1. Introduction

Each year, about 720,000 people die due to suicide, with 73% of these deaths occurring in low- and middle-income countries, 2021 [1]. There is a demographic and geographical differentiation in suicidal behaviour worldwide. While current crude suicide rates per 100,000 per population are generally lower in Africa than the global average, the trend reveals a consistent rise in suicide levels in the region. Global crude suicide rates fell from 11.8 in 2005 to 9.8 in 2015 and 9.2 in 2021, while Africa recorded a rise in suicide from 6.7 in 2005 to 6.8 in 2015 and 7.3 in 2021 [1]. Empirical evidence further shows that rural areas tend to have the highest suicide and suicidal ideation levels, with farmers being disproportionately at risk [2,3]. Across both the global North and South, recent evidence shows that agricultural workers have higher suicide ideation and suicide rates than populations in other occupations [4]. For Africa, suicide rates among farmers in rural areas may even be much higher than what is reported due to reporting and data limitations [5].

While there is a general well-established link between suicidal behaviours and mental disorders [6,7], many suicides occur spontaneously from situations of crisis due to the inability to deal with a sudden onset of stress from a life event [8]. For farmers in Africa who rely predominantly on rainfed agriculture, these life events can manifest through multiple sudden ecological risks, including extreme climate events such as floods, strong winds, drought, bushfires, crop raiding by livestock, and postharvest food loss. These sudden stressors can elevate stress and cause distress among farmers due to income loss, indebtedness, the inability to repay loans, and worry about future food insecurity [9–12]. Specific to the semiarid northern savannah region of Ghana, recent studies document heightening climate change-induced ecological grief, which sometimes manifests as “feelings of anxiety, powerlessness, helplessness, hopelessness, and sadness” [13,14]. The availability and timely access to productive resources at the household and community levels, including access to climate information, postharvest loss training, familial support, and access to remittances have been found to be protective factors against suicide ideation and overall mental health of farmers [15].

Food loss is a potential sudden shock in the life of a farmer that has the potential to trigger stress levels both in the immediate aftermath and in the long-term due to related pressures like food insecurity. Yet, globally, about one-third of all food produced by farmers is lost before it reaches the consumer. The rate of food loss is much higher in Africa (40%) [16], where a blend of factors, including the lack of storage infrastructure, poor transportation networks, and access to markets, remains a key driver [17,18]. For farmers who produce perishable crops like vegetables and fruits, the rate of food loss can go as high as 50%, with women farmers often bearing a disproportionate brunt of these losses [19,20].

Despite the scale and implications of food loss, limited attention has been paid to how such losses contribute to psychological stress and suicidal thoughts among farmers. This represents a critical gap in the literature. Postharvest food loss can undermine a farmer’s sense of productivity and personal agency, especially when such loss results in income shocks, unmanageable debt, or the inability to feed one’s household. These conditions can lead to chronic stress, hopelessness, and in some cases, suicidal ideation. Importantly, food loss is not only a material setback but also a deeply emotional one, particularly in contexts where farming is tied to identity, masculinity, and social status. Thus, understanding the food loss–mental health nexus is essential for designing responsive agricultural and public health interventions.

Food loss has the potential to trigger stress and suicidal ideation. In extreme cases where food loss levels are high or where farmers lose the entire harvest, stress can be sudden, and without immediate protective resources, can trigger suicidal attempts due to feelings of loss. Food loss may also trigger prolonged stress due to other pressures associated with its occurrence. For instance, food loss is known to drive financial and food insecurity, which in the long-term can lead to elevated stress and anxiety [19,20]. Further, postharvest food loss can trigger a market response whereby agricultural produce retailers offer lower prices to farmers for their remaining produce. This happens during the peak harvesting season when there is a produce glut in the market. In Ghana, for instance, the price of a basket of tomatoes plummets from about $150 in the lean season to just $10 in the peak harvesting season, and farmers must accept these low prices or risk losing the entire harvest due to the lack of cold storage technologies [21]. Selling at low prices means most farmers are unable to repay loans with evidence showing some farmers consider suicide to shorten the stress from indebtedness [5].

