AMERICAN JOURNAL OF INDUSTRIAL MEDICINE 57:1053–1063 (2014)
Evaluation of a Community-Based Participatory Farmworker Eye Health Intervention in the “Black Dirt” Region of New York State Giulia Earle-Richardson, PhD, MPH,1 Lynae Wyckoff, MA,1,2 Marilyn Carrasquillo,3 Melissa Scribani, MPH,2 Paul Jenkins, PhD,2 and John May, MD1,2
Background Eye irritation is a constant hazard for migrant and seasonal farmworkers, but there are few studies of the problem or how to address it. Researchers evaluated the effect of a community-based participatory eye health intervention on farmworker eye symptoms in the Hudson Valley, NY. Methods A randomized pre-post intervention with 2, 4-week follow-up periods was implemented with a sample of 97 farmworkers. Five eye symptoms were measured, along with utilization of protective eyewear and eye drops. Results Leading baseline eye symptoms were redness (49%), blurred vision (43%), itching (43%), and eye pain (29%). Significant reductions in eye pain (P ¼ 0.009), and nonsignificant reductions in redness were observed for the intervention group while controls experienced increases in both. Conclusions The intervention was effective in significantly reducing eye pain, and to a lesser extent, redness. Future eyewear promotion programs should offer a range of eye wear, tailor offerings to local climate and tasks, evaluate eyewear durability, and include eye drops. Am. J. Ind. Med. 57:1053–1063, 2014. ß 2014 Wiley Periodicals, Inc. KEY WORDS: eye health; migrant; agriculture; farmworker; occupational; community-based participatory research; participatory intervention development; evaluation
INTRODUCTION Migrant and seasonal farmworkers work long hours outdoors, often in extremely dusty, moldy or windy conditions, and eye irritation is a constant hazard. Exposure
1 New York Center for Agricultural Medicine and Health, Bassett Healthcare Network Bassett Healthcare Network, Cooperstown, New York 2 Bassett Research Institute, Bassett Healthcare Network, Cooperstown, New York 3 Hudson River Health Care, Alamo Health Center, Goshen, New York Correspondence to: Giulia Earle-Richardson, PhD, MPH, New York Center forAgricultural Medicine and Health,Bassett Healthcare Network Bassett Healthcare Network,One Atwell Rd., Cooperstown, NY13326. E-mail: [email protected]
Accepted 30 April 2014 DOI10.1002/ajim.22352. Published online 24 July 2014 in Wiley Online Library (wileyonlinelibrary.com).
ß 2014 Wiley Periodicals, Inc.
to ultraviolet light from the sun, agricultural chemicals and irritating plant materials can also damage the eyes. The relationship between sun, dust, and eye irritation has been established in connection with several common eye conditions, such as cataracts [Michael, 2000], pterygium [Nemesure et al., 2008; Maharshak and Avisar, 2009; West and Muñoz, 2009], and others [Lee et al., 1994; Longstreth et al., 1998; Islam et al., 2000; Lucas, 2011]. Wind has also been implicated in eye irritation from chemicals (fertilizers and pesticides) [Cameron et al., 2006; Behrens et al., 2012], and in eye micro-trauma from plant material traveling at high speed in windy conditions [Chang et al., 2002; Hwang et al., 2002; Villarejo and McCurdy, 2008]. Dust-related eye irritation has also been documented in a number of other occupations [Vedal et al., 1986; Melbostad and Eduard, 2001; Van Kampen et al., 2012]. To date, there are few studies of eye disease and nontrauma related conditions among migrant and seasonal
1054
Earle-Richardson et al.
farmworkers. What research there is suggests that eye pain and disease are common. For example, in one study of a predominantly Mexican cohort of 197 workers in North Carolina, eye pain (40%), redness (43%), itching (25%), and blurred vision (13%) were all reported [Quandt et al., 2001]. Other studies have found similar symptom prevalence [Cameron et al., 2006; Luque et al., 2012]. In another study, nearly one-quarter of workers (23%) were found to have pterygium, [Taylor et al., 2006]. Despite these problems, few workers report having visited an eye care professional [Quandt et al., 2008]. The frequency of reported eye problems among this population is particularly troubling because there are known effective prevention measures. Protective eyewear (goggles or safety glasses), not only protect the eye from injury, but also can protect against the drying effect of wind, blowing dust or debris, and if tinted, against the harmful effects of UV radiation [American National Standards Institute, 1997; Sakamoto et al., 2002; USDOL, 2013]. In addition, the use of lubricating eye drops can significantly reduce dry eyes and pain, remove debris, and reduce the damaging effects of particles remaining in the eye [Bhavsar et al., 2011]. Unfortunately, eye injury studies have demonstrated farm worker eyewear utilization to be at 10% and below [Quandt et al., 2001, 2008; Verma et al., 2011; Monaghan et al., 2012]. This disinclination to use eye protection is also found among other populations [Lombardi et al., 2009]. Reasons given include discomfort, inability to see as well when using, a lack of understanding of the benefit, undesirable appearance, slowing down the work pace, no mandate from employers, and not having them readily available [Quandt et al., 2008, Lombardi et al., 2009]. However, none of these studies reported including farmworkers in selecting the type of protective eyewear to be used. In 2004, researchers received research support from the National Institute for Environmental Health Sciences under its Environmental Justice Program to work with two agricultural communities to identify and address farmworker priority health issues. The Community Collaborations for Farmworker Health and Safety Project was designed to use a community-based participatory research (CBPR) approach to address farmworker health priorities. CBPR has gained prominence in recent years as a means of utilizing the lived experience of target audience members, acknowledging the value of what they know, and ceding control of the direction of the research [Green et al., 1995; Israel et al., 1998; O’Toole et al., 2003; Yu et al., 2007]. In this type of research, the community is part of the research team and the research is focused on their issues of greatest interest. Farmworker community coalitions were formed at two separate sites in the Northeast: Orange County in southeastern New York State (“Hudson”), and Washington County in eastern Maine. The coalition involved a multidisciplinary professional team consisting of a research team, primary care
