Food access determinants and food insecurity severity in the older population: evidence from a nationally representative, cross-sectional survey in Thailand

Abstract

Background

Food access is a critical public health issue closely associated with the growing problem of food insecurity (FI), which has shown a steady increase, particularly among older adults, a population group considered especially vulnerable. This study aimed to examine the situation and the influence of food access determinants at the individual, household, and community levels on the severity of food insecurity among Thai older adults.

Methods

This cross-sectional study utilised secondary data from the 2022–2023 round of the Survey on Food Consumption, Food Security, and Food Literacy in Thailand. The survey employed a multi-stage sampling design covering the four main geographic regions and Bangkok. Data were collected from all household members who met the inclusion criteria. For this analysis, the data were weighted and restricted to a sample of 2,641 individuals aged 60 years or older. Food insecurity (FI) was assessed using the Food Insecurity Experience Scale (FIES) and categorised into three severity levels. Predictor variables included socio-demographic characteristics, household-level food access, and community-level food access, all derived from the survey questionnaire. Ordered logistic regression was applied to examine the associations between the predictors and the severity of FI.

Results

Approximately one in four older adults (26.9%) experienced FI (i.e., 25.3% mild, 1.6% moderate-to-severe), while 73.1% reported none. Food access across individual, household, and community levels significantly predicted FI severity. Older Thais with low income and limited education were 1.84 and 3.45 times more likely to face severe FI. Receiving healthy food prepared by a household member reduced FI risk by 50%. Living in a rural area and having limited access to food outlet diversity increased severe FI risk by 1.30 and 2.88 times, respectively.

Conclusions

Improving food access for older adults requires a multilevel approach. The government should ensure inclusive food welfare, while communities should develop local systems to support older adults, especially those facing food shortages.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12877-026-07140-3.

Introduction

Food insecurity (FI) is a critical issue affecting the survival of people worldwide, and poses a significant barrier to achieving Sustainable Development Goal 2, which aims to end hunger and all forms of malnutrition [1]. Data from the Food and Agriculture Organization (FAO) indicate that FI is a global problem, including in Thailand, with an increasing trend. Between 2014 and 2016, Thailand experienced a moderate-to-severe FI prevalence of 4.7% [2], which rose to 7.2% in 2021–23 [3]. Similarly, reports on the prevalence of FI in the Thai population reveal that 4.4% of Thai households suffer from moderate-to-severe FI [4]. Among individuals age 15 years or older, the prevalence of mild FI is 23.7%, while moderate-to-severe FI affects 1.5% of the population [5]. Moreover, the severity of FI varies across regions and demographic groups, such as urban versus rural areas, gender, and age groups, with older adults exhibiting a higher prevalence of FI compared to other age groups [5–10].

FI among older adults is an important issue that Thailand needs to pay close attention to, especially as the country has become an ‘aged society’ with a rapidly increasing elderly population [11]. It is projected that, by 2040, Thailand will be a ‘super-aged society,’ where more than 28% of the total population will be age 60 years or above [12]. This growing elderly population has led to greater income inequality and increased government spending on healthcare services for older adults [13]. Additionally, most older adults are no longer working, have limited income, face higher health risks, and are more likely to live alone [14, 15]. Therefore, this demographic shift significantly raises the risk of FI among Thai older adults [16–18].

FI is closely linked to food access. According to the FAO’s definition of food security, food access is a critical component that enables individuals to obtain sufficient food, in quantity and quality, from available food sources before the food is utilized for consumption and nutrient absorption [3]. Food access is more than just physical availability of food, but also encompasses the economic and social means to acquire it [19–21]. Limitations in food access among older adults can lead to inadequate food intake, and contribute to FI [20, 21]. This is particularly supported by studies in low- and middle-income countries, which highlight that when older adults face financial hardships, poor living conditions, limited social support, and insufficient community infrastructure, these barriers reduce their ability to access food and increase the risk of FI [9, 17, 18, 22, 23].

Food access and FI are complex processes influenced by multiple factors across various dimensions, ranging from individual characteristics and interpersonal relationships to systemic and structural aspects of society [16, 24]. Therefore, studying food access and FI among older adults requires a comprehensive consideration of these factors at all levels. A review of previous research reveals that studies on food access affecting FI among older adults are limited [8, 16, 21, 22], particularly in Thailand, where research on FI is scarce and rarely addresses food access among the elderly [18, 25–27]. More importantly, there is currently no nationally-representative data on food access among food insecure older adults in Thailand.

Therefore, this study investigated the situation and the influence of food access factors at the individual, household, and community levels on the severity of FI among older adults in Thailand. The hypothesis is that perceived food access at these three levels (individual, household, and community) is associated with the severity of FI. The findings are expected to provide representative evidence on food access and FI among Thailand’s aging population, and to serve as a critical basis for informing policies and interventions that promote food security among older adults.