Notwithstanding these potential links between food loss and suicidal behaviours, this relationship remains underexplored in Africa, where suicide rates are observed to be rising. Suicides are preventable hence a better understanding of the risk and protective factors of suicide ideation such as food loss could be a useful entry point.

In this study, we focus on smallholder farmers, defined as individuals or households cultivating less than two hectares of land, using predominantly family labor, and relying primarily on farming as their main source of livelihood [19]. In this contribution, we draw data from a cross-sectional survey with smallholder farmers in the semi-arid northern Ghana to explore the association between postharvest food loss and suicidal ideation. Alongside highlighting the multiple pathways through which food loss is connected to suicidal behaviours, our contribution discusses other key risks and protective factors for suicidal ideation. We conclude with recommendations highlighting how these protective factors can be leveraged to minimize suicide risk in a proactive rather than a reactive policy environment.

2. Methodology

2.1. Study context

This study was conducted in Ghana’s Upper West Region (UWR), situated in the semi-arid northern savannah ecological zone of the country as shown in Fig 1. The region shares borders with Burkina Faso to the north and west, covering an area of approximately 18,476 square kilometers [22]. With a population of over 901,500 people, the region is predominantly rural, and around 80% of its residents rely on subsistence agriculture for their livelihoods [22,23]. It ranks as one of the most impoverished regions in Ghana, recording the third-highest Multidimensional Poverty Index (MPI) score of 0.348, markedly above the national average of 0.112 [23]. More than 65% of the region’s population lives in multidimensional poverty, with high incidences in districts like Wa West (61.9%), Wa East (48.7%), Lambussie (44.2%), Nadowli-Kaleo (40.6%), and Daffiama-Bussie-Issa (38.7%) [23].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Modified UWR map showing the study area.

https://doi.org/10.1371/journal.pstr.0000223.g001

The region is challenged with low industrial development, limited access to vital social services, and a dependence on rain-fed agriculture [22,23]. Situated within the Sudan and Guinea Savannah ecological zones, the area experiences a single annual rainy season from May to October. However, climate change has made rainfall increasingly unpredictable, causing longer dry periods, shorter growing seasons, and heightened challenges in postharvest handling [24]. Annual rainfall averages between 840 mm and 1,400 mm, while daytime temperatures during the dry season can rise above 35°C, conditions that contribute to the rapid deterioration of perishable crops in the absence of cooling infrastructure [25]. The region is also highly vulnerable to a range of climatic threats, including droughts, floods, storms, pest infestations, and diseases.

Postharvest loss is a widespread and persistent issue in UWR, exacerbating food insecurity and financial instability for smallholder farmers [26]. Losses often occur due to poor road infrastructure, lack of storage facilities, and inadequate access to markets and extension services [17]. Despite policies to improve agricultural outcomes, many farmers report receiving little to no tangible support, and inefficiencies, poor coordination, and insufficient funding undermine several assistance programs [23]. Psychological distress, including anxiety, depression, and suicidal thoughts, outcomes driven by recurrent losses, debt, social pressure, and the uncertainty of farming as a livelihood, often and silently drives many smallholder farmers in this context into poor mental wellbeing [27].

Nevertheless, these psychological challenges associated with agricultural stress remain largely invisible within policy discourse. In many rural communities in this study context, suicide is often linked to witchcraft instead of the psychosocial stress that smallholder farmers face. The region also lacks mental health services and trained professionals, making it nearly impossible for those in distress to seek or receive appropriate care [27]. Gender dynamics further compound these challenges. Women, who are integral to farming activities, are often disproportionately affected due to limited access to land, credit, and decision-making processes, while also bearing the brunt of household food provision and caregiving responsibilities [28]. These intersecting burdens place women farmers at a higher risk of both economic insecurity and mental health strain [27].

2.2. Data collection

Ethics statement.

