Wednesday, December 2, 2009

Bikeafer.com reviews the Thailand study

"Motorcycle Accident Causation And Identification Of Countermeasures In Thailand" was an influential study by Dr. Vira Kasantikul done between 1998 and 2001. The researchers were early adopters of the OECD methodology and worked closely with Head Protection Research Lab personnel in enhancing the original Hurt methodolgy. OECD was used in the Maids European study and the current US effort at OSU. It is important in understanding current research.

We singled out data on helmet use, alcohol, rain and rider training/skills issues as being relevant to the US situation.

The helmet data is extremely good, as the principal investigator, is a neurologist who performed an enhanced autopsy procedure on the head and neck of the dead riders. The data is instructive on neck injuries in particular, and is an effective counter to contentions that helmets increase neck injuries. The study did find that neck injuries are under-reported in most routine autopsies, but it is clear that serious cervical neck injuries are very few in relation to the reduction in head injuries claimed for helmets. Although the HPRL consultants are widely published on the helmet issue, the actual researchers were not significantly influenced by helmet-law issues and the information is of very high quality, in our opinion.

We did feel that injuries where helmets came off during the crash, which are reported separately in the report, should have counted towards the head injury totals, and we did take the liberty, as best we could, of trying to break out those injuries and adjusting the injury totals. We didn't count helmets which were incorrectly fastened in this crude adjustment. It is clear, even with these tweaks, that helmets provide significant protection when they are used. They reduce or prevent head injuries and death in more than 50% of crashes, and helmetless riders die 150% more often and have massive, disabling head trauma three times more often than their helmeted counterparts. The scope for additional severe neck injury because of helmet use is relatively small in comparison with the benefits of helmets.


The information on alcohol, unsurprisingly, has detail on the effects of alcohol on rider performance. Nothing we didn't know, but good information. Alcohol is a major killer of bikers.


On the rain issue, neglected by both Hurt and Maids, the Thailand practice of collecting exposure data by videoing and counting traffic at the crash sites in equivalent day-of-week, time-of-day and weather conditions successfully countered for the reduced ridership in the rain, and identified rain, principally due to its effect of rider vision, as a cause in two thirds of the rain crashes. We had taken this view but it is good to have our instincts on weather confirmed.

There are a few other notes of interest. The totally alien nature of the Thailand riding environment doesn't provide much additional data of interest in the US, but as the study is not widely available as a free download, we feel that our article, based on a copy kindly provided by one of the researchers, is a useful addition to the discussion on the OSU study and crash studies in general.

We feel that the contention of the OSU investigators that there were quality control issues in the Thailand study is unwarranted. Their enhanced autopsy procedures and methodology improvements in the exposure data over Maids are worthy of note. They used a control exposure population of 2100 for their 723 crash samples in the Bangkok part of the study, instead of the Maids approach of using the same number of control samples as crash samples, and their additional population controls obtained from their traffic count and video procedures proved to be of value.  The data was entered and coded by hand into Excel spreadsheets and crunched by SPSS, a standard statistical package, so by the nature of the work, some human error would probably have crept into the data and calculations.  The US-based pilot is using a customized Access database with enhanced data validation and built-in coding features.

We don't know what the study cost. It was financed privately by three Honda subsidiaries. Costs in Thailand would probably not convert easily to US conditions.

As we have said before in relation to the Maids study, we should consider ourselves very lucky to have a study this good in the USA.


0 Comments:

Post a Comment

Links to this post:

Create a Link

<< Home