 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ELECTRONIC ARTICLE |
* Alberta Research Centre for Child Health Evidence, Department of Pediatrics, University of Alberta, Edmonton, Alberta, Canada
Departments of Emergency Medicine and Community Health and Epidemiology, Queen's University, Kingston, Ontario, Canada
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Methods. Two reviewers independently screened studies and applied inclusion criteria on the basis of searches of 17 bibliographic databases (eg, Medline and Embase). We also screened reference lists of relevant studies and contacted experts in the area. Studies were included if they represented primary research, a comparison group was used, the study population included children or the intervention was directly applicable to children, and objective outcomes were reported. Two reviewers independently assessed the methodologic quality of included studies with the Downs and Black checklist. A qualitative analysis was performed because of extensive heterogeneity among studies.
Results. We included 23 controlled studies, ie, 4 randomized, controlled trials, 5 controlled trials, and 14 quasiexperimental or observational studies. Only 8 of the relevant studies were published in peer-reviewed journals. School-based programs appeared to be effective at increasing short-term knowledge acquisition; outcomes were enhanced with active, hands-on participation, as opposed to passive activities. Safety day camps showed positive results for knowledge acquisition. Tractor training programs and community- and farm-based interventions showed mixed results. Studies examining the North American Guidelines for Children's Agricultural Tasks suggested that uptake improves if dissemination is accompanied by a farm visit from a safety specialist or if information about child development principles is provided in conjunction with the guidelines.
Conclusions. There is a lack of randomized, controlled trials and community-based trials in this area. Studies primarily examined intermediate outcomes, such as knowledge acquisition; few studies evaluated changes in injury rates. The interventions targeted at children and youths that were evaluated focused on educational interventions. There is both the need and potential for the development and evaluation of injury control interventions for children, particularly programs addressing lethal injuries to young/preschool-aged children.
Key Words: wounds and injuries • agriculture • systematic review
Abbreviations: NAGCAT, North American Guidelines for Children's Agricultural Tasks • ROPS, rollover protection structures • QI, quality index • RCT, randomized, controlled trial • IQR, interquartile range
This review was conducted to gather evidence regarding methods to reduce the burden of injuries experienced by children on farms. The objective was to conduct a high-quality systematic review to synthesize the evidence on the effectiveness of interventions to prevent the occurrence of childhood injuries in agricultural settings.
The North American Guidelines for Children's Agricultural Tasks (NAGCAT) provide a good example of an educational intervention that may hold promise but has not been subjected to rigorous evaluation. The NAGCAT are a set of consensus guidelines developed to assist farm parents in assigning appropriate and safe work to their children 7 to 16 years of age (www.nagcat.org).1 Since their introduction in 1999, the efficacy of NAGCAT has been the topic of considerable debate. Nonetheless, NAGCAT have been widely distributed and used by agricultural communities across North America and Europe. This review therefore also focused on evidence surrounding the efficacy of NAGCAT and their effective dissemination.
Childhood agricultural injuries represent an important public health problem. Farm children experience high rates of premature death,2–4 morbidity,5, 6 and disability7 attributable to injury. Children account for 
19% of all agricultural injury fatalities and hospitalizations.8 The direct and indirect economic costs attributable to farm injuries are substantial. Leigh et al9 estimated the costs of agricultural occupational injuries in the United States in 1992 to be $4.57 billion annually, for all age groups combined. In Ontario, costs for children (0–14 years of age) represent 23.6% of all costs related to agricultural machinery injuries.10
Many pediatric farm injuries occur because children are exposed to specific occupational hazards.2, 6 There are 3 main approaches to the protection of children from workplace injuries: (1) occupational health and safety regulations; (2) modifications of the work environment (eg, engineering controls); and (3) education about known hazards.
In nonagricultural industries, there are regulations and work standards that outline appropriate work tasks for both adults and children. These are uncommon for childhood farm safety but have been shown to be efficacious in other occupational settings.11 Children from family farms are exempt from most regulatory policies aimed at occupational safety12, 13 and child labor regulations.14 Much of the work performed by children on farms is therefore unregulated. In addition, there are few regulations regarding bringing children into the farm workplace, even if they are not involved in farm work.
