Mental health disorders among Thai farmers: occupational and non-occupational stressors

Abstract

This cross-sectional study aimed to identify occupational and non-occupational factors that impact levels of stress, depression, and anxiety among farmers located in the northern part of Thailand, including the potential psychological impacts of pesticide use. The participants (N = 270) were interviewed with a survey adopted and modified from peer-reviewed articles and questionnaires. The survey consists of four parts, including demographic information and pesticide exposure; perceived farm stressors employing Farm Stressor Survey (FSS); mental health disorder utilizing Srithanya Stress Scale (ST5), Nine-Questions Depression- Rating Scale (9Q), and Depression Anxiety Stress Scale 21 (DASS21); in addition to COVID-related stress utilizing COVID Stress Scale (CSS). The participants were categorized into two groups i.e., Spray and No Spray based on their self-reported occupational use of pesticide spraying within the past year. No significant associations were observed between occupational pesticide exposures and mental health disorder scores. Being female, having a second job besides a farmer, having applied pesticides for greater than 20 years, and having a higher farm stressor perception showed a significant positive association with self-reported mental health disorders; while having a good agricultural practice and PPE use showed a significant negative association with those outcomes. This pilot study scrutinized expanded sources of stress in farm work and provided information for the development of more effective mental disorder intervention programs for Thai farmers.

Keywords:

• mental health
• farm stressors
• pesticides
• good health and well-being

Introduction

In Thailand, approximately 47% of the land (accounted for 150 million ria) is used for agriculture and approximately 32% of the total employment of the Thai population works in agriculture, compared to only 22% in the industrial sector (National Statistical Office 2023; O’Neill 2023). The northeastern region of Thailand is where the highest percentage (42%) of the total agricultural land is, followed by the northern (22%), central (21%), and southern (15%) regions, respectively (Office of Agricultural Economics 2023). Farmers work under challenging conditions that can expose them to a wide range of hazards that may impact their physical and mental health. Several studies have suggested that several factors related to farming contribute to poor mental health among farmers. Previous research into farmer mental health was mostly conducted in developed countries e.g., US, Australia, and UK, while those conducted in developing countries, e.g., Thailand, Nigeria, and Malaysia contributed to less than 1%. Key risk factors of farmer mental health have been limited scrutinized in previous studies; less than 5% have examined the association between mental health and risk factors, including weather conditions, governmental policies, problem with machinery, and working with family (Yazd et al. 2019). Mental health disorders depend on individuals’ context including their geographic location and culture, their socioeconomic status and associated lifestyle, accessibility to healthcare, as well as their health risk perceptions (Eberhardt and Pooyan 1990; Kearney et al. 2014; Truchot and Andela 2018). Stressors such as occupational hazards and injuries, personal finances, time pressures, economic conditions, employee-employer relationships, unfavorable environmental conditions, and other factors i.e., social isolation, governmental policies, and poor housing conditions potentially impact overall farmer mental health (Yazd et al. 2019; Rudolphi et al. 2020). Previous studies suggested an increase of mental health disorders among farmers in the low-income countries, e.g., India, Peru, and Nigeria, associated with hours of excessive farm work, financial problems, unpredictable weather, and uncertain future for agriculture (Kureshi and Somsundaram 2018; Olowogbon et al. 2019; Bazo-Alvarez et al. 2022; Belokar 2021). An increase of mental health disorders among farmers in high-income countries associated with personal finances, time pressures, workload, and poor weather (Rudolphi et al. 2020; Brennan et al. 2022). The farm stressor inventory survey has been used in studies in several locations including the US, New Zealand, Ireland, Nigeria, and Australia (Bin 1970; Kearney et al. 2014; Olowogbon et al. 2019; Rudolphi et al. 2020; Brennan et al. 2022); however, there have been no previous studies employing farm stressor inventory survey among Thai farmers. Therefore, further studies are needed in order to explore an expanded sources of stress among farmers in Thailand and other developing countries.

Agrochemical exposures and hazardous working conditions are two of the crucial factors associated with adverse mental health outcomes among farm workers. Pesticide exposures have been linked to farmer’s mental health and sleep disorders (Kim et al. 2013; Harrison and Mackenzie Ross 2016; Li et al. 2019; Buralli et al. 2020). In addition, a positive association between pesticide poisoning and mental health disorders was observed among Brazilian farmers (Campos et al. 2016). In Thailand in 2020, 19 million kilograms of insecticides, 15 million kilograms of fungicides, and 57 million kilograms of herbicides were imported (Department of Agriculture 2021). Some pesticides, including organophosphate and carbamate insecticides, are neurotoxic and directly impact the nervous system. Previous studies have suggested that lower acetylcholinesterase (AChE) activity is associated with higher depression/anxiety symptoms, as the cholinergic system has an important role in the central nervous system and may have modulatory effects on mood and behavior (Chen 2012; Altinyazar et al. 2016; Serrano-Medina et al. 2019; Suarez-Lopez et al. 2019). A study conducted in Thailand suggested that programs for preventing chemical intoxication and improving family communication skills decreased depressive disorders among rice farmers. The findings emphasized the significance of implementing good practices in chemical use in order to minimize occupational chemical exposures, resulting in lower self-reported depression symptoms (Kaewboonchoo et al. 2020). However, the epidemiological data determining the association between exposures to agrochemicals, particularly organophosphates and carbamates, and psychological distress are inconclusive. Therefore, further studies are needed (Khan et al. 2019).

The objectives of the present pilot study were 1) to explore the association between pesticide exposures and mental health disorders, i.e., stress, depression, and anxiety; and 2) to determine perceived occupational and non-occupational stressors employing a comprehensive Farm Stressor Survey collecting data on an expanded set of determinants of mental health, in addition to explore the association between those perceived farm stressors and mental health disorders among Thai farmers.