Materials and methods

Study design and participants

This study analyzed secondary data from a national, cross-sectional survey “The 2022–2023 Survey on Food Consumption, Food Security, and Food Literacy in Thai Population” [5]. The survey was conducted by the National Statistical Office (NSO) of Thailand using a multi-stage sampling design to ensure nationally representative coverage across demographic and geographic contexts. The sampling frame included Bangkok and the four main regions (Central, Northern, Northeastern, and Southern), with stratification by region and area of residence (municipal and non municipal). Provinces, districts, and enumeration areas (EAs) were selected sequentially, with EAs chosen using probability proportional to size. Households within selected EAs were then sampled using systematic sampling. Each EA comprised 20 households, a cluster size determined through standard sample size calculations that account for acceptable margins of error and the design effect, balancing statistical precision and operational feasibility [5]. Data were collected from all household members age 0 years or older during the period from August 2022 to September 2023 [5]. For this study, only individuals age 60 years or older were selected. The data were weighted by the NSO to ensure that the sample was representative of the national older adult population. A total of 2,641 older adults were included in the analysis, and no missing data were observed in the variables analyzed in this study.

Data collection

The secondary data used in this study were derived from a survey in which data collection was conducted through face-to-face interviews using the Qualtrics offline survey application. The questionnaire was developed based on an extensive review of both national and international literature and refined through expert consultation to ensure content and face validity of the questionnaire (see additional file 1). To ensure instrument reliability, a pilot test was conducted across all age groups This allowed for the calculation of internal consistency parameters (e.g., Cronbach’s alpha) for multi-item scales, such as the Food Insecurity Experience Scale (FIES), while confirming the stability and clarity of single-item manifest variables. Interviewers held at least a bachelor’s degree and received training, including mock interviews, to ensure data accuracy and minimal bias. Informed consent was obtained from all participants [5].

This manuscript was prepared in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines for cross-sectional studies (see additional file 2).

Measures

The variables included in the analysis capture factors related to food access and FI across multiple levels, including the individual, household, and community levels. Household- and community-level factors reflect aspects of food access as perceived and reported by older adults. All variables included in the analysis are described below.

Severity of food insecurity (FI)

FI was used as a dependent variable, and was assessed using Food Insecurity Experience Scale (FIES), developed and adapted by the FAO (and translated into Thai). The FIES consists of eight items asking whether, during the past 12 months, due to a lack of money or other resources: (1) “You were worried that you would run out of food; (2) You were unable to eat healthy and nutritious food; (3) You ate only a few kinds of foods; (4) You had to skip a meal because there was not enough money or other resources to get food; (5) You ate less than you thought you should; (6) Your household ran out of food; (7) You were hungry but did not eat; and (8) You went without eating for a whole day.” These items represent increasing severity of difficulty in accessing quality food due to insufficient income, financial hardship, or other resource constraints, and are a proxy indicator of psychological concern related to food access over the previous year [5, 28, 29]. The scale is intended for individuals age 15 years or older, and uses binary responses: “Yes” = 1 and “No” = 0. A “Yes” response to any item requires preceding items to also be answered “Yes.” FI severity is categorized into four levels based on total score (ranging from a potential of 0 to 8): food secure or no FI (0 points), mild FI (1–3 points), moderate FI (4–6 points), and severe FI (7–8 points) [5, 29]. Even mild FI is considered as an initial stage of public health concern, while moderate-to-severe FI indicates uncertain food access and insufficient intake that fails to meet basic dietary needs [3]. This study followed the categorization above but combined moderate-to-severe levels into a single group, creating three severity levels: no FI (0 point), mild FI (1–3 points), and moderate-to-severe FI (4–8 points).

Variables related to food access

The factors influencing food access among older adults are examined through the application of the concept of access [19] and the Social Ecological Model [30]. These frameworks provide a perspective for studying food access among food-insecure older adults in relation to factors at the individual, household, and community levels.

Individual-level food access

Socio-demographic characteristics, considered as individual-level factors associated with older adults’ food access [6, 7, 9, 10, 31–33], included the following variables: Sex was coded as male or female. Age was categorized into four groups: 60–64, 65–69, 70–74, and 75 years or over. Marital status was classified as single, married, or widowed/divorced/separated. Educational attainment was categorized as no formal education or some primary, primary, secondary, and bachelor’s degree or higher. Monthly income from employment was grouped into four categories: no income, 1–5,000 baht, 5,001–10,000 baht, and more than 10,000 baht.