This study received approval from the University of Western Ontario Non-Medical Research Ethics Board (NMREB) under Project ID 124838. All participants provided informed consent through an implied consent process, approved by the NMREB, aligning with inclusive global research practices (see S1 Checklist). This study draws on a cross-sectional survey of 1,033 smallholder farmers in Ghana’s UWR who grow cereals, legumes, and root and tuber crops. Data were collected between July 4, 2024, and July 4, 2025. For the current analysis, the dataset was accessed on July 30, 2025. Before completing the survey, participants received a Letter of Informed Consent that described the study’s purpose, procedures, possible risks and benefits, confidentiality measures, and their rights, including the right to withdraw at any time. By voluntarily completing and returning the survey, participants consented. To strengthen trust and transparency, this process was carried out in the presence of either a spouse, another adult household member, or a local community leader (such as an Assembly or Unit Committee member), who were also briefed on the study. No signatures or identifying details were collected to ensure participant anonymity, thereby creating a safe environment for them to share openly. This approach was deemed appropriate, given the minimal risks associated with the study, and met the approved ethical standards. The survey collected information on crop production and food loss for cereals, legumes, and root and tuber crops, with a particular focus on how food waste and loss can influence suicidal ideation, mental and physical health, postharvest management, food loss rates, and agroecological practices. It also included data on household demographics, socioeconomic conditions, agricultural activities, and climate resilience.

A multi-stage sampling design was used. First, five districts, Nadowli-Kaleo, Daffiama-Bussie-Issa (DBI), Lambussie, Wa East, and Wa West, were purposively selected due to their large populations of smallholder farmers, high poverty levels, and vulnerability to climate-related risks. Within each district, farming communities were chosen randomly to ensure broad representation. Because the total population was large and undefined, Cochran’s formula for an infinite population was applied to determine the sample size [29]. Within the selected communities, systematic random sampling was used to identify households: starting at the community entrance with a randomly chosen house, every fifth household was surveyed until the sample target was reached. In each selected household, the primary male or female farmer aged 18 or older responded on behalf of the household. The sample size was calculated using Cochran’s formula for an infinite population.

Where:

= Required sample size

Z = 1.96 (for 95% confidence level)

p = 0.5 (assumed proportion, as the actual proportion is unknown)

e = 0.03 (margin of error)

Substituting these values:

Thus, the required sample size was 1,067 households. To account for potential non-response and data inconsistencies, the sample size was further adjusted using the formula:

where is the expected non-response rate (assumed at 5% or 0.05):

Ultimately, the target sample size was set at 1,123 households. However, considering the availability and response rates, only 1,033 households were surveyed. Although this falls slightly short of the ideal sample size, it still accounts for over 96% of the required number and is statistically strong enough for the study’s goals.

2.3. Measures

The dependent variable for this study is “Suicidal Ideation.” In this study, suicidal ideation refers to the self-reported thoughts or considerations of ending one’s own life within the past 12 months. It is used as an indicator of psychological distress and mental health burden among rural smallholder farmers in northern Ghana, a population often exposed to multiple socio-ecological stressors such as climate shocks, food insecurity, and economic hardship. Participants were asked the question: “In the past 12 months, have you ever thought about taking your own life?” Responses were coded as: [0 = No – No reported suicidal thoughts and 1 = Yes – Reported suicidal thoughts]. This variable represents a critical endpoint in the mental health spectrum, reflecting an individual’s psychological resilience or vulnerability in the face of intersecting risks. In the socio-cultural context of the study area, where mental health is stigmatized and services are limited, suicidal ideation may go unnoticed or untreated.

Our focal independent variable is Postharvest Loss (PHL). PHL was quantified as a continuous variable derived from farmers’ self-reported estimates of the extent of crops lost after harvesting but before consumption or sale. Farmers were asked to quantify losses for three specific categories of crops: cereals (maize, rice, millet, and sorghum); roots and tubers (yams, cassava, and sweet potato); and legumes (groundnut, cowpea, soybean, and bambara beans). For each crop type, respondents answered the question: “What quantity of your total harvest for [crop type] was lost after harvest before it was consumed or sold?” The responses were recorded in quantity (kilograms), allowing for a quantifiable measure of the postharvest loss. The total PHL was calculated by summing the reported losses across all three crop categories, yielding an aggregate measure for overall food loss. “Self-Rated Ability to Manage Food Loss” is another key variable in our analysis. It was assessed through self-evaluation by farmers regarding their perceived capacity to manage, prevent, or reduce PHL. Respondents were asked to rate their ability to manage food loss using a six-point Likert scale with the following options: Excellent, Very Good, Good, Fair, Poor, and Very Poor. To facilitate analysis and focus on contrasting levels of perceived competence, the responses were recoded into a binary variable: Responses of Excellent, Very Good, and Good were coded as 0 = Good, indicating a higher self-perceived ability to manage food loss. Responses of Fair, Poor, and Very Poor were coded as 1 = Poor, reflecting a lower self-assessed ability to manage postharvest losses. This dichotomous coding approach allowed for a simplified analysis of perceived management ability in relation to postharvest loss outcomes and associated mental health indicators. Guided by existing literature in the study context [17,27,30], we controlled for a range of conceptually relevant variables as shown in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Predictors of suicidal ideation among smallholder farmers.