Materials and methods

Participant recruitment

This was a cross-sectional study approved by the Ethical Review Committee for Human Research Faculty of Public Health, Mahidol University (COA No. MUPH2021-087). The study location was Amphoe Payuhakiri, Nakhon Sawan province, in the lower northern part of Thailand. The study location was selected based on the high possibility of participant recruitment in Nakorn Sawan province which is ranked the fourth highest in Thailand for agricultural land use and the 16th highest for the number of agricultural households (Office of Agricultural Economics 2023; Digital Government Development Agency 2023). The study participants were selected by purposive sampling technique. The local site officers including public health volunteers along with a community leader visited farmers at their homes to provide the study information for the participant recruitment. Thai farmers and their family members aged 18 years or older, engaged in agricultural activities as a first or secondary job at least one year prior to taking part in the study, were recruited. Participants with preexisting clinical diagnosis of mental disorders and health conditions, such as alcoholism and drug addiction, were excluded from the study. Each participant was interviewed by one of three trained interviewers during January to April, 2022.

Survey

Each participant answered questions of the four-part survey inquiring about their demographic information, perceived occupational and non-occupational stressors related to farm work, household environment and agricultural activities pertaining to pesticide exposures, in addition to their symptoms of mental health disorders, i.e., stress, depression and anxiety. Part 1 collected demographic and pesticide use information. The questionnaire has been extensively used in previous studies (Kongtip et al. 2018, 2019, 2020; Konthonbut et al. 2020). Part 2 collected information on sources of stress among farmers. The primary tool used was the Farm Stressor Survey (FSS) adopted from previous studies and translated into Thai (Kearney et al. 2014). The translated survey was reviewed/compared to the original survey by three experts in order to validate the consensus between the translated and original versions. The self-reported questionnaire consisted of 25 items including 14 farm-related stressor questions (e.g., weather, problems with machinery, farm accidents and injuries, and dealing with hazardous chemicals), six finance-related stressor questions (e.g., market prices, taxes, health care costs, and debt load), and five social-related stressor questions (e.g., geographical isolation and not enough time for family) with one open-ended question identifying any other items participants find stressful. Each item was scored based on the responders’ perceived stress as zero (not stressful), one (a little stressful), two (moderately stressful), and three (very stressful). The FSS investigated perceived stressors both occupational-associated stressors i.e., 14 farm-related questions and non-occupational-associated stressors i.e., six and five finance- and social-related questions. Part 3 collected data on the outcome of interest in this study, mental health disorders including anxiety, depression, and stress symptoms. Depression involves a loss of interest or pleasure in activities or depressed moods including feeling sad, irritable empty, and/or feeling very tired or low in energy. Anxiety is manifested an intense and excessive fear and worry about a specific or a broad range of everyday situations. Stress symptoms are defined as a state of worry or mental tension caused by particular situations in an individual’s life. In this study, the participants’ mental health disorders were identified by the level of their self-reported symptoms without medical diagnosis. Two questionnaires developed by the Department of Mental Health, Ministry of Public Health, Thailand were used including the Srithanya Stress Scale (ST5) for assessing stress symptoms and Nine-Questions Depression- Rating Scale (9Q) for assessing depression. In addition, this section also included the widely used Depression, Anxiety, and Stress Scale 21 (DASS21) for assessing depression, anxiety, and stress symptoms (Silpakit 2008; Oei et al. 2013; Department of Mental Health Thailand 2019; Isaradisaikul et al. 2021; Wamaloon et al. 2565). Each question of the ST5 captures the frequency of symptoms related to stress within the four weeks prior to the interview date. Each question of the 9Q captures the frequency of symptoms related to depression within the seven days prior to the interview date. The DASS21 contains 21 questions capturing symptoms within seven days prior to the interview date for stress symptoms (seven questions), depression (seven questions) and anxiety (seven questions). Each question of these questionnaires was scored as zero (never/rarely) to three (always/every time/very often). Part 4 collected data on stress symptoms related to the COVID-19 pandemic using the COVID Stress Scale (CSS) developed by Taylor and colleagues and utilized for assessing COVID-related stress symptoms among Canadian and US population (Taylor et al. 2020). This translated survey was reviewed/compared to the original survey by three experts in order to validate the consensus between the translated and original versions. The CSS captures metrics indicating stress symptoms related to 1) concern with catching the virus, spread of the virus from foreigners, spread of the virus in public areas and running out of essential needs (CSS-phobia); 2) trauma due to fear of the virus (CSS-stress); and 3) consistency in checking information about the status of the pandemic (CSS-checking). Each question of the CSS questionnaire was scored as zero (not at all/never), one (slightly/rarely), two (moderately/sometimes), three (very/often), and four (extremely/almost always).

In addition to content validation of the four parts of the survey by three experts, including one experienced in epidemiological research and two experienced in occupational health research, the surveys were further tested for reliability. Thirty participants living in the study location areas aged 18 years and above, able to communicate in Thai, and who did not have a primary or secondary job as a farmer nor engage with agricultural activities for work within the past year were asked to complete the surveys. Nonetheless, this group of participants engaged with agricultural activities as their hobbies and/or their own interests. Cronbach’s alpha coefficients were 0.60-0.63 (farm stressor survey), 0.81-0.89 (stress survey), 0.86-0.92 (depression survey), and 0.81 (anxiety survey), and 0.81-0.88 (COVID-related stress survey).