practitioners in the study region and migrant health program directors. Local community committees were then established, including migrant and seasonal farmworkers, farm owners, representatives from agricultural and farmworker service agencies and other interested community members. The central organizing role was that of the Community Coalition Coordinator, who in the case of the Hudson eye health program, was a Latina health center outreach worker with many years of experience in the farmworker community. In New York, the community coalition selected eye irritation that results from the very fine black soil in the region as the leading health concern. The topic selection process is described in detail elsewhere [Hawkes et al., 2007], but the steps are reviewed here briefly. Once a community committee, consisting of farmworkers, farm owners, clinicians, and other agency personnel was formed, the group gathered and assessed data on local occupational injury and illness patterns. Each team relied upon its medical professional and community input (key informant interviews, focus groups) in its prioritization tasks. The Hudson team was composed of Latino workers and the owners of several relatively small vegetable operations in the area. In order to facilitate participation, the project held meetings in conjunction with other agricultural community meetings and events. Haddon Matrix [Haddon, 1980] was introduced as a tool for more clearly specifying issues, for broadening the range of issues under consideration, and for systematically comparing one idea with another. After these presentations the community committee chose the priority farmworker health issue. One unique characteristic of the eye irritation issue (and one reason it was selected) is that it is also a health issue for farm managers and employers. Anyone who works in the farm environment can experience significant eye irritation, particularly in the “black dirt region” of the Hudson Valley (see Fig. 1). The soil is very rich, and small particles frequently blow across the fields, creating small dust storms. Once eye irritation and eye health were selected, a range of possible interventions was explored and considered by the committee in early 2005. Using the Intervention Decision Matrix [Fowler and Dannenberg, 2003], the group ranked the proposed interventions, based on effectiveness, feasibility, affordability, sustainability, political acceptability, social and political will, potential for additional benefits, potential for unintended risks, to arrive at a final priority ranking. From among a range of interventions: protective eyewear, eye wash, an educational video, in-person educational presentations, First Aid training, hand-washing stations, and others, a three-part intervention was selected. The community work team developed a three-part intervention consisting of: eye drops, protective eyewear, and an in-person presentation on eye health (including how to use the eye drops and eyewear). As part of the education component, researchers also produced and distributed a pocket card that demonstrated the use of eye protection and
Farmworker Eye Health Intervention Evaluation
FIGURE1. The Orange County,‘‘black dirt’’ region experiences significant dust storms, due to high winds and fine, powdery soil.
eyewash. While this type of approach has been used for eye injury prevention, there has been very little study of eyewear and eye wash use for the prevention of irritation and disease. During the summer of 2006, researchers implemented this intervention with a random selection of Orange County farms. While process evaluation results are presented elsewhere [Earle-Richardson et al., 2009], the current analysis seeks to answer three questions related to the success of the intervention: (1) is the baseline prevalence of eye symptoms elevated (validating the choice of eye symptoms as a topic); (2) do farmworkers who use the intervention report fewer eye symptoms than control farmworkers at the end of 8 weeks; and (3) what subjective feedback do farmworkers provide about the intervention?
MATERIALS AND METHODS
1055
health outreach contacts, fruit and vegetable grower associations, the New York State Department of Labor contacts and information from Cornell Cooperative Extension. Two geographic regions in Orange County, New York (northern Orange County, southern Orange County) were identified for the study. The use of these areas was intended to maximize the number of workers who could be enrolled but still be small enough in total area so that project staff could visit all of the farms within each region in a 2-day period. The two regions were randomly assigned to intervention and control region. All eligible farms within each region were assigned to be an intervention farm or a control farm according to its region. All of the vegetable farms employing migrant and seasonal farmworkers within these two regions were enumerated and recruited from a randomly ordered list. This was done because of the risk of contamination if farmworkers worked on more than one contiguous farm or if owners or farmworkers from neighboring farms communicated. After each farm was enrolled, all migrant and seasonal farmworkers were recruited to participate. After 60 workers on participating intervention farms were enrolled, no additional farms were recruited. The local Project Coordinator visited farms, described the project, and invited the farm to participate. Once farms had agreed to participate, individual workers were recruited. After a short presentation about the project, workers wishing to participate received an explanation of their rights and gave written informed consent. In order to participate, individuals could not have participated in any previous eyewear trials or have been part of the community work team at any time. They also must never have been diagnosed with glaucoma or wear corrective lens eyeglasses or contact lenses. The project was approved by the Bassett Institutional Review Board.