https://doi.org/10.1371/journal.pstr.0000223.t001

2.4. Data analysis

We employed a three-stage analysis. First, we conducted descriptive (univariate) statistics to summarize the distribution of the key study variables across three levels: individual, household, and farm. Second, we performed bivariate analyses to explore the independent association between each predictor variable and the dependent variable (suicidal ideation). Given the binary nature of the outcome variable (suicidal ideation), we employed a nested multiple logistic regression approach in the third stage of the analysis. To ensure that multicollinearity did not bias the regression estimates, we conducted a collinearity diagnostic using Variance Inflation Factors (VIF). All VIF values were below 2, indicating no serious multicollinearity among the independent variables. We specified three models:

• Model 1 (Individual-level model): This includes postharvest food loss, self-rated ability to manage food loss, and individual-level demographic controls such as age, gender, education, religion, marital status, marital structure, employment status, and debt.
• Model 2 (Household-adjusted model): This builds upon Model 1 by adding household-level variables such as household size, agricultural labor force, wealth quintile, and decision-making structure.
• Model 3 (Full model): This final model adds farm and contextual-level variables, including land size, distance to and from farm (round-trip), type of farm power, use of chemical inputs, farm visitation and transportation systems, cultivation season, extension access, exposure to climatic stressors, and district fixed effects.

This stepwise structure allows us to evaluate how the strength and significance of the association between postharvest food loss and suicidal ideation evolve as broader layers of explanatory variables are introduced. To assess model fit and compare the relative performance of the nested models, we report both Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) values. AIC and BIC decreased across the models, from AIC: 999.498 in Model 1 to 909.040 in Model 3, and from BIC: 1088.423 to 1156.052, indicating improved model fit as additional variables were included. The general logistic regression model, as proposed by Wang and Hu (2006), is expressed as [31]:

where denotes the probability of an observation belonging to the category of the dichotomous Y variable (where 1 indicates suicidal ideation). The exponential function is represented by “exp.” The symbol represents the intercept, denotes the coefficient of the first predictor variable, and represents the coefficient of the final predictor variable. Results are presented as odds ratios (ORs) with 95% confidence intervals (CIs). ORs greater than 1 (OR > 1) indicate a higher likelihood of suicidal ideation associated with a given predictor, while ORs less than 1 (OR < 1) suggest a protective effect. All statistical analyses were conducted using Stata 19.

3. Results

3.1. Univariate results

The univariate results are presented in Table 2. About 25% of smallholder farmers reported experiencing suicidal ideation, while 75% did not. Postharvest loss had a mean of 41.8 kilograms with a standard deviation of 121.6, reflecting substantial variation in reported losses across households. For Self-Rated Ability to Manage Food Loss, 33% of respondents rated their ability as good, while 67% rated it as poor. Most primary farmers are aged 30–39 (38%), with a smaller percentage (7%) over 60. Males represent 59%, and 54% have no formal education. Most are Christians (52%) and are married (86%). Households average 7.7 members, with 30% in the poorest wealth category and 16% in the richest. 19% of respondents reported receiving remittances, while 81% did not. For savings, 56% of respondents reported having access to savings, whereas 44% did not. Decision-making is primarily male-headed (46%) or joint (45%). Most farmers rely on manual labor (82%), and 51% utilize farm credit. A large majority (82%) use synthetic chemical inputs. For transportation, 52% rely on mechanical means, and 90% have a storage system. Most farmers (73%) report profits from crop sales despite limited postharvest training (57%). Most smallholder farmers cultivate only during the wet season (95%), with a few practicing dry season farming.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Demographic characteristics of smallholder farmers in Ghana.

