Introduction

Pesticides are substances used to control, eliminate, attract, or repel pests, like insects’ rodents, and parasites1. These include organochlorines, organophosphate and carbamates, which are used to enhance agricultural productivity and ensuring global food security2,3. Worldwide, around 5 billion kilo gram (kg) of pesticides are applied every year, preventing over 40% of potential crop loss from pest infestation4. Rural farm workers are directly exposed to these pesticides during activities such as mixing, application, weeding, harvesting, and collecting of vegetables5,6. However, their improper application of this pesticide in agricultural areas leads to the contamination of the natural environment, including water, soil, and air, putting farmers at risk of developing various health problems7,8.

The rise of this health problems among farmers, including headaches, dizziness, allergic reactions, cancer, cardiovascular diseases, and reproductive problems can be attributed to direct exposure to a contaminated natural environment through the consumption of food and water9,10,11,12,13. This exposure has resulted in a significant number farmers were died annually14. Globally, an estimated of 1–5 million cases and approximately 300,000 deaths each year9,10,15. Low and middle-income countries (LMIC) represent a significant proportion of these cases and death, as of these nations are responsible for over half of global pesticide usage16,17. This is further supported by studies indicating that approximately 99% of the reported cases occurred in LMIC17. Among LMIC, African farmers are primarily impacted by these health problems, as they are least prepared to protect themselves and their communities from pesticide use. This lack of preparedness were resulted from low levels of literacy, limited education, inadequate access to information, and poverty18.

Understanding and following safety practices during pesticide application in agricultural activities is crucial for mitigating health risks14,19. However, many rural farm workers in Ethiopia engage in unsafe handling of pesticides, which increases their potential health problems10,17,20,21,22,23,24,25. This unsafe handling practice, were increased by factors like insufficient knowledge, improper use of personal protective equipment (PPE), inaccessibility of pesticides with proper safety information19,21,22,26 and negative attitudes towards safety practice, leads for increasing morbidity and mortality related to pesticide exposure27. Moreover, study in Ethiopia found that many participants undervalued pesticide risks17,28.

This has led farmers to engage unsafe handling practice of pesticide, like, overspray, lack of personal protective equipment, improper storage of pesticides and their containers10,29,30,31, and reuse of washed pesticide containers for food and drinking water28. Moreover, several studies reported only a limited number of farmers practice safe handling of pesticide, for instance, studies in different parts of Ethiopia 26.6%32, 24.4%33, 35.2%34,36.2%35, in Uganda 61%36, in Iran 50.8%21, in and in Nepal 43.1%37 of farmers were practiced safe handling of pesticide.

To address these problems several legislative measures have been established in Ethiopia aimed at safeguarding workers and minimizing pesticide exposure22. Despite legislative efforts such as the Pesticide Registration and Control Proclamation No. 674/2010 and other regulatory frameworks, challenges continue to exist in ensuring safe pesticide handling practices33,38,39. Because, the enforcement of the legislation, health and safety practice in Ethiopia is relatively weak40. Furthermore, the research conducted in Ethiopia has predominantly concentrated on individuals employed in the irrigation and floriculture areas, which are subject to close government support and there limitation of studies on farmers working out of these area10,19,33. Additionally, there have been no studies conducted in Ethiopia by utilizing the health belief model (HBM).

The Health Belief Model (HBM) can be used to understand farmers’ intentions regarding pesticide handling. The first dimension is perceived susceptibility, which refers to farmers’ perception of their risk or likelihood of contracting a condition41. The second dimension is perceived severity, relating to the seriousness of contracting an illness and its potential consequences. The third dimension is perceived benefits, which involves the belief that a specific action will effectively reduce a threat. Lastly, perceived barriers refer to the potential challenges associated with adopting recommended behaviors42. In agricultural production, there is often misuse of pesticides. Previous studies have suggested examining farmers’ behaviors through the HBM approach, as it is widely recognized for identifying factors that influence safe pesticide handling practices and HBM-based predictors43,44.

