Academic Articles Information

Information and the Reduction in Child Injury Mortality

Child mortality from unintentional injury or accident has been dropping over the last thirty years. Among children under 5, accident rates resulting in death have fallen from 44 deaths per 100,000 population in 1960 to 18.6 deaths per 100,000 in 1990. For children aged six to twelve years old, the analogous fall is from 19.6 to 9.8 deaths per 100,000. Why is this the case? A 1999 NBER working paper by Sherry Glied (subsequently published in Medical Care Output and Productivity) analyzes this phenomenon.

Glied gives four possible explanations for the decline: 1) a change in the children’s living conditions, 2) changes in the professional child injury knowledge base, 3) changes in regulation, and 4) changes in the information imparted to parents. We will look at each of these in turn.

  1. Change in Living Circumstance. The U.S. Census Bureau shows that median family income has increased between 1959 and 1989 from $27,632 to $45,956 (in 1999 dollars). If safety is a normal good, this should help child safety. On the other hand, during this time period, a higher percentage of mothers were working. Glied assumes that this decrease in parental time spent with the child will increase the accident rate. Because of the offsetting income and time effects, no conclusion is reached regarding the impact of living circumstances on child safety.
  2. Change in the professional injury knowledge base. Even if there were the same amount of accidents in 1960 as in 1990, if health care technology decrease the probability of death from a given accident, then the researcher could find decreased mortality with the same accident rate. Glied claims that there were “…substantial increases in the availability of childhood trauma care during the 1980s…[but] trauma care can contribute less to saving children’s lives than to saving adult’s lives.” Drawing on the epidemiology literature, the author claims that mortality rates likely did decline for each type of accident, but the magnitude of these declines seems to be small.
  3. Changes in Regulation. There were significant changes in regulation during this time period as well. “The late 1960s and early 1970s were the heyday of the U.S. consumer protection movement.” Also, between 1977 and 1984 all states passed laws that children under five had to be in a child safety seat and some states passed mandatory seat beat laws for all individuals. Despite these efforts by state and federal government, Glied contends that the regulations were ineffective.
  4. Changes in the Information imparted to parents. The author claims that this is the major driver of the changes in infant mortality. Changes in the information available to parents is difficult to measure. Glied cites an example of the increasing amount of information contained in Dr. Benjamin Spock‘s parenting guidebook and the increase in the number of publications indexed in Medline regarding childhood injuries. A further example is taken from New York City. When the Department of Health found that 30-60 children under age 5 were dying each year from falling off buildings, the city led a campaign to inform parents that they should install guards on their windows.

The argument Glied makes is very interesting, but the empirical analysis she uses to back it up is less robust. The data used is the National Mortality Detail File from a variety of years between 1968 and 1990. A linear prediction is made using the early years of the data. The residual between the projected mortality data and the actual childhood mortality is assumed to be caused by the above cited informational issues. A number of covariates and major categories of deaths are included, but the data is time series in nature so it is difficult to distinguish which of the four suggested mechanisms are causing the change in childhood mortality.