Brief Overview

People who are tired make mistakes, sometimes huge costly, fatal mistakes.

Fatigue-o-meter is a computerized assessment that compares a person's current level of mental alertness against his personal baseline performance. The system knows what the candidate's typical alertness levels look like, and can immediately detect a drop in performance.

A Fatigue-o-meter assessment takes approximately 1 minute to complete. The result is available immediately. In addition the results are automatically added to your reference norm group as well as to the individual's personal norm. In this way the system systematically builds up an increasingly accurate profile for each individual, making it more and more accurate over time.

By testing for alcohol or drugs, you can only eliminate drowsiness or lack of alertness associated with these factors. Fatigue-o-meter can detect anomalous performance, regardless of the underlying cause.

If safety is a non-negotiable in your organisation, then no employee involved in risky work should ever go on shift without first being assessed on Fatigue-o-meter.

Why Fatigue-o-meter?

Introduction to Fatigue-o-meter

The Fatigue-o-meter human alertness monitoring system has been in ongoing development since 2006. The product recognizes a need for the development of human risk management tools the behavioral and psychological field. The system is used by the client organization to identify employees who may be temporarily mentally or perceptually fatigued. For many years it has been acknowledged that accident prevention programs should address three important variables which interact dynamically in the occurrence of an undesirable incident or occupational accident:

The working environment, task methodologies, materials and equipment.

The competencies, knowledge and skills of the worker.

The mental state, health, fatigue, well being and stress levels of the worker

The first two variables are abundantly addressed by the legal framework governing safety in the working area. Only recently has the third variable received serious attention in South Africa, and a significant step has been taken by the Minister of Labour issuing a notice in terms of section 87(1)(a) of the BCE Act (1997), which addresses a code of good practice which attempts to provide guidelines in this area.

According to these guidelines, attention should be given to the assessment of risks to the health of employees, the implementation of appropriate control measures to deal with fatigue, stress, health related conditions, etc and providing employees with information so as to take action in these areas. In the above notice a significant onus is placed on the employee to take reasonable steps to protect their own health and safety, as well as that of other employees.

Does it work?

Research Findings

Driver vigilance during advanced driver training

In cooperation with Subaru (SA) the writer had the opportunity recently to assess the reaction times of drivers who were being trained in advanced driver skills at the Gerotek center, as well as the Swartkops racetrack in Pretoria

Objectives of the research project:

Assess the degree to which the Fatigue-o-meter could distinguish the changes in vigilance of these drivers as the day progressed and they were exposed to physically and mentally taxing procedures. Determine the correlation between objective measures of driver ability and the various ratings of vigilance as recorded during the day.

The agenda during the day's training was as follows:

The drivers met at 8h00 for a general theoretical lecture and orientation. The first activity of the day consisted of skidpan exercises. These activities were relatively mild and consisted of wet road turns, slides and skid control. This session ended in a precision and speed trial during which the drivers had to complete a specified circuit on the skidpan, making their way through obstacles. Driver times were recorded for this session and used as objective standard of general ability and skill.

The drivers were given lunch and relaxed for 40 minutes

The final session consisted out of grueling high speed (+- 160kms/hr) circuits at Swartkops. Drivers were expected to learn how to handle their vehicles at high speed and generally completed 3 to 6 circuits around the 1.3 km track. Drivers were visibly tired and mentally taxed by this activity. The fatigue-o-meter tests were done at the end of each of the above stages. Reaction times were corrected by a constant for errors.


Reaction time scores were obtained over all four stages.

Correlation between reaction times at each stage and the objective test of precision conducted at stage 2.

The graph below ( Graph 1) illustrates the vigilance of the drivers at each stage:

Graph 1: Driver reaction times on the Fatigue-o-meter:

As can be seen on the above graph, the drivers started relatively slow and were probably still waking up (Monday morning) at the start of the day. At the end of the skid pan exercise they reached their baseline level of vigilance and were relatively stable.

Lunch brought about an improvement of 20%. The afternoon session showed a dramatic deterioration of 50% from their lunch time response speeds and clearly indicates the effect of extended strain and road tension on driver vigilance.

