Self-Reported Pesticide Exposure During Pregnancy and Pesticide Handling Knowledge Among Small-Scale Horticulture Women Workers in Tanzania. A Descriptive Cross-Sectional Study

1. Introduction

Around 40% of the agricultural workforce worldwide consists of women, majority being from the low- and middle-income countries. In sub-Saharan Africa, over 60% of all working women are in agriculture [1]. The population-weighted average from six African countries (Malawi, Tanzania, Uganda, Nigeria, Niger and Ethiopia) estimated the female labor share in crop production to be 40% [2]. In Tanzania, 67% of the total female labor force works in the agriculture sector [3]. Horticulture, the fastest-growing sector of agriculture in the world, contributes 33% of the total agricultural output [4], and women are estimated to be involved in up to 70% of the actual farm work [5]. In Tanzania, women account for approximately 65 to 70% of the labor force in horticulture farming [6]. Thus, women constitute a significant workforce in the horticulture sector.

Management of horticulture crops often involves the intensive use of pesticides—chemicals or substances used to control, repel or eliminate pests that can harm crops. While pesticides are valuable for protecting crops, they can have serious adverse health effects on humans. Existing scientific evidence suggests that pesticide exposure can have adverse health effects on the respiratory system, nervous system, skin and eyes of agricultural workers, nearby communities and consumers, resulting in endocrine, reproductive, carcinogenic, neurological and other disorders [7], [8]. Exposure of pregnant women to pesticides might cause adverse health effects for the fetus because some pesticides can cross the placenta [9]. Adverse pregnancy outcomes, such as impaired fetal functional immunity [10], increased risk of congenital anomalies [11], impaired neuropsychological development [12], preterm birth, small for gestational age and low birth weight [13], have been suggested to be caused by prenatal pesticide exposure.

Over a decade ago, a study assessed the levels of organochlorine pesticide metabolites in maternal blood samples obtained from healthy pregnant women in northern Tanzania. The study revealed the presence of organochlorine pesticide metabolites in all maternal blood samples [14]. Subsequently, a study conducted in the southern region of the country, focusing on women working in horticulture, revealed that children born to mothers exposed to pesticides exhibited impaired neurodevelopment [15]. These localized findings shed some light on the likelihood of pesticide exposure during the critical period of pregnancy, yet the specific practices associated with such exposures in the course of work remain unclear. A good understanding of these practices is crucial for further research on this topic.

The present study aimed to identify practices that predispose pregnant women working in horticulture to pesticide exposure. It also aims to assess women’s knowledge about pesticide handling. Unfolding these practices is essential for future studies and the implementation of interventions to protect pregnant women and their unborn children. This topic aligns with the pursuit of Sustainable Development Goal 3.9, which aims at reducing the number of fatalities and diseases caused by hazardous chemicals [16].

2. Materials and Methods

2.1. Study design and population

We conducted a descriptive cross-sectional study in three regions of Tanzania: Pwani, Morogoro, and Mbeya. One district was selected in each region: Bagamoyo, Mvomero, and Mbarali, chosen for their high engagement in horticulture. In each district, the two wards with the most households involved in horticulture were selected with the assistance of District Agriculture Officers (DAOs) (Figure 1). In Mvomero, however, the study was conducted in only one ward due to the sudden departure of a trained research assistant for family reasons, leaving insufficient time to train a replacement.

A priori power analysis was conducted using G*Power version 3.1.9.7 [17] to estimate the sample size, drawing on data from a study that examined self-reported pesticide exposure during pregnancy (N = 5,997) [18]. The study revealed that 12.6% of pregnant women reported using pesticides during the 1st and 3rd trimesters. Given a Type I error rate (α) of 0.05 and a power of 0.80, a minimum sample size of 391 was determined for this effect size. Therefore, the obtained sample size of 432 was sufficient for this study. The sample size was distributed to the three study areas depending on the size of land available for horticulture in the area.

In collaboration with Ward Agriculture Extension Officers (WAEOs), areas actively engaged in horticulture were identified, focusing on those with larger farmlands. Hamlet leaders were introduced to the study and asked to identify households with women meeting the inclusion criteria: those working in horticulture, having a child aged 4–6 years, and who continued working before conception and during pregnancy. From the identified households, every third household was selected for interviews conducted by trained research assistants at home or on the farm.