Intervention Protective eyewear The eyewear used in the evaluation trial was selected after extensive pretesting during the previous year. Farmworker and owner feedback during the pilot testing period indicated that the eyewear was helpful; however, workers reported that dust and debris were still getting into their eyes through the sides of the glasses. In addition, the tint in the trial eyewear reportedly made it difficult to determine if the crop was diseased. Based on this information the work team finally selected a goggle-type design with a clear lens that would completely exclude ambient dust (see Fig. 2).
Population Saline solution bottles Subjects were farmworkers employed during the spring and summer of 2006 by any vegetable farm in Orange County. A sampling frame of Orange county farms employing migrant or seasonal workers was developed from migrant
Initially, the work team considered eyewash stations for washing out foreign bodies and for hand-washing; however, it became apparent that there would be too many challenges to
1056
Earle-Richardson et al.
components of the intervention. During the baseline and two follow-up interviews, farmworkers were asked whether they had used any protective eyewear or eye drops during the previous 2 weeks, in order to identify any study contamination that might be occurring.
Study Outcomes Prevalence of eye symptoms
FIGURE 2. Protective eyewear, type 1.
implement this effectively. Farmworkers cover a wide area on any given day, making keeping a washing station nearby difficult. The team chose small bottles of saline solution instead, which could be used both for washing out dust and debris and rewetting the eyes.
Farmworker eye protection education The work team believed that one major barrier to eyewear and eye drop use was a lack of awareness of the relationship between wind, sun, and eye irritation. After considering a range of modes of delivery for the training, the community work team chose in-person training by the community Project Coordinator together with a medical professional. The training was a 10–15 min presentation, providing a demonstration of the harmful effects of dust on the eye, and then showing how to use the protective eyewear and eye drops. The training was supported by a fold-up accordion card and posters put up on the farm, containing three simple messages: wear eyewear when working, rinse your eyes whenever you sense dust or debris in the eye, and if you have a large foreign body or other injury to the eye, seek medical care immediately. The materials were created in Spanish and English, and women’s versions were made with women featured in the photographs. These materials were pre-tested with farmworkers for language and clarity during 2005. After the training was complete, the medical care provider was available to discuss specific eye problems and to schedule medical follow-up.
Control Condition On control farms, subjects were recruited in the same manner; however, they were not provided with any
In order to evaluate whether the participatory process resulted in the selection of an important health condition, researchers established a measure of high eye symptom prevalence using published studies of other worker populations. This was done because there is no published prevalence of the general population, and office workers were considered to be most similar to the general population. Two recent studies [Rios et al., 2009; Portello et al., 2012] found eye symptom prevalence between 31% and 33% among office workers. Therefore, researchers used >33% as the definition of high prevalence. Baseline prevalence of eye symptoms was measured by asking participants to report the frequency of five eye symptoms during the previous 2 weeks: eye pain, red eyes, blurry vision, itchy eyes and dry eyes on a 3-point scale of occurrence (1 ¼ never, 2 ¼ sometimes, 3 ¼ always).
Utilization of protective eyewear and eye drops Researchers were interested to know baseline utilization percentages for both protective eyewear and for eye drops. If these two intervention components were hypothesized to be the solution to the eye irritation problem, it would be important to document that they were not already in widespread use. In addition, along with the outcome measurement, it was important to assess extent of intervention compliance, contamination of controls (controls who were exposed to the intervention), as well as what change in utilization occurred over the intervention period based on the educational program and encouragement by project staff. Subjects were asked about their use of protective eyewear and eye drop use during the previous 2 weeks, at baseline and at both follow-up measurements. Sub-analyses were conducted to examine the impact of compliance and contamination on study results.
Feedback about the intervention In order to learn as much as possible about why the intervention did or did not work, the Project Coordinator who interviewed subjects also asked for their frank feedback about the protective eyewear and the eye drops. These were openended questions, and free-text responses were recorded.
Farmworker Eye Health Intervention Evaluation
Comments were aggregated and summarized into major themes.
Effect of intervention use on eye symptoms Participants reported on five eye symptoms: eye pain, red eyes, blurry vision, itchy eyes, and dry eyes on a 3-point scale of occurrence (1 ¼ never, 2 ¼ sometimes, 3 ¼ always). This was collected three times for each subject: at baseline, 4 weeks later, and 8 weeks after baseline.
Data Collection Baseline and follow-up data collection were completed by in-person interview by the Project Coordinator, using Spanish language interview form that had been pre-tested with farmworkers the previous year. Interview questions were administered in June when subjects first enrolled, then again in July (4 weeks), and then in August (8 weeks). Snacks were also provided during the interview, since subjects were not compensated monetarily. At the end of the study, intervention subjects were invited to keep the eyewear and drops, and control subjects were provided with the entire intervention. In addition to collecting data at 4 weeks, researchers sought to “boost” the intervention by repeating the educational component of the intervention. At each interview instance, the interviewer provided additional bottles of eye wash. Unexpectedly, interviewing at 4 weeks revealed that many of the intervention subjects were not using their assigned eyewear, most said that the closed sides resulted in the glasses becoming very hot and even fogging up the lenses. Although this variety of eyewear had been selected by farmworkers the previous year, many workers in the trial were not using them. In order to maximize the eyewear use, subjects who did not like the first style were provided with a new type of eye protection, one that provided more ventilation, and was not tinted (see Fig. 3). In the final interview, nearly all of the farmworkers who had not been using the first type of eyewear indicated that they were using the second type.