As discussed recently, the historical reasons for the lack of work standards on farms are complex.15 In general, the lack of standards is related to the value placed on independence and self-sufficiency in rural societies.16, 17 Consequently, there is a reliance on voluntary safety standards, which may or may not be effective.18 Farm groups have been known to resist actively the institution of rules and regulations.16 This has resulted in farm children being exempt from most health and safety legislation despite the fact that they live and work in an occupational environment with known hazards.
Engineering controls have great potential to be effective in preventing farm injuries, but only if they are voluntarily adopted by farmers. One of the most common forms of engineering control on farms is rollover protection structures (ROPS) on farm tractors; the use of ROPS in Scandinavia has historically reduced the tractor rollover-related death rate to 0.19 The American Society of Agricultural Engineers has introduced a voluntary standard for inclusion of ROPS as standard equipment on new tractors.20 However, substantial numbers of North American farm children are operating tractors that are not equipped with ROPS.21 This suggests that reliance on a voluntary system for engineering controls is insufficient to protect many farm children from known occupational hazards.
Educational initiatives are the most commonly used approaches in farm injury prevention. These are generally considered to be less effective than engineering and regulatory initiatives in preventing injuries. Educational approaches may be useful as a means to increase awareness of injury control issues and to change attitudes toward them. However, educational approaches are heterogeneous and may produce varied results. It is clear that not all farm parents embrace educational messages.15 Furthermore, even if these messages are embraced, the acquisition of knowledge may not lead to changes in safety practices. This finding is consistent with much of the summarized injury prevention literature on agricultural safety programs, not only those involving children.22
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Reference lists of the included studies and other related studies were checked for potentially relevant articles. Content experts and an author of each included study were contacted for information on additional studies. The conference proceedings of the American Public Health Association were searched for the years 2001 to 2003 (because this was expected to identify NAGCAT-related research). Finally, the listing of extramural research awards funded under the National Institute for Occupational Safety and Health Childhood Agricultural Injury Prevention Initiative was reviewed to identify other relevant projects (www.cdc.gov/niosh/childag/ChildAgExtramuralResearch.html).
Both published research and unpublished research were considered for inclusion. We identified completed as well as ongoing research. The search was not restricted according to language of publication.
Selection of Eligible Studies
The identification of relevant studies involved 2 steps. First, 2 reviewers independently screened the titles and (when available) abstracts from the search output. Second, we obtained the full text of potentially relevant studies, and the 2 reviewers assessed each study by using a standard form with predetermined inclusion criteria. We resolved discrepancies with discussion between the 2 reviewers and with a third reviewer if necessary.
Included studies met the following criteria: (1) primary research evaluating the efficacy or dissemination of NAGCAT or the efficacy of other interventions to prevent acute childhood farm injuries; (2) inclusion of a comparison group (eg, controlled trial, before/after, or cohort design); (3) study population involving any age group as long as the intervention was relevant to children; and (4) report of at least 1 objective quantified outcome (eg, occurrence of injuries, uptake of guidelines, changes in behavior or attitudes, or knowledge acquisition). Studies were excluded if they simply described a program, if they provided an evaluation with no comparison group, or if the intervention specifically targeted adults.
Quality Assessment
The methodologic quality of included studies was assessed with a partially validated checklist developed by Downs and Black.23 This tool measures quality in terms of reporting, external validity, internal validity (bias and confounding), and power, with a maximal quality index (QI) of 29. QI scores of >20 were considered good, 11 to 20 moderate, and <11 poor. Two reviewers assessed study quality independently and resolved discrepancies through discussion. Quality was assessed only for published reports, internal reports, and reports from conference proceedings. Abstracts, presentations, and Web sites were not assessed for quality, because they did not provide enough detail for accurate assessment.
