Development of a Comprehensive Household Food Security Tool for Families with Young Children and/or Pregnant Women in High Income Countries

Despite increasing rates of food insecurity in high income countries, food insecurity and its related factors are inconsistently and inadequately assessed, especially among households with young children (0–6 years) and pregnant women. To fill this gap, researchers from the U.S. and Australia collaborated to develop a comprehensive household food security tool that includes the known determinants and outcomes of food insecurity among parents of young children and pregnant women. A five-stage mixed methods approach, including a scoping literature review, key informant interviews, establishing key measurement constructs, identifying items and scales to include, and conducting cognitive interviews, was taken to iteratively develop this new comprehensive tool. The resulting 78-item tool includes the four dimensions of food security (access, availability, utilization, and stability) along with known risk factors (economic, health, and social) and outcomes (mental and physical health and diet quality). The aim of this novel tool is to comprehensively characterize and assess the severity of determinants and outcomes of food insecurity experienced by households with young children and pregnant women.


Introduction
Food insecurity impacts millions of people globally; nearly one in three people did not have access to adequate food in 2020; an increase of 320 million people in just one year [1]. While there has been a significant focus on hunger and food insecurity in low-and middle-income countries, there is an increasing recognition that 'hidden hunger' and food insecurity also exist in high income countries (HICs). In Australia, an estimated 3.4 million people (13.5% of the population) experience food insecurity [2]. In the United States (U.S.), an estimated 13.8 million households (10.5% of the population) are impacted [3]. In both countries, there is an inconsistent and inadequate assessment of food insecurity and related factors that families experience which hampers effective programmatic and policy responses.
According to the Food and Agricultural Organization of the United Nations (FAO), achieving food security rests on four dimensions. The first is the provision of enough food to support a healthy lifestyle (availability), the second is food being economically and physically accessible to all people (accessibility), the third is whether people are able to utilize the food they acquire (utilization), and the fourth is how stable the availability, accessibility, and utilization of food is (stability).
There is considerable interest, debate, and uncertainty surrounding the best way to measure food insecurity in HICs. Ashby et al. examined food insecurity measurement tools utilized in this context, assessing the extent to which these tools capture the four dimensions of food security: access, availability, utilization, and stability [4]. The review found that most tools only measured the access dimension, meaning that existing tools are unable to determine the ability of a household to be food secure beyond their ability to physically access and financially afford food. McKay et al. examined how food insecurity has been measured in the Australian context over the last fifteen years, revealing that most tools utilized only measured the access dimension of food insecurity [5]. The U.S. Department of Agriculture (USDA) Household Food Security Survey Module is widely used and validated in the U.S. context, however, this tool is limited as it also only assesses the financial access dimension of food security [4].
The determinants of food insecurity are multifactorial and can occur across multiple levels of the social-ecological model: intrapersonal, interpersonal, organizational, community, and public policy [6]. Determinants of food insecurity include poverty, social and economic disadvantage, individual characteristics (e.g., gender, ethnicity), and the impact of the political and social environment [7]. However, given the limited focus on the access dimension of food insecurity in the most common measurement tools, many estimates of food security do not capture these additional determinants. In addition, most research to date has focused on the general population and, as a result, an understudied population with regard to food insecurity are the parents of young children (ages 0-6) and pregnant women, who are known to experience negative consequences of food insecurity [8,9].
With a global focus on improving maternal and child nutrition through the 2030 Sustainable Development Goals, it is important to understand the experience of food insecurity among pregnant women and families with young children, as food insecurity during these life stages can have ongoing negative health consequences. Women are at risk of becoming food insecure due to entrenched societal power inequality and a range of socioeconomic conditions such as domestic violence, poor employment, and education [10,11]. Households with children are at increased risk of food insecurity, as such, mothers and, in particular, single mothers experience a higher rate of food insecurity compared to women without children [11]. Living in a food insecure household during pregnancy may increase the risk of excess gestational weight gain, disordered eating, chronic disease, and various pregnancy complications [12]. The impact of food insecurity on young children is particularly concerning given they are at a key stage of growth and development which can influence health during adolescence and even adulthood [13]. It is important to understand food insecurity among families with young children and pregnant women in HICs as there is limited evidence that examines factors associated with food insecurity among these populations. Identifying such factors is essential to understanding how and when strategies that specifically target food insecurity in this population group may be implemented.
To fill this gap, researchers from the U.S. and Australia collaborated to develop a comprehensive household food security tool that includes the known determinants of food insecurity among parents of young children and pregnant women and to extend existing tools that measure the access dimension of food security. The aim of this study was to develop a new comprehensive household food security tool to characterize and assess the severity of determinants and outcomes of food insecurity experienced by households with young children and pregnant women.