1057
Data Analysis Baseline symptom frequencies as well as eyewear and eye wash usage were compared between intervention and control groups using Chi square or fisher’s exact test. In addition, comparisons were made between subjects lost to follow-up and those who completed the study. The main analyses concerned the change in eye symptoms over the 8-week period among intervention versus control farmworkers. It was assumed that the subjects perceived the three response levels (1 ¼ never, 2 ¼ sometimes, and 3 ¼ always) as equally spaced, and therefore, the analysis used the numerical values (group means) to reflect responses. This approach was necessary because participants were not all symptom-free at the beginning of the study. Therefore, depending on their baseline symptom frequency, a participant could improve a little (say, from “always” to “sometimes”), a lot (from “always” to “never”), get somewhat worse (from “sometimes” to “always”), get significantly worse (from “never” to “always”), and so on. With so many possible directions and magnitudes of possible effect, scaling the responses into a score, although somewhat crude, was the simplest means of analyzing the intervention effect. The change in each of the five symptoms across the three observation periods (baseline, 4 weeks, and 8 weeks) was compared between treatment and control groups using twoby-three analysis of variance (ANOVA) models. In this model, the individual’s change over time was measured as a withinsubject (repeated measures) effect. The interaction term (treatment by time) was used to assess whether treatment group membership was associated with a change in symptom score for each participant. Twenty-three subjects with missing data for either one (13 subjects) or two of the observation (10 subjects) periods were excluded from these analyses. Age and gender were tested in these ANOVA models as covariates and blocking variables respectively. Stratified analyses for subjects experiencing one or more symptoms at baseline versus those who were symptom-free at baseline were also performed. A sub-analysis was performed using only fully compliant intervention subjects (reporting both eyewear and eye drop usage at both follow-up observations) versus controls not reporting using either drops or eyewear.
RESULTS Participant Demographics
FIGURE 3. Protective eyewear, type 2.
One hundred-twenty subjects were initially recruited for the study. Of these, 97 completed all three interviews (baseline, 4-week, 8-week). There were no significant demographic differences between these 97 subjects and the 23 who had missing data. The dropout rates for the
1058
Earle-Richardson et al.
intervention (18.6%) and control groups (19.7%) were virtually identical. (Typically, dropouts occurred because a worker had left the farm and could not be reached.) Therefore, the rest of the analyses pertain only to the 97 subjects who completed all three interview rounds. As shown in Table I, there was some degree of imbalance between intervention and control subjects with regard to eye symptoms at baseline, with intervention subjects having somewhat higher scores for one or more eye symptoms. ANOVA models (analyzing within-subject changes) were not affected by these differences. Six intervention farms (response rate 10%) were recruited, while 10 control farms (response rate 45%) were recruited. The difference in the number of participating farms in each arm reflected the fact that one intervention farm provided a large number of individual participants (n ¼ 32). Of the 16 participating farms, the mean number of workers per farm was 8, and the range was between 1 and 32 workers. This farm size mean and range is characteristic of farms employing migrant and seasonal farmworkers in this region.
Baseline Eye Symptom Prevalence As shown in Table I, the overall prevalence of eye symptoms at baseline were between 20% and 49%, with redness, blurred vision and itching the most common. Sixtyseven percent of respondents reported one or more symptoms. This is well above the 33% threshold for elevated prevalence.
Use and Commentary About the Intervention During the first follow-up period, the weather became very hot, and workers found the protective eyewear difficult to use. This particular model of eyewear (shown in Fig. 1) has closed sides in order to keep the maximum amount of dust and debris out; however, this lack of ventilation became a serious problem. Thirty-five of the 48 intervention subjects completing the study made comments that included some mention the eye wear having problems with “heat,” “sweat,” or “fogging up.” When the second type of eye protection was made available, the mean eyewear usage score (1 ¼ never, 2 ¼ sometimes, and 3 ¼ always) among intervention subjects increased from 1.85 (T1; SD ¼ 0.41) to 2.35 (T2; SD ¼ 0.64) (P < 0.0001 by one-way ANOVA). By the end of follow-up, eyewear usage among intervention farmworkers rose from 6.8% to 91.7%, and eye drop usage from 28.8% to 75.0%. Most of the farmworker feedback related to the difficulty encountered with the heat and the first set of eyewear, and how much more comfortable the second pair was (36/48 commented that the “gray glasses” were “good” or “better”). Some farmworkers commented that the black pair was better for use in the field (especially if they could be removed periodically), while the gray pair was better for the packing house. Nearly all of the comments about the eye drops were positive (11/12).