Household-level food access

At this level, the variables describe household food sources, the living arrangements of older adults [7, 10], as well as the financial or resource support from family members related to their access to food [6, 8, 9, 21, 22]. The variables include living alone in the household, which is based on household composition data indicating whether the older adult lives alone (1 = yes, 0 = no); food prepared by the household, assessed by the question “Does the household prepare meals that help you eat healthier food?” (1 = yes, 0 = no); homegrown vegetables and fruits, based on the question “Do you grow vegetables and fruits in your home garden for personal consumption?” (1 = yes, 0 = no); and received financial or resource support, measured by the question “Does any family member provide financial or resource assistance to purchase healthy food?” (1 = yes, 0 = no).

Community-level food access

Community-level variables, with the geographic scope of “community” defined in this study as the village level in rural areas or community level in urban settings rather than the subdistrict level, reflect older adults’ perceptions of local food outlets, food prices and quality, awareness of community food assistance, as well as geographic characteristics of the community [6–10, 20, 34]. Specifically, the residential area was coded as rural or urban [6, 9], and regions of community were categorized into Central, Northern, Northeastern, Southern, and Bangkok Metropolitan Area [6, 9]. The number of food outlet types was derived from the question, ‘What types of stores selling fresh foods, vegetables, and fruits are available in your community?’ For response, participants selected from predefined options, including grocery stores, modern trade, convenience stores, supermarkets or shopping malls, fresh or street markets, mobile vendors, and direct purchase from food producers. Response was categorized into three groups: one type, two types, and three or more types. Perceived food prices were based on the question, “In general, how are the prices of fresh food, vegetables, and fruits in stores in your community?” with response options including unaware, lower, similar, and higher than other areas. Perceived food quality was assessed by the question, “In general, how is the quality of fresh food, vegetables, and fruits in stores in your community?” with responses categorized as better, average/unsure, and worse than other areas. Perceived food assistance was derived from the question, “Does your community help you gain better access to healthy food?” (1 = yes, 0 = no).

Statistical analysis

This study employed descriptive statistics (e.g., frequencies, percentages) to describe the situation of food access among Thai older adults in relation to the three levels of severity of FI. Subsequently, Chi-square was used to test for bivariate associations between food access factors at the individual, household, and community levels and the severity of FI without accounting for potential confounding variables.

To assess the influence of these food-access factors on the likelihood of FI severity, ordered logistic regression analysis was performed. The analysis used a logit model based on a non-linear probability model framework, as the dependent variable (severity of FI) was measured on an ordinal scale, and all independent variables were categorical (dummy coded as 0 and 1). The assumption of no multicollinearity among the independent variables was confirmed (Tolerance > 0.2, VIF < 5; see additional file 3). The proportional odds assumption test was conducted to assess whether the relationship between the independent variables and the log-odds of the outcome was consistent across all thresholds of the ordinal dependent variable. Independent variables were entered in sequential blocks at the individual, household, and community levels to compare their predictive ability for FI severity among older adults, using the likelihood ratio test and model goodness-of-fit indicators. In the final model, factors that were significantly associated with FI severity were reported as an adjusted odds ratio (AOR) with 95% confidence intervals (CIs), after adjusting for potential confounders such as sex, age, and income. To ensure robustness, a sensitivity analysis was conducted by comparing the results of the final ordered logistic regression model with those from binary logistic regression analysis (see additional file 4). A p-value of ≤ 0.05 (two-tailed) was considered statistically significant. All statistical analyses were performed using SPSS software version 22.

Results

In Thailand, older adults experienced higher FI (26.9%) than other age groups, including those age 15 to 59 years (24.9%). According to the FIES, 26.9% of older adults worried about having enough food due to limited money or resources, while the most severe FI (i.e., involuntarily going a whole day without eating), was reported by 0.3% in the past 12 months (Fig. 1).

Fig. 1.

Food insecurity among the Thai population by item of FIES and age group [5]

The analysis of FI severity among 2,641 older Thais (Table 1) showed that over one in four (26.9%) experienced FI, with 25.3% reporting mild FI and 1.6% moderate-to-severe FI, while 73.1% reported no FI. FI was most prevalent among females (30.2%), those age 65 to 69 years (31.5%), widowed divorced or separated individuals (28.1%), those with no formal or incomplete primary education (31.1%), and those with a monthly income of 1 to 5,000 THB (37.5%).

Table 1.