Materials and Methods
A five-stage mixed methods approach was undertaken to iteratively develop a comprehensive household food security tool. Deakin University provided human research ethics approval (2020-038) and included the University of Nebraska (0642-20-EX). The study was designed and has been reported in accordance with the COREQ checklist for qualitative studies [14]. The first two stages of this study are summarized below and have been published with full details available [15,16] with the next three stages reported in full detail for the first time below.

Stage 1
In Stage 1, a scoping literature review was conducted to identify the factors associated with food insecurity among pregnant women and households with young children aged 0-6 years living in HICs. Scoping reviews can be used to identify knowledge gaps, scope a body of literature, and identify key characteristics or factors related to a concept [17]. The scoping review was conducted to investigate factors that influence food insecurity among pregnant women and households with young children (aged 0-6 years) in HICs.
A comprehensive systematic search informed by Peters et al. [18] was conducted in four databases: Medline complete, Embase, Global Health, and CINAHL. These databases were chosen to provide coverage of public health nutrition and nursing and allied health literature in HICs. Search terms were relevant to food insecurity, determinants, pregnancy, and family. The search strategy involved combining the search terms and all terms were searched in the title and/or abstract.
Inclusion criteria were original research articles published in peer reviewed journals in the English language conducted with families or households incorporating pregnant women and/or caregivers of young children aged 0-6 years in high income countries (defined by the Human Development Index). Outcome measures included food security (measured in any way) and/or other aspects of food security as defined by the Food and Agriculture Organization (FAO) such as affordability, access, utilization, or stability.
Relevant information was extracted using predetermined categories including country of study, setting/population, study design, food insecurity measurement tool(s) utilized, and outcomes. Consistent with the approach taken in scoping reviews, articles were not assessed for quality [17]. The range of factors was identified and grouped into 13 overarching constructs. Detailed methods and findings of this scoping review have been published elsewhere [15].

Stage 2
In Stage 2, qualitative interviews were conducted with 41 pregnant women or households with young children under 6 years of age who were experiencing or who were at risk of food insecurity in Omaha, U.S. (n = 19) and Melbourne, Australia (n = 22). A detailed description of the methods and results for these qualitative interviews have been reported [16].
A multi-faceted approach was used to recruit participants. Nutrition and food organizations in Melbourne and Omaha were invited as recruitment sites. One maternity hospital in each city was also approached along with federally subsidized preschools in Omaha and online groups for Melbourne mothers. For those willing to assist with recruitment, fliers were provided to distribute via social media platforms and/or in hard copy. Potential participants contacted the researchers and were provided further information and telephone/online interviews via Zoom [19] were arranged when convenient.
A semi-structured interview guide was based on the findings of the scoping review (stage 1) and designed to gain perspectives on the constructs that influence household food security, key coping strategies, and social, economic, and health conditions that buffer from or exacerbate the experiences of food insecurity. The questions invited participants to provide information about their households (including children, partners, and other household members) and their experiences of feeding a family on a budget. Additional probes were used for pregnant women to elicit pregnancy-specific outcomes.
NVivo 12-assisted coding [20] was completed to establish themes and subthemes from the data, informed by Braun and Clarke's six-step process [21]. First, all interviews were transcribed iteratively by professional services in each city and checked for accuracy. Secondly, initial codes were developed in line with the interview guide topics and with input from all authors. Authors met regularly to discuss coding and emerging insights and to identify data saturation.

