 9BUTTON response button becomes backlighted.Continuous Performance Test (CPT; sustained attention): A series of stimuli are presented one at a time for several minutes, typically 5-10.Subjects are instructed to press a key when a target is presented.Digit Span Forward (DSF; attention and memory): A series of numbers is presented sequentially on the computer screen, and the subject is instructed to reproduce the sequence of numbers by pressing the numbered 9BUTTON buttons in the same order (forward).The number of digits increases until a failure criterion is met.Digit Span Backward (DSB; attention and memory): Same as DSF except that subjects press the numbered 9BUTTON in the reverse order.A 10×10 matrix of blocks is followed by three choices, among which one is the same as the sample stimulus.7.Selective Attention (SAT; sustained attention): A small dot is briefly presented Bull High Inst Public Health Vol.37 No.2 [2007] inside or outside one of two squares, one on the left and one on the right half of the screen.The subject is instructed to press one button when a dot appears in the square on the left, a different button when the dot appears on the right, and to not press a button when the dot appears outside of either square.8. Finger Tapping (TAP; response speed coordination): The subject is instructed to press (tap) button(s) as rapidly as possible using the index finger of one or both hands on one or two buttons.Responses incrementally increase the height of a dark bar to suggest progress to the subject.

Differences were statistically significant in 6 out of 8
administered tests (Digit span backward and Continuous performance were not statistically significant).On the other hand, performance of the US population was significantly better than Egyptian pesticide workers in all the administered tests.Non-exposed US population showed statistically significant better performance than non-exposed Egyptian workers in all tests except for tapping and reaction time tests (Table study assessed the reliability of the Arabic computerized neurobehavioral test battery among Egyptian population (pesticide non-exposed workers) over oneweek interval.Furthermore, performance of a group of Egyptian pesticide exposed workers was compared to both Egyptian and US non-exposed populations.The Arabic computerized neurobehavioral tests were reliable as shown by absence of significant differences in any of the administered tests between time 1 and time 2 among Egyptian non-exposed participants.Additionally, mean differences of performance of the Egyptian nonexposed workers from time 1 to time 2 did not show any statistically significant differences as compared to the US population.These data strongly support the use of the Arabic computerized neurobehavioral tests as reliable screening measures where short-term evaluation may be required (e.g., within one-week) for the assessment of exposures to different neurotoxicants at the workplace (e.g., pesticides).Furthermore, it supports the reliability of the Arabic computerized neurobehavioral tests among different populations with different backgrounds and educational levels.The significant differences in test-retest of finger tapping and symbol digit among US population have been attributed by the authors to the effect of practice 16 .Meanwhile, the magnitude of practice effects among current Egyptian nonexposed workers was fairly small in all measures as reported by absence of statistical significance of paired t-tests.Relatively, absence of practice effects among Egyptian non-exposed workers can be attributed to the use of alternate forms in the Arabic BARS.Computerized

Figure 2 .
Figure 2. Two of the participants during testing

Table 2 : Mean and Standard Deviation of Test, Retest, and Test-retest Mean Differences of the non-exposed Egyptian and US populations.
* data obtained fromFarahat et al., 2003

Table 3 : Analysis of Variance (ANOVA) to compare performance on neurobehavioral tests among exposed and non-exposed populations.
* p value.