Food insecurity by severity level and respondent characteristics among older Thais (n = 2,641)

Characteristics	Total	Prevalence of severity of FI			Χ2(p-value)
		No FI /Food-secure	Mild	Moderate-to-severe
	% (n)	% (n)	% (n)	% (n)
Severity of FI	100.0 (2,641)	73.1 (1,931)	25.3 (667)	1.6 (43)
Sex
Female	58.0 (1,531)	69.8 (1,068)	28.3 (434)	1.9 (29)	21.11(0.000)
Male	42.0 (1,110)	77.7 (863)	21.0 (233)	1.3 (14)
Age group(years)
60-64	31.9 (843)	74.3 (626)	24.2 (204)	1.5 (13)	14.58(0.024)*
65-69	25.3 (669)	68.5 (457)	29.1 (195)	2.4 (17)
70-74	20.0 (529)	73.2 (387)	25.3 (134)	1.5 (8)
75+	22.7 (600)	76.8 (461)	22.3 (134)	0.9 (5)
Marital status
Single	4.4 (115)	74.1 (85)	25.0 (29)	0.9 (1)	7.68(0.104)
Married	67.6 (1,787)	73.6 (1,315)	25.2 (450)	1.2 (22)
Widowed/divorced/separated	28.0 (739)	71.9 (531)	25.4 (188)	2.7 (20)
Educational attainment
No formal education/some primary	72.1 (1,906)	68.9 (1,313)	29.2 (557)	1.9 (36)	68.25(0.000)***
Primary	14.9 (394)	80.7 (318)	17.8 (70)	1.5 (6)
Secondary	8.0 (210)	86.8 (182)	12.9 (27)	0.5 (1)
Bachelor’s degree or higher	5.0 (131)	90.1 (118)	9.9 (13)	0.0 (0)
Monthly Income from Employment (baht)
No income	69.7 (1,840)	73.6 (1,354)	24.9 (459)	1.5 (27)	33.53(0.000)***
1-5,000	13.0 (343)	62.5 (214)	34.6 (118)	2.9 (11)
5,001-10,000	9.2 (242)	75.7 (184)	23.0 (56)	1.2 (3)
>10,000	8.2 (216)	83.3 (180)	15.7 (34)	0.9 (2)
Living alone
Yes	5.2 (137)	73.9 (101)	23.9 (33)	2.2 (3)	0.38(0.826)
No	94.8 (2,504)	73.1 (1,830)	25.3 (634)	1.6 (40)
Food prepared by household members
Yes	52.3 (1,380)	78.9 (1,088)	19.9 (275)	1.2 (17)	47.93(0.000)***
No	7.7 (1,261)	66.9 (843)	31.1 (392)	2.0 (26)
Homegrown vegetables/fruits used for meals
Yes	30.2 (796)	64.1 (510)	35.1 (280)	0.8 (6)	62.22(0.000)***
No	69.8 (1,845)	77.0 (1,421)	21.0 (387)	2.0 (37)
Received financial or resource support
Yes	57.4 (1,517)	71.5 (1,083)	27.0 (411)	1.5 (23)	6.31(0.043)*
No	42.6 (1,124)	75.4 (848)	22.8 (256)	1.8 (20)
Community residential area
Rural	55.5 (1,467)	67.9 (996)	30.2 (443)	1.9 (28)	45.99(0.000)***
Urban	44.5 (1,174)	79.7 (935)	19.1 (224)	1.3 (15)
Region of residence
Central	29.8 (788)	73.2 (577)	24.4 (192)	2.4 (19)	102.81(0.000)***
Northern	12.7 (336)	61.3 (206)	37.2 (125)	1.5 (5)
Northeastern	28.8 (761)	68.7 (523)	30.1 (229)	1.2 (9)
Southern	18.2 (480)	78.4 (377)	21.4 (102)	0.2 (1)
Bangkok	10.5 (276)	89.5 (248)	6.9 (19)	3.6 (10)
Perceived number of food outlet types
Single type	37.9 (1,000)	65.1 (652)	32.3 (322)	2.6 (26)	76.39(0.000)***
2 types	30.8 (813)	72.9 (593)	25.3 (206)	1.7 (14)
3+ types	31.3 (828)	82.9 (686)	16.8 (139)	0.4 (3)
Perceived food prices compared to other residential areas
Unaware	22.2 (586)	77.3 (453)	21.7 (127)	1.0 (6)	79.00(0.000)***
Lower	8.5 (224)	85.2 (190)	13.9 (31)	0.9 (2)
Similar	38.6 (1,018)	76.8 (782)	21.9 (223)	1.3 (13)
Higher	30.8 (813)	62.2 (506)	35.2 (286)	2.6 (21)
Perceived food quality compared to other residential areas
Better	15.1 (400)	67.5 (270)	31.0 (124)	1.5 (6)	13.04(0.011)*
Average/unsure	81.3 (2,146)	73.8, (1,583)	24.7 (529)	1.6 (34)
Worse than	3.6 (95)	83.2 (79)	14.7 (14)	2.1 (2)
Perceived existence of food assistance compared to other residential areas
Yes	33.9 (896)	68.1 (610)	30.3 (272)	1.6 (14)	18.57(0.000)***
No	66.1 (1,745)	75.7 (1,321)	22.6 (395)	1.7 (29)

Pearson Chi-Square

* p-value < 0.05, ** p-value < 0.01, *** p-value < 0.00

For variables related to household-level food access, FI was reported among those not living alone (26.9%), whose household members did not prepare healthy food (33.1%), who engaged in home gardening of fruits and vegetables (35.9%), and who received financial or resource support for food purchases (28.5%). At the community level, FI was higher among older adults who lived in a rural area (32.1%) and in the Northern region (38.7%). Regarding perceived community food access, FI was found among those reporting only one type of food retail outlet (34.9%), perceiving local food prices as higher than other areas (37.9%), perceiving food quality as better than elsewhere (32.5%), and reporting food-related assistance in the community (31.9%) (Table 1).

