You are here

Article Text

Article menu

Download PDFPDF

Original article
Acute occupational injury among adolescent farmworkers from South Texas
Free
  1. Eva M Shipp1,
  2. Sharon P Cooper2,
  3. Deborah J del Junco3,4,5,6,
  4. Charles J Cooper7,
  5. Ryan E Whitworth8
  1. 1Department of Epidemiology and Biostatistics, School of Rural Public Health, Texas A&M Health Science Center, College Station, Texas, USA
  2. 2San Antonio Regional Campus, The University of Texas Health Science Center at Houston School of Public Health, San Antonio, Texas, USA
  3. 3Departments of Surgery and Pediatrics, The University of Texas Health Science Center at Houston School of Medicine, Houston, Texas, USA
  4. 4Division of Epidemiology, Human Genetics and Environmental Sciences, The University of Texas Health Science Center at Houston School of Public Health, Houston, Texas, USA
  5. 5Center for Translational Injury Research, University of Texas Health Science Center at Houston, Houston, Texas, USA
  6. 6Center for Clinical and Translational Sciences, The University of Texas Health Science Center at Houston, Houston, Texas, USA
  7. 7Coordinating Center for Clinical Trials, The University of Texas Health Science Center at Houston School of Public Health, Houston, Texas, USA
  8. 8Statistics/Epidemiology Unit, Research Triangle Institute, Durham, North Carolina, USA
  1. Correspondence to Eva Monique Shipp, Department of Epidemiology and Biostatistics, School of Rural Public Health, Texas A&M Health Science Center, 210 SRPH Administration Bldg, College Station, TX 77843-1266, USA; eshipp{at}srph.tamhsc.edu

Abstract

Purpose This combined cross-sectional/cohort study addressed research gaps by estimating the rate of non-fatal occupational injury and identifying potential determinants among a population of adolescent farmworkers who are largely Hispanic and migrant.

Methods The cohort included 410 farmworkers (aged 13–19 years) attending high school in South Texas along the border with Mexico. Data collection involved a self-administered, Web-based survey that solicited information on demographics, farm work variables including person-time at risk, occupational injury, health status and health risk behaviours. Cox regression was used to identify potential risk factors for non-fatal injury events experienced during a 9-month recall period.

Results Depending on the definition of injury, the rate of non-fatal injury ranged from 27.0–73.6/100 full time equivalents. Variables with an increased and statistically significant HR in an adjusted Cox model included: age groups <15 years-old (5.82) and 16 years-old (4.47), usually sleeping <8 h during the week (2.10), feeling tense, stressed or anxious sometimes/often (2.25), not watching TV (2.65), working around ditches (2.01) and detasseling (2.70).

Conclusions The high observed rates of non-fatal injury combined with the potential negative consequences and cost of these injuries signifies a compelling need for injury prevention efforts targeting adolescent, Hispanic, farmworkers.

https://doi.org/10.1136/injuryprev-2012-040538

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Background

    Although agriculture is among the most hazardous industries in the USA, a number of minors labour in this industry as migrant and seasonal farmworkers. Based on the National Agricultural Workers Survey, minors comprise 6% of the entire hired farmworker population, which is largely Hispanic (83%), foreign-born (75%) and often temporary (ie, migrant or seasonal).1 The precise number of minors working as farmworkers is not known due to difficulties enumerating a mobile, temporary, underage and potentially undocumented workforce.1 Crude estimates are in the hundreds of thousands.2

    These young workers contend with a number of occupational hazards (eg, exposure to chemicals, sharp implements, long work hours)3 that could increase their risk of injury especially when combined with their development stage.4 However, these youth are protected by less stringent regulations (eg, minimum age requirements, hazardous task and hour restrictions) than their peers in non-agricultural industries. Recently, efforts were underway to address this inequality by revising the Agricultural Child Labour Hazardous Occupations Orders, but proposed changes were not adopted.5

    The high rates of fatal injury among young workers in agriculture illustrate its hazardous nature. The rate among those aged 15–24 years in the US from 1998–2007 was 21.3 deaths per 100 000 full time equivalents (FTE), second only to construction, and nearly seven times the rate for all industries combined (3.6 deaths per 100 000 FTE). Further, the rate of fatal injury for young Hispanic workers was nearly twofold the rate for young non-Hispanics.6 Despite these high rates for fatal injury, little is known about non-fatal injury or its prevention among farmworkers in general and minors specifically.

