The publication of Agriculture at Risk: A Report to the Nation in 1989 was the result of an organized effort by a coalition of concerned health and safety professionals to raise awareness among legislators and the public about the high rate of occupational injury and illness in agriculture. This document outlined research and prevention goals for agricultural health and safety, and gave rise to ten NIOSH-sponsored agricultural health and safety research centers around the country. Before that time, OSHA and a handful of other federal agencies had attempted to take action in three areas of agriculture: agricultural injury prevention, pesticide management and migrant farmworker health, but with mixed results (M. Myers, 1992). The publication of Agriculture At Risk: A Report to the Nation represented the first national, coordinated effort to address occupational safety in agriculture.
When this effort began, the focal point was production agriculture as a whole. The main message was that agriculture is one of the most dangerous industries in America, and is the only one that federal and state government was taking almost no action to protect. The bulk of the data presented related to white farm owners and their families, and had little to say about other races, ethnicities or occupations within agriculture.
While minority research in agricultural health and safety is limited, national data on agricultural fatality by race suggests that black workers are at higher risk for occupational fatality than whites (Hard et al., 1999; Kraus, Lightstone, & McArthur, 1998; J. Myers & Hard, 1995). However, when stratified by race and occupation within agriculture, fatality patterns are less clear. J. Myers and colleagues find white farmers to be at substantially higher risk (1995), while later work by the same NIOSH group presents data indicating excess risk among black farmers (Hard et al., 1999).
Data from occupational research in several other industries also suggest that minorities are at increased risk of occupational injury and illness (Loomis & Richardson, 1998; Wright, 1992; Robinson, 1989). While currently there are no definitive answers, there is enough data suggestive of increased occupational risk among minority workers to warrant further study.
In 1996 NIOSH published its National Occupational Research Agenda (NORA), and included as one of its priority research areas, "special populations at risk," defined as, "workers with specific biologic, social, and/or economic characteristics who are more likely to have increased risks of work-related diseases and injuries." In particular, the document named children, women, the elderly and minority workers as special populations. This was a milestone, since it officially named minority research as a priority area for future funding.
1990 appears to have been a turning point in the publication of research related to minorities in agriculture. From 1987 through 1990, a total of thirteen articles were published, whereas in the four years following, the rate more than doubled to 41 published articles. That trend continues, with roughly eight articles published each year.
In addition, data from the NIOSH-supported agricultural health and safety centers show steady increases in center activity devoted to minority occupational health. In a survey, the Centers were asked to estimate the number of studies and prevention programs ongoing in 1990 (or the year the center was established if after 1990) and again in 2000. The most frequent response for the baseline year was 1-2 projects in research and also 1-2 projects in prevention. By 2000, the most common response was 3-5 projects for both. Similarly, the most common comment received by the survey was that minority research and prevention has risen in importance among the Center's priorities. At the same time, most respondents commented on the inherent difficulty of carrying out research in this area. This comment was typical:
A major challenge is the temporary nature of employment either due to migration, or due to the workers actually settling in an area to do seasonal crop work during planting and harvesting, while at other times working in non-agricultural jobs. An equally major challenge in studying, or serving minority workers is that they are difficult to access due what seems to be a perception of adverse consequences from divulging information, or from just being observed (Kennedy, 2001).

However, five years after NORA, published health and safety research on minorities in agriculture is still scarce compared to the quantity of data that exists on white farm populations. If each racial or ethnic minority group in agriculture is considered separately, then the imbalance in the quantity of available research with respect to white farm owners is even more extreme.
The research that does exist concentrates on a few populations in a few locations, such as Hispanic migrant and seasonal farmworkers in the West and Southeast, and African-American farmers in the South. Certainly more research on all aspects of minority agricultural health and safety are needed, especially among groups as yet unstudied.
In considering current and future research and prevention, it is important to look at more than just the quantity of studies, and the groups represented, but also at what questions the research seeks to answer, and how data are collected. Ultimately, the manner in which research is conceptualized, conducted, and interpreted in minority occupational health and safety will determine whether these efforts actually result in improved health and safety of minority workers.