As shown in Table 1, the Chi-square test indicates that several food access factors were significantly associated with FI severity among older adults. At the individual level, sex, age, education, and monthly employment income were significant. At the household level, preparation of food by a household member (or not), use of homegrown vegetables/fruits (or not), and receiving financial/resource support (or not) were significantly associated with FI. At the community level, residential area, region, perceived number of food outlet types, perceived food prices, perceived food quality, and perceived food assistance were all significantly associated with FI severity (Table 1).

In Table 2, all three models significantly predicted FI severity (p < 0.001). Compared to the baseline Model 1 (which has only individual level variables), the inclusion of the household-level factors in Model 2, and community-level variables in Model 3 resulted in a substantial increase in the Nagelkerke R². These increases indicate that household and community variables enhance the model’s explanatory power. Estimates from Model 3, with the highest Nagelkerke R², are further discussed.

Table 2.

Adjusted odds ratio (AOR) from ordered logistic regressions by respondent characteristics and FI severity level (n = 2,641)

Characteristics	Model 1 – ILa			Model 2 – IL, HLa		Model 3 (Final) – IL, HL, CLa
	AOR		95% CI	AOR	95% CI	AOR	95% CI
Sex (Ref.= Male)
Female		1.37**	1.13, 1.67	1.11	0.89, 1.37	1.03	0.82, 1.30
Age group (Ref.= 75 + years)
60–64		1.37*	1.05, 1.80	1.14	0.87, 1.51	1.20	0.89, 1.61
65–69		1.68***	1.30, 2.19	1.38*	1.05, 1.81	1.40*	1.05, 1.87
70–74		1.25	0.95, 1.65	1.11	0.84, 1.48	1.19	0.89, 1.60
Marital status (Ref.= Widowed/divorced/separated)
Single		1.02	0.64, 1.61	1.03	0.65, 1.66	1.19	0.73, 1.95
Married		1.02	0.83, 1.26	0.99	0.79, 1.23	0.90	0.72, 1.14
Educational attainment (Ref.= Bachelor’s degree or higher)
No formal education/some primary		3.90***	2.15, 7.07	4.30***	2.37, 7.84	3.45***	1.84, 6.47
Primary		2.06*	1.09, 3.90	2.27*	1.20, 4.31	1.90	0.97, 3.69
Secondary		1.45	0.72, 2.93	1.57	0.77, 3.20	1.58	0.76, 3.29
Monthly Income from Employment (Ref.= > 10,000 baht)
No income		1.59*	1.08,2.34	1.57*	1.06, 2.31	1.56*	1.03, 2.35
1–5,000		2.47***	1.61, 3.79	2.28***	1.48, 3.51	1.84**	1.17, 2.90
5,001–10,000		1.37	0.85, 2.19	1.38	0.86, 2.23	1.48	0.89, 2.43
Living alone (Ref.= No)
Yes		-	-	0.87	0.57, 1.33	0.71	0.45, 1.12
Food prepared by household members (Ref.= No)
Yes		-	-	0.55***	0.45, 0.68	0.50***	0.40, 0.62
Homegrown vegetables/fruits used for meals ((Ref.= No)
Yes		-	-	1.64***	1.35, 1.99	1.43**	1.16, 1.76
Received financial or resource support (Ref.= No)
Yes		-	-	1.01	0.83, 1.22	0.88	0.72, 1.09
Community residential area (Ref.= Urban)
Rural		-	-	-	-	1.30*	1.05, 1.61
Region of residence (Ref.= Bangkok)
Central		-	-	-	-	2.16**	1.37, 3.40
Northern		-	-	-	-	4.01***	2.44, 6.59
Northeastern		-	-	-	-	2.94***	1.82, 4.73
Southern		-	-	-	-	1.73*	1.04, 2.87
Perceived number of food outlet types (Ref.= 3+ types)
Single type		-	-	-	-	2.88***	2.24, 3.71
2 types		-	-	-	-	1.74***	1.35, 2.26
Perceived food prices compared to other residential areas (Ref.= higher)
Unaware		-	-	-	-	0.68**	0.51, 0.90
Lower		-	-	-	-	0.30***	0.19, 0.46
Similar		-	-	-	-	0.63***	0.50, 0.79
Perceived food quality compared to other residential areas (Ref.= worse than)
Better		-	-	-	-	1.98*	1.07, 3.67
Average/unsure		-	-	-	-	1.63	0.91, 2.93
Perceived existence of food assistance compared to other residential areas (Ref.= No)
Yes		-	-	-	-	1.53***	1.24, 1.89
Model Fit		-2LL = 684.18, 2 = 127.91, df = 12, n = 2,641, p-value < 0.00		-2LL = 1,607.00, 2 = 190.91, df = 16, n = 2,641, p-value < 0.00		-2LL = 2,925.95, 2 = 403.45, df = 29, n = 2,641, p-value < 0.00
Pseudo R-square		Nagelkerke R2 = 0.07		Nagelkerke R2 = 0.10		Nagelkerke R2 = 0.20