    A number of factors complicate our ability to study non-fatal injury among farmworkers including: fear of retribution for reporting problems at work, difficulty enumerating a mobile workforce, temporary employment at multiple worksites each season, which can hinder calculation of person-time at risk (ie, number of hours spent working), and employment on small farms that are exempt from Occupational Safety and Health Administration regulations.7–9 A general lack of access to healthcare, a lack of clinicians trained in agricultural health, and language barriers may further minimise reporting. The meaning of ‘work-related injury’ in English does not translate adequately into Spanish.3

    Common criteria for injury definitions used in occupational research (eg, loss of ≥4 h of work time or medical treatment) may be too restrictive.3 The majority of farmworkers are impoverished, with a median annual family income of <$17 500, and lack employer-provided health insurance.1 They may not be able to pay for medical treatment or take time off from work. Accordingly, surveillance systems based solely on data from patient records, employer-reports (eg, Bureau of Labour Statistics Survey of Occupational Injuries and Illnesses), or worker compensation records, may yield substantial underestimates of injury risk.9–11

    To overcome these barriers and address research needs, we conducted a 3-year, population-based, combined cross-sectional and cohort study of adolescent farmworkers from the Texas-Mexico border, home to a number of migrant and seasonal farmworkers and their families. This study was school-based and supported by community stakeholders and educators. Participants reported their own injury experiences so that we could estimate the rate of non-fatal injury, by taking person-time at risk into account, and identify potential risk factors.

    Methods

    Study population

    The study population is located along the Texas-Mexico border in Starr County (pop. 60 968 in 2010), which is largely Hispanic (>95%).12 The county has three independent public school districts (ISD), each with one high school. At the beginning of this study, Rio Grande City Consolidated ISD enrolled 59.6% of the county's students. The next largest district was Roma ISD (38.9%) and the smallest was San Isidro ISD (1.6%).13 ,14 Starr County students are primarily Hispanic (99.6%) and of low-socioeconomic status (86.0%).14

    Sampling and recruitment

    We recruited all high school students from the three ISDs who were in grades 9–12 and enrolled in an English class on a main campus during the 2003–2004 school year (N=3584). This sampling frame ensured coverage of more than 90% of the enrolled student body. Three waves of letters inviting parents to allow their children to participate along with a parent consent form were sent home. All students who submitted written parental consent, or who were 18+ years-old, and who provided their own consent electronically could participate. Students received school spirit towels for returning a signed parental consent form, whether or not their parent consented to their participation. They received a school t-shirt for agreeing to participate.

    Data collection

    Data collection took place September 2003–January 2004. The data collection instrument was a self-administered, confidential, online survey in English or Spanish. A bilingual staff member was available to answer questions. Administered during English classes, the survey duration was approximately 45 min or one class period.

    Participants answered items on demographics, work history (eg, approximate dates of employment, hours per day, days per week, type of employer, crop type and tasks), work hazards (eg, use of knives), indicators of work organisation. The occupational injury section measured the nature, body part, location (eg, field, packing shed), approximate date, crop, task, lost time from work or usual activities, and treatment or medical care.3 Additional sections pertained to health status (eg, self-reported height and weight) and health risk behaviours (eg, alcohol and tobacco use, television watching) that were based on the Centres for Disease Control Youth Risk Behaviour Surveillance System.15 The majority of items referred to a 9-month recall period (January 1–September 31) to capture the typical migration period of farmworkers with a home base in Starr County, Texas and to be able to anchor questionnaire items using New Year's Day. Participants used calendars with icons that identified major holidays to assist with recall of work details.16 The survey was constructed in English, translated into Spanish, back translated and differences adjudicated. The survey was pilot-tested with bilingual (ie, Spanish and English) high school students, interviewers and other adults in the community.