Graph 2: Correlation between reaction times and objective measure of accuracy:

The above graph shows the correlation between reaction times and the measure of individual driving precision recorded at the end of the skid pan exercise. This line of exponentially increasing correlations indicates that two factors are at work:The driver's individual level of vigilance is constantly and positively correlated with driving precision.

The driver's individual deterioration in vigilance tends to produce a variable influence on precision.

General comments:

The above data adds practical credence to the Fatigue-o-meter ability to determine driver vigilance deterioration accurately and suggests that mild levels of negative driver vigilance can be restored (+- 20%) by rest and relaxation. The correlation analysis data suggests that the vigilance level of drivers (present as well as applicants) may be used to determine general driver precision. It furthermore indicates that drivers who are cautioned for lack of vigilance, may more often lack precision and vehicle control and may be more inclined to have driving accidents.

In the work-place

IR Procedures

The Fatigue-o-meter assesses the state of perceptual alertness of any employee at a specific point in time and against their own personal baseline of test scores established under controlled conditions before routine monitoring testing is commenced. We recommend that the baseline test is performed at least 2 times and preferably 3 times over the space of 3 days prior to starting the routine testing program.

Care should be taken to ensure that testing is done when the employee is rested, calm and relaxed. An effort should be made by the employee to do as well as they can and to choose their responses carefully and accurately. The baseline data on record records the statistical pattern of:

Accuracy (making a mistake about the hand orientation of the test figure)

The average number of errors made by the employee (the mean errors) is extended by 1 standard deviation plus 1, to determine the lower and upper range limits of expected accuracy from this particular employee.

Response time (the amount of time required to make a choice on a test figure)

The time which elapses from the moment when the figure is presented to the employee, to the point when a response is recorded, is taken and a norm is calculated for each employee in respect of all test items. This figure is extended by one standard deviation to determine the range limits for response time.

Whenever the test is performed, the data is compared to the baseline as follows:


If the employee makes more errors than the upper range limit (1 sd. Plus1), he/she will be flagged for too many errors and the number of errors is shown on their report (see reports. latest test profile) , compared to their mean baseline errors.

Response time:

If the employee takes longer than the upper range limit of elapsed time to decide on more than 5 items, she/he will be flagged for response time and the number of items slower than the upper range limit shown as a number (see reports. latest test profile).The Fatigue-o-meter is not a psychometric test, but monitors the actual level of vigilance displayed by the individual employee, against their own personal baseline. The individual results are never subjected to a norm table, or assessed against population statistics.

The following recommendations should under no circumstances be construed as taking precedent over the internal organizational practices, IR environment and agreements in place and are presented as general good practice guidelines for possible incorporation into the internal IR procedure of the client organization.

Proposed procedure when an employee is flagged:When an employee is flagged for the first time for accuracy or response time, the latest test profile should be printed out and the results discussed with the employee counseling session aimed at determining the source of fatigue. Care should be taken to distinguish slight deviations (accuracy and/or speed is just below baseline), from gross variations. Three causal conditions are normally presented and require different further action:

The employee shows no behavioral evidence (alcohol on the breath, staggering, bloodshot eyes, flu symptoms, etc.) and cannot explain the condition.The employee is clearly incapable to do their work or acknowledges that they are fatigued because of self induced actions or failure to rest properly. The employee is incapable or claims to be incapable because of factors beyond their control (an accident or injury, domestic problems, sickness, anxiety, shift pattern discomfort, etc). Cause (a) could lead to the following actions (in all cases record an incident (File Recording and incident) and make a hard copy printout for signature Reports List reported incidents for a candidate):

First timer and negligible variation warning to go to work but take care.First timer and more serious variation compulsory rest, relaxation, (show safety video), light duties and warning to take care. Gentle repeat trend pattern (flagged 1 2 times a month) 1st warning. Consideration may be given to achievement motivation stimulation for employees who suffer from poor work motivation. Repeat pattern (flagged routinely on Mondays, every 3rd or 4th testing event) 2nd warning.