^a: IL Individual level, HL Household level, CL Community level, LL Log Likelihood

* p-value < 0.05, ** p-value < 0.01, *** p-value < 0.00

In Model 3, which included food access variables at the individual, household, and community level, predictive ability for FI severity among older adults was strongest (χ2 = 403.45, -2LL = 2,925.95, Nagelkerke R2 = 0.20, df = 29, p < 0.00) compared to Models 1 and 2 (Table 2). The Test of Parallel Lines indicated that slope coefficients for FI severity levels were equal across response categories (p = 0.91), confirming the model’s suitability for Ordered Logistic Regression. Significant food access factors at the individual, household, and community level are described as follows:

At the individual level, older adults (age 65 to 69 years) were 1.40 times (95% CI: 1.05–1.87) more likely to experience greater FI severity than those age 75 + years. Those with no formal or incomplete primary education were 3.45 times (95% CI 1.84–6.47) more likely to experience higher FI severity than those with a bachelor’s degree or higher. Older adults with no employment income or earning 1 to 5,000 baht per month were 1.56 (95% CI 1.03–2.35) and 1.84 times (95% CI 1.17–2.90) more likely, respectively, to experience greater FI severity than those earning over 10,000 baht per month. Sex and marital status were not significantly associated with FI severity.

At the household level, older adults receiving healthy food prepared by a household member had 50% lower odds of higher FI severity (AOR 0.50,95% CI: 0.40–0.62) compared with those who did not. Those in households growing vegetables and fruits for consumption were 1.43 times (95% CI 1.16–1.76) more likely to experience severe FI. Living alone and receiving financial or resource support for food purchases were not significantly associated with FI severity.

At the community level, older adults living in a rural area were 1.30 times (95% CI 1.05–1.61) more likely to experience severe FI than those in an urban area. All regions had higher FI severity compared with Bangkok, with those in the Northern region 4.01 times (95% CI: 2.44–6.59) more likely. Perceiving only one or two types of food outlets increased the likelihood of higher FI severity by 2.88 (95% CI 2.24–3.71) and 1.74 times (95% CI 1.35–2.26), respectively, compared with three or more outlets. In contrast, perceiving community food prices as lower, the same, or being unaware reduced the likelihood of severe FI by 70% (95% CI 0.19–0.46), 37% (95% CI 0.50–0.79), and 32% (95% CI 0.51–0.90), respectively, compared with perceiving prices as higher. Perceiving food quality as better than elsewhere was associated with 1.98 times higher odds (95% CI 1.07–3.67) of increased FI severity. Similarly, perceiving food assistance in the community was associated with 1.53 times higher odds (95% CI 1.24–1.89) of increased FI severity.

Furthermore, the consistency in the predictive results between the ordered logistic regression (final model) and the binary logistic regression models reinforces the credibility and appropriateness of using ordered logistic regression for analyzing the severity of FI among older adults (see additional file 4).

Discussion

This study examined the situation and the influence of food access factors at the individual, household, and community levels on the severity of FI among older adults in Thailand. The findings demonstrate that food access is a significant factor influencing the severity of FI among older adults. This association is shaped not only by individual socio-demographic characteristics but also by household and community-level factors. These results are consistent with previous studies that applied the Social Ecological Model [6, 7, 9, 30] and emphasize the role of family and social support in addressing FI [6, 10, 17, 22, 35]. In particular, Neves Freiria et al. [9] highlighted that FI among older adults in low- and middle-income countries is not solely driven by limited financial resources, low educational attainment, social isolation, or physical limitations, but also by interpersonal barriers, lack of family and community food-related support, and gaps in food policy [9]. These findings underscore the importance of enhancing food access at multiple levels. Given the limited research in Thailand that has explored multilevel determinants of FI severity among older adults, this study was the first to examine food access among older Thais by linking individual, household, and community factors to FI severity. The study provides an important contribution to the national evidence base, and should help inform the development of effective policies and targeted interventions to improve food access for the ageing population.

