Improvement in Quantity and Quality of Prevention Measurement of Toddler Injuries and Parental Interventions

Injury is the leading killer of children in the United States, yet little research has focused on this vital subject. One of the distinct barriers to injury prevention is the absence of effective assessment devices. Epidemiological aspects of injury have been assessed, but these provide little information at a level suffi cient to allow conclusions about potential behavioral prevention methods. This paper describes an alternative, the Participant Event Monitoring (PEM) system. In this paper, the PEM system is used to examine a sample of 170 toddlers (ages 18 to 36 months), over a 6-month period, resulting in data on over 4,200 injuries, 1,000 proactive interventions, and 300 reactive interventions. PEM involves a structured interview, producing detailed information concerning measures of antecedents, events, and consequences of injury. Sample PEM data are included. Its ultimate goal is to guide effective interventions to decrease childhood injury. Injuries are the leading killer of children in the United States, taking more lives than the next nine leading threats to children’s health (Rivara, 1982). There are 14 million children injured every year (National Institute for Health Care Man271 We fi rst need to cite our funding agency, NICHD Grant #5R01HD25414-07. Connie Popkey assisted in multiple drafts of the paper and over 30 undergraduates assisted in coding the data. They deserve our sincere thanks. The interviewers Bonnie Benson, Jennifer Collins, Amy Damashek, Julie Long, Shannon Stanton, Lisa Wischmeyer, and Wendi Marien, M.S., all did consistently superb work as interviewers. The University of Missouri Pediatrics Department assisted us in recruiting, as did day care centers and parenting groups (such as “Parents as Teachers”). Jeff Crowson assisted in initial data entry and cleaning. Finally, our greatest thanks go to the women who spent such time and effort as participant observers. Without their training and willingness to accurately report on their children’s behavior, such research would not be possible. 272 PETERSON, DILILLO, LEWIS, & SHER IN BEHAVIOR THERAPY 33 (2002) PREVENTION MEASUREMENT OF TODDLER INJURIES 273 agement, 2000) and 600,000 of these children are hospitalized, 30,000 are permanently disabled, and 22,000 die from injuries (Rodriguez, 1990). Costs for treating unintentional childhood injury are estimated at over $14 billion in medical costs, $1 billion for other resources, and $66 billion in present and future work losses (Miller, Romano, & Spicer, 2000). Yet, prevention of childhood injuries has rarely been the subject of health care research. The majority of past work has been conducted by public health experts and epidemiologists, and has linked such factors as geography, ethnicity, and socioeconomic status to injury (Baker, O’Neill, Ginsburg, & Li, 1992). Although providing a necessary description of the broad demographic risk factors associated with injury, such research has rarely yielded the specifi city of data needed to understand the mechanisms of injury that could guide effective, family-level, behavioral methods of reducing childhood injury. One way of conceptualizing the prevention of childhood injuries is to place them on a continuum beginning with societal-level interventions involving no caregiver effort to interventions relying solely on caregiver effort. To understand this continuum, it is helpful to consider examples at each of fi ve levels: ● legislative changes (e.g., mandating that manufacturers produce only cribs with slats suffi ciently close together that an infant’s head could not fi t in between them; Consumer Product Safety Commission, 1979); ● product changes with minimal caregiver effort (e.g., encouraging parents to set water heaters one time to 125°F to insure prevention of scalding bums to children); ● environmental changes with some degree of repeated effort by caregivers (e.g., installing a child gate that must be closed each time it is used but requires only that and no subsequent supervision); ● learning safe behavior involving much caregiver or child effort and supervision (e.g., assisting parents in the proper use of child safety seats in cars and in their current use each time the child rides in the car); and fi nally, ● efforts involving maximum caregiver effort (e.g., prevention of child drowning in a bathtub or pool by continuous supervision; never leaving the child even for a single minute near water). The lower levels on the continuum are universally known to be the most effective, as no additional caregiver efforts are required, only the alteration of product safety. However, such interventions may be perceived by some as unnecessarily limiting individual freedom. Lobbying by powerful political groups has undermined the utilization of many effective barrier interventions (e.g., trigger locks on hand guns, the banning of mobile baby walkers; Peterson & Roberts, 1992). Individual freedom in the absence of government regulation is one of our country’s most cherished convictions, but for child safety this means the caregiver assumes total responsibility for children’s safety; most methods of maintaining child safety, such as continuous supervision, are so diffi cult to conduct in terms of cost of consistent parental effort (Canadian Children’s Safety Network, 1996) that they are unlikely to be used with peak effi ciency. Thus, there exists a constant tension between caregivers maintaining their complete right to make child-care decisions and the right of the community to protect its children by mandating caregiver behavior. Such decisions have often been infl uenced by the costs of intervention, with communities being more willing to impose rules on the caregiver if the demands of the rules are low. The tension can be described in this fashion: On one end of the scale, injury prevention can be achieved effectively, but with some cost to personal freedom; at the other end of the scale, high levels of personal freedom are preserved, but currently, because of our lack of knowledge concerning ways of ensuring consistent caregiver effort, some degree of effectiveness in maintaining children’s safety may be sacrifi ced. It is only recently that behavioral (Scheidt & Workshop Participants, 