Consideration may be given to achievement motivation stimulation for employees who suffer from poor work motivation. Severe repeat pattern (flagged almost every second testing) Final warning. Cause (b) could lead to the following actions (in all cases record an incident (File Recording and incident) and make a hard copy printout for signature Reports List reported incidents for a candidate): Follow the same procedure as would be applied when an employee is guilty of Dod, AWOL, damage to property or equipment. Cause (b) could lead to the following actions (in all cases record an incident (File Recording and incident) and make a hard copy printout for signature Reports List reported incidents for a candidate):

HR counseling, medical examination or appropriate actions.

Depending on the findings, personal counseling or medical boarding may be applied.

Temporary light duties may be considered.

Consideration may be given to achievement motivation stimulation for employees who suffer from poor work motivation.

Rolling out Fatigue-o-meter and some case studies

Learn more about the use of Fatigue-o-meter

The test consists of a randomized series of simple object requiring spatial judgments, which have to be made by the employee at speed. The test requires the subject to concentrate, decide and record their responses for a period of 2-5 minutes. The sophisticated statistical properties built into the system measures response time, accuracy of judgment and overall variability.

The performance of the employee is compared to his/her own performance during previous and baseline trials. If the subject's performance is significantly below par, or is more variable than usual, this is identified as cause for concern. You may then decide what to do about this situation (e.g.. you may decide to suspend the person from duties, determine the possible causes for the temporary impairment by an interview and behavioral inspection, breathalyzer, etc.).

It is clear from our research to date, that individual perceptual alertness varies normally across the population, therefore you may at point of selection, decide to select the most alert individual by reviewing their fatigue scores, however thereafter, individual negative variance analysis provides the most meaningful way of monitoring whom to trust with your assets. You will appreciate the value of spotting the individual who is impaired every Monday, after payday, after traumatized incidents, etc.

Psychometric theory indicates that combining judgment and reaction time, provides a valid system for measuring psychomotor efficiency. By adding the measurement of variability, the Fatigue-o-meter assesses the psychomotor performance deterioration which takes place as a result of perceptual fatigue and loss of concentration.

Reliability of the instrument:

The instrument has been validated on 7200 subjects and demonstrates a high reliability at (KR 20) 0.84. Subsequent usage of the instrument during more than 1 300 000 assessment cycles have repeatedly confirmed its reliability.

What to do if you want to use the system:

You will need:

A standard Web browser, we recommend the use of Chrome. If you wish to test more than one person simultaneously, you can use multiple PC's or android devices (smart-phones or tablets). The system is supported online.

Upon registration:

You need to decide how many employees you wish to test on the system. Based on this figure, your start-up pack will be prepared and installed by our representative. Upon installation, your system will be activated from our web site. The start up pack will provide you with a 1- month testing window period for the identified and registered employees. Our representative will provide support and training to your staff.

Each employee must complete 5 baseline trials on the system, to record their normal reactions and provide a start-up norm base.

Hereafter you may have these employees tested any number of times, before, after or during their working shift.

You may access reports on these employees at any time, but we will generate regular monthly performance reports via our website as well.

You will be billed monthly in advance for the registered number of employees on your system.

When staff leave your service, you simply remove them from the database.

As new staff join you, you register them on the system.

You will be provided with regular reports to review your staff profiles, comparing your employees with the general population norms, identifying staff who are consistently showing low scores, etc.

You may terminate your contract at any stage, by providing us with 7 days written notice. Your service will then be terminated at the end of the current period of credit.

It will cost you:

Our price depends on the number of employees registered as well as the number of sites registered, please Contact us for a formal quotation.

Consider the following:

Traffic related and occupational accidents have an enormous impact on a developing country such as the RSA in terms of human loss, lower productivity as well as loss of valuable resources and capital assets. Occupational accidents were calculated in 1996 as having cost the South African economy more than R17b alone. At a 6% annualized wage and medical expense increase, this figure was estimated to be R28b in 2005. The compensation Commissioner, according to Cosatu, pays out more than R993m to accident victims in industry every year (this figure obviously does not include loss of productive capacity or assets).