The analysis of socio-demographic factors at the individual level of food access indicates that Thais who are in the early stages of aging, with no formal education or low educational attainment, or with low income were significantly more likely to experience higher levels of FI. In particular, low income is directly linked to limited purchasing power and restricted access to food outlets. This aligns with previous studies demonstrating that older adults with limited income have fewer financial resources to secure sufficient food, whereas those with higher income are better positioned to cover food-related expenses, such as transportation or food delivery services, thereby increasing their access to a wider variety of food sources [20, 32, 36]. These sources include supermarkets, shopping centers, convenience stores, and outlets offering healthier food options [33]. In addition, older Thais age 65–69 years were in a period of life when they had already stopped working for some time after the typical retirement age of 60. Some continued to serve as household heads responsible for family expenses, which reduced the savings available for food consumption [7, 16] and those with lower educational attainment were found to have reduced employment opportunities, leading to lower income [17, 37]. Moreover, limited or no formal education was associated with a lack of knowledge about accessing food sources and difficulty in making decisions regarding food prices and quality [21, 32, 33, 36], which in turn contribute to insufficient food availability and an increased risk of FI [17, 37].

Household-level food access has a statistically-significant association with the severity of FI among older Thais. Importantly, the provision of healthy meal preparation by family members reflects the availability of food and the presence of household support, which contributes to a reduced likelihood of experiencing FI [35]. Such support is particularly important for older adults with physical limitations, as it facilitates food access, promotes adequate consumption, and enhances dietary diversity compared to a situation where older adults must prepare and consume food independently [35]. These findings suggest that living arrangements in which older adults reside with household members may contribute to strengthening their food security. Additionally, the finding that households of older adults who cultivate fruits and vegetables for consumption have a higher likelihood of experiencing severe FI contrasts with evidence from studies conducted in Pacific Island countries, where home gardening has been shown to promote household food security [38]. This discrepancy may be partly explained by the FIES, which focuses on food access in relation to financial and material resources rather than food production itself. Resource-constrained households may therefore both experience FI and engage in home cultivation as a coping strategy. Moreover, contextual differences between Thailand and Pacific Island settings may further account for these divergent findings. In this study, home gardening likely reflects an adaptive response to pre-existing FI rather than an indicator of adequate food availability.

Community-level food access is particularly noteworthy, as factors at this level have greater predictive power for the severity of FI among older Thais than those at the individual or household level. Notably, urban-rural differences play a significant role: Older adults living in a rural area are more likely to experience severe FI compared to those in an urban setting. This finding aligns with previous research indicating that basic transportation infrastructure in rural areas is often limited, inefficient, and unsafe, hindering mobility -- especially in geographically-remote regions and, thereby, reducing access to food sources. Additionally, access to food assistance programs is generally more difficult in rural communities than in urban [20, 21]. Furthermore, rural residents typically have lower incomes than their urban counterparts [17], which exacerbates barriers to purchasing food, including travel expenses to reach food outlets, leading to insufficient food intake and greater risk of FI among older adults [20, 21]. Older Thais residing in geographic regions outside Bangkok were more likely to experience FI than their counterparts in Bangkok, reflecting broader urban–rural disparities. This issue was most pronounced in the North, where lower household incomes and higher poverty rates prevail [39]. Moreover, the region’s predominantly mountainous terrain may create additional barriers to transportation which further constrain access to food markets [26, 40, 41].

Another important finding concerns the diversity of food retail outlets in the community. Older adults living in communities with a wider variety of food sources such as fresh markets, mobile vendors, grocery stores, convenience stores, supermarkets, and shopping malls have increased food security. This suggests that the availability and readiness of diverse food outlets in the community facilitate greater access to food for older adults. Conversely, communities with only one type of food outlet may limit food access due to restricted options and distances, reducing the availability and variety of food types [20].

Moreover, older adults who perceive local food prices as lower than elsewhere may experience greater food affordability, and thus experience a higher likelihood of food security compared to those who perceive prices as higher [20, 42, 43]. However, this study found that older Thais who perceive the quality of food in their community as better than elsewhere are more likely to experience severe FI. This may reflect that higher-quality foods are often more costly, limiting access for the lower-income older Thais. Consequently, the presence of high-quality foods in a community does not necessarily ensure adequate access or consumption for all older adults [2, 44].

An additional noteworthy finding relates to the perception of food assistance within the community. Older adults who are aware of food aid programs are often those experiencing FI, consistent with prior studies showing that individuals with insufficient income or financial resources tend to rely on community food assistance or participate in government welfare programs [6, 20, 22, 35]. Moreover, awareness of such assistance is associated with lower educational attainment, which may reflect limited access to information and ineffective communication channels to obtain community-based food support [37].