Statistical analysis

Demographic characteristics of the study population were determined. An individual’s level of pesticide exposures was classified based on self-reported engagement of occupational pesticide spraying within the past year i.e., the Spray group was presumed to be relatively high exposure (participants reported engaging in pesticide spraying within one year prior to the interview date) and the No Spray group was presumed to have lower exposure to pesticides (participants reported no pesticide spraying within one year prior to the interview date). Another three parameters relevant to pesticide exposures are agricultural practices, personal protective equipment (PPE) use, and PPE use in the past were also considered. The agricultural practice related questions include not eating/drinking during working, washing hands before breaks, standing upwind during pesticide spraying, changing contaminated clothes, and washing immediately after the work shift. The PPE use related questions capture the list of used PPE e.g., boots, gloves, masks, long-sleeved shirts and pants. Each item of the parameters was assigned scores and computed as score fractions. Agricultural practice score fractions were computed by dividing the summation of the scores obtained by the total scores of all agricultural practice items. Each item was scored based on the frequency of performing good practices as zero (never), one (sometimes), two (every time). PPE score fractions were computed by diving the summation of scores obtained by the total scores of appropriated PPE used. Higher score fractions represent better practices.

The perceived farm stressors among participating farmers were calculated as a percentage of participant stress levels ranging from not stressful to very stressful for each stressor. Self-reported farm stressors were categorized as not/a little stressful and moderately/very stressful. The Chi-squared test was conducted to explore the correlation between each perceived farm stressors (not/a little vs. moderately/very stressful) and mental health disorders (none/little vs. moderate/severe symptoms). The score fractions of three groups of perceived farm stressors i.e., farm-related, finance-related, and social-related were computed by dividing the summation of the scores obtained by the total scores of each prong of the farm stressor inventory questionnaire. A higher score represents higher stress. The total scores and score fractions of mental health parameters indicating stress, depression, and anxiety were calculated and summarized for different levels of pesticide exposures (Spray vs No Spray).

The participants’ total self-reported scores of each mental health disorder parameter were summed and divided by the total scores of each parameter and documented as mental health disorder score fractions. Higher scores represent worse mental health. The descriptive statistics of the mental health disorder scores i.e., mean, range, and standard deviation were reported. The correlations between continuous data of each mental health disorder parameter score fractions obtained from the different questionnaires, i.e., ST5-stress and DASS21-stress; 9Q-depression and DASS21-depression, were explored utilizing Pearson correlation. Pearson correlation coefficients (r) were computed and documented as a high degree of correlation (r ≥ 0.8), a moderate degree of correlation (0.5 ≥ r > 0.8), and a low degree of correlation (0.2 ≥ r > 0.5). Self-reported mental health outcomes were combined and categorized into no/little level of mental health disorders and moderate/severe level of those outcomes. Univariate logistic regression analyses of five mental health disorder parameters i.e., ST5-stress, DASS21-stress, 9Q-depression, DASS21-depression, and DASS-anxiety categorized into the binary outcomes i.e., no/little and moderate/severe levels of symptoms were performed to explore the factors driving mental health disorders among the study population. In addition, multiple logistic regression analyses were performed utilizing the stepwise method for five mental health disorders. The dependent variables were symptoms of mental health disorders including ST5-stress, 9Q-depression, DASS21-stress, DASS21-depression, and DASS21-anxiety. In response to the first objective, to explore the association between pesticide exposures and mental health disorders, the independent variable was occupational pesticide exposures associated with engaging in pesticide spraying (Spray vs. No Spray). In response to the second objective, to determine perceived occupational and non-occupational stressors among Thai farmers and explore the association between occupational stressors i.e., farm-related stressors and mental health disorders, the independent variable was occupational stressors, which was defined as 14 items of work-related farm stressor indicators including farm injuries and accidents, workloads and work responsibilities, work relationships, work support and supervisions. The multiple logistic regression models were adjusted with potential confounding variables including demographic variables i.e., gender and having a secondary job; PPE use; agricultural practice; historical years of engaging in pesticide spraying; and COVID-related stress. All statistical analyses were conducted using the R program (RStudio, MA).

Results

Participant demographic and pesticide exposures

Of the 270 farmers who participated in the study, 146 were in the Spray group and 124 were in the No Spray group, accounting for 54% and 46%, respectively. The demographic characteristics of the study population are summarized in Table 1. Participants most frequently held primary and secondary school degrees, with the mean work experience in farming of 20 years and the mean farming work duration of 29 h per week. The majority of the Spray group were male (66%), listed farming as their primary job (95%), were the household head (64%), and were indebted (62%). Smaller proportions of participants in the No Spray group listed farming as their primary job (74%), were the household head (32%), and were indebted (44%). A higher proportion of the No Spray group listed farming as a second job (26%) compared to that of the Spray group (5%). The No Spray group was predominately female (64%), had farmed for significantly fewer years and were more likely to live within a 1-km radius of areas where pesticides were applied. The No Spray group also had significantly fewer participants who smoked and drank alcohol than the Spray group.

Table 1. Demographic characteristics of the participants.