Change in Eye Symptoms TABLE I. Subject Characteristics, Baseline Symptoms, and Behaviors n ¼120 All subjects Total subjects 120 Mean age 32.6 Male 98 (81.7%) Hispanic 120 (100%) In past 2 weeks, N (%) experienced Eye pain 34 (29%) Eye redness 58 (49.1%) Blurred vision 51 (43.2%) Eye itching 51 (43.2%) Eyes feel dry 23 (22.0%) In past 2 weeks, utilized Safety glasses 10 (8.4%) Eye drops 33 (27.5%)
P (control vs. intervention)
Intervention
Control
59 31.0 54 (91.5%) 54 (100%) N (%)
61 34.1 44 (72.1%) 44 (100%) N (%)
0.18 0.006 ç
24 (41.4%) 37 (62.7%) 29 (49.2%) 30 (50.9%) 14 (23.7)
10 (17%) 21 (35.6%) 22 (37.3%) 21 (35.6%) 9 (15.3%)
Evaluation of a Community-Based Participatory Farmworker Eye Health Intervention in the “Black Dirt” Region of New York State Giulia Earle-Richardson, PhD, MPH,1 Lynae Wyckoff, MA,1,2 Marilyn Carrasquillo,3 Melissa Scribani, MPH,2 Paul Jenkins, PhD,2 and John May, MD1,2
Background Eye irritation is a constant hazard for migrant and seasonal farmworkers, but there are few studies of the problem or how to address it. Researchers evaluated the effect of a community-based participatory eye health intervention on farmworker eye symptoms in the Hudson Valley, NY. Methods A randomized pre-post intervention with 2, 4-week follow-up periods was implemented with a sample of 97 farmworkers. Five eye symptoms were measured, along with utilization of protective eyewear and eye drops. Results Leading baseline eye symptoms were redness (49%), blurred vision (43%), itching (43%), and eye pain (29%). Significant reductions in eye pain (P ¼ 0.009), and nonsignificant reductions in redness were observed for the intervention group while controls experienced increases in both. Conclusions The intervention was effective in significantly reducing eye pain, and to a lesser extent, redness. Future eyewear promotion programs should offer a range of eye wear, tailor offerings to local climate and tasks, evaluate eyewear durability, and include eye drops. Am. J. Ind. Med. 57:1053–1063, 2014. ß 2014 Wiley Periodicals, Inc. KEY WORDS: eye health; migrant; agriculture; farmworker; occupational; community-based participatory research; participatory intervention development; evaluation
INTRODUCTION Migrant and seasonal farmworkers work long hours outdoors, often in extremely dusty, moldy or windy conditions, and eye irritation is a constant hazard. Exposure
1 New York Center for Agricultural Medicine and Health, Bassett Healthcare Network Bassett Healthcare Network, Cooperstown, New York 2 Bassett Research Institute, Bassett Healthcare Network, Cooperstown, New York 3 Hudson River Health Care, Alamo Health Center, Goshen, New York Correspondence to: Giulia Earle-Richardson, PhD, MPH, New York Center forAgricultural Medicine and Health,Bassett Healthcare Network Bassett Healthcare Network,One Atwell Rd., Cooperstown, NY13326. E-mail: [email protected]
Accepted 30 April 2014 DOI10.1002/ajim.22352. Published online 24 July 2014 in Wiley Online Library (wileyonlinelibrary.com).
ß 2014 Wiley Periodicals, Inc.
to ultraviolet light from the sun, agricultural chemicals and irritating plant materials can also damage the eyes. The relationship between sun, dust, and eye irritation has been established in connection with several common eye conditions, such as cataracts [Michael, 2000], pterygium [Nemesure et al., 2008; Maharshak and Avisar, 2009; West and Muñoz, 2009], and others [Lee et al., 1994; Longstreth et al., 1998; Islam et al., 2000; Lucas, 2011]. Wind has also been implicated in eye irritation from chemicals (fertilizers and pesticides) [Cameron et al., 2006; Behrens et al., 2012], and in eye micro-trauma from plant material traveling at high speed in windy conditions [Chang et al., 2002; Hwang et al., 2002; Villarejo and McCurdy, 2008]. Dust-related eye irritation has also been documented in a number of other occupations [Vedal et al., 1986; Melbostad and Eduard, 2001; Van Kampen et al., 2012]. To date, there are few studies of eye disease and nontrauma related conditions among migrant and seasonal
1054
Earle-Richardson et al.
farmworkers. What research there is suggests that eye pain and disease are common. For example, in one study of a predominantly Mexican cohort of 197 workers in North Carolina, eye pain (40%), redness (43%), itching (25%), and blurred vision (13%) were all reported [Quandt et al., 2001]. Other studies have found similar symptom prevalence [Cameron et al., 2006; Luque et al., 2012]. In another study, nearly one-quarter of workers (23%) were found to have pterygium, [Taylor et al., 2006]. Despite these problems, few workers report having visited an eye care professional [Quandt et al., 2008]. The frequency of reported eye problems among this population is particularly troubling because there are known effective prevention measures. Protective eyewear (goggles or safety glasses), not only protect the eye from injury, but also can protect against the drying effect of wind, blowing dust or debris, and if tinted, against the harmful effects of UV radiation [American National Standards Institute, 1997; Sakamoto et al., 2002; USDOL, 2013]. In addition, the use of lubricating eye drops can significantly reduce dry eyes and pain, remove debris, and reduce the damaging effects of particles remaining in the eye [Bhavsar et al., 2011]. Unfortunately, eye injury studies have demonstrated farm worker eyewear utilization to be at 10% and below [Quandt et al., 2001, 2008; Verma et al., 2011; Monaghan et al., 2012]. This disinclination to use eye protection is also found among other populations [Lombardi et al., 2009]. Reasons given include discomfort, inability to see as well when using, a lack of understanding of the benefit, undesirable appearance, slowing down the work pace, no mandate from employers, and not having them readily available [Quandt et al., 2008, Lombardi et al., 2009]. However, none of these studies reported including farmworkers in selecting the type of protective eyewear to be used. In 2004, researchers received research support from the National Institute for Environmental Health Sciences under its Environmental Justice Program to work with two agricultural communities to identify and address farmworker priority health issues. The Community Collaborations for Farmworker Health and Safety Project was designed to use a community-based participatory research (CBPR) approach to address farmworker health priorities. CBPR has gained prominence in recent years as a means of utilizing the lived experience of target audience members, acknowledging the value of what they know, and ceding control of the direction of the research [Green et al., 1995; Israel et al., 1998; O’Toole et al., 2003; Yu et al., 2007]. In this type of research, the community is part of the research team and the research is focused on their issues of greatest interest. Farmworker community coalitions were formed at two separate sites in the Northeast: Orange County in southeastern New York State (“Hudson”), and Washington County in eastern Maine. The coalition involved a multidisciplinary professional team consisting of a research team, primary care