Overview Table I shows the quantity and type of studies published on various minority worker populations in agriculture. A second table (Table II) was created to summarize the unpublished outreach and education activities undertaken by the NIOSH-supported Agricultural Health and Safety Research Centers in each region. Table II is not intended to represent all education and outreach efforts being undertaken for minority agricultural workers, but simply to provide examples of current activities, organized by group, from around the country.
In reviewing Table I, some generalizations about agricultural health and safety research on minority populations can be made. First, the bulk of the research concentrates on Hispanic migrant and seasonal farmworkers in California and in the Southeast, with the Hispanic farmworkers in the Northeast a distant third. Speaking very generally, research in these three leading regions is similar in terms of study question type and methods. All three look at injury, communicable disease, environmental exposures and social factors. In addition, Hispanic farmworker research in both California and the Southeast includes general health assessments. In California we also find evaluations of ergonomic interventions.
In terms of findings, the table shows several elevated prevalence rates for Hispanics, African-Americans, and Caribbeans in communicable disease (Ciesielski, Seed, Ortiz, & Metts, 1992; Ciesielski, Seed, Esposito, & Hunter, 1991; Ciesieleski, Handzel, & Sobsey, 1991; Hibbs,Yeager, & Cochran, 1989; Jacobson, Mercer, Miller, & Simpson, 1987). In terms of fatal injury, the data are inconclusive. Some data show white farm owners at highest risk for death (J. Meyers & Hard, 1995), whereas other research suggests that African-American and Hispanic farmers and farmworkers have higher fatal injury risk than white farmers (Hard et al., 1999; Stallones, 1999). In addition, several common health problems are identified for many of the racial/ethnic populations, but they are not in rates, which lend themselves to comparison with the general population. There are also a wide variety of qualitative descriptive findings presented throughout the research.
Because of the diversity of the outcomes considered by each study, even within the groupings of race/ethnicity, occupation, and region, their results cannot be combined to paint a clear picture of the overall occupational health of the populations within any of these strata. This may simply be a function of the field being relatively young.
The following discussion of published studies will focus on how effectively researchers have addressed each of four key areas: a) the definition of the study population; b) definition of the "occupationally healthy" worker; c) measures of occupational health; and d) identification of possible causal factors.
a) definition of the study population In any research, careful consideration must be given to how the study population is to be defined. Traditionally, researchers looking at minorities in all fields have tended to group minorities into broad categories such as "African-American" or "Hispanic," without any thought to whether these labels actually constitute a group in any meaningful way. Unless one is to argue that the tint of a worker's skin or his eye color is causally related to his risk of occupational injury or illness, other factors than these must be considered in the design of studies aimed at establishing causal relationships between risk factors and morbidity. Certainly, the tasks the worker is undertaking will play a large role in risk. So will cultural and social factors associated with ethnic or racial group membership, but only if that group actually functions as a group or is viewed as a distinct group in society.
In addition, the geographic location of a group is an important and often overlooked factor, since the political and historical influences on minority occupational safety will certainly vary in different regions of the country. In reviewing and presenting the existing literature, the authors grouped published research along the three dimensions mentioned above: race/ethnicity, job type, and geographic location. While this is not intended as a hard and fast rule of group definition, each group should be defined following some explicit rationale. Clear and defensible definitions of study populations are particularly important in minority research, since minority health research has at various times in the past been appropriated for racist purposes in an attempt to show white superiority (Krieger, 1987).
In the case of minority farm owners, research tends to focus on specific groups using the race/ethnicity, occupation, and location definition advocated in this review. Arcury (1995) describes African-American farmers in detail, looking at, for example, geography, farm size, commodity, and history of farm ownership. Describing the study population in detail may be easier to do with some farm populations than others, since not all workers are permanent, legal residents, with distinct racial identities, as this population is.
For example, in studies looking at migrant and or seasonal farmworkers, definitions are often incomplete or extremely broad. This becomes problematic when attempting to interpret results. Researchers need to state precisely who is represented in the results, and what work task, cultural, political or social conditions may be associated with the findings.
