Study Population and Setting
This study is a pre-post intervention study nested in a longitudinal study of adolescent and young adult pesticide applicators in Egypt. The original longitudinal study started in 2014 and continued to 2017, with follow-up testing in 2018 and 2019. Adolescents under the age of 19 who worked for the Ministry of Agriculture to spray pesticides on cotton in the Nile delta were recruited in 2014 and 2015 from four field stations (Quesna, Shohada, Tala, and Berket El-Sabe’) in Menoufia Governorate, Egypt. A small subset of applicators had been enrolled in a previous longitudinal study (n=13, 8.4%) and were invited to participate in the current study (14). Results were replicated excluding these individuals and no difference in the results were found. Applicators had the following job responsibilities surrounding pesticide use: mixing pesticides, filling backpack sprayers, application of pesticides, and cleanup procedures.
Agriculture is one of the largest employers in Egypt (18). The primary agricultural product in Egypt is cotton, and because of its national economic importance, the use of pesticides on that crop is highly regulated by the Egyptian Ministry of Agriculture (MOA) (19). The national government purchases and sells the country’s entire cotton production, and once farmers agree to plant cotton applications of chemicals on those fields come under the control of the Ministry of Agriculture. Thus, all pesticides, equipment, and calibration procedures are standardized across the Governorate. Adolescents are hired by the MOA as seasonal workers to apply pesticides and may work for repeated seasons.
Intervention Description
We worked with the MOA to identify feasible and appropriate methods to reduce pesticide exposure. Initial workplace observations had identified behaviors during mixing, loading, and applying pesticides that increased adolescents' contact with pesticides (e.g., mixing with hands, contact with pesticides during loading and applying, reentry into sprayed fields) and therefore exposure. Additionally, self-reported hygiene practices indicated variability in time to change clothes or bathe after applying pesticides. We also found that increased urinary metabolite levels were associated with increased time applying, and lower urinary metabolite levels were associated with bathing immediately after work and using a stick to mix pesticides (instead of hands) (20). Focus groups held separately with officials from the MOA, adolescents, and parents presented study findings and discussed feasible methods to reduce exposure during application. We found that while some PPE is supplied by the MOA (masks, gloves, glasses) there is not enough for all workers, and some workers had concerns about the quality and effectiveness of the supplied PPE. Although the MOA has adopted procedures to reduce exposure (e.g., point nozzle downward, maintain distance between applicators, consider wind direction when applying), applicators reported receiving no formal training. Most workers reported bathing after work, however, there was variation in how frequently work clothes were washed (e.g., daily, monthly). The intervention incorporated feedback from the focus group in the intervention.
To decrease the burden of pesticide exposure among adolescent applicators, an educational intervention was designed and provided to them. The intervention included a training for participants about the dangers of and potential preventive measures for the use of pesticides. The training also included information about pesticides, e.g., their types, hazards they pose on the adolescents’ developing bodies, as well as the sources and modes of exposure among this group of adolescents. Three behaviors were targeted: staying out of fields recently sprayed, using a stick (instead of hands) to mix pesticides, and bathing/wearing clean clothes. The goal of the intervention was to impact the direct activities taken by the adolescents themselves. This allows the adolescent to reduce their exposure to pesticides regardless of their workplace or available resources.
Training was conducted once in each of the field offices in four districts in May 2017 before the beginning of the 2017 application season. The training was presented in a classroom format and lasted for approximately 1 hour. To study the effect of the training, a pre and post intervention survey was administered on the same day. Participants were asked about their views of pesticide safety and the effectiveness of targeted behaviors to reduce exposure to pesticides before and after the intervention. A follow-up survey was administered 8 months later to assess retention. Additionally, practices of pesticide applicators were assessed before and after the training using an observational checklist administered by trained research staff. This assessment occurred in August 2016 and August 2017. Research staff were present during pesticide application throughout the multiyear study. Participants were unaware of when specific elements of the observational checklists were collected. Observational checklists included personal protective equipment worn before and during pesticide application, mixing procedures, and hygiene after pesticide application. Consent of human subjects was obtained from participants and procedures approved by the University of Iowa Institutional Review Board and the Medical Ethics Committee at Menoufia University.
All participants that attended the training (n=119) completed the pre and post-training pesticide safety and behavior survey, while 95 responded to the survey 8 months after the training. Among the 119 who attended the training, 87 had completed observational checklists of pesticide application during August 2016 and 92 had observational checklists completed during August 2017. There was a total of 71 participants that completed the intervention training and had observational checklist data for 2016 and 2017.