Policy implications

The findings highlight important policy areas to improve food access and reduce FI among older Thais. Policies should address determinants of food access at the individual, household, and community levels, especially given the rising FI prevalence in Thailand’s aging population. At the national level, the government should develop welfare programs to help older adults (especially those experiencing FI) secure adequate and accessible food through coordinated efforts among relevant ministries. Examples include food voucher programs offering discounts at participating vendors, policies empowering local government and community initiatives, and support for research to comprehensively strengthen food security. Such food policies should also consider the coverage, quality, and continuity of assistance. At the local level, efforts should focus on enhancing food access among older adults and encouraging local agencies to adopt the FIES as a tool to assess FI, thereby identifying those at risk. Local agencies should also integrate their work with relevant stakeholders to strengthen support systems and improve food welfare programs, ensuring that state-sponsored initiatives are both effective and accessible to all older adults, especially those with FI. These actions could be supported through existing local Community Health Security Funds to facilitate effective implementation. Such support may include the establishment of centers for older adults facing a food shortage, either in normal or emergency situations, combined with screening, counseling, care, and recovery services. Communities can develop senior-friendly food markets connecting various food outlets to improve access to affordable, safe, and suitable food. These markets should serve mobile older adults, and provide delivery for those with limited mobility. Vocational training should be offered to older, low-income Thais to improve income and food access. Household education on food management and meal preparation should be promoted to ensure adequate nutrition. Finally, rural infrastructure such as transportation, lighting, and roads should be improved to enable safe and convenient access to food, especially in remote areas.

Limitations

This study has some limitations. First, the cross-sectional design of the survey (which was the source of the data) permits identifying statistically-significant associations, but does not allow testing for causality between FI and other factors. In addition, the literature indicates that other factors may be associated with food insecurity and explain greater variability, including chronic conditions, BMI, financial management capacity, functional ability and self care, ownership of agricultural assets or land, dietary diversity, distance to food outlets, modes of food access, and state support policies [6, 8, 10, 31, 34]. These factors could not be included due to limitations of the secondary data. Future longitudinal or qualitative studies are needed to clarify any causal relationships. Second, food access data at the household and community level were based on individual perceptions rather than information directly collected or aggregated at the community level. This may undermine the collection of accurate data on the number, density, or location coordinates of stores, food outlets, and local food production sites [45]. However, FI data from the FIES also reflect personal experiences over the past year, aligning with this subjective nature. Third, although food access was examined at the individual, household, and community level, multilevel modeling was not possible due to sample limitations that did not meet statistical assumptions [46, 47].

Conclusions

Food insecurity (FI) among older adults is an increasingly important issue as populations worldwide, including Thailand, continue to age. This study demonstrates that food access across the individual, household, and community characteristics, is significantly associated with the severity of FI. Older Thais who have lower income and less education are more likely to face serious FI, whereas those who get help from family members in preparing healthy meals are less at risk. People who live in a rural area, especially where there are few places to buy food, also have a higher chance of suffering FI. These findings highlight the need for designing or improving policies that enhance food access for older Thais at all levels, including through national programs such as food support and welfare initiatives. Programs that consider the important role of families in providing healthy food for older adults within households should be implemented. Community-based action is also essential. Strengthening food systems and support services, both in everyday life and during emergencies, is crucial to ensure that older Thais have adequate food.

Abbreviations

FI

Food insecurity

FAO

Food and Agriculture Organization

NSO

National Statistical Office

EA

Enumeration area

FIES

Food Insecurity Experience Scale

AOR

Adjusted odds ratio

CI

Confidence interval

IL

Individual level

HL

Household level

CL

Community level

Authors’ contributions

J.S. conceptualized, designed the study, conducted statistical analysis, and wrote the manuscript. S.P., S.S., R.G., A.C., and U.P. reviewed the methodology. All authors reviewed and approved the final version manuscript.

Funding

Open access funding provided by Mahidol University. This study was supported by the 60th Year Supreme Reign of His Majesty King Bhumibol Adulyadej Scholarship for the 2023 academic year, and the Scholarship for Thesis Promoting SDGs Policies for the 2025 fiscal year (Contract No. MUGR SDGs68-10), both from the Faculty of Graduate Studies, Mahidol University. The scholarships had no role in the design, conduct, or reporting of this research.

Data availability

Data are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The authors affirm that all research methods and procedures employed in this study complied with the principles and requirements outlined in the Declaration of Helsinki. Ethical approval for this study was obtained from the Institutional Review Board of the Institute for Population and Social Research, Mahidol University (IPSR-IRB), Certificate of Approval No. 2025/04–020, dated April 24, 2025. All data were analyzed in aggregate form without identifying individual respondents, and no direct contact with participants was involved.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.