The community studies by Griffith (1995a) are very detailed and descriptive, but they define groups by geography (city) mainly, and then make inferences about migrant and seasonal workers, under the assumption that they make up a large proportion of the population. The authors attempt to determine racial/ethnic percentages of their study populations, but do not present results stratified along racial/ethnic lines.
Many studies (Hard et al., 1999; Arcury & Quandt, 1998b; Crandall, 1997; J. Myers & Hard, 1995) do not specify exactly what is meant by the term, "farm worker". It is not clear whether this group includes temporary harvest labor, family members or full-time, paid employees. These are very distinct groups which should be identified.
One important aspect of group definition relates to not only naming the occupation of the group within agriculture, but in identifying as closely as possible the actual job tasks undertaken, since they are the fundamental sources of injury and illness risk. Isaacs, (1995) gives a very informative listing of major job tasks of migrant and seasonal farmworkers by commodity. Ergonomic studies (Meyers, 1999; Miles, 1996) contain highly detailed descriptions of work task. Hopefully, subsequent ergonomic research will use this detailed information to correlate certain occupational injuries or illnesses with certain commodities and tasks. This type of task description would also be helpful in other types of studies.
Several studies look at migrant and seasonal farmworkers as an occupational group without selecting one ethnicity/race as the main population of interest (Garcia, 1996; Earle-Richardson, 1998; Chi, 1991a-d; Jones, 1991; McDermott, 1990). This may be the result of the research interest being mainly on all workers doing a certain type of work, (such as "migrant farmworker", or "grape worker" in the case of Samet (1992) and Gamsky, Schenker, McCurdy, & Samuels (1992)), or with the idea of making cross-group comparisons.
The choice may be made to simply group by job category, discounting race/ethnicity as a major factor in occupational health, which may or may not be appropriate. It may be best used as a first exploratory step to determine if any health effects may be present. If indicated, this can then be followed later by more detailed analysis by group. One might certainly argue that in some cases race/ethnicity does not define the group nor its occupational risk.
While it is legitimate to collect data on several different identified ethnic/racial groups with the idea of making comparisons between them, there are some risks in undertaking multiple group research. Some studies employing this methodology lacked adequate statistical power due to limited sample size within certain ethnic strata (Ford, 2000).
While comparative data is an important aspect of meaningful research, limiting the comparison groups to one or two, and/or oversampling the main group of focus may be necessary in the face of financial and logistical constraints. In addition, there may be other benefits to studying one group extensively. For example, when researchers acquire an in-depth knowledge of the cultural history and sociological factors related to a particular group, they are better able to conduct research and interpret the findings related to that group. This is typically best accomplished by the assembly of a multidisciplinary team.
In reviewing the agricultural health and safety literature, it was found that most studies addressed one of two broad categories of workers: a) minority farm owner-operators and b) migrant and seasonal farmworkers. It is likely that more occupational categories exist (such as farm manager, full-time employee, or crew chief), but as yet these do not appear in the research literature to any great extent. Certainly other worker types exist (e.g., Mexican dairy workers, chicken processing workers, and year-round orchard workers). Further research on these groups is needed.
In sum, much of the current minority health and safety research in agriculture could benefit from a more complete and precise definition of the study population. The reviewers advocate the use of a study population definition based on race/ethnicity, occupation, and geographic location. This will allow for more meaningful analyses and interpretation of study results.
b) defining the "occupationally healthy" worker The next important step is to define what occupational health actually means for a given population under study. The identification and prioritization of attributes of a healthy worker is important because it does two things. First it identifies the priority areas for research, and second, it provides some kind of yardstick by which to measure study results.
For example, consider the Mexican migrant farmworker who picks ground crops in Southern California, perhaps travels north to Washington for the apple harvest, and then goes back to Mexico at the end of the harvest season. How does one conceptualize this individual at his optimum occupational health? Would occupational health dictate that he be disease and injury-free, or just not beyond a certain level or frequency? Must fields be free of pesticides, or is the use of protective gear adequate? Similar questions must be answered with regard to behavioral risk factors such as those for chronic disease and poor mental health.