Stage 3
Stage 3 involved mapping the themes and subthemes from the qualitative interviews to the constructs found within the scoping review. One researcher (A.B.) conducted the initial mapping which was reviewed by co-authors (C.P., F.H.M., P.v.d.P., and S.M.). Further analysis of the food insecurity themes and constructs was conducted by comparison with known food security frameworks [15,22] to explore how the themes and constructs aligned. Discussion and a consensus on themes and constructs were reached amongst co-authors.
A conceptual basis for the tool was agreed upon by organizing the mapped food security constructs and themes into three components of the household food insecurity experience relevant to families with young children and pregnant women. These were (i) individual or household risk factors for food insecurity (e.g., economic, health, and social), (ii) the four dimensions of food security (e.g., access, availability, utilization, and stability), and (iii) the health outcomes or consequences of food insecurity.

Stage 4
In Stage 4, a comprehensive multi-dimensional measurement tool was drafted based on approaches used previously in public health, allied health, and medicine, including reviewing the literature, identifying and/or writing items, and subsequent field testing [23].
Using the agreed conceptual basis (i.e., risk factors, dimensions, and outcomes), appropriate items to include in the tool were identified from reviewing the literature and current measurement tools. These items were then reviewed against the following criteria for selection: use of validated instruments for the risk factors, dimensions, and outcomes of interest; use of brief or short items to reduce participant burden; and use of items used commonly in Australia or the U.S. and/or tested in this population group and/or would allow comparison with national health surveillance.
An extensive list of possible items to include in the tool was drafted and refined by assessing each item against the selection criteria. If no existing items could be found for a construct or theme, the research team developed item(s) taking into consideration the factors suggested by de-Vet et al. [23] for item development, including target population, the purpose of measurement, the difficulty of the items, the application in practice, and response options.

Stage 5
In Stage 5, cognitive interviews using the drafted tool were conducted with pregnant women or households with children aged 0-6 years who were experiencing or at risk of food insecurity in Melbourne, Australia (n = 11). Budget constraints limited the piloting to one setting. Cognitive interviewing is a psychologically oriented method for empirically studying the ways in which individuals mentally process and respond to survey items [24]. It can be useful in pretesting items and determining how they should be modified to make them easier to understand and answer [25].
Recruitment for the cognitive interviews included contacting participants who had been involved in the semi-structured qualitative interviews conducted in Melbourne (Stage 2) and who had indicated a willingness to be contacted for this additional research component. In total, 9 out of 22 women indicated they were willing to be contacted and 4 agreed to participate in the cognitive interview. The study also was advertised on social media via new mothers' Facebook groups, and seven women were recruited via this method. Participants completed an online consent form and were contacted by a researcher (A.B.) via phone or email to arrange a convenient interview time.
Interviews were conducted online via Zoom [19] during a two week period (late May, early June 2021) and were, on average, 39 min in length (range from 23 min to 76 min). Interviews were audio recorded and cross-checked against field notes captured by interviewers upon completion of the interview. Participants received $40 in supermarket vouchers as compensation for their time.
During the cognitive interviews, respondents completed the comprehensive household food security tool for families with young children and/or pregnant women with an interviewer while being asked open-ended and probing questions about their responses.
Respondents were asked to read and answer the questions out loud and to talk through their decision-making process for each question. This allowed the researcher to discern the participant's comprehension of items and answer choices.
Respondents were also invited to give general feedback on questions and specific probes were developed for questions by the research team. For example, at the end of the questions on demographics, respondents were asked to reflect and answer the following questions: How hard or easy were these questions to answer? Were any of these more difficult? If so, why? What were you thinking about as you answered? After the question on their main sources of income, respondents were asked: How easy or difficult was it for you to identify your main source of income? Were any of the items confusing or needing further explanation? This allowed for further exploration of how the questions were constructed and suggestions for how these items could be improved.
The interviews were conducted by researchers (A.B. and R.L.). Feedback from cognitive interviews was collated across each item and discussed with co-authors (C.P., F.H.M., and P.v.d.P.). Suggested changes resulting from this feedback were discussed by the research team and the resulting changes and finalization of the tool were agreed upon by all authors.

Results
The full details of the scoping literature review and qualitative interviews have been published elsewhere (References [15,16]). In brief, findings from the qualitative interviews identified a range of prominent themes associated with food insecurity. The constructs from the scoping review (Stage 1) were mapped against the themes and subthemes arising from qualitative interviews (Stage 2) to identify 13 food security constructs and themes in pregnant women or households with young children aged 0-6 years old (Table 1). Table 1. Constructs and themes associated with food security in pregnant women or households with young children aged 0-6 years old.