Parameters		Spray (%)	No Spray (%)	p-value*
Gender	Male	66	36	<0.0001
	Female	34	64
Highest education level obtained	Less than primary school	2	7	NS
	Primary	49	44
	Secondary/vocational	47	43
	Bachelor’s or higher	3	6
Marital status	Single	26	24	NS
	Married/living together	68	69
	Married/not living together	1	1
	Widow	4	6
	Divorced/separated	0	1
Farming as the primary job	Yes	95	74	<0.0001
	No	5	26
Having a secondary job with a primary job as a farmer	Yes	21	8	0.0495
	No	74	65
	Did not have a primary job as a farmer (has a secondary job as a farmer)	5	26
	Missing	1	2
Household members	1-2	25	24	NS
	3-4	53	56
	5-6	15	16
	7-8	7	4
Head of household	Yes	64	32	NS
	No	36	66
	Missing	1	0
Household income per month (THB)	≤1000	0	1	NS
	1001-5000	19	15
	5001-10000	28	27
	10001-25000	36	47
	25001-50000	16	8
	>50000	1	2
	Do not know/do not want to respond	1	2
Satisfaction with household income/expenses	Satisfied with saving	9	10	NS
	Satisfied with no saving	59	55
	Not satisfied	30	35
	Do not know/do not want to respond	0	0
	Missing	2	0
Indebtedness	Yes	62	44	0.0102
	No	36	53
	Do not want to respond	2	3
Having health insurance	Yes	93	94	NS
	No	7	6
Smoking status	Yes	22	9	0.0074
	No	76	90
	Quit	2	2
Alcohol consumption status	Yes	55	27	<0.0001
	No	43	72
	Quit	1	1
Living within a 1-km radius from areas where there is pesticide application	Yes	38	52	0.0268
	No	62	48
Household pesticide application	Yes	90	88	NS
	No	10	12
		Spray mean ± sd	No Spray mean ± sd	p-value*
Age (years)		48 ± 16	46 ± 18	NS
Farming years (years)		22 ± 15	18 ± 16	0.0314
Farming hours (hours)		29 ± 14	28 ± 15	NS
Agricultural practice score fractions (higher = better practices)		86.73 ± 9.10	85.63 ± 12.94	NS
PPE use score fractions (higher = better practices)		55.16 ± 18.79	50.58 ± 17.83	0.0428
PPE use score fractions in the past (higher = better practices)		53.46 ± 20.13	26.06 ± 31.72	<0.0001
Farm-related perceived stress score fractions (higher = more stress)		15.00 ± 12.63	14.52 ± 11.99	NS
Financial-related perceived stress score fractions (higher = more stress)		35.84 ± 18.52	30.82 ± 15.86	0.0171
Social-related perceived stress score fractions (higher = more stress)		8.63 ± 12.94	10.59 ± 13.11	NS

p-values were calculated using Chi-squared test and t-test (No Spray vs. Spray groups), *p-value < 0.05; NS = not statistically  significant.

The participants of this study performed several agricultural activities. Observations suggested agricultural tasks were similar among both the Spray and No Spray groups with the exception of two activities which most farmers in the Spray group performed but the farmers in the No Spray group did not i.e., sorting agricultural products and applying fertilizers. Among the Spray group, insecticides and herbicides were commonly used/sprayed; fungicides were rarely applied but they were purposely used for soaking seeds and sprouts. The majority of the participants grew one type of plant; mostly rice, followed by cassava and sugarcane, respectively. Participants’ history of pesticide application was documented. A higher percentage of the Spray group (98%) reported that they engaged in pesticide spraying in the past compared to the No Spray group (45%). No participants reported any history of physician-diagnosed pesticide poisoning within the past year. The top three work-related adverse health outcomes reported were dizziness (29%), perspiration (32%), and fatigue (32%).

Participant perceived farm stressors

The perceived farm stressors among the participating farmers showed that most of both the Spray and No Spray groups reported no/little stress on work-related and social-related matters, whereas they reported moderate/very stressful for finance-related matters (Table 2). The top four farm stressors perceived as moderate/very stressful among the participants include poor weather (69%), inadequate crop market prices (69%), not enough money for day-to-day expenses (48%), and debt load (47%), respectively. Interestingly, a higher percentage of the No Spray members (14%) reported moderate/very stressful for illness related to agricultural chemical exposures compared to those of the Spray group (10%). Several perceived farm stressors among the participants were statistically significantly associated with higher (worse) scores for mental health disorders (Table 3) including: 1) work-related Injuries and illnesses; 2) balancing the many roles as a family member and a farmer; 3) working with extended family members on the farm operation; 4) having too much work for one person; 5) dealing with incompetent help, finding and supervising help; 6) distance from shopping centers/schools/recreation, etc.; and 7) not enough leisure time to spend with family.

Table 2. Comparison of perceived farm stressors between the Spray and No Spray groups of farmers.

Farm Stressor Indicators	Spray members (%)		No Spray members (%)		p-value*
	Not/Little Stressful	Moderately/ Very Stressful	Not/Little Stressful	Moderately/ Very Stressful
Work-related
Farm injuries and accidents	85	15	84	16	NS
The weather (inadequate/drought/flood, etc.)	26	74	37	63	0.065
Having too much and intense overall agricultural workload	92	8	93	7	NS
Concern over the future of the farm/ranch	93	7	95	5	NS
Not having the manpower to operate the farm	95	5	96	4	NS
Government policies	82	18	86	14	NS
Injuries caused by dealing with machines, mechanical equipment, and sharp objects	91	9	90	10	NS
Illnesses caused by dealing with hazardous chemicals (pesticides, chemical fertilizers)	90	10	86	14	NS
Problems with balancing the many roles I perform as a family member and a farmer	95	5	94	6	NS
Problems with machinery (purchases, repairs, breakdowns)	90	10	90	10	NS
Problems with crops (illness, disease, noxious weeds, rodents)	62	38	76	24	0.013
Working with extended family members in the farm operation (parents, in-laws, children)	99	1	98	2	NS
Having too much work for one person	97	3	96	4	NS
Dealing with non-relative help (incompetent help, finding good help, supervising help)	98	2	99	1	NS
Finance-related
Market prices for your crops	28	72	35	65	NS
Not enough money for day-to-day expenses	47	53	58	42	0.067
High debt load	50	50	57	43	NS
Taxes	97	3	98	2	NS
Not enough cash/capital for unexpected problems (illness, health care, breakdowns, etc.)	78	22	86	14	0.084
Financing for retirement	79	21	87	13	0.078
Social-related
Distance from shopping centers/school/recreation centers, etc.	90	10	90	10	NS
Lack of close neighbors	97	3	94	6	NS
Limited social interaction opportunities	95	5	91	9	NS
Not enough time to spend together as a family in recreation	90	10	93	7	NS
Distance from hospitals, clinics, doctors	98	2	98	2	NS

p-values were calculated using Chi-squared test (No Spray vs. Spray groups), *p-value < 0.1; NS = not statistically significant.

Table 3. Correlation between binary categories of perceived farm stressors (not/little vs. moderately/very stressful) and binary levels mental health disorders (no/little vs moderate/severe symptoms) among the participants.