According to a study published by the CSIR's Transportek division in 2004, the total cost of road accidents (including the cost of vehicle damage, as well as human casualty), amounted to R42.8b in 1998. Working on a conservative increase of 6% p/a, this figure is estimated to have been R77b in 2005.According to an article published by Drive Alive in 2006, a United Kingdom study indicates that 20% of road accidents in the UK are caused by driver fatigue. It is suggested by the authors that this figure may be considerably higher (as high as 60%) in respect of the South African motorist.

The following graphical analysis shows how perceptual alertness (measured in accuracy and response times) of a sample group of bakery van drivers varied when tested by the Fatigue-o-meter:

Figure 1: Driver perceptual alertness

As can be observed from the above graph, 60% of the vehicle accidents recorded over the previous 6 months in the sample group, occurred with the drivers falling into the high error, slow response time group!

Case study - comparing two results and accident prevention on two sites

The following case study illustrates the latest findings ex two current sites, which clearly illustrate the results which may be achieved from a well managed system:

Monthly summary driver vigilance was recorded in respect of two sites in Gauteng . These two sites employ a total of 150 (80 and 70 respectively) drivers and operate bulk carriers making deliveries on a national basis. The objective of the research project was to determine whether:

1. The operational success in applying and managing the Fatigue-o-meter system differed from site to site, and

2. Whether a relationship exists between the effective management of the site Fatigue-o-meter system and a direct reduction in site vehicle accident costs.

The following statistics were collected on a monthly basis:

Ave Resp is the average response speed for the total group tested in the month.

Av Err is the average number of erroneous responses given by each group.

Resp Var is the degree of response variability caused by individual drivers being slower that their baseline response patterns.

Err Var is the degree of error variance caused by drivers making more errors than their baselines.

These results were graphically presented in order to determine whether consistent management monitoring and cautioning of drivers were having the effect of coaching the drivers to show faster response patterns, less error and less individual variance (erratic-ness) around their baselines.

The hypothesis contained in this analysis suggests that effective management of drivers vigilance leads to a convergent and steadily improving response pattern and therefore should lead to more attentive drivers on the road and a consequent reduction in accident costs.

The first graph below illustrates the well management site results. As can be seen, the various graphs show a convergent pattern and all graphs indicate a downward movement (in the direction of improved response vigilance and a reduction in cautions):

The second graph below indicates the poorly managed site, where the initial graphical direction was similar as the above site, but in month 3 the management control was lost and instead of converging, the site results show a divergent trend line, with a generally worsening of driver vigilance. The only graph showing an improving line is average response speed (likely caused by the drivers realizing that no action would be taken on errors):

The writer obtained accident statistics from both sites and the well managed site was compared with the poorly managed one. The following table shows the results from each site:

Accident analysis over the last 4 months of monitoring

Well managed site

Poorly managed site

1. Drivers on record 80 70
2. Drivers with accidents 5 21
3. Cost of accidents (R) over the period 70 000 400 000
5. Cost of accidents per driver 14 000 19 000

As can be seen, the site which shows a well managed vigilance profile, achieved an accident improvement ratio as follows:

4 times fewer accidents than the poorly managed site

82.5% less cost of total accidents.

A reduction of 26% in average financial impact (seriousness) per accident.

Years in operation
Fatigue tests
where did Fatigue-o-meter come from

About Us

The first version of The Fatigue-o-meter was released in 2006 by PSI SYSTEMS cc in conjunction with MANSPEC SELECTION & DEVELOPMENT SERVICES (PTY) LTD. The system is used by the client organizations to identify employees who may be temporarily perceptually fatigued.


Tel: +27 82 552 4614

E-mail: mwmanspc@mweb.co.za

Psi Systems

Tel: +27 82 551 9908

E-mail: jan@psisystems.co.za

Font generated by flaticon.com.

Under CC: Freepik, Dreamstale, Eucalyp, Smashicons