    Definitions of farm work, migrant, and work-related injury

     We defined farm work as any work that involved an aspect of food production in the USA (eg, field work, farm-based or commercial packing sheds, butchering) for pay or not for pay.3 We defined migrant farm work as work that required spending the night away from their home in Starr County. Non-migrant farm work referred to work in or near Starr County that did not require an overnight stay. It should be noted that these definitions are not those used in the National Agricultural Workers Survey (NAWS). The definition of migrant workers in NAWS is persons who travel at least 75 miles during a 12-month period to obtain a farm job.1 We asked the farmworkers to report detailed information on their most severe, acute injury event that occurred while working on a farm during the recall period. Farmworkers selected from the following injury categories: major cut or scrape, bruise or crush, sprain or strain, acute muscle spasm, puncture or jab, torn ligament or muscle, heat exhaustion, skin rash or eye problems due to pesticides, sunburn, major insect bite or sting, snake bite and other (specify). Additional items measured loss of time from work or usual activities, medical treatment, location of the injury (eg, field, farm building or shed), approximate date, body part, crop and task. Exclusions included: chronic musculoskeletal symptoms and superficial cuts, scrapes or other injuries, which did not require at least a bandage. A second tighter definition of injury (ie, most severe events that resulted in 4+ hours of lost time from work, school or usual activities or medical treatment) was also used to facilitate comparison with other studies of youth and adults.3 ,17 ,18

    Statistical analysis

    We used SAS system for Microsoft Window V.9 series and Intercooled STATA V.11.0 for all analyses. Computation of descriptive statistics included the rate of injury per 100 FTE based on person-time at risk and our injury definition. FTEs were computed based on 2000 work hours per year. We evaluated time to most severe injury using Cox regression. First, we considered potential primary determinants or strong confounders based on prior work and published literature using bivariate hazard rate ratios and 95% CIs. Variables significant at the p≤0.05 levels with the greatest point estimates and most precise CIs were considered to be the strongest variables and the basis of an intermediate model. Additional potential determinants, confounders or effect modifiers were examined next by including them individually in the intermediate model if they were significant at the p≤0.25 level. Criteria for remaining in the model were variables statistically significant at the p≤0.05 level or those variables that instituted a 15% change in the hazard rate ratio for one or more of the strongest variables.19 Next, variables formerly eliminated at the p≤0.25 level in the unadjusted models were entered independently into the model to ensure that all essential variables were included. The final step was to assess all first-order interactions. Thus, the final model included significant determinants (p≤0.05 level), confounders (those causing a 15% or more shift in the hazard rate ratio of the strongest variables), and effect modifiers (first order interaction terms significant at p≤0.05 level). We assessed the fit of the final model using Cox-Snell residuals analysis, influence/leverage analysis,20 and a global test based on Schoenfeld residuals to assess violation of the proportional hazards assumption.21

    Results

    Demographics and health behaviors

    The response rate was 83.9% (n=1247) in Roma ISD, 61.2% (n=1243) in Rio Grande City Consolidated ISD and 67.6% (n=46) in San Isidro ISD, the smallest district, for an overall response rate of 70.8%. There was not a statistically significant difference (p≤0.05) between participants and non-participants with respect to grade level, the only variable available for comparison. During the 9-month recall period, 410 did farm work. The distribution of demographic variables and selected health behaviours is shown in table 1.

    View this table:
    Table 1

    Unadjusted associations between demographics and health behaviours and work injury

    Due to unexpected problems with computer servers during data collection, records for 61 (14.9%) of the 410 farmworkers lacked some of the non-work exposure variables. Demographics and injury status for farmworkers with and without data were compared using Pearson χ2 statistics to assess potential bias. The only variable that differed significantly was gender. Approximately, 80% of farmworkers without data were male compared with 55.9% of farmworkers with data (1 df; χ2=12.3; p=0.000). The reason for this difference is unclear.

    Employment and workplace hazards

    The vast majority of farmworkers (79.8%) engaged in migrant farm work during the recall period. The remaining 20.2% only held farm jobs in or near Starr County. On average, farmworkers laboured 5.1 days/week and over 8.3 h/day for an average of 1.9 farm employers (range: 1–6). The average duration of employment was 1.7 months (9-month period). The largest proportion (36.5%) reported working for a small farm owner only while 23.6% reported working for a contractor only, 12.9% reported working for a commercial owner or grower only, and 27.0% reported working for other or a combination of employer types. Farmworkers laboured in 29 different states of which the top 10 were California, Illinois, Indiana, Michigan, Minnesota, Nebraska, North Dakota, Oklahoma, Texas and Washington. The largest proportion (48.6%) worked only within Texas (table 2). The most frequently worked crops and performed tasks are displayed in table 2.