Data Extraction and Synthesis
Data were extracted directly into Microsoft Excel (Microsoft, Redmond, WA) by 1 reviewer and were checked for completeness and accuracy by a second reviewer. The following information was recorded for each study: year of publication, country where the study was conducted, study design, objectives, intervention and comparison groups, characteristics of the study population, number of participants, outcomes, type of injury targeted, study setting, results, authors' conclusions, and study limitations reported by the authors.
Our approach to analysis was qualitative rather than quantitative. A meta-analysis was not conducted because of extensive heterogeneity in terms of the interventions, study populations, and outcomes. Results were organized according to the 2 review questions.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
|
Two randomized, controlled trials (RCTs) examined the effects of different dissemination strategies for NAGCAT.25, 26 One study evaluated parental use and knowledge of NAGCAT at 6 and 15 months after intervention,26 whereas the other study measured injury incidence during a 2-year surveillance period.25 One study was published in a peer-reviewed journal in 2002 and was of rigorous methodologic quality (QI: 29 of 29),26 and the other study is ongoing.25 The studies showed that enhanced or active dissemination of NAGCAT improved outcomes, compared with standard dissemination strategies.
An additional 2 ongoing studies involving NAGCAT were identified through the National Institute for Occupational Safety and Health list of extramural research. One study is exploring factors that might explain the discrepancy between the content of the guidelines and parents' practices concerning assignment of tasks to children.27 The second study will evaluate the impact of adapting NAGCAT for ethnic communities.28
Effectiveness of Other Interventions to Prevent Childhood Agricultural Injuries
We identified 20 studies that evaluated the effectiveness of interventions to prevent childhood agricultural injuries.29–48 The studies varied widely in the interventions and outcomes studied, the types of injury targeted by the interventions, and the study design and methodologic rigor (QI range: 11–20). Although all studies evaluated educational interventions, their focuses varied; 7 were designed for, or examined in, the school setting,29–35 12 were community-based,36–47and 1 was designed for use on the farm.48
The most commonly assessed primary outcome was knowledge acquisition.29–32, 34, 36–38, 40, 44, 45, 47 The timing of this measure varied, ie, immediately after the intervention (n = 2),36, 37 1 day after the intervention (n = 1),31 1 week after the intervention (n = 1),32 4 weeks after the intervention (n = 1),45 1 year after the intervention (n = 1),37 and not specified (n = 7).29, 30, 34, 38, 40, 44, 47 Ten studies measured changes in safety attitudes and/or behavior as 1 of their principal outcomes.29, 35, 39, 40, 42–45, 47, 48 Many of these studies evaluated short-term changes; 2 studies reported outcomes measured the year after training,42, 43 and 1 study evaluated effects on behavior at 1 year after the intervention for a small subset of the original sample.35 One study measured recall of farm hazards at 1, 14, and 28 days and the extent to which children made comparisons with their own farms.33 Another study measured changes in hazard scores.47 Finally, 1 study examined changes in the numbers of fatalities.41
Two studies were classified as cluster RCTs,35, 47 5 were controlled trials,29, 33, 41, 46, 48 4 used cohort designs,39, 40, 44, 45 1 was described as a quasiexperimental, separate-sample, pretest/posttest, control group design,34 and 8 were before/after studies.30–32, 36–38, 42, 43 The studies were generally of moderate methodologic quality (median QI: 18; maximal score: 29; interquartile range [IQR]: 15–19). Generally, most performed well on components assessing reporting (median score: 8 of 11; IQR: 6–9) and bias (median: 6 of 7; IQR: 5–6). The studies performed poorly on components related to external validity (median: 1 of 3; IQR: 1–2), confounding (median: 2 of 7; IQR: 2–3), and power (median: 0 of 3; IQR: 0–0).
Seventeen studies were conducted in the United States and 3 were conducted in Australia.33, 44, 45 Seven studies were published in peer-reviewed journals between 1999 and 2002,32, 33, 35–37, 39, 46 and 1 was published in a book.48 Four studies were identified as abstracts,30, 31, 40, 47 4 were presentations,29, 41–43 3 were reports,38, 44, 45 and information for 1 was found on a Web site.34
School-Based Interventions
Seven studies evaluated educational programs administered within the school setting.29–35 These studies consistently reported positive results in terms of short-term knowledge acquisition or changes in attitudes toward farm safety. Results showed enhanced outcomes with active, hands-on participation, as opposed to passive activities.