practitioners in the study region and migrant health program directors. Local community committees were then established, including migrant and seasonal farmworkers, farm owners, representatives from agricultural and farmworker service agencies and other interested community members. The central organizing role was that of the Community Coalition Coordinator, who in the case of the Hudson eye health program, was a Latina health center outreach worker with many years of experience in the farmworker community. In New York, the community coalition selected eye irritation that results from the very fine black soil in the region as the leading health concern. The topic selection process is described in detail elsewhere [Hawkes et al., 2007], but the steps are reviewed here briefly. Once a community committee, consisting of farmworkers, farm owners, clinicians, and other agency personnel was formed, the group gathered and assessed data on local occupational injury and illness patterns. Each team relied upon its medical professional and community input (key informant interviews, focus groups) in its prioritization tasks. The Hudson team was composed of Latino workers and the owners of several relatively small vegetable operations in the area. In order to facilitate participation, the project held meetings in conjunction with other agricultural community meetings and events. Haddon Matrix [Haddon, 1980] was introduced as a tool for more clearly specifying issues, for broadening the range of issues under consideration, and for systematically comparing one idea with another. After these presentations the community committee chose the priority farmworker health issue. One unique characteristic of the eye irritation issue (and one reason it was selected) is that it is also a health issue for farm managers and employers. Anyone who works in the farm environment can experience significant eye irritation, particularly in the “black dirt region” of the Hudson Valley (see Fig. 1). The soil is very rich, and small particles frequently blow across the fields, creating small dust storms. Once eye irritation and eye health were selected, a range of possible interventions was explored and considered by the committee in early 2005. Using the Intervention Decision Matrix [Fowler and Dannenberg, 2003], the group ranked the proposed interventions, based on effectiveness, feasibility, affordability, sustainability, political acceptability, social and political will, potential for additional benefits, potential for unintended risks, to arrive at a final priority ranking. From among a range of interventions: protective eyewear, eye wash, an educational video, in-person educational presentations, First Aid training, hand-washing stations, and others, a three-part intervention was selected. The community work team developed a three-part intervention consisting of: eye drops, protective eyewear, and an in-person presentation on eye health (including how to use the eye drops and eyewear). As part of the education component, researchers also produced and distributed a pocket card that demonstrated the use of eye protection and
Farmworker Eye Health Intervention Evaluation
FIGURE1. The Orange County,‘‘black dirt’’ region experiences significant dust storms, due to high winds and fine, powdery soil.
eyewash. While this type of approach has been used for eye injury prevention, there has been very little study of eyewear and eye wash use for the prevention of irritation and disease. During the summer of 2006, researchers implemented this intervention with a random selection of Orange County farms. While process evaluation results are presented elsewhere [Earle-Richardson et al., 2009], the current analysis seeks to answer three questions related to the success of the intervention: (1) is the baseline prevalence of eye symptoms elevated (validating the choice of eye symptoms as a topic); (2) do farmworkers who use the intervention report fewer eye symptoms than control farmworkers at the end of 8 weeks; and (3) what subjective feedback do farmworkers provide about the intervention?
MATERIALS AND METHODS
1055
health outreach contacts, fruit and vegetable grower associations, the New York State Department of Labor contacts and information from Cornell Cooperative Extension. Two geographic regions in Orange County, New York (northern Orange County, southern Orange County) were identified for the study. The use of these areas was intended to maximize the number of workers who could be enrolled but still be small enough in total area so that project staff could visit all of the farms within each region in a 2-day period. The two regions were randomly assigned to intervention and control region. All eligible farms within each region were assigned to be an intervention farm or a control farm according to its region. All of the vegetable farms employing migrant and seasonal farmworkers within these two regions were enumerated and recruited from a randomly ordered list. This was done because of the risk of contamination if farmworkers worked on more than one contiguous farm or if owners or farmworkers from neighboring farms communicated. After each farm was enrolled, all migrant and seasonal farmworkers were recruited to participate. After 60 workers on participating intervention farms were enrolled, no additional farms were recruited. The local Project Coordinator visited farms, described the project, and invited the farm to participate. Once farms had agreed to participate, individual workers were recruited. After a short presentation about the project, workers wishing to participate received an explanation of their rights and gave written informed consent. In order to participate, individuals could not have participated in any previous eyewear trials or have been part of the community work team at any time. They also must never have been diagnosed with glaucoma or wear corrective lens eyeglasses or contact lenses. The project was approved by the Bassett Institutional Review Board.
Intervention Protective eyewear The eyewear used in the evaluation trial was selected after extensive pretesting during the previous year. Farmworker and owner feedback during the pilot testing period indicated that the eyewear was helpful; however, workers reported that dust and debris were still getting into their eyes through the sides of the glasses. In addition, the tint in the trial eyewear reportedly made it difficult to determine if the crop was diseased. Based on this information the work team finally selected a goggle-type design with a clear lens that would completely exclude ambient dust (see Fig. 2).