One must also consider social and legal factors which may affect the migrant's overall well-being. For example, does the worker have his own car to travel, and all the legal documentation he needs to safely and easily migrate? Does he have long-term relationships with employers, for job security? Depending on one's point of view, some or all are important indicators of occupational health.
Moreover, one must consider how the definition of occupational health relates to the individual's health status on entering the occupation: Should a worker with initially poor health get better, or just not get worse? Some might even take the position that good health is totally incompatible with stoop labor. The authors do not advocate for one definition over another, however, it seems clear that one cannot effectively research problems in occupational health and safety, without first defining occupational health. The issue of definition takes on special importance in the case of minorities, since minority workers (particularly immigrants) frequently undertake unpleasant and difficult work, and because many minority workers enter the workforce in poor health.
This broad issue of what constitutes good occupational health and safety must be carefully applied to each individual research question that is asked. Unfortunately, this was not found to be the case among the articles reviewed. Rather, this definition must be deduced from the research question. For example, Arcury (1995) makes a connection between farmers' views of work-related hazards and injury risk. We then must deduce that his definition of the occupationally healthy worker includes the belief that threats of injury on the farm are high. In other injury studies, it can be inferred that the injury-free worker is the occupational health gold standard (Ford, 2000; Lyman, 1999; Ciesielski, Hall, & Sweeney, 1991). Perhaps many published researchers have considered this question in developing their research, but it would clarify the purpose and meaning of the study if it were discussed explicitly in the results.
If one is comfortable inferring definitions of good occupational health from research questions, one sees a wide variety of definitions in the published minority agricultural health and safety literature: being disease-free (Jones, 1991), having social support (Schoonover Smith, 1987), taking precautions against pesticide exposure (Arcury, 1998), getting cancer screening (Lantz, 1994), and even having the ability to promote ones own health (Kerr, 1990).
One problem that these widely varying definitions illustrate is that individual studies focus on one problem without placing it in the context of overall occupational health for the study population. This fragmented view is evident when looking at the research as a whole.
c) measures of occupational health. Once some kind of "gold-standard" has been constructed, measures of good occupational health can be easily identified. While many of the studies reported traditional measures of morbidity and mortality, such as incidence and prevalence, some other indicators were encountered, which demonstrate varying definitions of good occupational health among researchers. Some examples are: worker awareness of danger in agriculture, existence of physical hazards, presence of safety programs, farmer behavior, such as alcohol use and fatigue, seroprevalence of various infectious diseases, degrees of emotional support, prevalence of violence, subject's ability to promote one's own health, willingness to have medical tests, degree of fear of cancer, and the belief in survivability of cancer. It is also worth noting that the majority of these measures relate to individual behavior, rather than environmental, social or legal determinants of health.
Clearly the wide divergence in occupational health measurements springs from the different conceptualizations of good occupational health. With greater consistency among researchers addressing a.) and b.), there would be greater consistency among occupational health measures, and more opportunity for comparison between studies.
d) Identification of possible causal factors Before any measurements are taken and analyzed, a wide variety of possible causal factors must be considered. Among them are: the physical and social work environment, the political context of the job, cultural factors, and measurements, if available, on white or other minority groups doing these same jobs.
Much of the research seems to focus on identifying and quantifying poor outcomes rather than considering the reasons why. Again, this may simply be an indicator of a young field of study. Research by Griffith (1995a), while not strictly occupational health-related, has been included in this review as an example of the sociological approach which could provide the needed contextual data to assist in interpreting the meaning of current minority research. A number of complex social and economic phenomena are described and explained in this work.