https://doi.org/10.1371/journal.pstr.0000223.t002

3.2. Bivariate results

Table 3 presents the bivariate results. Postharvest loss is positively associated with suicidal ideation. For each additional kilogram of PHL, the likelihood of experiencing suicidal ideation increases by 1.1%. Several control variables were also associated with suicide ideation. Smallholder farmers who rated their ability to manage food loss as poor (OR: 2.293, p < 0.00) were more likely to report suicidal ideation compared to those who rated their ability as good. Similarly, individuals aged 30–39 were almost twice as likely to report suicidal thoughts compared to the reference group of 18–29 years (OR: 1.973, p < 0.001). Widowed, divorced, or separated individuals also had increased odds of suicidal ideation (OR: 1.787, p < 0.001) compared to married individuals. Despite the potential protective role of marriage, marital structure revealed some interesting dynamics. Those in polygamous relationships were more likely to experience suicidal thoughts (OR: 1.757, p < 0.001) than monogamous individuals. Smallholder farmers who were not employed in any other income-generating activity were significantly more likely to report suicidal ideation compared to those engaged in other employment activities (OR: 5.578, p < 0.001). Furthermore, the amount of debt owed in the past year was directly associated with suicidal ideation (OR: 1.000, p < 0.001). Additional household characteristics also impacted these odds; for instance, household sizes and agricultural labor forces were correlated with increased likelihoods of suicidal ideation (OR: 1.069, p < 0.001; OR: 1.084, p < 0.001, respectively). On the other hand, farmers from wealthier households showed a reduced likelihood of suicidal ideation. Those in the middle wealth category were less likely to report suicidal thoughts (OR: 0.526, p < 0.001), and this likelihood further decreased for poorer (OR: 0.396, p < 0.001) and poorest households (OR: 0.357, p < 0.001) compared to the richest households. Although counterintuitive, farmers with access to remittances (OR: 6.579, p < 0.001) and savings (OR: 1.923, p < 0.001) were more likely to experience suicidal ideation compared to those without access. Moreover, joint decision-making by the primary male and primary female in the household heads significantly reduced the odds of suicidal ideation (OR: 0.258, p < 0.001). For each additional kilometer in distance to the farm, the thoughts of suicidal ideation increased by 20.6% (OR: 1.206, p < 0.001), and the use of synthetic chemical inputs (OR: 2.570, p < 0.001). In contrast, access to farm credit (OR: 0.339, p < 0.001) and postharvest training (OR: 1.560, p < 0.001) were associated with lower suicidal thoughts. Lastly, experiencing climatic stressors, such as droughts (OR: 1.655, p < 0.001) and floods (OR: 3.259, p < 0.001), was significantly linked to increased suicidal ideation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Bivariate logistic regression analysis of the association between suicidal ideation and the predictor variables.

