Identifying the Key Predictors of Occupational Fatigue among Long-Distance Truck Drivers in East Africa: A LASSO-Regularized Regression Approach

This study utilized LASSO-regression analysis on 207 long-distance truck drivers along the Kenya-Uganda corridor to reveal a 51.7% prevalence of occupational fatigue, identifying deadline pressure, stimulant use, and excessive shift lengths as key predictors that necessitate urgent regulatory and organizational interventions.

The Gist

Imagine a long-haul truck driver in East Africa as a marathon runner who is forced to run a race that never ends, on a track that is often broken, while being chased by a relentless coach screaming, "Faster! You're late!"

This research paper is a deep dive into why these "marathon runners" (truck drivers) are so exhausted that they are a danger to themselves and everyone else on the road. The researchers didn't just ask, "Are you tired?" They used a high-tech detective tool to figure out exactly why they are tired.

Here is the story of the study, broken down simply:

1. The Setting: The Great East African Highway

Think of the road connecting Kenya and Uganda as a giant, busy conveyor belt carrying goods to the rest of the continent. The drivers are the workers keeping this belt moving. The researchers stopped at two major border crossings (Busia and Malaba) and interviewed 207 drivers.

The Big Discovery: More than half of them (51.7%) were suffering from severe occupational fatigue. That's like saying if you walk into a room of 100 truck drivers, 52 of them are running on empty batteries.

2. The Detective Tool: The "LASSO" Filter

The researchers had a huge list of possible reasons why drivers might be tired:

• Are they too old?
• Do they have too many kids?
• Do they know the traffic laws?
• Do their bosses care about them?
• Do they drink coffee or chew khat (a stimulant plant)?
• Are they under pressure to meet deadlines?

To find the real culprits, they used a statistical method called LASSO.

• The Analogy: Imagine you have a messy room full of 100 different objects (clues). You want to find the two things that actually caused the fire. LASSO is like a super-smart vacuum cleaner that sucks up all the useless clutter (like "did they eat breakfast?" or "what color is their truck?") and leaves only the two most important items on the table.

3. The Two Main Culprits

After the "vacuum" cleaned up the data, only two suspects remained standing, and they were the heavy hitters:

Suspect #1: The "Deadline Pressure" Monster

• What it is: The feeling that you must deliver the cargo by a specific time, no matter what.
• The Impact: Drivers who felt this pressure were 20 times more likely to be fatigued.
• The Metaphor: Imagine a runner who is told, "If you don't finish in 2 hours, you lose your job." They will push their body past its limits, ignoring the warning signs of exhaustion. In this study, the pressure to meet deadlines was the single biggest driver of fatigue.

Suspect #2: The "Stimulant Trap"

• What it is: Using substances like caffeine, energy drinks, or khat (a local leaf that keeps you awake) to stay alert.
• The Impact: Drivers who used these substances were 14 times more likely to be fatigued.
• The Metaphor: This is like trying to fix a flat tire with duct tape. The stimulant gives you a quick burst of energy (the tape), but it doesn't fix the underlying problem (the flat tire). In fact, it often makes the tire worse later. These drivers were using stimulants to stay awake, which disrupted their sleep, making them even more tired the next day, creating a vicious cycle.

4. The Surprising "Not Guilty" Verdicts

The study looked at other things people thought were the problem, but the data said otherwise:

• Knowing the Rules: Drivers who knew the laws about rest breaks weren't necessarily less tired.
  • Why? Because knowing you should rest doesn't help if your boss won't let you stop. It's like knowing you need to eat healthy, but having no money to buy food.
• Paperwork: Whether drivers kept official logs of their rest times didn't matter much.
  • Why? Because a company can have perfect paperwork while still forcing drivers to drive 18 hours straight. The paperwork was just a "prop," not a real safety net.

5. What Does This Mean for the Real World?

The researchers are saying: We can't just tell drivers to "sleep more."

If you try to fix this by just giving drivers a pamphlet on safety or asking them to fill out more forms, it won't work. The system is broken.

The Solution requires a team effort:

1. The Bosses (Trucking Companies): They need to stop treating drivers like machines. They need to create realistic schedules that don't require drivers to drive 24/7.
2. The Government: They need to enforce the rules. If a truck is driving too long, there needs to be a real consequence, not just a warning.
3. The Drivers: They need support to say "no" to unsafe schedules without losing their jobs, and they need help breaking the cycle of using stimulants to stay awake.

The Bottom Line

Fatigue in East African trucking isn't just about being tired; it's about a system that pushes people to the breaking point. The study found that pressure and substance use are the two main engines driving this fatigue. To fix the road safety crisis, we have to turn down the pressure and stop the cycle of using drugs to stay awake, rather than just blaming the drivers.

Technical Summary

1. Problem Statement

Driver fatigue is a critical global public health and occupational safety issue, particularly in the transport sector, where it significantly impairs vigilance, reaction time, and decision-making. In Low- and Middle-Income Countries (LMICs), the burden of road traffic fatalities is disproportionately high, exacerbated by limited enforcement of occupational health standards, irregular driving schedules, and poor infrastructure.

Despite the economic importance of the Northern Corridor (linking the Port of Mombasa to landlocked East African nations like Uganda, Rwanda, and South Sudan), there is a scarcity of empirical data regarding occupational fatigue among truck drivers in this region. Existing policies (e.g., Kenya's Road Safety Policy) recommend structured rest breaks and driving limits, but implementation is hindered by inconsistent enforcement. This study addresses the gap by investigating the prevalence and specific determinants of occupational fatigue among long-distance truck drivers operating at the Kenya-Uganda border points (Busia and Malaba).