Farm Stressor Indicators	Chi-squared p-values*
Work-related	ST5-stress	DASS21-stress	DASS21-depression	DASS21-anxiety
Farm injuries and accidents	NS	NS	0.025	0.099
The weather (inadequate /drought/flood, etc.)	NS	NS	NS	NS
Having too much and intense overall agricultural workload	0.081	NS	NS	NS
Concern over the future of the farm	<0.001	NS	0.027	NS
Not having the manpower to operate the farm	0.001	NS	0.027	NS
Government policies	0.008	NS	NS	NS
Injuries caused by dealing with machines, mechanical equipment, and sharp objects	<0.001	0.041	0.018	0.005
Illnesses caused by dealing with hazardous chemicals (pesticides, chemical fertilizers)	0.079	NS	0.041	0.019
Problems with balancing the many roles I perform as a family member and a farmer	0.001	NS	0.002	0.018
Problems with machinery (purchases, repairs, breakdowns)	0.007	0.045	0.003	0.032
Problems with crops (illness, disease, noxious weeds, rodents)	NS	NS	NS	NS
Working with extended family members in the farm operation (parents, in-laws, children)	0.075	0.058	0.010	0.001
Having too much work for one person	<0.001	0.007	0.007	0.003
Dealing with non-relative help (incompetent help, finding good help, supervising help)	0.006	0.044	0.010	NS
Finance-related
Market prices for your crops	NS	NS	NS	NS
Not enough money for day-to-day expenses	0.064	NS	0.016	0.004
High debt load	0.041	NS	0.015	0.010
Taxes	NS	NS	NS	NS
Not enough cash/capital for unexpected problems (illness, health care, breakdowns, etc.)	0.054	NS	0.047	0.064
Financing for retirement	NS	NS	NS	NS
Social-related
Distance from shopping centers/school/recreation centers, etc.	NS	NS	NS	NS
Lack of close neighbors	0.009	NS	NS	NS
Limited social interaction opportunities	0.002	NS	NS	NS
Not enough time to spend together as a family in recreation	<0.001	0.002	<0.001	<0.001
Distance from hospitals, clinics, doctors	0.002	0.073	0.001	0.065

The 9Q-depression was not included in the Chi-squared test as all responses reported in the 9Q questionnaire were no/little of symptoms, with zero response reported moderately/severe symptoms. *p-value < 0.1; NS = not statistically significant.

Participant mental health disorders

Participants were asked to complete the survey questions related to stress, depression, and anxiety symptoms. The prevalence of mental health disorders i.e., stress, depression, and anxiety among participating Thai farmers was found to be 3% for moderate to severe depression symptoms, 9% for moderate to severe anxiety symptoms, and 12% for moderate to severe stress symptoms. The mean mental health disorder scores of the Spray and No Spray groups were not significantly different. The COVID-related stress scores also were not significantly different among these two groups (Table 4). The correlations between participants’ mental health disorder score fractions from different questionnaires were computed. Both stress score fractions from ST5 and DASS21 and depression score fractions from 9Q and DASS21 had some degree of correlation with r = 0.5.

Table 4. Comparison of mental health disorder scores between No Spray and Spray groups.

Mental Health Disorder Parameters [Total Scores]	% of farmers reported moderate/severe levels of mental health symptoms	Scores mean ± sd [range]			p-value*
		Total (N = 270)	No Spray (n = 124)	Spray (n = 146)
ST5-stress [15]	12	2.1 ± 2.3 [0-14]	2.2 ± 2.5 [0-14]	1.9 ± 2.0 [0-9]	NS
DASS21-stress [21]	4	1.6 ± 2.4 [0-19]	1.7 ± 2.6 [0-19]	1.6 ± 2.1 [0-8]	NS
9Q-depression [27]	0	2.1 ± 2.5 [0-13]	2.3 ± 2.7 [0-13]	1.9 ± 2.4 [0-10]	NS
DASS21-depression [21]	3	1.0 ± 1.8 [0-15]	1.2 ± 2.2 [0-15]	0.9 ± 1.5 [0-7]	NS
DASS21-anxiety [21]	9	1.0 ± 1.6 [0-11]	1.1 ± 1.7 [0-11]	1.0 ± 1.5 [0-6]	NS
COVID Stress Scores (CSS) [92]	54	26.1 ± 9.6 [2-58]	26.7 ± 9.5 [5-58]	25.6 ± 9.7 [2-58]	NS
CSS-phobia [44]	44	17.6 ± 6.9 [0-37]	17.9 ± 6.6 [0-34]	17.2 ± 7.1 [0-37]	NS
CSS-stress [24]	29	1.3 ± 2.3 [0-9]	1.4 ± 2.3 [0-9]	1.3 ± 2.3 [0-15]	NS
CSS-checking [24]	40	7.2 ± 3.1 [0-22]	7.4 ± 3.3 [0-22]	7.1 ± 2.9 [0-16]	NS

p-values were calculated using t-test (No Spray vs Spray groups), *p-value < 0.1; NS = not statistically significant.

Potential factors associated with mental health disorders among participants

Univariate analyses of mental health disorder scores were conducted and documented in Table 5. The 9Q-depression was not included in the analyses as all responses reported in the 9Q questionnaire had no/a little for symptoms, with zero responses reporting moderate/severe symptoms. Therefore, the univariate logistic regression analysis could not be performed. Factors that showed a significant positive association with levels of self-reported mental health disorders including being female, having a second job (if reporting farmer as the primary job), and having applied pesticides for greater than 20 years. Factors that had a significant negative association with levels of self-reported mental health disorders included having higher (good) agricultural practice scores, and having the higher (good) PPE use scores. In addition, it was observed that with an increase in the perceived farm stressor score fractions, the odds ratio of self-reported mental health disorder scores increased. Selected multiple logistic regression models were documented in Table 6. The estimated odds ratio of self-reported mental health disorder symptoms suggested no significant differences between the Spray and No Spray groups. The significant differences in self-reported anxiety (DASS21) were observed between the Spray and No Spray groups after adjusted for parameters including gender, agricultural practice score fraction, and PPE use score fraction. The estimated odds ratio of self-reported mental health disorder symptoms suggested with an increase in the perceived farm-related perceived score fractions, the odds ratio of self-reported mental health disorder scores increased.

