Household food insecurity (FI) has been defined as limited or uncertain availability of nutritionally adequate and safe foods or limited or uncertain ability to acquire acceptable foods in socially acceptable ways [1, 2]. In wealthy countries, it has been estimated that 11.1% of households are food insecure, with FI higher than 35% among the poor [1]. FI and related indicators, food insufficiency and hunger, have been associated with decreased adult caloric intake [3–5], psychosocial dysfunction in children [6], overweight in young children [7], increased body weight in women [8, 9], sociofamilial dysfunction [10], and lower self-reported physical and mental health [11–14].

FI, food insufficiency, and hunger are generally thought to be substantially more prevalent in poorer countries than they are in richer countries. Although the literature about the consequences of FI is sparse in less wealthy countries, the extant evidence suggests that the consequences of FI extend well beyond undernutrition [15, 16]. Ethnographic and epidemiologic studies from sub-Saharan Africa (SSA) suggest that FI is associated with poor mental health [15, 16]. Hadley and Patil showed that household FI was associated with symptoms of anxiety and depression in rural Tanzania [15], and that seasonal increases in FI were associated with increases in symptoms of depression and anxiety [16]. Hadley and colleagues showed that FI was associated with higher symptoms of depression, anxiety, and post-traumatic stress [17].

One of the potential correlates of FI is neurologic dysfunction (ND). It is estimated that 1 in 9 people worldwide die of a ND, and that ND accounts for 28% of years lived with disability [18]. Several studies have considered the prevalence of ND in Ethiopia. Tamrat and colleagues [19] found that the prevalence of all-cause disability in three rural areas in northern Ethiopia was 4.9%. In particular, the prevalence of epileptic seizure was 0.7%, of vision dysfunction was 1.5% and of walking difficulty was 1.7%. In a similar study, Fitaw and Boersma [20] assessed the prevalence of disability in one urban and three rural areas in northwestern Ethiopia. They found that the prevalence of disability was 3.8%; 47.0% of those had lower locomotive disability, 28.6% had vision dysfunction, 16.1% had upper motor disability, 10.3% had cognitive impairments, and 8.3% had hearing loss, respectively. Tekle-Haimanot and colleagues [21] assessed the incidence of specific neurologic diagnoses in a rural population in central Ethiopia. They found high incidence of the following disorders: epilepsy (520/100,000), postpoliomyelitis paralysis (240/100,000), mental retardation (170/100,000), peripheral neuropathy (150/100,000) and deaf-mutism (130/100,000). The incidence of hemiparesis was 62/100,000, that of cerebral palsy was 20/100,000, 16/100,000 for optic atrophy, 12/100,000 for perceptive deafness, 10/100,000 for tropical spastic paraparesis, 7/100,000 for Parkinson's disease and 5/100,000 each for motor neuron disease, ataxia, and chorea/athetosis.

Although there is a paucity of research about specific social and demographic determinants of ND, both behavioral, as well as physiological studies suggest that malnutrition in early neurologic development is a risk factor for ND in later life [22, 23]. Considering the determinants of ND in less wealthy countries, the World Health Organization (WHO) suggested that malnutrition, perinatal complications, and advanced parasitic infection are among the most important worldwide determinants of ND. Because structural causes underlying some of the leading risk factors for disabling disorders of the nervous system, such as malnutrition, poor sanitation, and lack of healthcare access are most prevalent in less wealthy countries, populations in these countries are especially at risk [18, 24].

In light of the extant research, it is plausible that FI could be a determinant of ND. Alternatively, it is also plausible that those with ND are more likely to be disabled and therefore, less capable of providing themselves and their families with a consistent food supply. This is especially true in rural SSA where subsistence agriculture is the primary means of food accrual. In this analysis, we use data from the Gilgel Gibe Growth and Development Study (GGGDS), a population-representative sample in southwest Ethiopia to assess the relation between FI and symptoms of ND, including spinal pain, extremity weakness, extremity numbness, seizure, vision dysfunction, aphasia, tremors/ataxia, carpal tunnel syndrome, and spinal pain.

Complete data were available for 900 individuals (i.e. 450 men, and 450 women) for analysis. The majority of the sample (61.8%) was between the ages of 20 and 30. According to the median cut-off of the asset index, 47.3% of the sample was classified as having assets below the median number reported. Severe food insecurity was prevalent among 24.7% of the sample.

Overall, vision dysfunction was the most prevalent aggregate of symptoms with 60.9% of the sample reporting diplopia or temporary vision loss in the last week. Seizure was the least prevalent ND with only 6.3% of the sample reporting a seizure in the past week. 48.1% of the sample reported symptoms falling within two or more symptom clusters within the past week.

Using data from a population-based representative sample of households in a rural area in Ethiopia, we found that FI was associated with a range of neurologic symptoms including self-reported seizure, extremity numbness, movement abnormalities, speaking difficulties, carpal tunnel syndrome, spinal pain, and vision dysfunction. FI was also associated with a higher prevalence of ND symptom comorbidity. We found that FI was generally more predictive of ND symptoms among men than among women. Finally, ND symptoms in this population were highly comorbid.