Community- and Farm-Based Interventions
The community-based interventions varied widely in their method of delivery and the type of injury targeted. The interventions included theater (n = 1),36 camps (n = 2),37, 38 tractor safety initiatives (n = 5),39–43 and multifaceted interventions (n = 4).44–47
A study evaluating the effect of theater on farm health and safety knowledge showed an increase in short-term knowledge.36 Two studies that evaluated the effectiveness of farm safety camps showed improved short-term safety knowledge,37, 38 and 1 study showed some evidence that knowledge was maintained over the longer term (
1 year).37
Five studies evaluated tractor safety programs.39–42 One study evaluated a community-wide intervention and found that the number of fatalities decreased and the number of ROPS sales increased during a 2-year period.41 The other studies evaluated tractor training programs and produced inconsistent results. Two studies evaluated the tractor training program in Wisconsin, using the 1992 and 1993 cohorts.42, 43 Both studies found little difference in safety behaviors 1 year after training, although the first study noted that parents observed an increase in safety knowledge.42, 43 A study of the 4-H Tractor Program showed positive changes in safety behaviors but no change in attitudes (which is inconsistent with what is known about the progression of developing changes in behaviors),39 whereas a study of the tractor training program in Ohio reported a positive change but did not present data to support the findings.40
The remaining studies evaluated a variety of community-based44–47 or farm-based48 initiatives. The studies found mixed results with respect to the youth interventions. One RCT evaluated the Partners for a Safer Community program and found no differences in terms of knowledge, attitudes, practices, and leadership skills.47 Landsittel et al46 compared several interventions, including self-audit, youth intervention, and community coalition. The effectiveness of the interventions varied according to the baseline farm hazard scores. Overall, the self-audit group showed the best results, whereas the youth intervention group showed little change. The community coalition intervention showed positive results among farms with low baseline hazard scores. Two studies evaluated different components of the Child Safety on Farms (Giddy Goanna) Project in Australia. The first study evaluated a farm safety booklet targeted to young children and found significant improvements in children's knowledge, attitudes, and behavior, compared with the control group.44 The second study evaluated the impact of demonstration days that were part of the same multifaceted program, and the study found a significant increase in knowledge and improvement in attitudes.45 Finally, Hawk et al48 evaluated the Farm Safety Walkabout program and found that changes in hazard scores among the 2 intervention groups were not significantly different from those of the control group.
Research in Progress
We identified 3 ongoing studies evaluating prevention interventions. A pretest/posttest study is evaluating the effect of teaching teen farm workers about agricultural health and safety through school-based classes on English as a second language.49 The purpose of the second study is to develop, implement, and evaluate (with pretests and posttests) curricula for adolescents that focus on hazard recognition and problem-solving skills in the agricultural work setting.50 The third study will evaluate a health and safety curriculum for ninth grade students in rural Minnesota schools, using a cluster-randomized, nested, cohort design.51
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Effectiveness and Dissemination of NAGCAT
To date, there is little tangible evidence with which we can make practical recommendations on whether NAGCAT are likely to lower rates of injury or improve knowledge and attitudes regarding agricultural practices. The 1 study reviewed showed that an important number of injuries could be prevented through the application of NAGCAT.
The 2 studies that examined the effectiveness of dissemination strategies found that simple things can be done to improve the uptake of the NAGCAT guidelines. One study showed that making the guidelines available would lead to a degree of uptake by motivated farm families.26 Provision of information outlining the scientific basis for the guidelines and personal visits to the farm by a trained safety professional were shown to improve rates of uptake and decrease rates of injury, respectively.25, 26
Effectiveness of Other Prevention Interventions
The 20 studies relevant to this issue varied immensely in their method of delivery and target population. School-based programs consistently showed positive results in terms of short-term knowledge acquisition or changes in attitudes toward farm safety. We speculate that the success of these programs results from the age of the target group; that is, younger children may be easier to influence. Among the school-based programs, studies showed that outcomes were enhanced with active, hands-on participation, as opposed to passive activities. However, it is not known whether or how this increase in knowledge had any influence on behavior and ultimately injury rates.