https://doi.org/10.1371/journal.pstr.0000223.t003

3.3. Multivariate results

This section presents the results of the nested logistic regression analysis. We employed three models, each building on the previous one by progressively adjusting for additional variables across individual (Model 1), household (Model 2), and farm/community (Model 3) levels. This nested structure allows us to assess how the relationship between postharvest loss and suicidal ideation changes as broader contextual factors are introduced, and to identify which types of variables have the most significant influence in moderating or reinforcing this association. The results are presented in Table 4. Postharvest Loss was consistently associated with increased odds of suicidal ideation across all models after adjusting for relevant sociodemographic and farm-level variables. In Model 1, each additional percentage of postharvest loss made respondents 0.7% more likely to experience suicidal ideation (OR: 1.007, p < 0.001). This effect persisted in Model 2 (OR: 1.004, p < 0.001) and Model 3 (OR: 1.003, p < 0.001), though with slightly reduced magnitude. Consistent with this finding, smallholder farmers who rated their ability to manage food loss as poor were more likely to report suicidal ideation than those who rated their ability as good. Similarly, farmers who reported a poor capacity in dealing with food loss were 2.5 times more likely (OR: 2.504, p < 0.001), in Model 2, they were nearly twice as likely (OR: 1.957, p < 0.001), and in Model 3, they were over twice as likely (OR: 2.258, p < 0.001). In terms of age, farmers aged 30–39 were still more likely to report suicidal ideation at the multivariate level compared to those aged 18–29 (OR: 2.065, p < 0.01). Marriage still had a potential moderating impact on suicide ideation at the multivariate level. Smallholder farmers who were widowed, divorced, or separated were more likely to report suicidal ideation compared to those who were married (OR: 2.673, p < 0.01). Remarkably, smallholder farmers with access to remittances were still more likely to report suicidal ideation at the multivariate level compared to those without access (OR: 2.887, p < 0.001). Shared decision-making by the primary male and female was associated with a lower likelihood of suicidal ideation (OR: 0.415, p < 0.001). A kilometer increase in distance to and from the farm was associated with a higher likelihood of suicidal ideation (OR: 1.149, p < 0.001). Access to farm credit was associated with less likelihood of suicidal ideation (OR: 0.659, p < 0.01). Farming only during the wet season was associated with less likelihood of suicidal ideation compared to farming in both dry and wet seasons (OR: 0.356, p < 0.01). Lastly, experiencing specific climatic stressors also showed significant associations. Smallholder farmers who experienced droughts (OR: 1.148, p < 0.01), floods (OR: 2.243, p < 0.001), and erratic rainfall, storm surges, dry spells, pests, or diseases (OR: 1.138, p < 0.01) were more likely to report suicidal ideation compared to those who did not report any climatic stressor. Smallholder farmers who used synthetic chemical inputs emerged more likely to experience suicidal ideation than those who did not at the multivariate level (OR: 2.838, p < 0.001).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Nested multiple logistic regression analysis of the suicidal ideation and postharvest food loss among smallholder farmers in Ghana.

https://doi.org/10.1371/journal.pstr.0000223.t004

4. Discussion

Suicide among farmers has received ample research attention yet remains one of the less understood issues. The World Health Organization (WHO) identified farming-related job stress as a frequent precursor to poor mental health and a risk factor for suicide [32]. Empirical studies focusing on farmer suicides have identified depression from shocks, hazardous work environments, easy access to pesticides, reduced access to emergency services, and social isolation as some of the key triggers [2,33]. This contribution builds on this budding body of work by highlighting the empirical connection between postharvest food loss and suicidal ideation. Postharvest food loss is a complex risk factor due to its sudden nature and the possibility for it to drive future events like food insecurity. The shock from a sudden loss of the harvest can push farmers into considering taking their lives due to the immediate feeling of hopelessness and indebtedness. Research shows that in situations where farmers financed their activities with loans, the thought of not being able to service the debt with lenders and the accompanying legal implications and social ridicule push farmers into considering or completing suicide [5,34]. The complexity of food loss as a possible trigger of suicide ideation may also stem from the long-term implications of its incidence. Beyond the immediate shock, postharvest loss can push farming households into food insecurity, especially in semi-arid environments like northern Ghana, where there is a single growing season and farmers must live with the impact of sudden crop losses for a year until the next harvest. This period of lack may trigger continuous feelings of hopelessness and eventually lead to depression and suicide ideation. This immediate and long-term manifestation of food loss-related stress highlights the need to closely monitor the sudden psychosocial and long-term material stress it produces. Thus, addressing the psychosocial needs of farmers who experience postharvest loss should be an immediate process following such sudden events, along with material interventions that alleviate the economic pressure caused by these losses.

The study also revealed that various demographic, household, and farming-related factors are linked to suicidal ideation among smallholder farmers in rural settings. A key finding is how a farmer’s ability to handle food loss is associated with suicidal ideation. Farmers who assessed their capacity to handle food loss as poor were more likely to report suicidal ideation. Given that postharvest management is a resource-intensive process shaped by access to efficient transportation and storage technologies, a lack of access to these resources could predispose a farmer to greater losses. In the Upper West Region of Ghana, in particular, storage options are limited, with most farmers relying on their personal traditional strategies to manage produce postharvest, which frequently fail [17]. The failure of these personal strategies is known to be connected to mental health struggles. For instance, in Benin and elsewhere, research shows that farmers who felt they had little control over postharvest issues also experienced high stress levels and alcoholism [35,36].