Risk Behavior Diagnosis Scale
The results of the survey were used to categorize individuals using the Risk Behavior Diagnosis (RBD) Scale based on pathways from the Extended Parallel Process Model (EPPM) (22, 23). The EPPM has been widely used in health promotion and disease prevention to develop interventions among diverse international populations (24-29).
The EPPM pathways identify how individuals are likely to control a health risk based on threat perception and efficacy of controlling that threat (23). These pathways have been used to group individuals into four quadrants of behavior using the RBD Scale: responsive (high threat-high efficacy), pro-active (high threat-low efficacy), avoidant (low threat-high efficacy), and indifferent (low threat-low efficacy) (30, 31). The questions used to assess threat and efficacy were related to the targeted behaviors of the interventions — mixing pesticides with a stick (not your hand), not entering fields, and hygiene around pesticide application. An average response of agree (4) on a 5-point scale of strongly disagree (1) to strongly agree (5) was used to place individuals into levels of high or low efficacy and high or low threat. Individuals were surveyed before the intervention, immediately following the intervention and 8 months post-intervention. In addition to the questions used to construct the RBD Scale, several questions about pesticide safety were asked at the same time periods.
Internal Reliability of Perceived Efficacy and Perceived Threat
The questions used to assess perceived efficacy and perceived threat are given in Table 1. Evaluation of the internal consistency of the RBD scale is given in Table 2. The internal consistency of perceived efficacy was calculated using Cronbach’s alpha and found to be 0.641. Cronbach’s alpha estimates reliability for measurement scales used in research (32). No absolute cutoff for reliability exists, but a minimum for fair reliability of 0.60 and 0.65 has been suggested for studies of less than 100 participants and studies of 100 to 300 participants respectively (33). Removing any individual question did not meaningfully change the internal consistency (<1% change) of perceived efficacy and removing multiple measures only marginally increased consistency, so the final scale included all 6 questions (removal of Q11 and Q12 increase Cronbach’s alpha to 0.661). The measures of perceived threat had a Cronbach’s alpha of 0.639. Removing Q13 and Q14 improve the internal consistency and were excluded from the scale in analysis (Cronbach’s alpha = 0.690, Table 2).
Table 1: Risk Behavior Diagnosis Scale Questions to classify Adolescent and Young Adult Pesticide Applicators Perceived Efficacy and Threat
Scale
|
Question
|
Perceived Efficacy
|
Q11 Using a stick instead of my hands to mix pesticides will prevent me from getting sick
|
Q12 I am able to use a stick instead of my hands to mix pesticides
|
Q15 Staying out of fields that were recently sprayed will prevent me from getting sick
|
Q16 I can stay out of fields that were recently sprayed
|
Q19 Changing into clean clothes after working with pesticides will prevent me from getting sick
|
Q20 I can wear clean clothes with no pesticide residues
|
|
|
Perceived Threat
|
Q9 Using hands to mix pesticides can cause sickness
|
Q10 If I mix pesticides with my hands it could make me sick
|
Q13 Entering fields that were recently sprayed will cause sickness
|
Q14 If I enter a field that was recently sprayed I am likely to get sick
|
Q17 Wearing clothes with pesticide residues on them for several hours will cause sickness
|
Q18 If I wear clothes with pesticide residues for several hours, I am likely to get sick
|
Table 2: Internal consistency of the Risk Behavior Diagnosis Scale for Adolescent Applicators in Egypt
Scale
|
Cronbach’s alpha
|
Adjustment Made
|
Median Score (range)
|
Perceived Efficacy
|
0.641
|
|
24 (17-30)
|
|
|
|
|
Perceived Threat
|
0.639
|
|
|
|
0.627
|
Remove Q13a
|
|
|
0.631
|
Remove Q14b
|
|
|
0.690
|
Remove Q13&14
|
16 (10-20)
|
aQ13 – Entering fields that were recently sprayed will cause sickness
bQ14 – If I enter a field that was recently sprayed I am likely to get sick
Statistical Methods
The internal reliability of the RBD scale was assessed using Cronbach’s alpha coefficient. The scale was then used to place individuals into the four RBD quadrants of health risk behavior. Differences in demographics and pesticide application practices were compared across the quadrants using the Chi-squared test. Change in behavior post-intervention and 8 months later was compared using the McNemar test to detect a difference in the proportion of individuals placed in each RBD quadrant. Similarly, changes in patterns of PPE used found during the 2016 and 2017 observational checklist and changes in feelings regarding pesticide safety were compared with the McNemar test. Data analysis was completed with SAS 9.4 (SAS Institute, Cary, NC) with an alpha of 0.05 used for testing significance.