Food Security Construct
Constructs and Outcomes Identified from Scoping Review (Stage 1)

Themes and Outcomes Identified from Qualitative Interviews (Stage 2)
Income and employment Low income, job loss, and payment schedules; not receiving welfare; low social economic status (i.e., education, occupation, and household income); living below the poverty line; and mothers as homemakers Employment; government assistance (e.g., accessing programs, the trade-off between earning income/losing assistance, and running out of assistance); competing expenses (e.g., other bills, children's activities, time of year, and special occasions) Coping strategies Stretching food, going without food, and skipping meals; cutting back on the variety of foods consumed; going to bed hungry; cooking whatever is available, buying cheaper food, shopping at value stores, and using coupons; reducing money spent on children's education and activities; borrowing money; and food and social supports Utilizing resources (e.g., food pantries, utility/bill aids, and other non-government programs); budgeting skills (e.g., couponing, bargain shopping, and buying cheaper foods); family and friend support (e.g., food, money, or other resources provided; social support); rationing (e.g., making food last all month, limiting intake); nutrition knowledge and skills (e.g., being a good home cook)  A conceptual basis for the tool was established by organizing the identified food security constructs and themes into three components of the household food insecurity experience. These were organized into: (i) individual or household risk factors for food insecurity (economic, health, and social), (ii) four dimensions of food security (access, availability, utilization, and stability), and (iii) health outcomes or consequences of food insecurity (Table 2). Table 2. The three components of the household food security experience-a conceptual basis for comprehensively measuring household food security in families with young children and pregnant women.

Individual or household risk factors for food insecurity
Economic: income and employment Health: stress and mental health, chronic health conditions Social: demographics including education, ethnicity, and household composition 2. The four dimensions of food security Utilization: resilience and coping strategies, kitchen facilities, nutrition skills and literacy, and participation in food assistance programs Access and availability: physical and financial access to foods and stores, food insecurity screener items Stability: annual competing expenses and challenging times of year 3. Health outcomes or consequences of food insecurity Health: stress and mental health, chronic health conditions Diet quality Using the three identified components described in Table 2, 60 suitable items were identified from the literature and existing measurement tools: 25 items on individual or household risk factors for food insecurity, 13 items on the four dimensions of food security, and 22 items on health outcomes or consequences of food insecurity.
Cognitive interviews with a sample of the population (n = 11) led to amendments, resulting in 56 improvements to the measurement tool. These are documented in Appendix A, Table A1. Changes included improving the overall readability (2), changes to question responses offered to make them more relevant (10), changes to question content to make them clearer (22), and adding new items (20) or deleting items (2). Examples of these types of changes are presented in Table 3. The resulting product is a comprehensive household food security tool for families with young children and/or pregnant women that includes 78 items that span across the three components described above. These can be grouped into 1 screening item, 27 items on individual or household risk factors for food insecurity, 27 items on the four dimensions of food security, and 23 items on health outcomes or consequences of food insecurity.

