Objective and perceived food environment and household economic resources related to food insecurity in older adults living alone in rural areas

A cross-sectional study was conducted in rural areas of South Korea: Yangpyeong County in Gyeonggi Province and Hongcheon County in Gangwon Province. These two regions were selected for review by geographers, nutritionists, and public health professionals to demonstrate diverse characteristics of rural areas such as land and mountain areas. Yangpyeong County is 45 km east of Seoul, the capital of South Korea. It is a designated agricultural area with regulatory exemptions under the Environment-Friendly Agriculture Fosterage Act. Hongcheon County is 81 km northeast of Seoul. It is a mountainous region that flows from the Bakdu Mountain in the center of the Korean peninsula and forms a canyon. In this study, Yangpyeong County represents the land area while Hongcheon County represents the mountain area. Both regions have similar age distributions to other rural areas. Subjects were recruited by nurses of the National Home Healthcare Services (NHHS). The NHHS is a service that allows home-visiting nurses in public centers to visit vulnerable population with health problems and manage their health [22]. Priority registration for the NHHS applies to households with a monthly income below 50% of the median income [22].

Among subjects in the NHHS, home visiting nurses recruited older adults by the following inclusion criteria: 1) aged 65 years or more; 2) health conditions without cognitive impairment; 3) living at home at the time of the assessment; and 4) provided informed consent. All subjects received a telephone call or a face-to-fact contact. They were recruited by their home visiting nurses of public health centers in Yangpyeong and Hongcheon counties between November and December 2013. These two study areas had similar household economic characteristics, including food insecurity (40.0% in Yangpyeong County and 46.3% Hongcheon County, data not shown) and average monthly income for the last year. Upon agreement to participate in this study, all subjects received a phone call or a text message at least 3 times inviting them to attend a comprehensive fact-to-face interview. The base population was about 900 households with older adults receiving the NHHS (9 towns from Yangpyeong and 2 towns from Hongcheon). A total of 170 rural seniors without disability living alone finally completed the survey. Therefore, our study samples covered 18.9% of the study population. A well-trained survey team was composed of professionally experienced interviewers in this study. To ensure uniformity, two head interviewers underwent 1 day of training by a principal investigator of the study with a standardized study protocol. All interviewers were required to attend 2 days of training by head interviewers before the survey was initiated. Informed written consent for participation was obtained from each individual. This study was approved by the Institutional Review Board of Sangmyung University, Seoul, Korea (approval number: BE2013–8).

Detailed data on demographic characteristics, household economic resources/expenditure, food environmental factors, and food insecurity were collected for each study participant. Demographic variables were age, sex, education, employment, and beneficiary of national basic livelihood. The national basic livelihood is a service of monetary payment for daily necessities such as food, clothing, and fuel for low-income families. It is defined as the amount of the reported household income deducted from the minimum level of livelihood wage (30% of the median income). Objective and perceived measures of economic resources and food environments are described as below.

Objective household economic indices of monetary incomes and expenditures per month for food and non-food (housing, heating and medical expenses) were collected in Korean Won (the exchange rate of currency: 1125 Korean won = 1 US dollar). The distribution of household consumption expenditure (percentage of total) from each monetary term was calculated. Perceived indicators of household economic experience were measured by asking whether they reduced food expenditure due to the burden of non-food expenditures such as housing fee and heating costs during a winter season.

Food environment conceptualizations in this study were developed based on the concept model of Glanz and colleagues [23]. This model organizes food environment features into community food environment (the distribution of food sources within a community such as number and accessibility of food outlets) based on an ecological model of health behaviors. The organizational food environment (the multiple settings where people eat or procure food such as home, school, work, and others) and consumer food environment (available healthy food options and the cost and quality of foods in local food outlets) are contributors to healthy eating patterns. These food environments could be moderated or mediated by social environment [24]. Especially, for older adults living in rural areas, limited food availability and accessibility due to financial limitations, physical limitations, and inability to drive should be addressed by social supports [25‐27].

Based on the literature, this model includes two aspects of availability and accessibility of food at household and community levels. The assessment of food availability was focused on the supply source of available food and the presence of social supports to help ensure food supplies at household level and the presence of various foods at community level. Household-level food accessibility included the presence of social supports helping access to foods while community-level food accessibility included geographic food access such as location of food store and ease of getting to that location. Because there was no best measurement concerning the definition of food environment, we measured food environment using questionnaires including both objective observational and perceived dimension as proposed in several previous studies [18, 28, 29].