Before the research assistant's visit, hamlet leaders informed households about the study and determined their willingness to participate. The next day, the research assistant, accompanied by a hamlet leader, visited consenting households. Informed consent was obtained verbally from participants, and the research assistants verified that the women met the inclusion criteria before starting the interview.

Interviews were conducted using a digital questionnaire on Samsung Galaxy Tab A tablets, with data collected offline and uploaded daily to the Kobo Toolbox platform for secure storage. The researcher cleaned the data daily, and it was stored in the database until the end of data collection, after which it was downloaded for analysis.

2.2. Background information and exposure assessment

The questionnaires guided women in recalling instances of pesticide exposure that occurred during pregnancy, providing researchers with information not documented in any farm or health facility registers. The questionnaire was divided into two sections. The first section collected women’s background information, including age, distance from the house to the horticulture farm, number of years lived in the area, number of years in horticulture, number of years in horticulture before her recent pregnancy and others. The second section included questions about pesticide handling. The section had nine key questions (Table 1). After each question, all women who responded “sometimes” or “often” had a follow-up question about gestation age.

2.3. Knowledge assessment

Knowledge of safe pesticide handling and use was assessed by two questions. The questions focused on assessing the mothers’ ability to comprehend pesticide information displayed on the packaging label of the pesticide used. Mothers were initially asked whether they knew that pesticide handling and information about how to use them was displayed on the label. Then, mothers were prompted with an open-ended question to recall and list all pertinent information communicated by the packaging label.

The list of items listed by participants during the interviews included eight (8) main items: mixing, spraying, first aid, toxicity, personal protection, pesticide name, storage and disposal. Each item mentioned was given a ‘yes’, and a ‘no’ was given to participants who did not mention the item. Later, the answers were scored. Scores of 1 and 0 were assigned for ‘yes’ and ‘no’, respectively. Hence, the score ranged from 0 (minimum) to 8 (maximum). We adopted score categorization from Monger et al., 2023 [19]. A cutoff level of ≤50% (≤4 points) was considered low, and >50% (5 or more points) was regarded as high knowledge.

2.4. Data analysis

The data collected was downloaded, cleaned and imported into IBM SPSS Version 23. Frequencies and percentages were used to summarize the sociodemographic variables. Pesticide exposure practices (dependent variables) were first dichotomized from a 4-point scale into a binary variable [20]. “Never” and “I don’t remember” were coded as 0 or did not practice, and “Sometimes” and “Often” were coded as 1 or practiced. Chi-square and Fisher’s exact tests were used to compare the distribution of sociodemographic and exposure variables among the three study areas. A P value of ≤ 0.05 was used as a measure of statistical significance.

3. Results

3.1. Participant Characteristics

A total of 432 women (99%) consented out of 436 who were asked to participate. Four women did not want to participate because they were not satisfied with the expected benefits of the study. The mean age of participants was about 34 years, with most (55.1%) aged 31–40 years. Women from Mbarali were significantly younger compared to the other study areas. The participants from the three study areas also differed in terms of years involved in horticulture work and gestation age when they decided to stop doing horticulture work for pregnancy’s sake (Table 2).

The Bagamoyo and Mbarali districts had the most women who lived in the area for 10 years or less. In the Mvomero district, most women have been living in the area for 20 years or more. Most women had worked in horticulture for less than five years, but more than 40% had worked there for more than six years before their recent pregnancy. Most women did not stop working immediately after they became pregnant, and they decided to stop working in the second and third trimesters.

3.2. Pesticide exposure during pregnancy

Women were asked about their involvement in activities that result in exposure to pesticides during pregnancy. Despite their awareness of being pregnant, some women persist in participating in activities that involve direct contact with pesticides. For instance, 37.7% of the surveyed women indicated involvement in pesticide spraying during pregnancy. In addition, a considerable number of women reported re-entry into fields within 24 hours post-spray, with 57.2% engaging in weeding and 22.5% participating in harvesting during pregnancy. Furthermore, 45.6% of the women acknowledged consuming horticultural crops within 24 hours after pesticide spraying during pregnancy (Table 3).

Except for the spraying of pesticides (p=0.290), a chi-square test of independence showed that there was an association between the study area and women’s engagement in activities involving direct contact with pesticides (p≤0.001). Years in horticulture (experience) showed a significant association with eating farm products within 24 hours of post-spraying (p=0.044) only.