We know of no other studies that have assessed the relation between FI and risk for neurologic symptoms. However, our findings are consistent with literature documenting associations between FI and adverse health outcomes in both wealthy and less wealthy countries. For example, in a cross-sectional analysis of 724 randomly chosen low-income single women participating in welfare in the state of Michigan, USA, Siefert and colleagues [11] found that food insufficiency was associated with poor or fair self-rated health and DSM-III criteria major depressive symptoms. In a sample of 1,006 men and women in the rural US, Pheley and colleagues [12] found that FI was associated with poorer health measures before and after adjustment for demographic and healthcare access variables. Using data from the National Population Health Survey in Canada, food insufficiency was associated with self-reported heart disease, diabetes, hypertension, food allergies, and overweight among men, as well as poorer mental and social health [13]. Among 1488 households in a rural region of the United States, FI was associated with lower self-reported physical and mental health as well as poorer health measures [32]. Another study, by Gao and colleagues, found that FI was associated with lower cognitive function in Hispanics living in the US [33]. Other studies, summarized earlier, have also shown there to be associations between FI and health in SSA [15, 17].

One of the central challenges in the literature on this topic is the directionality of the association between FI and adverse health. Helping clarify this question, three longitudinal cohort studies provide evidence that is suggestive that FI or food insufficiency is a causal determinant of worse mental and physical health. Siefert and colleagues [14] found in a longitudinal analysis of 753 female public assistance recipients in the US that persistent, recurrent food insufficiency over two years was associated with higher limitations in physical functioning as compared to having only experienced FI in the present period. Heflin and colleagues [34] found in the same cohort that increases in food insufficiency predicted increases in symptoms of major depression over a three year study period. In Tanzania, Hadley and Patil [16] found that among 173 caretakers, seasonal variation in FI predicted changes in symptoms of depression and anxiety. Biological evidence also suggests that malnutrition during early neural development can cause ND in later life [18, 24]. Conversely, however, it is also plausible that persons suffering from NDs are more likely to be disabled, and therefore unable to provide themselves with a consistent food supply. This is especially likely to be true in economies that are based heavily on subsistence agriculture. Further research will need to disentangle the causal relationship between FI and ND.

Our work also highlights the current paucity of and need for quality instruments for assessing the burden of ND in the rural SSA setting. We found considerable discordance between the prevalences of common symptoms of ND in our sample and those reported in comparable studies. For example, we found that 60.9% (not shown) of respondents reported vision dysfunction, which compares to 1.5% and 28.6% in comparable studies [19, 35]. Moreover, we found that 6.3% (not shown) of our sample reported seizure compared 0.7% in a comparable study [36]. Although there is no gold standard in the field, given that the prevalences of common symptoms of ND in our sample appear higher than those reported in similar studies, it is likely that our instrument overestimated the prevalences of common ND in our study population.

There are several limitations that should be considered when interpreting our findings. First, we recorded symptoms that could potentially have neurologic causes, rather than specific diagnoses of ND, and it is impossible to differentiate between neurologic, orthopedic, or other causes of many of the symptoms included in our instrument. Moreover, symptoms were not confirmed by neurologists. Therefore, the prevalences reported here should be interpreted very cautiously; it is impossible to draw conclusions about the prevalence of specific diagnoses within this population or about the specific etiologies of individual symptoms. Of particular note, for example, 60.9% of our sample reported vision dysfunction. However, because there is a high prevalence of organic eye diseases that affect vision in Ethiopia [37, 38], it is possible that a sizable proportion of the vision dysfunction in our population is not of neurologic etiology. Second, because we sampled only households with children, it is possible that our findings do not generalize to households without children. Third, our covariate set was limited and therefore, there may be residual confounding of the association between FI and each symptom aggregate. Of particular note, we did not collect data about dietary intake, and therefore, it is plausible that the observed associations could be confounded by malnutrition.

It has been estimated that 33% of people living in SSA are undernourished, a widely used indicator of FI. Unfortunately, in SSA, FI is likely to increase in the coming decades [39, 40]. Moreover, the global burden of ND is estimated to rise significantly by 2020 [18]. Therefore, studies that explore the correlates of FI and ND in SSA are needed to inform clinicians and public health practitioners about how to better serve the needs of populations in this region. Investigators interested in the social causes and effects of ND could fruitfully explore different causal mechanisms that relate FI and ND. Other avenues for future research include assessing the relation between FI and specific ICD-10 diagnosable ND and exploring the associations between FI and ND in other geographic contexts. Finally, quality instruments for measuring ND in populations in SSA are deficient, and researchers interested in the epidemiology of ND in SSA might consider creating and validating such instruments for use in future studies.

Among a population-based representative sample of households in a rural area in Ethiopia, FI was associated with several self-reported neurologic symptoms, including seizure, extremity numbness, movement abnormalities, speaking difficulties, carpal tunnel syndrome, spinal pain, and vision dysfunction, as well as higher prevalence of neurological symptom comorbidity. The directionality of these relations is unclear, and research about potential mechanisms relating FI and ND is needed. Finally, quality instruments for measuring ND in the SSA context are deficient, and research is needed to construct and validate such tools.