Development of a Road Accident Prediction Model to Enhance the Safety of Gravel Road Maintenance: A Case of Monduli District, Tanzania

Gravel roads managed by the Tanzania Rural and Urban Roads Agency (TARURA) in Monduli District serve pastoral, agricultural and tourism communities, yet their maintenance is scheduled by routine condition-based cycles rather than by safety risk. High-risk sections therefore receive no priority, and TARURA has no quantitative means of showing how maintenance reduces accidents. This study developed a road accident prediction model to enhance the safety of gravel road maintenance in Monduli District. A mixed-methods, cross-sectional design was used. A structured questionnaire administered to 83 stakeholders, comprising TARURA engineers, contractors, project managers, road safety stakeholders and road user representatives, rated twelve road condition variables, which were ranked with the Relative Importance Index (RII). Seven variables met the threshold of RII ≥ 0.75: maintenance quality (0.930), drainage adequacy (0.911), surface condition (0.880), clear zone (0.851), sight distance (0.819), horizontal alignment (0.790) and carriageway width (0.759). These were used as predictors in a multiple linear regression (MLR) model of a section Safety Index, an accident risk index on a 0–10 scale anchored to police and TARURA accident records, using 83 section-level assessment records from 25 sampled road sections split 70/15/15 into training, validation and test sets. The model explained 95.8% of the variance in the training set (F = 162.49, p < 0.001), with all seven predictors significant at p < 0.001; surface condition (β = 0.339) and clear zone (β = 0.291) had the largest effects. On the test set, R² was 0.949, and MAPE was 1.52%. Scenario testing showed predicted accident risk rising consistently from 11.2% under excellent conditions to 89.7% under very poor conditions. Accident risk on Monduli's gravel roads is therefore driven mainly by surface and maintenance-related conditions, and the model offers TARURA a practical tool for ranking sections by risk and prioritising maintenance.

Authors

Institutions

Publication Details

Journal
East African Journal of Engineering
Published
2026-11-21
DOI
https://doi.org/10.37284/eaje.9.2.6356
Primary Topic
Traffic and Road Safety
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Development of a Road Accident Prediction Model to Enhance the Safety of Gravel Road Maintenance: A Case of Monduli District, Tanzania

Ambele Mtafya, Janeth Erasto
East African Journal of Engineering
Traffic and Road Safety
article

Development of a Road Accident Prediction Model to Enhance the Safety of Gravel Road Maintenance: A Case of Monduli District, Tanzania

Ambele Mtafya, Janeth Erasto
article en

Abstract

Gravel roads managed by the Tanzania Rural and Urban Roads Agency (TARURA) in Monduli District serve pastoral, agricultural and tourism communities, yet their maintenance is scheduled by routine condition-based cycles rather than by safety risk. High-risk sections therefore receive no priority, and TARURA has no quantitative means of showing how maintenance reduces accidents. This study developed a road accident prediction model to enhance the safety of gravel road maintenance in Monduli District. A mixed-methods, cross-sectional design was used. A structured questionnaire administered to 83 stakeholders, comprising TARURA engineers, contractors, project managers, road safety stakeholders and road user representatives, rated twelve road condition variables, which were ranked with the Relative Importance Index (RII). Seven variables met the threshold of RII ≥ 0.75: maintenance quality (0.930), drainage adequacy (0.911), surface condition (0.880), clear zone (0.851), sight distance (0.819), horizontal alignment (0.790) and carriageway width (0.759). These were used as predictors in a multiple linear regression (MLR) model of a section Safety Index, an accident risk index on a 0–10 scale anchored to police and TARURA accident records, using 83 section-level assessment records from 25 sampled road sections split 70/15/15 into training, validation and test sets. The model explained 95.8% of the variance in the training set (F = 162.49, p < 0.001), with all seven predictors significant at p < 0.001; surface condition (β = 0.339) and clear zone (β = 0.291) had the largest effects. On the test set, R² was 0.949, and MAPE was 1.52%. Scenario testing showed predicted accident risk rising consistently from 11.2% under excellent conditions to 89.7% under very poor conditions. Accident risk on Monduli's gravel roads is therefore driven mainly by surface and maintenance-related conditions, and the model offers TARURA a practical tool for ranking sections by risk and prioritising maintenance.

East African Journal of EngineeringVol. 9(2)
Dar es Salaam Institute of Technology (TZ)
Traffic and Road Safety
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon
Development of a Road Accident Prediction Model to Enhance the Safety of Gravel Road Maintenance: A Case of Monduli District, Tanzania — Ambele Mtafya, Janeth Erasto · East African Journal of Engineering (2026) | TGRS Research Map | TGRS