As shown earlier, only 12% of the women decided to stop horticulture work due to pregnancy within the first trimester (Table 2). Approximately half (50.9%) of those who continued working engaged in weeding throughout the first and second trimesters, with only a few of them being able to continue with the task in the third trimester. Pesticide spray was also a common task that more than one-third of women reported continuing to use pesticides until the third trimester of pregnancy. Moreover, washing clothes with pesticide spray was the first-trimester task for 46.3% of the women who continued working during pregnancy (Table 4).

A chi-square test of independence showed that there was an association between study area and gestational age when women engaged in weeding within 24 hours after spray (p<0.001), spraying pesticides (p=0.002), harvesting within 24 hours after spray (p<0.001) and washing clothes used for pesticide spray (p<0.001) (Table 5). Years in horticulture did not result in any significant association.

3.3. Knowledge of pesticide handling among horticulture women

When women were asked if they were aware of the presence of important pesticide information displayed on the packaging label, approximately 70% of them (n=301) reported being aware. Pearson’s chi-square test showed that awareness was significantly different among women from the three study districts, with the Mvomero district having the highest proportion of women who were aware.

Regarding women’s knowledge of safe pesticide handling, a substantial majority, 93.3% (n=403), of women were classified as having low knowledge. The results of the Fisher’s exact test (p=0.862) do not indicate a significant association between years spent on horticulture (experience) and knowledge level, but the level of knowledge indicated a significant association between study areas (p<0.001), with Bagamoyo having the highest proportion of women with high knowledge, and gestational age (p=0.001) when individuals decided to stop horticulture work.

A summary of the frequency distribution for each category of label information mentioned by women is provided in Figure 2. Most women were aware that pesticide labels include directives for spraying (64.1%) and mixing (62.5%). However, a smaller percentage of women knew that the labels contain information on toxicity, disposal, storage, first aid, and personal protection.

4. Discussion

In this study we identified practices which predispose pregnant women working in horticulture to pesticide exposure. Our results indicate that women working in small-scale horticulture farms continue engaging in activities that involve direct contact with pesticides during pregnancy, exposing themselves and the fetus to hazardous pesticides. Common activities practiced by pregnant women include spraying pesticides, weeding and harvesting within 24 hours after spray, washing clothes used for pesticide spray and consuming horticulture crops within 24 hours after spray. The proportion of women practicing these activities differed across study areas and trimesters. Just over a tenth of women decided to stop early and about half of those who continued made it to a third trimester of pregnancy. We additionally found that the practices might have been influenced by knowledge of pesticide handling because most women in our study had low knowledge.

Despite the prevailing social construct that associates pesticide handling and application with men's roles [21], [22], this study presents evidence of women's involvement in these tasks. However, the proportion of women engaged in spray work is notably lower than that of men. For instance, in horticulture farms in Kenya, only 11.5% of women participated in pesticide spraying [23], which is lower than the percentage of women who, not only sprayed pesticides but also, sprayed pesticides during pregnancy reported in this study. In contrast, a recent study conducted in Thailand involving 78 mothers with children aged 0 to 72 months found that 70.5% of the mothers participated in pesticide spraying [24]. In both studies, most women were disproportionately subjected to pesticide exposure through other tasks on the farm, such as planting, weeding, and harvesting, which is similar to the findings of this study. Exposure to pesticides during field activities is considered high due to their occurrence in recently sprayed fields [25], and workers do not consider them to be risky [26]. This is why women usually perform re-entry activities without adequate protection [27]. Interestingly, a study conducted within the vegetable farming community in Ghana did not identify a statistically significant association between engaging in re-entry activities and alterations in serum concentrations of cholinesterase [28], suggesting that farm tasks alone might not be enough in establishing pesticide exposure.