    View this table:
    Table 2

    Description of most common work variables and unadjusted association with injury

    Analysis of time to injury event

    In total, 381 of the 410 farmworkers reported viable work records and contributed 165 796.2 person-hours of time at risk or 82.9 FTE. These farmworkers reported a total of 61 most severe injuries during the recall period. The rate of most severe injury was 73.6 per 100 FTE (95% CI 57.3 to 94.6). For those injuries that were medically attended or included ≥4 h lost work/activity time, the rate was 27.0 per 100 FTE (95% CI 18.1 to 40.3). Table 3 displays descriptions of farm work injuries by gender, nature, body part, lost time, and medical treatment.

    View this table:
    Table 3

    Description of farm work injuries

    Tables 1 and 2 display the unadjusted associations between our main exposure variables of interest and injury. Table 4 displays the adjusted Cox model. The following variables had statistically significant (p≤0.05) HRs: sleeping fewer than 8 h/night on weeknights on average (2.10), feeling tense, stressed or anxious sometimes/often (2.25), watching 0 h of television per week on average (2.65), working around ditches (2.01) and detasseling (2.70). Age was also included in the model as five categories: <15 years (5.82), 15 years (1.71), 16 years (4.47), 17 years (referent; 1.00) and ≥18 years (2.76). All graphical depictions based on the Cox-Snell residual analysis and influence/leverage analysis supported a reasonably well-fit model. Similarly, the global test based on Schoenfeld residuals suggests no violation of the proportional hazards assumption across time based on a χ2 distribution with 9° of freedom and the p value (0.67). Since data represented students from all three different public schools in Starr County, we also examined the potential impact of clustering of responses by school by introducing a term for school into the final model. We found no evidence that clustering of responses by school biased the findings.

    View this table:
    Table 4

    Final Cox model examining time to injury event

    Discussion

    A major strength of this study is the availability of person-time data at risk and estimated rate of non-fatal occupational injury among a large number of adolescent farmworkers who are Hispanic and largely engaged in migrant farm work. This is notable because the majority of published studies of non-fatal agricultural injury concentrate on farm owners and operators and their families.8 ,22–25 A few studies involved surveys of farmworkers, but these focus mostly on adults.3 ,10 ,17 ,26 ,27 Studies with adolescent agricultural workers often do not include Hispanic workers or have other limitations (eg, small sample size, lack person-time at risk).3 ,27–33

    Our estimated rate (27.0/100 FTE) for the most severe injury using the restrictive definition is similar to available studies of adolescents working in agriculture that employ similar definitions (20.8–28.1/100 FTE).18 ,28 ,30 The study with the most comparable population focuses on teenage agricultural workers from Washington State. In the study, the rate of injury among Hispanic teenagers was 20.8/100 FTE (95% CI 6.6 to 62.2).28 This and our estimate surpass the estimate (5.2/100 FTE) derived from NAWS data for hired farmworkers <20 years of age.34 NAWS data are collected at or near the worksite where respondents, fearing negative consequences, may not report injury events. This could partially explain the discrepancy.

    Using our broader injury definition, the rate increased to 73.6/100 FTE (95% CI 57.9 to 95.6), which underscores the need to collect comprehensive injury data in order to fully understand the injury experience of this vulnerable population. Six potential determinants of acute injury were identified. A non-linear pattern emerged for the rate of injury by age that was unexpected based on other studies of youth working in agriculture.35 For example, the risk of injury was lowest among 15 year-olds and 17 year-olds, but highest among those <15 years-old and those who were 16 years-old. Also unlike other studies, we did not find an elevated risk for males compared with females.8

    The association (HR=2.10) with sleeping <8 h per night during the week is consistent with studies of adolescent farm residents in Colorado36 and adolescent farmworkers in China.37 The association (HR=2.25) with general symptoms of anxiety (experienced sometimes or often) was similar to the study of adolescent farmworkers in China where the risk of injury was greater among youth who experienced family and school related stressors.37 Not watching television on a weekly basis was associated with an increased rate of injury (HR=2.65), potentially indicative of lower socio-economic status. The rate of injury was greater among those who reported working around ditches (HR=2.01), a possible surrogate for an unmeasured hazard or combination of hazards. Finally, those who reported detasseling as a job task were more likely to sustain an injury (HR=2.70). Detasseling corn involves working with or around sharp implements and irrigation systems, exposure to extreme heat and other harsh weather conditions, and maintaining hazardous postures that may increase the risk of injury.