Population Saline solution bottles Subjects were farmworkers employed during the spring and summer of 2006 by any vegetable farm in Orange County. A sampling frame of Orange county farms employing migrant or seasonal workers was developed from migrant
Initially, the work team considered eyewash stations for washing out foreign bodies and for hand-washing; however, it became apparent that there would be too many challenges to
1056
Earle-Richardson et al.
components of the intervention. During the baseline and two follow-up interviews, farmworkers were asked whether they had used any protective eyewear or eye drops during the previous 2 weeks, in order to identify any study contamination that might be occurring.
Study Outcomes Prevalence of eye symptoms
FIGURE 2. Protective eyewear, type 1.
implement this effectively. Farmworkers cover a wide area on any given day, making keeping a washing station nearby difficult. The team chose small bottles of saline solution instead, which could be used both for washing out dust and debris and rewetting the eyes.
Farmworker eye protection education The work team believed that one major barrier to eyewear and eye drop use was a lack of awareness of the relationship between wind, sun, and eye irritation. After considering a range of modes of delivery for the training, the community work team chose in-person training by the community Project Coordinator together with a medical professional. The training was a 10–15 min presentation, providing a demonstration of the harmful effects of dust on the eye, and then showing how to use the protective eyewear and eye drops. The training was supported by a fold-up accordion card and posters put up on the farm, containing three simple messages: wear eyewear when working, rinse your eyes whenever you sense dust or debris in the eye, and if you have a large foreign body or other injury to the eye, seek medical care immediately. The materials were created in Spanish and English, and women’s versions were made with women featured in the photographs. These materials were pre-tested with farmworkers for language and clarity during 2005. After the training was complete, the medical care provider was available to discuss specific eye problems and to schedule medical follow-up.
Control Condition On control farms, subjects were recruited in the same manner; however, they were not provided with any
In order to evaluate whether the participatory process resulted in the selection of an important health condition, researchers established a measure of high eye symptom prevalence using published studies of other worker populations. This was done because there is no published prevalence of the general population, and office workers were considered to be most similar to the general population. Two recent studies [Rios et al., 2009; Portello et al., 2012] found eye symptom prevalence between 31% and 33% among office workers. Therefore, researchers used >33% as the definition of high prevalence. Baseline prevalence of eye symptoms was measured by asking participants to report the frequency of five eye symptoms during the previous 2 weeks: eye pain, red eyes, blurry vision, itchy eyes and dry eyes on a 3-point scale of occurrence (1 ¼ never, 2 ¼ sometimes, 3 ¼ always).
Utilization of protective eyewear and eye drops Researchers were interested to know baseline utilization percentages for both protective eyewear and for eye drops. If these two intervention components were hypothesized to be the solution to the eye irritation problem, it would be important to document that they were not already in widespread use. In addition, along with the outcome measurement, it was important to assess extent of intervention compliance, contamination of controls (controls who were exposed to the intervention), as well as what change in utilization occurred over the intervention period based on the educational program and encouragement by project staff. Subjects were asked about their use of protective eyewear and eye drop use during the previous 2 weeks, at baseline and at both follow-up measurements. Sub-analyses were conducted to examine the impact of compliance and contamination on study results.
Feedback about the intervention In order to learn as much as possible about why the intervention did or did not work, the Project Coordinator who interviewed subjects also asked for their frank feedback about the protective eyewear and the eye drops. These were openended questions, and free-text responses were recorded.
Farmworker Eye Health Intervention Evaluation
Comments were aggregated and summarized into major themes.
Effect of intervention use on eye symptoms Participants reported on five eye symptoms: eye pain, red eyes, blurry vision, itchy eyes, and dry eyes on a 3-point scale of occurrence (1 ¼ never, 2 ¼ sometimes, 3 ¼ always). This was collected three times for each subject: at baseline, 4 weeks later, and 8 weeks after baseline.
Data Collection Baseline and follow-up data collection were completed by in-person interview by the Project Coordinator, using Spanish language interview form that had been pre-tested with farmworkers the previous year. Interview questions were administered in June when subjects first enrolled, then again in July (4 weeks), and then in August (8 weeks). Snacks were also provided during the interview, since subjects were not compensated monetarily. At the end of the study, intervention subjects were invited to keep the eyewear and drops, and control subjects were provided with the entire intervention. In addition to collecting data at 4 weeks, researchers sought to “boost” the intervention by repeating the educational component of the intervention. At each interview instance, the interviewer provided additional bottles of eye wash. Unexpectedly, interviewing at 4 weeks revealed that many of the intervention subjects were not using their assigned eyewear, most said that the closed sides resulted in the glasses becoming very hot and even fogging up the lenses. Although this variety of eyewear had been selected by farmworkers the previous year, many workers in the trial were not using them. In order to maximize the eyewear use, subjects who did not like the first style were provided with a new type of eye protection, one that provided more ventilation, and was not tinted (see Fig. 3). In the final interview, nearly all of the farmworkers who had not been using the first type of eyewear indicated that they were using the second type.