One example of how this more sociological approach of Griffith (1995a) could benefit current health and safety research is his finding that men traveling alone are increasing throughout the migrant workforce. This finding is not in itself an occupational health outcome, however, the fact that increasing numbers of males are working alone in the migrant and seasonal workforce is directly related to farmworker health and safety. As he explains, the forces that are creating this phenomenon also reinforce the entrenched powerlessness that is associated with poor working conditions and poor health outcomes:
... the separation of productive [farmworkers] from reproductive labor [non-working families and dependents] is so extreme that the costs of reproducing and maintaining workers during youth and old age are borne in separate geographical regions and often separate political entities. …[this] model undermines worker's attempts at assimilation and the power assimilation brings. (p. 280)

If U.S. laws and labor dynamics force men to travel without families, and never settle near their place of employment, then they will forever be disconnected from American society, and will not receive the benefits of social services, preventive health care, or labor union membership, which are all things that might lead to better health. It is this kind of contextual information that allows for more informed interpretation of health outcome data.
In looking at minorities at risk for occupational injury and illness in agriculture, two common but distinct scenarios appear. In the first, dangerous (or undesirable) jobs are filled by minority workers, and in the second, minorities and whites fill the same jobs, but something about being a minority increases the risk (Frumkin, Walker, & Friedman-Jimenez, 1999; Loomis & Richardson, 1998). It is important to make this distinction when thinking about the causes of poor occupational health outcomes among minorities.
Migrant and seasonal farmworkers are an example of the first scenario. Demographic shifts over the last two decades have resulted in this temporary labor force becoming almost entirely minority. However, unlike other sectors, it may be that it is the strenuousness, unpleasantness and low status of manual farm labor (versus farm management or ownership) that is relegating it to minority and immigrant work, rather than an inherent increased risk of fatality. At least one study has found the risk of occupational fatality to be higher among white farm owners than among minority farm workers (J. Myers et al., 1995).
An example of greater risks experienced by minorities holding the same jobs as whites is that of minority owned and operated farms. Some data suggests that black farm owners have higher injury and fatality rates than white or Hispanic owners. More research is needed to determine what social factors may account for these differences.
Many of the studies reviewed find that the study populations have elevated rates of some adverse health outcomes, but they do not address the question of what aspects of group membership may account for these differences. While one might argue that that this question can legitimately be left for subsequent research, it is recommend that hypotheses be put forward, even speculative ones, as to what may be causing poorer outcomes for a particular group. As in any field of science, these theories are critical to designing coherent research.
For example, the one study of Native Americans and farm fatality (Crandall, 1997) concludes that Native Americans have higher farm fatality rates than whites. But it is not evident to the reader whether all Native Americans have the same elevated risk, or whether different tribes might differ, or on the other hand whether Native Americans as a whole might be part of some larger group of workers who all experience higher risk. If we had some theories as to what puts these Native Americans in New Mexico at risk, it might be easier to interpret these results.
Another illustration of how the definition of a study population may be influenced by theories of causation is Stiles' (1999) study of culture and driver knowledge. In the introduction, she focuses on the driving behavior of Mexican migrant workers, citing excess motor vehicle crashes among Hispanics and also the assertion made by California law enforcement that driver performance among Mexican migrant workers is poor. This is believed to be due in to a lack of familiarity among this group with California traffic laws, particularly those related to the use of seatbelts and child safety seats. Clearly, this definition of study population is driven by the tentative assessment of what is causing the high rate of motor vehicle crashes.
Published literature reviews, policy papers, and commentary Despite the fact that published research is scarce, there does exist a large number of policy and commentary papers on different aspects of minority agriculture and safety; 17 papers on migrant and seasonal farmworkers alone were located. Seven literature reviews on migrant and seasonal farmworkers were also identified. One that this large number of policy articles and literature reviews may be the result of a growing interest in the poor working conditions of migrant and seasonal farmworkers coupled with frustration over the difficulty of conducting research on this group.
Conclusion and recommendations Over the last ten years the quantity and diversity of research in minority agriculture and health has expanded. However, the volume of research that provides incidence or prevalence data has not grown as steadily, possibly due to the difficulty inherent in enumerating and studying these often invisible and mobile populations.
First, more clarity with respect what group is being studied would strengthen minority research. This can be achieved through more detailed and careful definitions of the study population in the research. The authors suggest a definition of study population based on three factors: ethnicity/race, work task (or job) and geographic location. Similarly, the research would be more useful if studies regularly included detailed descriptions of subject's work tasks. Research leading to a job task classification system that can be used in all of agricultural safety and health research is also needed.