    The use of self-reported data in this study has both strengths and limitations. We were unable to validate the reporting of injury or work patterns against medical or work records. However, our injury rates, based on a more restrictive definition, are within the range of prior studies. The reported work patterns such as crops worked, tasks performed and state of employment are similar to a prior study and suggest that our method of data collection is reliable.3 Finally, bias may arise from collecting data pertaining to the previous migration season (2002–2003) prior to the formation of the study cohort in 2003. As a result, identified determinants of injury may be predictors of injury survival and recovery rather than the occurrence of injury events. To examine this potential bias, we compared the rates in this study to those from a subsequent analysis of injury across two additional years of follow-up. Rates were similar across all 3 years signifying that this potential bias, if present, is not dramatically impacting our conclusions.

    Conclusions

    This study extends our previous work with migrant farmworker families by demonstrating that meaningful injury and person-time data can be collected directly from adolescents. The difference in magnitude of the injury rates based on the comprehensive versus restricted injury definitions illustrates that commonly used criteria for injury in occupational health research (eg, medical treatment, loss of ≥4 h of work time) may grossly underestimate its occurrence in this adolescent population. In addition, Zaloshnja, Miller and Lee38 estimated that the annual cost of non-fatal injury among hired youth in agriculture was $15.4 million in terms of medical costs, work and household productivity loss, and quality of life loss (in 2005 dollars). Long term consequences, such as disability, are especially a concern if future participation in the workforce is threatened leading to economic hardships. The high observed rates of non-fatal injury combined with their potential consequences and costs illustrates a compelling need for injury prevention efforts targeting adolescent, Hispanic, farmworkers. A variety of prevention approaches are applicable to working youth and may be tailored to this population. These include programmatic interventions designed to reduce hazardous exposures as well as evidence-based policies that are implemented, and enforced.

    What is already known on the subject

    • Farmworkers often labour intensely for long hours while being exposed to a variety of occupational hazards that influence injury risk.

    • Although the rate of occupational injury in the agricultural industry is among the highest in the nation, little is known about injury in migrant and seasonal farmworkers in general and adolescents specifically.

    • National surveillance systems that are currently in place do not capture adequately injury among migrant and seasonal farmworkers.

    What this study adds

    • Estimated rates of non-fatal occupational injury among adolescent migrant and seasonal farmworkers that are based on a large sample size and person-time at risk.

    • An illustration of how standard definitions may not be sufficient for understanding the burden of occupational injury in this population. By applying a definition that is more appropriate for vulnerable, Hispanic, adolescent farmworkers, the rate of occupational injury was nearly three times the rate based solely on the standard definitions often used in research and practice.

    • Evidence that sleep and stress also may contribute to occupational injury in this population.

    Acknowledgments

    We thank our participants, RGCCISD, Roma ISD, San Isidro ISD, Yolanda Morado (Starr County AgriLife Extension), Ricardo Reyna, Jesus Guerra, Luis Garza, Noe Guzman, Abigail Ortega, Roel Gonzalez, Maria Elena Olivarez, Freddy Alvarez, Mike Garcia, Ana Garcia, Mario Alvarado, Barbara Adams MPH, advisory board members and survey administrators.