Therefore, considering the (HBM) as a strong foundation and widespread application in health interventions, this study employed the model to forecast farmers’ intentions to safe handling practice of pesticide, aiming to reduce health risks from pesticide exposure. The HBM, outline a two-step evaluative process determines the likelihood of engaging in health maintenance behaviors, first, assessing the perceived threat of the situation, and second, weighing the benefits against the costs associated with the health maintenance behaviors41. Therefore, understanding the HBM determinants of farmers’ intention towards safe pesticide handling practice enables to educational interventions or agricultural extension program in Ethiopia.

Methods and materials

Study area design and area

Community based cross sectional study was conducted among 603 farm workers, in Awi Zone, which is situated about 436 km northwest of Addis Ababa, Ethiopia, and part of the Amhara National Regional State. Bordered by Benishangul Gumuz Region, North Gondar Zone, and West Gojjam. Awi Zone comprises twelve districts three of them urban and nine rural where most of the residents rely on agriculture and use pesticides to enhance crop production.

In this study, we utilize the (HBM) to investigate the factors that deter farmers from adopting safe pesticide handling practices. This model posits that several elements affect farmers’ intentions to implement safety measures when working with pesticides, such as, Perceived Susceptibility: This refers to the farmer’s belief regarding the likelihood of experiencing health issues or poisoning due to pesticide exposure. Perceived Severity: This encompasses the farmer’s concerns about the seriousness of potential health consequences resulting from pesticide exposure. Perceived Benefits: This reflects the farmer’s perception of the effectiveness of safety measures in mitigating health risks. Perceived Barriers: This includes the negative aspects and challenges associated with safe pesticide handling practices. Cues to Action: These prompts encourage farmers to take appropriate actions when dealing with pesticides. Additionally, health motivation measures can be incorporated to predict health-related issues41. The figure below illustrates the proposed model of relationships among the variables in the HBM. According to this model, five components of HBM positively influence farmers’ intentions to implement safe handling practices for pesticides, while only perceived barriers negatively affect their intentions to adopt these safe practices during pesticide application (Fig. 1).

Fig. 1
figure 1

Hypothesized model of relations between the variables.

Full size image

Population

All farmers from selected Kebeles within the designated districts of the Awi zone during the survey period. Study unit includes systematically selected heads of households from each of the chosen Kebeles in the districts.

Inclusion and exclusion criteria

Respondents were individuals primarily engaged in farming. Only those farmers who had used pesticides within the year prior to the study were included. Farmers who had used only organic pesticides excluded from the study, as were those who were unable to communicate due to illness during the data collection period.

Sample size determination and sampling procedure

The required sample size for the study was calculated using a single population proportion formula. The minimum sample size (n) was determined with a critical value (Za/2) of 1.96 for a 95% confidence interval(CI), a degree of precision (d) set at 0.05, and a proportion of safe pesticide handling practices (P) of 61.3%, based on previous studies19. Considering 10% of non-response rate and of 1.5 design effect to reduce the bias on researcher during sampling method and the adjustments were made to the initial calculation. Consequently, the final calculated sample size was 603 farmers. A power of 80% was considered adequate to detect significant differences in population proportions.

Sampling procedure

A multistage sampling procedure was utilized to select households for the study. The research considered twelve districts within the zonal administration, from which three districts, Zigem, Guangua, and Azena, were selected using a simple random sampling technique. These districts have 20, 20, and 20 kebeles, respectively. The number of kebeles was allocated based on an assumption of 25% of the existing kebele size within each districts. Then 5 kebele were included for the study from each districts, the sample size was proportionally distributed among each selected kebele. Finally, households participating in the study were selected using simple random sampling.

Operational definition

Pesticide use

Pesticides are substances or mixtures used in agriculture or public health programs to protect plants from pests, weeds, and diseases45.

Safe pesticide handling practice

Respondents scoring at or below the mean on 10 related questions are classified as having a “poor level of practice,” while those scoring above are deemed to have a “good level of practice”35.