1057
Data Analysis Baseline symptom frequencies as well as eyewear and eye wash usage were compared between intervention and control groups using Chi square or fisher’s exact test. In addition, comparisons were made between subjects lost to follow-up and those who completed the study. The main analyses concerned the change in eye symptoms over the 8-week period among intervention versus control farmworkers. It was assumed that the subjects perceived the three response levels (1 ¼ never, 2 ¼ sometimes, and 3 ¼ always) as equally spaced, and therefore, the analysis used the numerical values (group means) to reflect responses. This approach was necessary because participants were not all symptom-free at the beginning of the study. Therefore, depending on their baseline symptom frequency, a participant could improve a little (say, from “always” to “sometimes”), a lot (from “always” to “never”), get somewhat worse (from “sometimes” to “always”), get significantly worse (from “never” to “always”), and so on. With so many possible directions and magnitudes of possible effect, scaling the responses into a score, although somewhat crude, was the simplest means of analyzing the intervention effect. The change in each of the five symptoms across the three observation periods (baseline, 4 weeks, and 8 weeks) was compared between treatment and control groups using twoby-three analysis of variance (ANOVA) models. In this model, the individual’s change over time was measured as a withinsubject (repeated measures) effect. The interaction term (treatment by time) was used to assess whether treatment group membership was associated with a change in symptom score for each participant. Twenty-three subjects with missing data for either one (13 subjects) or two of the observation (10 subjects) periods were excluded from these analyses. Age and gender were tested in these ANOVA models as covariates and blocking variables respectively. Stratified analyses for subjects experiencing one or more symptoms at baseline versus those who were symptom-free at baseline were also performed. A sub-analysis was performed using only fully compliant intervention subjects (reporting both eyewear and eye drop usage at both follow-up observations) versus controls not reporting using either drops or eyewear.
RESULTS Participant Demographics
FIGURE 3. Protective eyewear, type 2.
One hundred-twenty subjects were initially recruited for the study. Of these, 97 completed all three interviews (baseline, 4-week, 8-week). There were no significant demographic differences between these 97 subjects and the 23 who had missing data. The dropout rates for the
1058
Earle-Richardson et al.
intervention (18.6%) and control groups (19.7%) were virtually identical. (Typically, dropouts occurred because a worker had left the farm and could not be reached.) Therefore, the rest of the analyses pertain only to the 97 subjects who completed all three interview rounds. As shown in Table I, there was some degree of imbalance between intervention and control subjects with regard to eye symptoms at baseline, with intervention subjects having somewhat higher scores for one or more eye symptoms. ANOVA models (analyzing within-subject changes) were not affected by these differences. Six intervention farms (response rate 10%) were recruited, while 10 control farms (response rate 45%) were recruited. The difference in the number of participating farms in each arm reflected the fact that one intervention farm provided a large number of individual participants (n ¼ 32). Of the 16 participating farms, the mean number of workers per farm was 8, and the range was between 1 and 32 workers. This farm size mean and range is characteristic of farms employing migrant and seasonal farmworkers in this region.
Baseline Eye Symptom Prevalence As shown in Table I, the overall prevalence of eye symptoms at baseline were between 20% and 49%, with redness, blurred vision and itching the most common. Sixtyseven percent of respondents reported one or more symptoms. This is well above the 33% threshold for elevated prevalence.
Use and Commentary About the Intervention During the first follow-up period, the weather became very hot, and workers found the protective eyewear difficult to use. This particular model of eyewear (shown in Fig. 1) has closed sides in order to keep the maximum amount of dust and debris out; however, this lack of ventilation became a serious problem. Thirty-five of the 48 intervention subjects completing the study made comments that included some mention the eye wear having problems with “heat,” “sweat,” or “fogging up.” When the second type of eye protection was made available, the mean eyewear usage score (1 ¼ never, 2 ¼ sometimes, and 3 ¼ always) among intervention subjects increased from 1.85 (T1; SD ¼ 0.41) to 2.35 (T2; SD ¼ 0.64) (P < 0.0001 by one-way ANOVA). By the end of follow-up, eyewear usage among intervention farmworkers rose from 6.8% to 91.7%, and eye drop usage from 28.8% to 75.0%. Most of the farmworker feedback related to the difficulty encountered with the heat and the first set of eyewear, and how much more comfortable the second pair was (36/48 commented that the “gray glasses” were “good” or “better”). Some farmworkers commented that the black pair was better for use in the field (especially if they could be removed periodically), while the gray pair was better for the packing house. Nearly all of the comments about the eye drops were positive (11/12).
Change in Eye Symptoms TABLE I. Subject Characteristics, Baseline Symptoms, and Behaviors n ¼120 All subjects Total subjects 120 Mean age 32.6 Male 98 (81.7%) Hispanic 120 (100%) In past 2 weeks, N (%) experienced Eye pain 34 (29%) Eye redness 58 (49.1%) Blurred vision 51 (43.2%) Eye itching 51 (43.2%) Eyes feel dry 23 (22.0%) In past 2 weeks, utilized Safety glasses 10 (8.4%) Eye drops 33 (27.5%)
P (control vs. intervention)
Intervention
Control
59 31.0 54 (91.5%) 54 (100%) N (%)
61 34.1 44 (72.1%) 44 (100%) N (%)
0.18 0.006 ç
24 (41.4%) 37 (62.7%) 29 (49.2%) 30 (50.9%) 14 (23.7)
10 (17%) 21 (35.6%) 22 (37.3%) 21 (35.6%) 9 (15.3%)