Age also plays a role, with farmers between 30 and 39 years old more likely to have suicidal ideation than younger farmers. In many rural farming communities, this age group often bears a greater share of familial responsibilities, including caring for their families, earning a living, and contributing to their communities. When farming becomes difficult and unstable, these responsibilities can feel overwhelming. Corroborating with research by King et al [37], farmers who are aging tend to experience more mental health issues due to various factors, including the pressure to succeed in both work and family life.

Consistent with empirical evidence [38,39], marital status emerged as another critical factor. Farmers who were widowed, divorced, or separated were more likely to experience suicidal ideation than those who were married. Generally, marriage provides emotional and material support in meeting daily responsibilities and confronting sudden shocks. Not having this partnership, whether due to death, separation, or not having a spouse, can lead to loneliness and more hardship in the face of a shock like food loss. In the study context, this is especially tough for women, who may lose access to productive resources, land, and support when a marriage ends [40–42]. This highlights the importance of social support in managing agricultural losses. Indeed, our finding that those in households where both men and women make decisions jointly were less likely to have suicidal thoughts than those in households where decisions are made solely by males buttresses this discussion on social support. Empirical research shows that when couples share decision-making, it promotes trust, cooperation, and emotional support, all of which are good for overall mental wellbeing [43]. While our findings align with this perspective, other studies in 2025, such as one in Mozambique, have found weaker or inconsistent associations between joint decision-making and mental health outcomes [44].

Our findings also demonstrate the potentially beneficial role of livelihood diversification in lowering suicide ideation, as those without other sources of income outside farming were more likely to report suicidal ideation. In the Upper West Region, farming is predominantly seasonal and rain-fed, providing the primary source of income for many individuals. The absence of a second income can lead to financial difficulties especially when postharvest losses are high. Consistent with the literature, income diversification can provide a protective layer against sudden ecological and agricultural shocks, including postharvest loss [45,46]. A broader risk-spreading strategy for smallholder farmers could include diversifying to include other non-farm income-generating activities. Interestingly, farmers who received money from family members in the form of remittances were however, more likely to experience suicidal thoughts. Although remittances can ease financial pressure and have been linked to improved physical and mental wellbeing [27,47], remittances may also signal that the farmer is struggling. Specific to the sociocultural context and social construction of responsible adulthood, repeated or prolonged dependence on outside help can itself be viewed as a sign of incapability. This stigma can deepen when young relatives are those remitting. Thus, although remittances may provide short-term relief, emphasis should be on promoting personal financial freedom. support programs should not only provide financial assistance but also promote self-confidence and address concerns related to identity.

Some of the practical challenges of farming are also linked to suicidal ideation. Farmers who had to travel long distances to their fields were more likely to experience suicidal thoughts. Distance to the farm directly shapes postharvest management. Hauling farm produce manually on footpaths increases the levels of losses. During the peak rainy season, in particular, when most footpaths are inaccessible, harvested produce on farms far inland can decay in the field if not transported home on time for processing. When the distance to the farm is longer, the likelihood of rain beating harvested farm produce is high, which further exposes the harvest to aflatoxins and decay if there is no immediate access to sunlight. These issues are particularly compounded when the distance to the market is equally long [17]. Crops like maize and groundnuts start to germinate with such delays. Walking longer distances to the farm, especially alone, can also lead to feelings of isolation or anxiety. It is therefore not surprising that farmers who complete farm tasks with machines had lower odds of suicidal ideation than those who complete tasks manually [48]. Motorized transport has the potential to eliminate delays in produce transport and reduce crop losses.

Access to farm credit also helped reduce suicidal ideation. When farmers can borrow money, they can buy supplies, recover from setbacks, and avoid loan sharks. In the Upper West Region, where financial resources are often scarce, having access to credit can alleviate financial anxiety. Research found that credit programs (either through formal or informal sources like village savings and loans associations) lowered depression and food insecurity levels among farmers [16,49]. However, capacity building on financial literacy could position farmers to manage credit and investments effectively to avoid indebtedness, which in itself may trigger suicidal ideation. Credit facilities could be bundled with insurance products to provide farmers with an additional layer of protection in the event of significant postharvest losses.