Knowledge Level

The knowledge score was calculated from 10 questions with two responses: ‘yes’ coded as + 1, and ‘No’ coded as 0. The percentage of knowledge score for each Farm workers was categorized into two labels: <50% was labelled as “less knowledgeable” and ≥ 50% was labelled was labelled as “knowledgeable”35.

Attitude level

The Farm workers attitude was assessed using a 7-point Likert scale with 5 questions (strongly disagree (1), disagree (2), somewhat disagree (3), neutral/no opinion (4), somewhat agree (5), agree (6) and strongly agree (7)). The sum of their responses was calculated as attitude score between 7 and 35. Those Farm workers who scored less than the mean (50%) were considered to have “negative attitude” and those who scored ≥ 50% were considered to have “positive attitude”19.

Read and write(formally educated): Those who could read and write but did not attend school17.

Data collection instrument and procedures

Data was collected through interviews with randomly selected household spouses, using a structured questionnaire and an observational checklist. The questionnaire was developed in English, then translated into Amharic (local language), then back to English to ensure its consistency. The questionnaire contained sociodemographic characteristics, environmental variables and HBM items. Data collection was conducted, using KoboCollect (version 2022.4.4) twenty-nine.

Data quality control

Data quality was assured using a properly designed questionnaire adapted from literatures. Training was provided for both data collectors and supervisors on the purpose of the study, data collection technique and tool by the principal investigator for two days’ the data collector pretested the questionnaire, 5% of the sample size on the district where the study was not undertaken, and necessary amendment were taken based on the finding of the pretest. Every day after data collection, the principal investigator reviewed the questionnaires to ensure the completeness of each response. The principal investigator and supervisor closely monitored the data-collection process.

Data management and analysis

The data were collected using the KoboCollect toolbox and exported to SPSS version-23 for analysis. Data were cleaned (categorization for continuous variables and re-categorization for categorical variables) and descriptive statistics such as frequency distribution tables, means, and standard deviations were computed to describe the data. Crude and adjusted odds ratios (COR) with 95%CI were calculated to measure the degree of association between independent variables and dependent variable. A p-value < 0.05 was considered as a level of statistical significance in multivariable logistic regression analysis. And 5-point Likert scale responses from (HBM 30 items) related to each HBM component were coded into a binary scale (yes/no), the coding responses of 5 (strongly agree) and 4 (agree) as yes (1), while other responses were coded as no (0). All predictor variables (30 HBM items) were entered simultaneously using the enter method to check the predictor items form each construct of (HBM). Independent t-tests was conducted to compare composite scores of HBM components between farmers intending to practice safe handling and those not intending safe handling of pesticide (P value < 0.05). A p value of < 0.25 was used as a criterion during bivariate analysis to retain variables for the multivariable logistic regression model. Crude and adjusted ORs (AORs) with 95% CI were calculated to measure the degree of association between independent variables and composite scores of HBM components of farm workers intending to practice safe handling. A p value of < 0.05 was considered as a level of statistical significance in multivariable logistic regression analysis.

Results

Socio demographic characteristics

Out of 603 total sample size, 594 (98.0%) farmers participated in the study, with ages ranging from 22 to 70 years and a mean age of 31.6 years (± 5.4) years. Most participants (89.9%) were male, and 352 (59.3%) had no formal education, while only 115 (19.4%) completed secondary education or higher. Approximately 69% reported having five or more family members, and 89.7% had been farming for at least two years. Regarding land tenure, 80.3% were landowners, while 19.7% were renters. Among pesticide users, 38.6% applied pesticides two or more times, and about 40.7% had 5–10 years of pesticide use experience and access to safety materials was reported by 278 (46.8%), (Table 1).