Table 5. Univariate logistic regression analysis of four mental health disorder scales where survey scores categorized as moderate/severe vs. no/little.

Parameters	ST5-stress		DASS21-stress		DASS21-depression		DASS21-anxiety
	OR (95%CI)**	p-value*	OR (95%CI) **	p-value*	OR (95%CI) **	p-value*	OR (95%CI) **	p-value*
Gender (male vs. female)	3.966 (1.788,9.732)	0.001		NS	3.450 (1.003,15.814)	0.070		NS
Education (primary school and less vs. secondary/vocation school and above)	0.345 (0.146,0.748)	0.010		NS		NS		NS
Having a secondary job with a primary job as a farmer (yes vs. no)	4.914 (2.060,11.615)	<0.001		NS	6.571 (1.662,27.718)	0.007	6.532 (2.331,18.679)	<0.001
Farming Hours		NS		NS	1.036 (1.001,1.070)	0.032	1.030 (0.998,1.056)	0.057
Living within a 1-km radius of an area where there is pesticide application (yes vs. no)		NS		NS	4.036 (1.173,18.507)	0.040		NS
Agricultural practice score fractions	0.033 (0.002,0.803)	0.032		NS	0.008 (0.000,0.790)	0.030	0.008 (0.000,0.339)	0.009
PPE use score fractions	0.498 (0.224,1.070)	0.077		NS	0.291 (0.063,1.032)	0.073	0.145 (0.033,0.455)	0.003
Historical years of engaging in pesticide spraying	5.095 (2.130,11.853)	<0.001		NS	5.842 (1.635,19.562)	0.004	5.973 (2.161,15.898)	<0.001
Farm-related stressor score fractions	1.075 (1.047,1.107)	<0.001	1.082 (1.028,1.145)	0.003	1.071 (1.035,1.111)	<0.001	1.075 (1.044,1.111)	<0.001
Financial-related stressor score fractions	1.030 (1.009,1.053)	0.006	1.064 (1.010,1.125)	0.022	1.045 (1.012,1.081)	0.007	1.045 (1.018,1.074)	0.001
Social-related stressor score fractions	1.061 (1.036, 1.089)	<0.001	1.084 (1.023,1.158)	0.007	1.059 (1.024,1.097)	<0.001	1.063 (1.034,1.094)	<0.001

A variety of demographic, farming practice and farm stress indicators were examined as predictors.

p-values were calculated using logistic regression, *p-value < 0.1; NS = not statistically significant, **estimated odds ratio.

Table 6. Selected multiple logistic regression models of four mental health disorder scales where survey scores categorized as moderate/severe vs. no/little among the participants.

Adverse Mental Health Outcomes	Variables	Estimated odds ratio (95%CI)	p-values	AIC
ST5-stress	Spray vs. No Spray	1.030 (0.496,2.166)	0.937	204.25
	Spray vs. No spray adjusted for gender (male vs. female) and having a secondary job with a primary job as a farmer (yes/no)	1.608 (0.638,4.255)	0.322	154.12
	Spray vs. No spray adjusted for gender (male vs. female), having a secondary job with a primary job as a farmer (yes/no), and COVID-related stress scores	1.626 (0.618,4.506)	0.333	144.54
	Farm-related perceived stress score fractions	1.075 (1.047,1.107)	<0.001*	174.41
	Farm-related perceived stress score fractions adjusted for gender (male vs. female) and Agricultural practice score fraction (≤ mean score fraction vs. > mean score fraction)	1.069 (1.037,1.106)	<0.001*	158.04
	Farm-related perceived stress score fractions adjusted for gender (male vs. female), Agricultural practice score fraction (≤ mean score fraction vs. > mean score fraction), and COVID-related stress scores	1.069 (1.033,1.110)	<0.001*	160.04
DASS21-stress	Spray vs. No Spray	0.278 (0.014,2.201)	0.270	44.15
	Spray vs. No Spray adjusted for gender (male vs. female) and Historical years of engaging in pesticide spraying (≤ 20 vs. >20 years)	0.164 (0.005,2.097)	0.224	45.95
	Spray vs. No Spray adjusted for gender (male vs. female), Historical years of engaging in pesticide spraying (≤ 20 vs. >20 years), and COVID-related stress scores	0.171 (0.004,2.336)	0.248	45.69
	Farm-related perceived stress score fractions	1.082 (1.028,1.145)	<0.001*	37.24
	Farm-related perceived stress score fractions adjusted for gender (male vs. female) and Historical years of engaging in pesticide spraying (≤ 20 vs. >20 years)	1.075 (1.021,1.141)	0.007*	40.72
	Farm-related perceived stress score fractions adjusted for gender (male vs. female), Historical years of engaging in pesticide spraying (≤ 20 vs. >20 years) and COVID-related stress scores	1.078 (1.006,1.161)	0.045*	42.71
DASS21-depression	Spray vs. No Spray	1.199 (0.373,4.142)	0.762	102.09
	Spray vs. No Spray adjusted for gender (male vs. female), and having a secondary job with a primary job as a farmer (yes/no)	2.475 (0.515,17.971)	0.293	74.46
	Spray vs. No Spray adjusted for gender (male vs. female), having a secondary job with a primary job as a farmer (yes/no), and COVID-related stress scores	2.444 (0.458,19.514)	0.329	67.10
	Farm-related perceived stress score fractions	1.071 (1.035,1.111)	<0.001*	87.79
	Farm-related perceived stress score fractions adjusted for gender (male vs. female) and Historical years of engaging in pesticide spraying (≤ 20 vs. >20 years)	1.057 (1.014,1.103)	0.007*	69.07
	Farm-related perceived stress score fractions adjusted for gender (male vs. female), Historical years of engaging in pesticide spraying (≤ 20 vs. >20 years), and COVID-related stress scores	1.028 (0.977,1.081)	0.264	66.94
DASS21-anxiety	Spray vs. No Spray	1.634 (0.647,4.475)	0.312	145.33
	Spray vs. No Spray adjusted for gender (male vs. female), Agricultural practice scores (≤ mean score fraction vs. > mean score fraction), and PPE use score fraction (≤ mean score fraction vs. > mean score fraction)	2.877 (1.000,8.937)	0.056*	121.08
	Spray vs. No Spray adjusted for gender (male vs. female), Agricultural practice scores (≤ mean score fraction vs. > mean score fraction), PPE use score fraction (≤ mean score fraction vs. > mean score fraction), and COVID-related stress scores	2.783 (0.947,8.821)	0.051*	119.03
	Farm-related perceived stress score fractions	1.075 (1.044,1.111)	<0.001*	123.73
	Farm-related perceived stress score fractions adjusted for gender (male vs. female), Agricultural practice scores (≤ mean score fraction vs. > mean score fraction), and PPE use score fraction (≤ mean score fraction vs. > mean score fraction)	1.068 (1.030,1.112)	<0.001*	112.16
	Farm-related perceived stress score fractions adjusted for gender (male vs. female), Agricultural practice scores (≤ mean score fraction vs. > mean score fraction), PPE use score fraction (≤ mean score fraction vs. > mean score fraction), and COVID-related stress scores	1.062 (1.020,1.112)	0.005*	113.92