1988), more active (Shields, 1997) methods have received increased attention as methods of increasing effectiveness and concurrently decreasing effort or costs. This article describes a method of considering all levels of prevention, but unlike most past articles on prevention, focuses especially on the high end (maximum parental effort but attempting to also ensure maximum effectiveness) of the scale. Those few studies investigating spontaneous intervention efforts used by caregivers typically focus on situations in which parental action is the only way in which children’s safety can be maintained. The few data that do exist suggest that parents who are aware they are being observed rely on relatively continuous supervision of ongoing child behavior and they focus on children’s risk behaviors as a more frequent target for intervention than other behaviors (e.g., more on safety than on social behaviors such as “please,” or self-care such as brushing teeth). There may also be a shift toward decreasing such interventions from younger (1 year) to older (2 to 3 years of age) children. For example, Gralinski and Kopp (1993) asked mothers of toddlers to complete behavior checklists identifying child behaviors that elicited their typical parental interventions. They reported that 2 to 4 times more behavioral interventions with children concerned safety rules in comparison with other elicitors of parental attention, but the number of safety interventions unexpectedly decreased in children from 13 to 30 months of age. Similarly, Power and Chopieski ( 1986) observed more maternal restrictions placed on child injury behaviors than other behavior problems, with an average of 4.9 restrictions in each 45-minute session. Studies like these suggest that, left to their own devices, parents tend to rely on brief behavioral interventions to keep children safe from physical harm. In describing their interventions they anecdotally often seem relatively unskilled. Past research has shown that behavioral methods focusing on teaching a single preventive skill and using a strong consequence can be very effective (Peterson, Mori, Selby, & Rosen, 1988). There are, however, few studies currently in the literature demonstrating a widespread approach to prevention that has been effective. 274 PETERSON, DILILLO, LEWIS, & SHER IN BEHAVIOR THERAPY 33 (2002) PREVENTION MEASUREMENT OF TODDLER INJURIES 275 Behavioral changes that would involve altering general parenting strategies such as supervision, distraction from hazards, etc., at least in theory, can be very effective (Finney et al., 1993) and can also be broad, in terms of application to different types of injuries that can be prevented. Why have such broad tactics failed to be the subject of research? Many researchers have suggested that the absence of a reliable and sensitive measurement methodology that might reveal the situations in which changes in parenting are required is a major defi ciency in the area (Deal, Gomby, Zippiroli, & Behrman, 2000); this defi ciency may limit the possibility to effectively study broad-scale behavioral prevention. Perhaps if there were a more comprehensive system of assessing the situations in which injuries take place, the task of matching injury risk with type of preventive intervention would be more straightforward. Prior methods, however, have failed to produce the behavioral specifi city of information necessary to understand the complex behavioral mechanisms thought to underlie childhood injuries, mechanisms that represent some of the most promising points of intervention for future preventive efforts. The present paper describes a methodology focused on achieving this goal. By having caregivers record in detail the primary behavioral and environmental events that precede, coincide with, and follow the injury event, heretofore unknown factors contributing to injury risk may be identifi ed. Current Sources of Data Examining Etiology of Injury Random Community Samples Some data relevant to childhood injury occurrence that may infl uence caregivers’ decisions about appropriate behavior underlying injury have been obtained by simply selecting a random community sample. The data collected from these samples may reveal what drives the types of preventive interventions conducted by caregivers. For example, beliefs about the extent to which injuries are a threat to children may infl uence all levels of the continuum, from altering the environment to increasing supervision. For example, Eichelberger, Gotschall, Feely, Harstad, and Bowman (1990), in a national sample, and Peterson, Farmer, and Kashani (1990), in a smaller community-based sample, found that parents were concerned with a variety of child hazards. However, when asked about specifi c hazards, they listed concerns such as drug use and stranger abduction, which are legitimate concerns, but which harm far fewer children than the leading cause of death, physical injury. Such priorities are likely to result in parents who do not give injuries the resources and effective preventive efforts they deserve. Parents may have inaccurate beliefs about the extent to which children are at risk for injury , both because of inaccurate beliefs stemming from sources such as the media, health care professionals, and friends, but also because they do not accurately recall the number of injuries their own child has sustained, if the injuries were not serious. Parents seem to recall that a medically attended injury occurred, but they do not recall the incident in suffi cient detail, and this makes it diffi cult for them to plan to prevent future injuries. Further, in community samples parents do not seem able to readily recall even the occurrence of nonmedically attended injuries, let alone the details surrounding the injury (Peterson, Moreno, & HarbeckWeber, 1993). The use of large-scale community samples remains rare, partially because of their potential for inaccuracy (individuals feeling a need to report more safety habits than they actually practice or having a poor memory for relevant events) and the expense of such data. Because the present method produces high accuracy concerning the occurrence of all events (and thus yields records of a large number of events), a very large number of participants is unnecessary to study a sizeable sample of injuries.