In addition to defining the study population and the work they do more completely, minority occupational research in agriculture would benefit greatly from a clearer definition of the occupationally healthy minority worker. Although it is a subjective construct, it is needed for research and prevention to have any direction toward healthier workers. This definition should also include some process of prioritizing issues of concern. The current literature contains a wide variety of topics, which may or may not reflect the most pressing health issues of the populations.
The existing research appears to center around a few groups - many groups are not represented at all. Similarly, there is anecdotal evidence that other job categories of minority agricultural workers exist that may be at high risk for occupational injury and illness. These categories (e.g., Mexican dairy workers, chicken processing workers, and year-round orchard worker) should be more clearly identified and considered for further study.
To move from documenting poor health outcomes to understanding why, there is a need for a multi-disciplinary approach to research. Quantitative and qualitative methods should be combined in order to have prevalence data and descriptive data together. The body of literature could benefit from more sociological research into what social factors may be contributing to poor outcomes and why. More basic demographic research on minority populations would be useful.
Much of the research and discussion in policy papers consider minority groups nation wide without any consideration of region. Caution must be used when generalizing findings outside of the region and group of study, certainly until researchers have a better understanding of the most important contextual factors in determining occupational health and safety among minority populations.
On a more methodological note, communication and collaboration within the agricultural health and safety community could be improved by developing a better system for cataloguing published and unpublished agricultural research and prevention activities. Currently, there is no one repository of agricultural health and safety research, let alone a source dedicated to such research for minorities. There are only a few journals dedicated to health and safety issues in agriculture. The remainder of the published research is spread throughout journals of occupational health, public health, industrial hygiene, industrial medicine, epidemiology, education, sociology and anthropology, requiring researchers to search a large number of literature databases. Furthermore, a great deal of relevant information tends to be in government reports or academic holdings and do not appear in commercial databases.
Research and prevention in the area of agricultural health and safety and minorities has grown significantly over the last ten years. With the support of NORA and the advocacy of committed researchers and safety specialists, a wide range of questions are now being asked relating not only to injury and disease, but also to issues of equity, minority representation and perceived well-being of the workers themselves. This represents a maturing of the field, and brings us to a point where new, more challenging questions must be addressed, such as what constitutes a minority group, and how we define occupational health. The authors recognize the enormous challenges presented by the recommendations presented above for all researchers in the field. However, we believe these steps are necessary if minority workers in agriculture are to realize the benefits of this research with improvements the quality of their own lives.
Methodological Note: the authors are part of the Northeast Center for Agricultural and Occupational Health, one of 10 centers around the country funded by NIOSH to study and promote occupational health and safety in agriculture. Consequently, our perspective on the state of agricultural health and safety is heavily influenced by the work of this center and the other nine centers. Due to the accessibility of data from these centers, their work is discussed whereas the work of many other institutions with whom the Northeast Center is not affiliated is not. The authors are certain that many unpublished activities are occurring around the country worthy of mention, however, the agricultural health and safety centers as a group span the country and are featured here only as an example of activities undertaken by a variety of local, state and federally-supported institutions.

The authors wish to gratefully acknowledge the directors and staff of the following institutions for their assistance providing information about past and current minority agricultural health and safety activities: the Southeast Center for Agricultural Health & Injury Prevention, Lexington, KY; the Deep-South Agricultural Health and Safety Center, Tampa, FL; the National Farm Medicine Center, Marshfield, WI; the Southwest Center for Agricultural Health, Tyler, TX; the Great Plains Center for Agricultural Health, Iowa City, Iowa; the High Plains Intermountain Center for Agricultural Health and Safety, Fort Collins, Colorado; the UC Agricultural Health & Safety Center, Davis, California; the Pacific Northwest Agricultural Safety and Health Center, Seattle, Washington; the California Institute for Rural Studies, and the National Children's Center for Rural and Agricultural Health and Safety, Marshfield, WI.