p-values were calculated using logistic regression, *p-value < 0.05.

As known, COVID-19 impacted individuals’ mental health; therefore, pandemic-related stress was considered. The CSS scores among the Spray and No Spray groups were not significantly different. Of most concern among participants was the fear of catching the COVID virus and not receiving treatment and healthcare services after being infected, followed by concerns about not receiving updated information related to COVID etiology, transmission, symptoms, treatment, and available healthcare services and resources. Most participants (79-84%) reported no mental health disorder symptoms related to the COVID pandemic.

Discussion

The uniqueness of the present study is the investigation of the farm stressor inventory among Thai farmers. Mental health disorders are complicated and could be driven by several factors including an individuals’ geographic location and culture, their socioeconomic status and associated lifestyle, accessibility to healthcare, as well as their health risk perceptions. Our pilot study validated the feasibility of employing the FSS among Thai farmers. The FSS expanded an exploration on determinants of mental health i.e., occupational- and non-occupational- related stressors among farmers. Previous studies in the US, Ireland, and Nigeria suggested several factors causing personal stress among farmers and ranchers including concerns of bad weather, poor agricultural product market prices, time pressures, and personal finances which are aligned with the results observed in the present study (Kearney et al. 2014; Olowogbon et al. 2019; Rudolphi et al. 2020; Brennan et al. 2022). These findings also reinforce the results from a previous study suggesting significant associations between climate change risk perception and farmers’ distress and anxiety (Howard et al. 2020). The top four farm stressors perceived as moderate/very stressful among the participants include poor weather (69%), inadequate crop market prices (69%), not enough money for day-to-day expenses (48%), and debt load (47%), respectively. Interestingly, a higher percentage of the No Spray members (14%) reported moderate/very stressful for illness related to agricultural chemical exposures compared to those of the Spray group (10%). This could imply that the farmers having a concern over illnesses related to chemical exposures were less likely to engage in pesticide spraying.

The present study examined the prevalence of three of major mental health disorders i.e., stress, depression, and anxiety capturing different mental health states among Thai farmers. The prevalence of mental health disorders i.e., stress, depression, and anxiety among participating Thai farmers was found to be 3% for moderate to severe depression symptoms, 9% for moderate to severe anxiety symptoms, and 12% for moderate to severe stress symptoms. Among three mental health parameters, the prevalence of stress was highest, followed by anxiety and depression. Stress is a natural human response that typically caused by external triggers i.e., short-term and/or long-term individuals’ life challenges and threats. Depression and anxiety are beyond regular mood changes and feelings about everyday life. Individuals can experience these two symptoms over an extended period even though there are no current stressors present (American Psychology Association 2022; World Health Organization 2023a, 2023b). This could imply that participants were likely encountered any particular situations in life e.g., occupational and non-occupational stressors in addition to COVID pandemic causing stress. Then, some participants developed mental health symptoms over an extended period. The results of the present study show some degree of alignment with previous studies. According to the World Health Organization, approximately 2% of Thai population experience depression (World Health Organization 2023c). Another study conducted in Chiang Mai province, in northern Thailand, reported the prevalence of the top-three ranked neurotic symptoms and probable mental health disorders i.e., sleeping problems (8%), headache (6%), and lack of appetite (4%) (Ong-Artborirak et al. 2022).

The association between occupational pesticide use and mental health disorders was evaluated in the present study. Previous studies have suggested that agrochemicals including pesticides are a crucial factor causing both short-term and long-term adverse health outcomes (Hu et al. 2015; Kim et al. 2017; Tudi et al. 2022). In the current study, some adverse symptoms that could be related to pesticide exposures were reported such as dizziness, perspiration, and fatigue. However, these symptoms could be caused by other factors including working for long hours in the hot environment and engaging in an intense workload, as well as working under stressful circumstances. Participants neither reported adverse health effects specifically associated with pesticide exposures nor reported that they were clinically diagnosed with pesticide poisoning within one year prior to the interview date. Our study only captured participants’ self-reported adverse health effects within the past year in order to minimize recall bias. Neither clinical examinations nor pesticide biomarker monitoring were conducted, therefore, asymptomatic health effects and/or biological indicators of pesticide exposures were not identified.

