Lies, damned lies and statistics.

Thing is, KSI (killed or seriously injured) means a large range of things. Lumping it all together makes it largely meaningless.

For example, a 'serious injury' that makes the stat for KSI is a broken thumb. Whereas a broken finger does not.A broken thumb (a few weeks off work, some physio) is viewed identically to a spinal fracture that puts you in a wheelchair for the rest of your life.

The second article actually says this :

But these headline figures don’t tell the whole story.

According to the DfT, reporting practices have changed since the last time they were compiled

But it’s the way injuries and deaths are being reported that might be skewing results. The claim is that more seriously injured casualties are being recorded now and those that were once thought to be minor injuries are being promoted to particularly-nasty-and-needing-a-little-more-attention

As Rob JP says the SI part of KSI is absurdly broad.

I was involved in a cycling accident in 2012. Broke neck of femur requiring pinning and week in hospital followed by six weeks on crutches.

I've since climbed 3000 foot mountains and live life as before.

Statistically I've same value as someone with life changing amputation, paraplegia or brain injury.

For the statistics, however, you need some sort if figure to make comparisons. If you were just to take the number of fatalities, improvements in critical care would give the impression of improvements in traffic safety.

Indeed - the same can be said for safety improvements in vehicles over the years: seat belts, side impact bars and front crumple zones, anti-lock brakes (as well as a whole host of other three-lettered acronymed safety features), etc etc. Tyres especially have come a long way, particularly in terms of performance in poor weather and now the coming of all-season tyres as well as a greater use of winter tyres.

Some of the above (e.g. ABS and similar) I think have actually contributed to more accidents, because people driver faster, closer to vehicles in front and brake later because they believe the technology will do the job. Unfortunately it is often a driver's poor judgement of road and weather conditions as well as an over-confidence of the performance of this tech (as well as their own abilities) in all weathers that leads to most accidents.

The other thing is, of course, people are driving cars (and motorbikes, vans/lorries) that are more powerful and faster than they were 20+ years ago - more momentum to stop, often leading to worse accidents, especially if the driver loses control and ploughs into pedestrians. The huge increase in the number of high-powered motorbikes is, I'm sure, also a factor in the number of road deaths plateauing.

Quite a lot of fatal or (prorper) serious-injury accidents in my area are down to mainly (not in any order):

1. Increases in traffic levels, especially out of the normal rush hour times;
2. More higher powered vehciles and bikes around, often they take more risks overtaking slower-moving (often not that far under the speed limit, if at all) vehicles, especially lorries;
3. People under more time pressure (mainly job-related) than was the case 20+ years ago. The 'lifting' of the recession hasn't improved this - many people are still having to work under severe pressure for workload/deadlines, often having to accomplish far more in the working day than a few years ago with no more resources (as well as pay). This encourages people to speed and IMO take far more risks driving than before;
4. Poor signage and layouts of junctions, especially of rural country lanes joining/crossing fast-moving roads such as national speed limit single lane roads (overtaking problems and not seeing traffic pulling out/crossing) and dual carriageways (no decent slip-ons and traffic levels high/fast, leading to frustrated drivers taking high risks crossing traffic. Often such junctions aren't lit either, making the dark rush hour times in winter the most dangerous, especially when bad weather is around.