    References

    1. ↵
      1. Carroll D,
      2. Samardick RM,
      3. Bernard S,
      4. et al
      . Findings from the national agricultural workers survey (NAWS) 2001–2002: a demographic and employment profile of united states farm workers. Washtington, D.C: US DOL Office of the Assistant Secretary for Policy Office of Programmatic Policy, 2005. Report No.: 9.
    2. ↵
      US General Accounting Office. Child labor in agriculture: changes needed to better protect health and educational opportunities. Washington DC: US General Accounting Office, 1998. Report No.: GAO/HEHS-98-193.
    3. ↵
      1. Cooper S,
      2. Burau K,
      3. Frankowski R,
      4. et al
      . A cohort study of injuries in migrant farm worker families in south texas. Ann Epidemiol 2006;16:313.
      OpenUrlCrossRefPubMed
    4. ↵
      1. Perry M
      . Children's agricultural health: traumatic injuries and hazardous inorganic exposures. J Rural Health 2003;19:269.
      OpenUrlPubMedWeb of Science
    5. ↵
      1. Miller ME
      . Historical background of the child labor regulations: strengths and limitations of the agricultural hazardous occupations orders. J Agromed 2012;17:163–85.
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      Occupational injuries and deaths among younger workers–united states, 1998–2007. MMWR Morb Mortal Wkly Rep 2010;59:449.
      OpenUrlPubMed
    7. ↵
      1. Cooper SR,
      2. Cooper S,
      3. Felknor S,
      4. et al
      . Nontraditional work factors in farmworker adolescent populations: Implications for health research and interventions. Public Health Rep 2005;120:622–9.
      OpenUrlPubMedWeb of Science
    8. ↵
      1. McCurdy SA,
      2. Carroll DJ
      . Agricultural injury. Am J Ind Med 2000;38:463–80.
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Runyan CW,
      2. Zakocs RC
      . Epidemiology and prevention of injuries among adolescent workers in the united states. Annu Rev Public Health 2000;21:247.
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Earle-Richardson G,
      2. Brower M,
      3. Jones A,
      4. et al
      . Estimating the occupational morbidity for migrant and seasonal farmworkers in new york state: a comparison of two methods. Ann Epidemiol 2008;18(1):1.
      OpenUrlCrossRefPubMed
    11. ↵
      1. Leigh JP,
      2. McCurdy SA,
      3. Schenker MB
      . Costs of occupational injuries in agriculture. Public Health Rep 2001;116:235–48.
      OpenUrlPubMedWeb of Science
    12. ↵
      State and county QuickFacts: Starr county, TX (Internet).; 16 August 2010. http://quickfacts.census.gov/qfd/states/48/48427.html (accessed 19 December 2010).
    13. ↵
      Texas Education Agency. Disaggregation of PEIMS student data: Rio Grande City CISD. 2010 March 20, 2009.
    14. ↵
      Texas Education Agency. Region one education service center disaggregation of PEIMS data-student as of october 31, 2003. 2004.
    15. ↵
      1. Brener N,
      2. Kann L,
      3. Kinchen S,
      4. et al
      . Methodology of the youth risk behavior surveillance system. Morb Mortal Wkly Rep. Recommendations and reports 2004;53:1–13.
      OpenUrlPubMed
    16. ↵
      1. Zahm SH,
      2. Colt JS,
      3. Engel LS,
      4. et al
      . Development of a life events/icon calendar questionnaire to ascertain occupational histories and other characteristics of migrant farmworkers. Am J Ind Med 2001;40:490–501.
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. McCurdy S,
      2. Samuels S,
      3. Carroll D,
      4. et al
      . Agricultural injury in california migrant hispanic farm workers. Am J Ind Med 2003;44:225.
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Munshi K,
      2. Parker D,
      3. Bannerman-Thompson H,
      4. et al
      . Causes, nature, and outcomes of work-related injuries to adolescents working at farm and non-farm jobs in rural minnesota. Am J Ind Med 2002;42:142.
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Rothman KJ,
      2. Greenland S,
      3. Lash TL
      . Modern epidemiology. 3rd edn. Philadelphia: Wolters Kluwer & Lippincott Williams & Wilkins, 2008.
    20. ↵
      1. Klein JP
      . MML. Survival analysis: techniques for censored and truncated data. New York: Springer-Verlag, 1997.
    21. ↵
      STATACorp LP. Survival analysis and epidemiological tables reference manual release 11. College Station, TX: STATA Press, 2009.
    22. ↵
      1. Atrubin D,
      2. Wilkins JR,
      3. Crawford JM,
      4. et al
      . Self-reported symptoms of neurotoxicity and agricultural injuries among ohio cash-grain farmers. Am J Ind Med 2005;47:538.
      OpenUrlCrossRefPubMed
    23. ↵
      1. Layne LA,
      2. Goldcamp EM,
      3. Myers JR,
      4. et al
      . Youth living on hispanic-operated farms: Injuries and population estimates in the U.S., 2000. J Agric Saf Health 2009;15:377.
      OpenUrlPubMed
    24. ↵
      1. Marlenga B,
      2. Berg R,