Studies evaluating farm safety camps showed improved short-term safety knowledge, and 1 study showed some evidence that knowledge was maintained at 1-year follow-up assessments. Again, the success of these programs may be the result of the ages of the study participants. Studies evaluating tractor safety programs and community-based or farm-based initiatives showed mixed results.
Comments on the State of the Literature
We screened 5822 unique abstracts from 17 bibliographic databases. The vast majority of abstracts fell into 1 of several categories, ie, descriptive epidemiologic studies, descriptions of injury control programs without any evaluation or with only process evaluations, anecdotal reports of various approaches to farm safety prevention, or studies that had only limited relevance to the pediatric farm injury problem. The lack of controlled evaluations of farm safety interventions may be attributable in part to difficulties in conducting research in this area (eg, difficulty in identifying and recruiting farms or farm workers).
The volume of published literature evaluating prevention interventions does not reflect the amounts of public funds that have been committed in recent years to research on the childhood farm injury problem. This could reflect the fact that many of these projects are still in progress, but it could also reflect the need for more rigorous standards for the funding of such projects.
Much of the literature in this area is either unpublished or not indexed in databases, which makes it difficult to identify. Unpublished studies may be of lower quality because they have not gone through a rigorous peer-review process. This may affect the validity of results; however, we did not observe a notable difference in results for published versus unpublished studies.
Publication bias, or the selective publication of studies with statistically significant results, is important to consider for this body of literature. The dissemination of only positive results has implications for decision-makers, researchers, and program planners. There are ethical and scientific responsibilities to share research findings so that subsequent programs, research, and policies can build on past experiences, regardless of whether they are positive or negative.
Only 8 of the relevant studies were published in peer-reviewed journals, with low to medium impact factors. This is another indication of the methodologic quality of studies in this area. There may be reason to think that this pattern in changing. For instance, we searched from the year 1980 onward and found nothing of relevance until the 1990s. The studies published in peer-reviewed journals were all from 1999 onward.
Of the 23 relevant studies, 4 were RCTs, 5 were controlled trials, and the others used quasiexperimental or observational methods. The lack of RCTs and community-based trials in this field is problematic, because these study designs represent the best standards of evidence for the effectiveness of interventions.52 Modifications of the standard RCT design that may be more appropriate in this setting include the cluster RCT, in which groups are randomized rather than individuals, and randomization of the introduction of prevention initiatives to various regions.
Irrespective of study type, there was a lack of methodologic rigor in many studies. Recurrent weaknesses were related to issues of confounding, statistical power, and generalizability. Few observational studies adequately controlled for potential confounders, reducing our confidence that the results reflect the interventions in question. Few studies reported a calculation for sample size or statistical power. Without such information, we were unable to determine whether the studies were large enough to detect important effects. Findings indicating a positive intervention effect might have been missed. The external validity (generalizability) of studies was also generally poor, which makes the application of results in the planning of injury control interventions challenging.
The majority of studies evaluated short-term process outcomes (eg, knowledge acquisition or changes in attitudes or behavior). Few studies evaluated injury occurrence. The timing of outcome assessment was also less than optimal. Outcomes were typically measured immediately or shortly after the intervention was delivered. It would be useful to determine the long-term effects of these interventions. Moreover, studies did not distinguish between programs directed at youths who work on farms and children who live and play around the farm worksite.
Content of Existing Research
Evaluative studies do not appear to address directly 1 of the leading injury prevention priorities for farm children. Approximately one-half of fatal farm injuries among children occur among those 0 to 6 years of age (toddler and preschool ages).53 Few studies were explicitly aimed at this age group.
