Food security, nutritional status and mental health are strongly connected. Food insecurity, defined here as poor nutritional intake, insufficient amounts of food eaten and being unable to afford nutritious food is a problem affecting many households worldwide, including in Australia [1–5]. Food insecurity has been shown to be independently associated with higher levels of psychological distress, psychiatric disorders and poor child development [6–9].

Climate change is an important contemporary global health issue [10–14]. Internationally and in Australia, climate change is increasing food insecurity, with risks for nutrition and health. Through the increased frequency and duration of adverse weather events such as drought, cyclones and flooding, climate change can reduce agricultural productivity and the viability of agricultural support industries [15–18]. These climate impacts can affect food yields, household livelihoods and food prices, which influence dietary habits through food availability and affordability pathways [17, 19–21]. The 2001–2008 Australian drought, termed the ‘Big Dry’, and other extreme weather events are reported to have affected agricultural yields and food prices: between 2005 and 2007, vegetables in Australia increased in price by 33% and fruit by 43% [22].

Climate change is also having an impact on mental health [23–25]. Climate change can affect mental health directly by exposing people to the psychological trauma associated with higher frequency, intensity and duration of climate-related disasters [24], including extreme heat exposure [26], and also by harming landscapes, which diminishes the sense of belonging and solace that people derive from their connectedness to the land [27]. Indirectly, climate change may affect community wellbeing through damage to the economic and, consequently, the social fabric of communities [24, 25].

Recent Australian reports on the health consequences of climate change have identified an urgent need to understand how drought and long-term drying affect population health and wellbeing [10]. It is possible that exposure to drought may exacerbate the food and mental health relationship via potential negative impacts on the availability, livelihoods, food price and consumption of nutritious food, thereby creating acute and chronic stress.

It is likely that drought and long-term drying will exacerbate existing inequities in health risks [28]. Previous Australian analyses have shown that extreme drought is associated with psychological distress and rural/remote areas but not urban [29]. International and Australian evidence suggests that food security is socio-economically [2, 30, 31] and geographically distributed [32–42], with low-socioeconomic status and people living in rural areas more likely to report food insecurity.

While the links between exposure to drought and food insecurity, and drought and mental health have been made in the literature, we are unaware of any studies that have explored the association between drought exposure, food security and mental health. This study will address this knowledge gap by examining the associations between indicative measures of food security and mental health in Australia during the last great drought, and exploring whether these associations are sensitive to levels of drought exposure across urban and rural locations.Figure 1 describes our hypothesized food-related pathways from drought exposure to mental health outcomes. These pathways are based on the existing literature and variables available in the main dataset used in the analysis. We hypothesize that the long dry in Australia has reduced local agricultural productivity and the viability of agricultural support industries. The knock-on effects of this has been, we hypothesize, a reduction in agricultural yields and therefore household income and increased food prices. These impacts have affected household purchasing patterns due to changes in the volume and types of food available for consumption and also the relative affordability of different foods. In some instances the impact on households’ income from the drought pressures will have meant missing meals due to cost. The impacts therefore on mental health arise from potential negative effects on food habits and nutrition, and from the acute and chronic stress of having to miss meals due to not having enough money (due to food price increases and/or reductions in household income).

Relatively small numbers of people (1.6%) reported going without meals due to financial pressures (Table 2). Almost two-thirds consumed higher-than-average regular consumption of discretionary foods, regardless of location. The mean psychological distress score for the study participants was borderline none-to-moderate distress and did not differ between rural/urban locations.

We describe the findings of a large population-based study examining the associations between drought, food insecurity and people’s mental health in urban and rural settings in Australia.

Our findings support the strong evidence from other settings that food insecurity is associated with psychological distress [6–9]. Study participants who reported missing meals, or consumed below-average levels of core foods or above-average levels of discretionary foods reported moderate-high distress levels.

The findings also shed further light on the emerging international evidence regarding the relationship between climate change and food insecurity (increased during drought), and between climate change and mental health (harmed by weather-related disasters such as drought). Our main observation was the significant association between missing meals and exposure to drought, with urban dwellers who lived in ‘constant dry’ weather conditions more likely to miss meals due to financial pressures. This pattern was similar among rural dwellers but was not statistically significant, probably due to small numbers. Similarly, our study substantiates some of the conceptual analysis linking climate change and mental health [24, 26]: rural people who had been exposed to ‘constant and long dry’ drought reported a moderate level of distress while those living in any other category reported little distress. In urban areas, where there may be less disadvantage and more access to products and services, people reported a low level of distress for all drought categories.