      3. Linneman J,
      4. et al
      . Changing the child labor laws for agriculture: impact on injury. Am J Public Health 2007;97:276–82.
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Sprince NL,
      2. Zwerling C,
      3. Lynch CF,
      4. et al
      . Risk factors for agricultural injury: a case-control analysis of iowa farmers in the agricultural health study. J Agric Saf Health 2003;9:5–18.
      OpenUrlCrossRefPubMed
    26. ↵
      1. Earle Richardson G,
      2. Jenkins P,
      3. Scott E,
      4. et al
      . Improving agricultural injury surveillance: a comparison of incidence and type of injury event among three data sources. Am J Ind Med 2011;54:586–96.
      OpenUrlCrossRefPubMed
    27. ↵
      1. Brower M,
      2. Earle-Richardson G,
      3. May J,
      4. et al
      . Occupational injury and treatment patterns of migrant and seasonal farmworkers. J Agromed 2009;14:172.
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Bonauto DK,
      2. Keifer M,
      3. Rivara FP,
      4. et al
      . A community-based telephone survey of work and injuries in teenage agricultural workers. J Agric Saf Health 2003;9:303.
      OpenUrlPubMed
    29. ↵
      1. Vela Acosta MS,
      2. Sanderson M,
      3. Cooper SP,
      4. et al
      . . Health risk behaviors and work injury among hispanic adolescents and farmworkers. J Agric Saf Health 2007;13:117.
      OpenUrlPubMed
    30. ↵
      1. Chapman LJ,
      2. Newenhouse AC,
      3. Meyer RH,
      4. et al
      . Musculoskeletal discomfort, injuries, and tasks accomplished by children and adolescents in wisconsin fresh market vegetable production. J Agric Saf Health 2003;9:91–105.
      OpenUrlPubMed
    31. ↵
      1. Chapman L,
      2. Taveira A,
      3. Karsh B,
      4. et al
      . Work exposures, injuries, and musculoskeletal discomfort among children and adolescents in dairy farming. J Agromed 2009;14:9–21.
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. McCurdy S,
      2. Samuels S,
      3. Carroll D,
      4. et al
      . Injury risks in children of California migrant hispanic farm worker families. Am J Ind Med 2002;42:124.
      OpenUrlCrossRefPubMed
    33. ↵
      1. Wilkins JR,
      2. Crawford JM,
      3. Stallones L,
      4. et al
      . Using participant event monitoring in a cohort study of unintentional injuries among children and adolescents. Am J Public Health 2007;97:283.
      OpenUrlCrossRefPubMedWeb of Science
    34. ↵
      1. Wang S,
      2. Myers J,
      3. Layne L
      . Injuries to hired crop workers in the united states: a descriptive analysis of a national probability survey. Am J Ind Med 2011;54:734–47.
      OpenUrlCrossRefPubMed
    35. ↵
      1. Schulman MD,
      2. Evensen CT,
      3. Runyan CW,
      4. et al
      . Farm work is dangerous for teens: Agricultural hazards and injuries among north carolina teens. J Rural Health 1997;13:295–305.
      OpenUrlPubMedWeb of Science
    36. ↵
      1. Stallones L,
      2. Beseler C,
      3. Chen P
      . Sleep patterns and risk of injury among adolescent farm residents. Am J Prev Med 2006;30:300.
      OpenUrlCrossRefPubMedWeb of Science
    37. ↵
      1. Postel MW,
      2. Jaung MS,
      3. Chen G,
      4. et al
      . Farm work-related injury among middle school students in rural china. J Agric Saf Health 2009;15:129.
      OpenUrlPubMed
    38. ↵
      1. Zaloshnja E,
      2. Miller T,
      3. Lee B
      . Incidence and cost of nonfatal farm youth injury, united states, 2001–2006. J Agromed 2011;16:6–18.
      OpenUrlPubMedWeb of Science

    Footnotes

    • Contributors SPC, DJDJ and EMS conceived and designed the project. SPC, EMS, CJC and REW were responsible for data collection. CJC and REW were responsible for data management and quality assurance and quality control measures. SPC, DJDJ, EMS, CJC and REW were involved in the data analysis and interpretation of results and development of the manuscript, its review and approval of the final draft.

    • Funding This work was supported by Cooperative Agreement No. U50 OH07541 to the Southwest Center for Agricultural Safety and Health at The University of Texas HSC at Tyler from CDC/NIOSH, NIH/NCRR Grant No. UL1 RR024148, and Southwest Center for Occupational and Environmental Health, a NIOSH-funded Education and Research Center Grant No. T42OH008421. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of CDC/NIOSH or NIH.

    • Competing interests None.

    • Patient consent Obtained.

    • Ethics approval This study was approved by the Institutional Review Board at Texas A&M University (protocol #2002-250) and the Committee for the Protection of Human Subjects at The University of Texas Health Science Center at Houston (protocol #HSC-SPH-07-0495).

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data sharing statement We are currently analysing the data collected from 2004–2006 during the follow-up surveys in Roma, Texas.