The questionnaire-based measurement of food environment was as follows. Household food availability was obtained objectively by food acquisition methods such as food purchase, farming or home gardening, tangible private food, and beneficiaries of public food assistance programs. Household food accessibility was measured by asking subjects whether there were intangible supports related to food purchase from a family and/or neighbors existing for subjects enrolled for this study.

Community food availability and accessibility were examined using objective or perceptional-based measurements. Community food availability was measured by asking them whether subjects purchased foods mostly and by asking them in detail their perceptions on whether the nearest food store accommodated various purchasable food items to meet their needs. Community food accessibility was measured objectively by transportation and distance to get to the nearest food stores. Perceived community food accessibility was measured by determining whether subjects experienced any difficulty in food purchasing due to long distance to food stores from home, a bus stop being located in a remote location from home base, or inconvenient bus routes.

Household food insecurity was measured using the validated Korean Household Food Security Survey Module (K-HFSS) from the Korea National Health Examination and Nutrition Survey [30, 31]. The K-HFSS was based on the 18-item US Household Food Security Survey Module (HFSSM). This 18-item questionnaire consisted of 3 household-referenced questions, 7 adult-referenced questions, and 8 child-referenced questions. In the present study, the adult food security survey module consisting of 10-item questionnaire (3 household-referenced questions and 7 adult-referenced questions) was completed by each household. A score of 1 was allocated to affirmative responses to each item and a score 0 was assigned otherwise. Subjects with scores of 3–10 were classified as having food insecurity.

Demographic characteristics are presented in Table 1. The prevalence of food insecurity was 34.7% (n = 59) in the study sample. The mean age of subjects was 77.6 years. Most participants were women (84.7%) and unemployed (91.8%). Majority (90.6%) of them had less than a middle school education. About 44% of participants were beneficiaries of the national basic livelihood. There was no significant difference in age, sex, education, employment, or beneficiaries of national basic livelihood between food-secure and food-insecure households.

Economic characteristics of single senior households according to food insecurity are shown in Table 2. For objective economic indices, monthly housing fee in real terms ($66.4 vs. $31.3, P < 0.01) and percentage compared to total expenditure (29.4% vs. 14.1%, P < 0.01) was higher for food-insecure households than that for food-secure households. In contrast, monthly medical expenditure in monetary amount ($30.9 vs. $14.7, P < 0.05) and percentage compared to total expenditure (18.3% vs. 9.5%, P < 0.05) was higher for food-secure households than that for food-insecure households. For subjective economic indices, the experience of reducing food expenditure resulting from burden of housing fee (p < 0.01) and heating costs during winter (p < 0.01) were higher for households with food insecurity than those for households without food insecurity.

Table 3 shows food environmental characteristics at household and community levels according to food insecurity. Among household food availability and accessibility factors, household with public food assistance program had higher proportion in household with food insecurity than that in household with food security (66.1% vs. 45.1%, p < 0.01). The proportion of household that acquired foods by using farming or home gardening resources (24.3% vs. 11.9%, p = 0.0532) and the proportion of household with intangible support for food purchasing from family (12.6% vs. 1.7%, p = 0.0502) tended to be greater in households with food security than those in households with food insecurity. Among relevant community food accessibility factors by perceived measurements, the proportion of households with difficulties in food purchasing due to food stores being located far from home was greater in households with food insecurity (44.1% vs. 16.2%, p < 0.001) whereas the proportion of households with difficulties in food purchasing due to inconvenient bus routes was greater in households without food insecurity (16.2% vs. 3.4%, p < 0.05).

The results of stepwise logistic regression to select the most explainable economic and food environmental factors related to food insecurity are shown in Table 4. Among economic factors, the percentage of total expenditure on housing fee was positively related to food insecurity (OR = 1.021, 95% CI: 1.008–1.034). Among food environmental factors, having difficulties in food purchasing due to food stores being located far from home (OR = 14.487, 95% CI: 5.139–40.842) and non-purchasing of foods regardless of such difficulties (OR = 5.946, 95% CI: 1.659–21.311) were positively related to food insecurity while foods purchasing at super supermarket (OR = 0.398, 95% CI: 0.166–0.951) and having difficulties in food purchasing due to inconvenient bus routes (OR = 0.083, 95% CI: 0.015–0.460) were negatively related to food insecurity.