We have shown that exposure to drought affects the food and mental health association, possibly by reducing the availability, affordability and consumption of nutritious food, which are each linked to increasing mental health problems [58, 59]. However, the relationship is not straightforward - the type of drought experienced can have positive or negative effects. Drought appears to moderate the association between food insecurity and mental health in the following ways:

First, compared to the average unadjusted distress score, people experiencing ‘very dry’ drought reported a significantly higher level of distress whereas people living in ‘constant and long dry’ drought reported the lowest. These findings suggest that living in constant and long dry drought affects the association between missing meals and mental health. We can hypothesize that when drought is a predictable outcome, households have some capacity to plan expenditures thereby preventing them from being caught short financially. In addition, persons who choose to remain in long term drought areas may have financial reserves to buffer their consumption against unexpected price and income changes. In contrast, in areas where drought occurs but usually not every year, households may spend based on “normal” income or prices for the locale and the onset of drought may come as a surprise.

Second, people consuming below-average levels of core foods reported greater distress than those consuming above-average levels, this association is only statistically significant among people experiencing ‘zero and moderate dry’ drought. This suggests that being in some kind of extreme drought disrupts the association between core food consumption and psychological distress that is observed in times of relatively normal weather conditions. While we do not have data to empirically explore the reasons for this, it may be that hunger is suppressed due to the heat associated with drought.

Third, compared to the average unadjusted distress score, significantly higher levels of distress were observed in people who consumed above-average levels of discretionary food among those who experienced ‘zero or moderate’, ‘very dry’, and ‘constant and long dry’ drought. The magnitude of the mean difference for the two drought categories are twice as big as the others, suggesting that these type of drought experiences amplified people’s distress by increasing people’s consumption of discretionary unhealthy food.

These relationships are not simple or straightforward. There are many interconnected factors that affect food consumption and mental health, including the social and built environments in which people live [60]. While such factors may or may not be directly related to drought, they will moderate and mediate its impact. Additionally, we do not know from these data whether the food insecurity effects of drought are through its impact on local food prices or through the impact of drought on incomes earned by persons in the affected area, or if food insecurity is an outcome of chronically insufficient income, or due to high volatility of purchasing power of income. The nature and scope of the cross-sectional data does not allow us to interrogate the complexity of these relationships and warrants further investigation.

Limitations of our study reduce the generalisability of our findings. First, we selected participants (~2/5 total sample) who did not move in the period measured by the drought variable (2001–2007). Disadvantaged Australians tend to move house more often [61], are more likely to experience poor mental health, and are much more often lost to follow-up than other respondents and were thus under-represented in the final datasets [62, 63]. This may explain why the observed levels of going without meals due to financial stress are lower than those reported in other Australian studies [2, 30]. Australia is a vast, sparsely populated continent in which the large majority lives on the seaboard, where most of the rain occurs. As the drought definitions used in this study rely on precipitation deviations from long-term norms, greater deviations will be found where precipitation is greater (i.e. where there is more variance from which to deviate). Thus, it is less likely that rural and remote areas will register as deviating significantly from their norms (because they tend to have less precipitation in the first place) and they will – incorrectly – appear to have lower rates of experiencing drought. Our study thus likely underestimates the prevalence and impact of drought in these areas. In addition, the HILDA Survey does not sample from remote and very remote locations.

Another limitation of our study is the use of three indicative measures of food insecurity. Ideally we would have used internationally validated food security assessment scales. However, as this was a secondary analysis of an existing dataset, we were constrained by the food measures recorded in the dataset. Finally, the analysis is based on self-reported food insecurity and mental health, likely causing poor mental health to be underreported [64, 65], and indicating underestimation of associations found. Despite these limitations, we conducted, to our knowledge, the first nationally representative study examining modelled drought conditions, food insecurity and mental health. We have shown that some types of drought can be harmful to mental health because they increase food insecurity and that these circumstances tend to be more common in rural areas.

Drought is an increasing public health problem globally, likely to exacerbate existing health risks and conditions. The impact of drought on the well-established and strong association between food insecurity and mental health has received little attention. In this study, exposure to Australia’s ‘Big Dry’ moderates the association between measures of food insecurity and psychological distress, generally increasing the distress level. Climate adaptation strategies that consider social, nutrition and mental health impacts are needed.