Consistent with the literature, using chemical products like pesticides, including herbicides, insecticides, and weedicides, was linked to higher odds of suicidal ideation. Studies show that exposure to pesticides can lead to poisoning and elevated risk of neuropsychiatric sequelae, which manifest in mood disorders, depression, and anxiety [50,51]. Such disorders may impair the cognitive judgment of farmers, leading to suicide attempts and completed suicide. Further, in the study context, as in many parts of the Global South, pesticides are often used as the main chemical in self-poisoning or suicide. In situations of stress and cognitive impairment, the availability of pesticides and weedicides to farmers may contribute to suicidal thoughts and completed suicide [52,53].

Another interesting finding was that farmers who only farmed during the rainy season were less likely to report suicidal thoughts compared to those who practiced dry season farming in addition. While farming in both seasons can bring more income, dry season farming is particularly exposed to risks stemming from the lack of irrigation systems, pest infestations, and higher post-harvest losses. Farmers rely on hand-dug wells, which can run dry at any point in the growing cycle. Pest infestations are also more intense during the dry season, as only a limited number of farmers are cultivating. Generally, dry-season agriculture in the study context is limited to vegetables such as tomatoes and peppers, which have a relatively short lifespan and are more prone to postharvest losses. In similar regions in Ghana and Burkina Faso, year-round farming without enough support has also been shown to increase pressure and burnout [54,55]. Consistent with our finding on the potential protective role of income diversification in reducing suicide ideation, policy support for irrigation and extension could help unlock the potential of dry-season farming as an income diversification option in such semi-arid contexts.

Finally, our findings show that exposure to extreme environmental events such as floods, droughts, pests, and diseases was strongly linked to suicidal ideation, with floods having the most significant impact. These events can simultaneously destroy crops, homes, and income. Many farmers in the Upper West Region lack insurance or emergency assistance, leaving them to cope with the consequences of these events on their own. These findings corroborate empirical evidence, which shows that climate disasters can negatively impact mental health [56–58].

This study offers valuable insights into suicidal ideation among smallholder farmers in Ghana’s Upper West Region, but it is not without some limitations worth highlighting. It relies on self-reported data, which may be biased due to stigma or fear, potentially underestimating prevalence. Its cross-sectional design limits our interpretation to associations, requiring longitudinal studies to establish causality. Despite these issues, the research provides a foundation for policy and future research on mental health in smallholder agricultural settings.

5. Conclusions and policy directions

While suicidal ideation does not always result in completed suicide, it remains a critical indicator of potential suicide. For farmers who are exposed to sudden risk from environmental events such as food loss, addressing stress from these events remains crucial in addressing suicide among farmers. We recommend a two-pronged approach focused first on identifying suicide ideation and, second, addressing it. Agricultural events, though sudden, are largely seasonal. For instance, food loss mostly occurs during the harvesting season, hence scarce mental health resources can be channelled to provide intense farmer counselling during these seasons including targeting farmers who report mage crop losses postharvest. In Ghana and other parts of the Global South, where mental health resources are limited, such efforts can be mainstreamed into existing primary health care arrangements. Health resources can be consolidated during the postharvest season to provide routine screening and counselling sessions in geographies with high food loss risk. Crisis lines devoted to farmer mental wellbeing and counselling, especially for rural populations, may be timely in further addressing the challenge of ensuring coverage, particularly where farm settlements are remote. In such hard-to-reach environments, an emphasis on peer support and training of gatekeeper farmers may also help provide culturally sensitive support in the absence of healthcare professionals. Addressing suicide risk must be both psychosocial and material. Aside from providing counselling to farmers at risk of suicide ideation, welfare policies to compensate farmers will be timely in addressing other related risk modifiers, such as food insecurity, that are closely associated with food loss. While these measures could help address the immediate needs of farmers exposed to the risk of suicide ideation, proactive policy efforts are needed to target the long-term strengthening of agricultural systems to deliver postharvest management technologies and effective early warning systems to improve the capacity to deal with agricultural shocks.

Supporting information

Acknowledgments

We thank the smallholder farmers, research assistants, and community leaders for their invaluable support throughout the research process.