The results from the present study found that most participants in both the Spray and No Spray groups reported no/little symptoms of mental health disorders. In addition, no significant association was observed between occupational pesticide use via reported pesticide spraying and mental health disorders among the study population. However, the logistic regression for Dass21-anxiety at a moderate/severe level was significant for the Spray vs. No Spray groups when controlled for confounding by gender, agricultural practice, and PPE use. Several previous studies have suggested occupational exposures to a range of types of pesticides and exposure levels were associated with farmworker mental health disorders as well as sleep disorders (Kim et al. 2013; Harrison and Mackenzie Ross 2016; Li et al. 2019; Buralli et al. 2020). The limited findings in this study may be explained by pesticide exposure misclassification. In the present study, the levels of occupational pesticide exposure were classified based on self-reported engagement of occupational pesticide spraying within the past year. However, other work-related activities, i.e., agricultural practices, personal protective equipment (PPE) use, PPE use in the past, and household pesticide use, possibly impacted individuals’ overall pesticide exposure levels. These could over or underestimate the degree of occupational and non-occupational pesticide exposure. The total amounts of household chemical pesticides used were not captured, nor were pesticide exposure concentrations or biomarker levels quantified. We observed that 50% of the No Spray group reported that they engaged in pesticide mixing for other farmers who performed pesticide spraying. Moreover, approximately 70% of the No Spray group reported that they were sometimes and/or often working in the areas while pesticide spraying was occurring. These could enhance levels of pesticide exposures among the No Spray members, resulting in misclassification. On the other hand, both the Spray and No Spray groups had moderate PPE use score fractions during the past year, PPE use could lessen the total occupational pesticide exposures and lower levels of mental health disorders. In the case of historical PPE use (prior to the past year), the Spray group’s mean PPE use score fraction was approximately two times higher than that of the No Spray group. This could imply that the Spray group members’ long-term cumulative exposure concentrations of occupational chemicals (including pesticides and fertilizers) were likely to be the same or lower than those of the No Spray group members. Good work practices and PPE use could minimize exposures to agrochemicals among the participants, particularly the Spray group members, and may have decreased the impacts of pesticide exposures on mental health overall. These findings agree with a previous study conducted in Thailand ​​by Kaewboonchoo and colleagues suggested that programs on preventing chemical intoxication and improving family communication skills decreased depressive disorders among rice farmers. That study emphasized the significance of implementing good practices in chemical use in order to minimize occupational chemical exposures resulting in lower self-reported depression symptoms (Kaewboonchoo et al. 2020). Our study observed that most of the participants reported no/little stress which was not in agreement with a previous study conducted in another area in Thailand where a high prevalence of occupational stress was reported among Thai rice farmers (Pintakham et al. 2019). That study focused on job-related content under the framework of the Demand-Control-Support Model which covered six aspects of work i.e., job control; psychological job demand; physical job demand; job security; social support; and hazards at work. The authors suggested that a high percentage of participating farmers reported stress from job control and psychological job demand. Our study incorporates expanded stressors potentially impacting overall individuals’ mental health including those that are work-related i.e., occupational and environmental hazards, as well as financial and social-related factors. Our results highlighted that weather conditions and finance-related factors were perceived as the most stressful factors among the participants, while other farm and social-related factors were perceived as less stressful. These perceptions could, in turn, lessen the association between occupational stress and mental health disorders overall. However, the univariate and multiple logistic regression analyses suggested that the perceived farm stressor had a significant association with mental health disorders. The results suggested that the participants with relatively high perception of the farm stressors were likely to have mental health disorder symptoms. Nevertheless, the use of the farm stressor inventory shows promise for creating effective programs/policies for preventing mental health disorders and improving overall well-being among farmers and their community.

The present study has a number of limitations including potential pesticide misclassification as described above as well as the measurement of individuals’ mental health disorders utilizing self-reported surveys. We assumed that the participants clearly understood the questions, as they were interviewed by research assistants experienced in collecting interview survey data, however, these self-reported data have not been validated by other measures i.e., using biological information, medical examinations, or any type of personal journal. The participants were purposively sampled and located in a specific area of Thailand; therefore, the results might not be generalizable to the whole country. However, the data from this pilot study has provided instructive information for future development of effective mental disorder intervention programs for Thai farmers. Future studies should incorporate quantitative measurements of pesticide exposure levels, in addition to determining resilience, the ability of successfully cope with stress, in Thai farmers across broader locations. Therefore, extensive information will be obtained in order to develop appropriate exposure/health risk mitigation measures for improving farmers’ mental health and overall well-being.

Conclusions

This study is one of the first to investigate farm stressors among Thai farmers. Individuals’ mental health disorders are complicated. This pilot study explored and determined several occupational and non-occupational factors that potentially impact mental health disorders among Thai farmers. The findings suggested the top farm stressors perceived as moderately or very stressful among the participants include bad weather conditions and finance-related matters i.e., inappropriate crop market prices, not enough money for day-to-day expenses, and being indebted. Good work practices and PPE use could be one of the essential factors minimizing occupational chemical exposures and lowering the potential of having mental health disorders among Thai farmers. Further studies that identify and evaluate farm stressors and farmer mental health in other regions in Thailand are needed.

Author contributions

Saisattha Noomnual conceptualized the core idea of the research, conducted data collection and data analyses, and prepared the drafts of this manuscript. Pajaree Konthobut participated in survey development and data collection, in addition to reviewed and formatted the manuscript. Pornpimol Kongtip and Susan Woskie supervised the research, finalized the idea of data analyses, reviewed, and edited the manuscript.

Acknowledgements

Research reported in this publication was supported by the Fogarty International Center of the National Institutes of Health under Award Number U2RTW010088. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Disclosure statement

No potential conflict of interest was reported by the author(s).