This study was designed to explore the combinational effect of the roadway and traffic characteristics of national rural roads on fatal and serious-injury crashes in Namibia. Namibia, like many other countries in Southern Africa, experiences high numbers of high-severity crashes, particularly on its rural road network. The study applied Generalised Poisson regression models to analyse the study dataset. The results indicate that high-severity crashes are almost always influenced by a combination of factors, not simply the presence of a single factor. For example, on higher-order roads, a higher proportion of heavy vehicles in the traffic stream had the highest absolute impact on high-severity crash levels, but this effect was heightened when hilliness, wider lanes, surfaced shoulders and higher operating speeds were also present at the same time. On lower-order roads a higher proportion of light vehicles in the traffic stream had the highest absolute impact on high-severity crashes, and again this was compounded in the presence of wider unpaved shoulder widths and higher operating speeds. Overall, wider lane widths and wider unpaved shoulder widths, especially when found in combination with each other, were linked to a subsequent increase in higher-severity crash levels on higher-order roads. The study shows conclusively that the development of safer roads relies not on simply implementing minimum safety standards for individual design elements, but on understanding the relationship between features of road design, traffic parameters and road safety, to recognise which coincidences of factors affect crash risks on different classes of road. This knowledge can result in more careful road design so that crash likelihood is reduced.