Few of the studies relevant to our review evaluated structural changes (eg, ROPS and seatbelts, machinery with appropriate guarding/shut-off devices, correction of fall hazards in haylofts, and animal containment systems) that can be introduced to farms to make them safer for all people, including children. Most studies directly addressed training and behavior of the affected groups of children. The lack of literature identified in this area may reflect the scope of our review and the criteria for inclusion.
There were no studies that evaluated the efficacy of programs (eg, enhanced day care, "child-free zones," or safe play areas) that aim to keep young children from coming into contact with farm worksite hazards. Programs aimed at providing childcare for farm children exist but have not been rigorously evaluated; therefore, their effectiveness is unknown. We did not identify any studies that examined the issue of supervision of children, whether toddlers or young workers, in the agricultural setting. There were also no studies of the effectiveness of incentive programs to improve safety training and/or the safety of farm worksites.
Challenges to Enhancing the Safety of Farm Worksites
An important consideration in the prevention of farm injuries among children is the fact that the farm worksite has not been subject to the imposition of occupational health and safety regulations. Dramatic changes in injury occurrence in other industrial worksites have come with the introduction of health and safety standards. We need to understand more about the reasons for this situation and how to invoke widespread changes in the attitudes of farmers toward the value of regulations in their workplace. Controlled studies that evaluate interventions to effect such changes, both individually and culturally, will be important for the implementation of effective agricultural injury control interventions.
There is evidence that the farm safety community is beginning to think about interventions that go beyond educational programs delivered to children. Some recent literature has addressed the lack of efficacy of educational interventions and the need for reassessment of other approaches (ie, regulatory).54 This has also been reflected in statements from groups such as the American Academy of Pediatrics.11
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Most of the existing literature either is unpublished or appears in peer-reviewed journals of modest rank. There is little evidence available from this literature that is helpful in making decisions about what programs to implement to optimally prevent the occurrence of pediatric farm injuries. Moreover, an important omission in this body of literature is that of evaluations of interventions aimed at preventing injuries among toddlers and preschool-aged children. The interventions targeted at children and youths that have been evaluated focus solely on education. Although the importance of educational initiatives is obvious and some are efficacious, educational initiatives by themselves appear to be insufficient to eliminate recurrent injury control problems on farms. We did not identify any studies of regulatory approaches that specifically targeted children.
|
|
|
ACKNOWLEDGMENTS |
|---|
Colette Freitas and Philip Berry (Alberta Research Centre for Child Health Evidence) assisted with administrative support and the collection of articles. Kris Cramer and Michele Tubman (Alberta Research Centre for Child Health Evidence) assisted with quality assessment and data checking, respectively. Deborah Emerton and the Office of Research Services at Queen's University provided administrative support. Catherine Isaacs (Canadian Agricultural Injury Surveillance Program) provided background injury surveillance data. We are grateful to the authors of the primary studies and other experts in the field who reviewed the preliminary list of included documents in an effort to identify additional studies.
|
FOOTNOTES |
|---|
Address correspondence to Lisa Hartling, MSc, Alberta Research Centre for Child Health Evidence, Aberhart Centre One, Room 9424, 11402 University Ave, Edmonton, Alberta, Canada T6J 2G3. E-mail: hartling{at}ualberta.ca
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. Marlenga, R. L. Berg, J. G. Linneman, R. J. Brison, and W. Pickett Changing the Child Labor Laws for Agriculture: Impact on Injury Am J Public Health, February 1, 2007; 97(2): 276 - 282. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. B Reed, D. T Claunch, H. P Cole, and J. M Mazur Characteristics of instructors at Farm Safety 4 Just Kids day camps Health Education Journal, June 1, 2006; 65(2): 180 - 192. [Abstract] [PDF]
|
|
|
|
 |
|
 W Pickett, R J Brison, R L Berg, J Zentner, J Linneman, and B Marlenga Pediatric farm injuries involving non-working children injured by a farm work hazard: five priorities for primary prevention Inj. Prev., February 1, 2005; 11(1): 6 - 11. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2004 American Academy of Pediatrics. All rights reserved. |
||
|
||
