The two papers that follow report the results of painstaking epidemiological work on pedestrian accidents involving children. Because these papers relate the age of the child to specific behavior patterns that result in accidents and to specific injury patterns that result from such accidents, they would appear to offer a sound empirical base for the development of a variety of countermeasures, whether such countermeasures are intended to reduce accidents or to limit the severity of the resulting injuries.
Unfortunately, however, the ability to generalize such data is extremely limited. A given pattern of child-pedestrian behavior is the result of a highly complex set of variables: the social class and ethnicity of the child, the ecological characteristics of the neighborhood and the broader community, the characteristics of the traffic pattern and the types and density of the vehicles that make it up, the current enforcement policy, climatic conditions, road and highway characteristics, and a host of other conditions that combine and interact in various ways. Consequently, few localities are sufficiently similar to justify the application of data from one to another. The injury patterns described by Ryan, for example, are those produced by Australian vehicles on children walking, playing, and cycling in a specific Australian city. It seems quite unlikely that the injury patterns produced in an American urban environment would resemble those that Ryan reports. Indeed, data from another Australian city might show striking differences.
For the same reasons, such data as Read presents cannot be used reliably as a before-and-after measure to assess the effectiveness of a specific countermeasure. Any significant changes in the data after the introduction of a countermeasure might well be attributable not to the countermeasure itself but to changes in traffic patterns, ecology, the weather, or other events which operated to reinforce or counteract the countermeasure in question.
The overwhelming obstacle to systematic research in vehicular accidents and on the assessment of countermeasure effectiveness is the investigator's inability to control the numerous variables that affect both the incidence and the consequences of accidents. The present papers, despite their inherently interesting data, are seriously limited by this problem.
The effects on mortality and the associated financial costs of wood heater pollution in a regional Australian city