2. Methodology

Study Design and Population

• Design: Cross-sectional analytical study.
• Setting: Busia and Malaba border towns, key entry/exit points on the Kenya-Uganda corridor.
• Participants: 207 exclusively male long-distance truck drivers.
• Inclusion Criteria: Drivers aged ≥28 years, holding valid Class C or E licenses, with ≥4 years of long-distance driving experience, originating from Mombasa.
• Exclusion Criteria: Drivers on medication affecting sleep or alertness.
• Sampling: Systematic random sampling (every fourth truck) during the customs clearance waiting period.

Data Collection and Variables

• Outcome Variable: Occupational fatigue, measured using the Chalder Fatigue Questionnaire (CFQ-11). Scores were dichotomized (binary method); a score of ≥4 indicated the presence of fatigue.
• Predictor Variables: Socio-demographics (age, education, marital status), occupational factors (shift length, deadline pressure, record-keeping), health behaviors (substance use, sleep habits), and psychosocial factors (social support, financial strain).
• Data Collection: Structured, interviewer-administered questionnaires (translated to Kiswahili) by trained research assistants.

Statistical Analysis

The study employed a rigorous machine learning-enhanced statistical approach to handle potential multicollinearity and overfitting:

1. Univariable Screening: Variables with $p < 0.2$ were retained for further modeling.
2. LASSO Regression: A Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression was performed using the glmnet package in RStudio.
   • Purpose: To select the most relevant predictive features from a larger set of variables by applying a penalty function, shrinking coefficients of less important variables to zero.
   • Parameter Selection: The optimal penalty parameter ($\lambda$) was determined via 10-fold cross-validation, selecting the $\lambda_{1SE}$ (one standard error rule) for a parsimonious model.
3. Multivariable Logistic Regression: Variables retained by LASSO were entered into a final logistic model to calculate Adjusted Odds Ratios (aOR) and 95% Confidence Intervals (CI).
4. Model Validation:
   • Discrimination: Assessed via the Area Under the Receiver Operating Characteristic Curve (AUC).
   • Internal Validation: Performed using bootstrap resampling (1,000 iterations) to assess model stability and generalizability.

3. Key Results

Prevalence

• The overall prevalence of occupational fatigue among the participants was 51.7%.

Predictor Selection (LASSO)

The LASSO procedure reduced the initial set of candidate variables down to two significant independent predictors:

1. High Deadline Pressure
2. Substance Use (including stimulants like khat, caffeine, and alcohol).

Other variables, such as employer medical support, knowledge of ILO regulations, and work-rest record-keeping, were excluded by the LASSO model as they did not demonstrate independent predictive power once the primary drivers were accounted for.

Multivariable Model Findings

The final logistic regression model demonstrated exceptional performance:

• High Deadline Pressure: Drivers experiencing high pressure had 20.50 times higher odds of fatigue compared to those with low pressure (aOR = 20.50; 95% CI: 1.57–31.00; $p < 0.001$).
• Substance Use: Drivers using substances to maintain alertness had 14.37 times higher odds of fatigue (aOR = 14.37; 95% CI: 1.62–26.00; $p = 0.017$).

Model Performance

• Discrimination: The model achieved an AUC of 0.98, indicating near-perfect ability to distinguish between fatigued and non-fatigued drivers.
• Bootstrap Validation: Across 1,000 bootstrap samples, the mean AUC remained stable at 0.987 (95% CI: 0.983–0.991).
• Accuracy: Mean sensitivity was 85% and mean specificity was 86%.

4. Key Contributions and Significance

Methodological Contribution

This study is notable for applying LASSO regularization in the context of occupational health research in East Africa. By effectively addressing multicollinearity and preventing overfitting, the authors derived a highly parsimonious model that isolates the most potent risk factors, moving beyond simple correlation to identify robust predictors.

Theoretical and Practical Insights

• Systemic vs. Individual Factors: The study challenges the notion that fatigue is solely an individual issue (e.g., lack of knowledge or poor sleep hygiene). Instead, it highlights that organizational and operational pressures (deadline pressure) are the root causes, which subsequently drive maladaptive coping mechanisms (substance use).
• The "Knowledge-Action" Gap: The finding that knowledge of ILO regulations was not a significant predictor suggests that awareness alone is insufficient without the agency to enforce rest periods against economic pressure.
• Symbolic vs. Substantive Compliance: The non-significance of work-rest record-keeping implies that in this high-pressure environment, paperwork often serves as a formality rather than a reflection of actual safety practices.

Policy Implications

The findings necessitate a shift in intervention strategies:

1. Enforcement: Urgent need for enforceable regulations on driving hours and deadline management, moving beyond voluntary guidelines.
2. Corporate Responsibility: Transport companies must shift from viewing drivers as expendable assets to valuing them as human capital, implementing wellness programs that address substance use through rehabilitation rather than punishment.
3. Targeted Interventions: Public health messaging should focus on the dangers of stimulant use as a coping mechanism for work pressure, rather than generic safety education.

Conclusion

The study concludes that occupational fatigue among East African long-distance truck drivers is a multifactorial crisis driven primarily by intense deadline pressure and the resulting reliance on substance use. The high predictive power of the LASSO-derived model underscores the urgency for multi-stakeholder interventions that address the systemic operational pressures of the transport corridor to safeguard both driver well-being and public road safety.