Table 1 Socio-demographic characteristics of study participants in selected districts in Awi zone, Northwest ethiopia, 2024.
Full size table

Knowledge and attitude of farmers on pesticide handling practices

he knowledge assessment of 594 respondents regarding safe pesticide handling practices revealed a split in awareness. Among them, 291 (49.1%) had good knowledge, while 303 (51%) had poor knowledge. A significant 67.2% were aware of the health effects of pesticides on humans, and 75.6% recognized their impact on non-target organisms. However, only 17% understood the environmental consequences of pesticide use. While 63.3% acknowledged how weather conditions affect application, 60.8% were unaware of the need to read safety instructions. Additionally, 343 (57.7%) had a negative attitude toward pesticides, indicating a critical need for enhanced education on pesticide safety and environmental impacts.

Pesticide handling practices (SPHP) among farmers

These finding reveal concerning practices among farmers regarding pesticide handling. Overall, 40.6% (95% CI: 36.5 − 45%) of participants exhibited good pesticide handling practices, while 59.4% displayed poor practices. Among the ten components of handling practices, only 4.2% of respondents reported always using personal protective equipment (PPE) during spraying, with a significant 31.3% never using it. Although seeking expert advice was common, 48.0% did so sometimes, while 22.7% never sought guidance. Alarmingly, 48.0% of farmers did not consistently follow application instructions. Safe storage practices were inadequate, with 30.8% never storing pesticides in separate containers. Additionally, only 16.8% participated in training to enhance handling practices. While some farmers-maintained equipment regularly (21.9% always), only 23.4% consistently recorded pesticide use information, highlighting crucial areas for improvement in safety practices, (Fig. 2).

Fig. 2
figure 2

pesticide handling practices farm workers in the selected districts of Awi Zone, North West Ethiopia, 2024.

Full size image

Sociodemographic factors influencing farmers’ pesticide handling practice

The logistic regression analysis in this study examined various socio-demographic and environmental factors influencing farmers’ intentions to practice safe handling (SHP) of pesticides. Among the nine variables analyzed, four demonstrated a significant relationship with SHP intentions: education level, accessibility of safety materials, participation in training, and farmers’ knowledge status. Farmers with primary education or higher were 1.97 times more likely to practice SHP (OR = 1.97, 95% CI (1.33, 2.9)). Those with easy access to safety materials (OR = 1.82, 95% CI (1.24, 2.67)), participants in workshops (OR = 1.97, 95% CI (1.33, 2.9)), and those with good knowledge (OR = 1.732, 95% CI (1.2, 2.5)) were also more likely to engage in safe practices (Table 2).

Table 2 Sociodemographic factors associated with pesticide handling practice among farmers, in awi, zone North West Ethiopia 2024.
Full size table

Reliability of HBM constructs

For each construct, the reliability of the items was cheeked using reliability analysis. The reliability analysis of the Health Belief Model constructs showed strong internal consistency for perceived susceptibility (α = 0.85), perceived severity (α = 0.88), health motivation (α = 0.84), and cues to action (α = 0.80), while benefits (α = 0.74) and barriers (α = 0.64).

Association between HBM constructs with farmers intention with SHP practice

In the context of the Health Belief Model (HBM), a two-regression analysis was conducted to identify predictors of farmers’ intentions regarding safe pesticide handling practices (SHP). The first regression model assessed the impact of 30 individual HBM items as independent variables. The analysis revealed that while none of the perceived susceptibility items significantly influenced intentions, two motivation items had a positive impact: the desire to change behavior due to health risks associated with pesticides (OR = 2.639, p = 0.001) and the belief in the importance of maintaining good health (OR = 2.048, p = 0.048). Additionally, one severity item, indicating that pesticide exposure can cause acute health effects, was significant (OR = 2.228, p = 0.026).

The benefits component also contributed positively, with the belief that safe handling can reduce risks associated with pesticides (OR = 4.755, p = 0.000). Furthermore, two cues to action items showed significant positive relationships: participation in training on safe pesticide use (OR = 4.236, p = 0.000) and receiving safety information from neighboring farmers (OR = 2.530, p = 0.001). Conversely, several barrier items negatively affected intentions, including difficulties in obtaining safety materials (OR = 0.551, p = 0.012) and the unpleasantness of disposing of containers and using personal protective equipment (OR = 0.529, p = 0.020), highlighting the challenges faced by farmers in adopting safe practices (Table 3).

