 |
Did the Program Work?
The methods described in this paper and now recommended by the US Centers for Disease Control and Injury Prevention, assume that bicycle helmet use is the most appropriate proxy indicator of preventable bicycle-related head injuries. This is based on the 74 - 85 percent effectiveness of helmets in reducing the risk of brain injury in a bicycle crash. It would take years to collect enough injury and death data to achieve adequate numbers for analysis because brain injuries are a relatively rare event in a community mounting a prevention program.
Recommendations:
- Helmet use should be measured, not helmet ownership.
- When using self-report surveys, ask about helmet use during a recent, brief period. Define use as “always use” and define non-use as all other categories.
- Even though the inherent problems of such surveys may affect the results, this is acceptable if the same survey is used for both pre- and post- observations and during any subsequent surveys of the same population.
- Observational surveys are the best method to measure the outcome of helmet use.
- A mix of observation sites needs to represent all riders in the target population and individual sites need to yield high numbers of riders and be duplicated in follow-up observations.
- An efficient community sampling frame can de developed by consulting community informants.
- Control groups are problematic. Instead the focus should be on the intervention community, attempting to measure the degree of change accurately and ensuring that the outcome is a direct consequence of the intervention.
- More observations taken during each time period and more cycles of observations are better than fewer ones. Migratory observations combine the best features of mobile and stationary methods.
- Informed consent and protecting the rights of subjects are generally not issues in evaluating publicly displayed behavior change.
- Serial observations should be conducted at the same sites, the same time of day, under the same weather conditions, and preferably by the same observers each time.
- A well-designed evaluation might include one or more observations before the intervention, followed by observations several weeks after, then every few months for at least 6-12 months.
The above approach rests on the prerequisite that the relationship between the countermeasure and the outcome of interest is known, so that it is possible to estimate how many injuries can be prevented for each incremental increase in the use of the countermeasure. The approach also intends to standardize the measurements needed, although validity and reliability issues remain. Evaluation of behavior change is still a “sticky wicket”, but lessons learned through bicycle helmet program evaluation can be applied to other public health issues in which the key outcome is publicly displayed behavior.
Editorial note: This paper, only briefly outlined here, gives an excellent discussion of the variables when evaluating bicycle helmet use programs. Read the entire paper for a more complete understanding of these factors. The suggested methods could also be useful for evaluating motorcycle helmet use programs, as well as other injury prevention programs.
Schieber RA, Sacks JJ. Measuring community bicycle helmet use among children. Public Health Reports March-April:116;113-121:2001.
back
to Headlines
|
|
|