Occupational exposure to pesticides during pregnancy is largely understudied, as most available studies tend to focus on domestic and environmental sources of exposure. Considering that half of the global labor force is employed in agriculture [29], it becomes evident that occupational exposure in this sector is a critical area requiring examination. We found that some women engaged in activities involving exposure to pesticides until the third trimester of pregnancy, mirroring findings from Thailand [30], where nearly half of the women continued working in agriculture and more than a quarter applied pesticides during pregnancy. The findings of this study also support what has been reported by Maritano et al. (2022), who studied pregnant women from the Italian Nascita e Infanzia: gli Effetti dell’Ambiente (NINFEA) birth cohort, with the proportion of women engaged in agricultural activities decreasing from 21.3% in the 1st trimester to 13.7% in the 3rd trimester [18]. In contrast to the findings for women from the NINFEA cohort, where only 4.7% of women reported using pesticides in the 3rd trimester, in this study, more than one-third of women reported using pesticides in the 3rd trimester. Although hospital-based studies report small proportions of women involved in occupations that expose them to pesticides during pregnancy [31], they also report high proportions from domestic exposures [32], [33], [34]. Therefore, these reports suggest that pesticide exposure throughout pregnancy is common and not limited to occupational exposure. Thus, protective interventions are necessary both at home and at work.

It was observed in this study that most women have low knowledge of pesticide handling. Only a few women knew that the pesticide packaging label displayed first aid, toxicity, and personal protection information. The level of knowledge in this study was lower than that of previous studies [35], [36], possibly because of its unique focus on women working in small-scale horticulture farms and its approach of using information on pesticide package labels, which is often ignored by pesticide users [37]. However, these results agree with the findings reported by Pandiyan et al. (2023), who demonstrated that most farm workers who attended a tertiary cancer care hospital in Hyderabad, India, did not have any knowledge of the toxicity signs displayed on the pesticide container [38]. Generally, male farmers are reported to have better knowledge of pesticide use and handling than females [39], possibly because literacy rates are either higher for males than for females [40] or because most pesticide-related educational programs are not gender sensitive, as they are biased toward men [4]. Evidence shows that educational programs have a significant impact on the knowledge and pesticide safety practices [24], [41], hence, gender-inclusive educational initiatives should be introduced to enhance women's understanding and awareness of pesticide safety.

In its Safety and Health in Agriculture Convention (No. 184), adopted in 2001, the International Labor Organization (ILO) highlighted the dangerous nature of agricultural employment. The convention calls for national agricultural safety and health policies to eliminate, minimize, or regulate dangers in the agricultural working environment to prevent accidents and injuries to health deriving from, related to, or happening during work. In particular, special provisions have been made for women working in the sector in Article 18, which states that “measures shall be taken to ensure that the special needs of women agricultural workers are taken into account in relation to pregnancy, breastfeeding and reproductive health” [42]. The convention entered into force in September 2003, but after twenty years of existence, only 21 countries have ratified. This undermines efforts toward safeguarding women working in the agricultural sector.

Our study had a cross-sectional design. We only collected data at a single point in time, offering a glimpse of a population's features or connections between variables. This design did not provide an opportunity to monitor the study population over time, and we had to rely on the memory of the interviewed women to describe their previous experiences with the pesticides. Nevertheless, the recall bias was low as many of the practices under investigation were not isolated events but rather ongoing activities, making them more memorable. Also, there was a notable disparity in the sample size between the Bagamoyo and Mbarali districts in comparison to the Mvomero district. The smaller sample size in Mbarali can be attributed to the fact that a significant proportion of women in this region regarded horticulture as a secondary pursuit, prioritizing rice farming. Consequently, the number of women who met the predefined inclusion criteria was lower than initially anticipated, but this did not affect the outcome of the study. In the case of Bagamoyo, data collection coincided with a dry season, during which the water sources required for the irrigation of horticulture crops had dried. Hence, many women transitioned to alternative activities, rendering them less accessible during data collection. Despite these differences, the findings are consistent across all three areas, suggesting that the findings are robust and valid.

5. Conclusions

The data from this study suggest that women working in horticulture engage in activities that expose them to pesticides during pregnancy and have low knowledge of pesticide handling. Women reported to be involved in spraying pesticide, re-entry few hours after pesticide application for weeding and harvesting, and washing clothes and equipment used during spraying. Some women reported to eat horticulture crops sprayed with pesticides within 24 hours. All these practices expose pregnant women and their offspring to hazardous pesticides. Unfortunately, no serious local or international efforts are directed toward protecting women in dangerous workplaces such as agriculture. The findings from this study should constitute a wake-up call for all responsible parties to spearhead policy dialogs in line with the ILO Convention No. 184.