Table 3 Multivariate logistic regression of individual HBM components associated with farmers’ intention to safe handling practice of pesticide, Awi zone, North wets Ethiopia 2024.
Full size table

Regression model of HBM components

A multiple logistic regression analysis was conducted to examine the relationship between farmers’ intentions to engage in safe handling practices (SHP) of pesticides and the constructs of the Health Belief Model (HBM). This model explained 68% of the variance (R² = 0.68). The analysis included six composite components: susceptibility, severity, cues to action, health motivation, benefits of SHP, and barriers to SHP. The findings revealed significant factors positively influencing farmers’ intentions to practice safe pesticide handling. Specifically, higher perceived cues to action (OR = 1.624, p = 0.013) and greater perceived benefits of safe handling practices (OR = 1.809, p = 0.013) were associated with increased intentions, while lower perceived barriers to safe handling (OR = 0.435, p = 0.004) negatively influenced these intentions (Table 4).

Table 4 Multivariate logistic regression of composite HBM components associated with farmers’ intention to safe handling practice of pesticide, Awi zone, North wets Ethiopia 2024.
Full size table

Comparison of HBM components

An independent t-test compared farmers intending to practice safe handling of pesticides with those not intending to do so. The results revealed that perceived severity of health effects, perceived cues to action, and perceived benefits of safe handling practices were significantly higher among respondents planning to safe pesticide handling practice (p < 0.05). Conversely, perceived barriers to using safety measures were significantly lower in those intending to practice SH (p < 0.05), (Table 5).

Table 5 Comparison of HBM components among farmers’ intender and non-intended on safe pesticide handling practice, Awi zone North wets ethiopia, 2024.
Full size table

Discussion

This study was conducted to determine the intention of farm workers on safe pesticide handling practice and its HBM predictors associated with these practices. The study revealed that only 40.6% (36, 44. 6) of farm workers had safe pesticide handling practice in the study area. This finding predominantly highlights the unsafe pesticide handling practice among farmers, these practices include farmers disposing of pesticide spray solution leftover improperly, washing sprayers in their home yards or in river canals29, dumping waste into water sources, and inadequately wearing personal protective equipment (PPE)31. Additionally, numerous studies have reported that 53% and 49.2% of farmers exhibit a low level of risk perception and have negative intentions toward safe pesticide use46,47, which contributes to unsafe handling of pesticides.

The current finding is in line with the result reported by a study in Ethiopia (36.2%)35. The possible explanation might be there is similarity by sociodemographic characteristics and their attitude towards pesticide. In contrast, this finding is lower than what was reported by a study conducted in Ethiopia (61.3%)19, in Uganda (61%)36, in Iran (50.8%)21,46, of farmers were practice safe pesticide handling. This discrepancy may resulted from differences in study area, accessibility of safety equipment, and training availability. Notably, the finding of this study was higher than study conducted in Fogera, Ethiopia(24.4%)2, in Malaga Ethiopia(35.5%)34, and in Iran (34.6%)48. The possible explanation for this might be resulted from the study period among studies.

In this study, having a primary education or higher was positively associated with safe pesticide handling practices. This finding is supported by several studies conducted in Ethiopia35,49,50, the possible explanation may be educated farmers possess knowledge about the harmful effects of pesticides and willing to accept and implement safe pesticide handling practice. Additionally participating in different training and having good knowledge regarding pesticide application were statistically significant factors for safe pesticide handling practices. This is supported by study conducted in Nepal and Ethiopia34,49. This might be safety training enhances farmers’ awareness, provides essential knowledge, and promotes practices that are more effective. Furthermore, the availability of safety equipment was positively correlated with safe pesticide handling practice. And access to safety materials emerged as a significant predictor of farmers’ intentions to adopt safe pesticide handling practices19.

On the behalf of (HBM) constructs indicates that, half of these constructs were associated with farmers’ intentions to engage in safe pesticide handling practices. Notably, perceived benefits of pesticides and cues to action were significantly associated with farmers’ intention to safe pesticide handling practice. This suggests that the (HBM) is somewhat effective in identifying the determinants of farmers’ behavior on safe pesticide handling practice This finding was supported previous studies conducted in different parts of Iran41,42,51,52, and in Nepal52 all of these identifies perceived benefit and cue to action as determinant factors for safe pesticide handling practice. The possible explanation might be the perceived benefits motivate farmers to adopt safe pesticide practices by highlighting health and economic advantages, while cues to action, such as training and peer influence, reinforce the importance of safety measures. Additionally, the current study identifies perceived barrier as a significant predictors of safe pesticide handling practice. This constructs of (HMB) was also supported by study from Iran41 and Nepal52. Perceived barriers, such as lack of knowledge and resources will restrict safe pesticide handling practices.

On the other hand, perceived susceptibility, perceived severity, and health motivation were not significant predictors of this practice. Because, a significant number of farmers in Ethiopia lack awareness of the potential health risks associated with pesticides. As of research shows that around 85.3%19 and 87.5%10 of farmers identified only one type of pesticide-related problem. Moreover, merely 6.310 were informed about the possible health issues linked to pesticide use. Consequently, only 3%38 of farmers recognize the importance of undergoing medical check-ups.

The regression analysis of individual elements of each (HBM) components revealed that, various statements were significantly associated with farmers’ intention to safe pesticide handling practice. These includes,, “Pesticide exposure can result in various acute health effects, such as nausea, dizziness, vomiting, headaches, stomach pain, and eye problems,” This finding was supported by research done in Iran41. Additionally, “I have attended a training course on the safe use of pesticides” and “I have heard about safety measures for pesticide handling from neighboring farmers” were significant within the cue to action component, the finding in line with previous studies done in different area41,52. Moreover, four elements of perceived barriers negatively influence farmers’ intentions to practice safety measures: discomfort from heat or harassment caused by safety equipment hampers their ability to work quickly and comfortably, preparing personal protective equipment (PPE) is time-consuming, compliance with safety measures incurs costs, and the awkwardness of using safety measures deters them. This is supported by research on farmers’ intentions regarding PPE use41. The results of this study further demonstrate that these barriers are significant predictors of safety measures in pesticide use.

Moreover, items related to the benefits of safety measures, such as the idea that safe handling practices for pesticides are an effective way to mitigate potential hazards from various pesticide formulations, found to increase farmers’ intentions to adopt safe handling practices. This is supported by a study conducted among farmers and retailers in Nepal52. The regression analysis of the participants’ sociodemographic variables supports this, showing that access to safety materials and farmers’ experience in attending different training programs are significant predictors of safe pesticide handling practice. In addition, there was some evidence suggesting that extension materials, training courses, or product labels could affect farmers’ willingness to adopt safety measures while handling pesticides5,35,41.

Limitation of the study

The main limitations of this study from the binary coding of Likert-scale responses, which can lead to significant information loss, obscuring nuanced participant attitudes. Additionally, the cross-sectional design fails to establish temporal sequences, making it difficult to determine causal relationships. Furthermore, the reliance on self-reported data introduces biases such as self-reporting bias, recall bias, and social desirability bias, potentially skewing results.

Recommendation

Based on the study findings, stakeholders should implement targeted interventions like workshops and educational programs to raise awareness among farm workers about pesticide risks and safety measures. Subsidies for safety materials and their availability at local agricultural centers will encourage safe practices. Community engagement and peer learning can further strengthen safety culture. For regulatory bodies, mandatory safety training for all workers involved in pesticide handling should be integrated into certification processes. Additionally, clearer pesticide labeling, including visual aids, is essential for better understanding. Enforcement of safety regulations should be strengthened through regular inspections and audits to ensure adherence to safety standards.