Discussion
This study developed a comprehensive household food security tool to examine social demographics and household characteristics, food security status, health, and dietary outcomes for parents and households with young children. The five-stage mixed method research was designed to devise a new tool that would expand the measurement and understanding of household food security beyond financial access. Food security is complex, as illustrated by the many frameworks that attempt to explain its multifactorial determinants and the number of evolving definitions that aim to articulate its various dimensions [6,7]. With our evolving understanding of food security, there is a need to develop and refine measurement tools that capture the diverse aspects of food security. Previous tools such as the USDA household food security survey module, Cornell Child Food Security Measure, Hager two-item screen, and Girard four-point tool have largely focused on the economic or financial aspect of food insecurity [4]. While this is certainly an important pillar of food security, it overlooks the physical accessibility to food, availability of healthy and affordable food, people's ability to utilize food, and stability of this experience across weeks, months, years, and even generations. This comprehensive household food security tool is unique as it includes items that measure the other dimensions of food security and is specific to households with children under the age of 6 years. Broadly speaking, questions on food literacy and coping skills uncover the ability to utilize food, questions on fruit and vegetable availability and the distance to nearest shops examine the availability dimension, while questions on how various events throughout a year impact food security examine the stability dimension.
The inter-household dynamics between parents/caregivers and their young children are known to be complex in a food insecure context, and hence, a specific tool to investigate these components among an understudied population is warranted. The conceptual basis developed during this research and used to underpin the comprehensive household food security tool highlights the complex nature of food security. Risk factors for food insecurity are commonly attributed to economic variables such as low income and unemployment [26][27][28]. Health risk factors such as maternal depression have also consistently been found to be higher among women experiencing food insecurity [29]. This relationship may be bidirectional as poor mental health has been associated with a transition into food insecurity [30] and food insecurity has been observed to precede depression [31,32]. Further, having a household member with chronic health needs and the associated health care costs is also a risk factor associated with food insecurity [33]. The third type of risk factor contributing to food insecurity is social and includes sociodemographic factors such as ethnicity, education level, marital status, family size, acculturation, and urban life stressors [33][34][35]. The known consequences or outcomes of food insecurity include poorer dietary quality [36,37], associations with diet-related chronic conditions such as diabetes and obesity [38,39], and elevated poor mental health including depression and anxiety [40]. This comprehensive household food security tool could be useful in postnatal, maternal, and child health settings to explore a more comprehensive set of components of food security. Women who report being food secure may in fact present with several economic, health, or social risk factors that, if identified, could enable supports to be implemented in these households to prevent the transition into food insecurity. Furthermore, as the comprehensive household food security tool captures health and dietary outcomes, the inter-relationships between food insecurity and these outcomes can be further explored to better understand the food insecurity experience in families with young children and/or pregnant women.
One strength of this research is that it combines multiple types of evidence into the iterative design of the final instrument. The scoping literature review followed by qualitative and then cognitive interviews with the target population allowed grounding of the tool in both lived experience and evidence. Impacted populations provide their own forms of evidence (knowledge, experience, ideas, and opinions) that aid in expanding the understanding of a given issue [41]. The information gleaned from the qualitative interviews helped to contextualize the constructs found within the scoping literature review to develop the final conceptual framework that underpins the comprehensive household food security tool. Further, the robustness of the tool comes not only from the efforts to generate constructs grounded in people's daily experiences but also from the efforts to select and generate high-quality items to include in the tool [42].
This tool and study have some imitations. Firstly, the tool is in English. Populations with English as a second language or who do not speak English at all are often at increased risk of food insecurity. Race, ethnicity, and acculturation were identified as factors associated with food insecurity within the literature [27,43], as such, there is a need to develop and test the tool in languages other than English so that food insecurity among diverse populations can be better understood and addressed. Secondly, in 2020, the fifth and sixth dimensions were proposed by the High-Level Panel of Experts on Food Security and Nutrition and include whether people have the ability to make choices and control their engagement with the food system (agency) and whether the food system is environmen-tally, economically, and socially sustainable (sustainability) (HLPE 2020). As this research commenced prior to the FAO high-level panel of experts recommending the agency and sustainability dimensions, the tool overlooks these two dimensions [44]. Thirdly, while the tool is comprehensive, its length might mean its use is limited to research or to settings where there is sufficient time available to screen.
Future research could explore agency and sustainability and incorporate items around them into the tool to expand the understanding of how these aspects impact food security specifically in this population group. This tool is at its prototype stage and several steps need to be devised before implementing the tool at a population level. These include the development of a scoring system to rate food insecurity status based on social, cultural, and economic risk factors and the severity of food insecurity based on answers to the availability, access, utilization, and stability dimensions of food security. The next step for this measurement tool is reliability and validity testing. Once these steps are undertaken, the tool could be useful in screening families at risk of food insecurity in a range of settings. Future research may also explore how researchers and practitioners may utilize specific scales and items from the comprehensive tool depending on their interests and goals. It could be useful in better measuring and understanding food insecurity in this population group so that policy and programs to address this issue can be devised.

Conclusions
There is a need for further research beyond the economic dimension of food security to truly understand this complex issue and be able to better identify and support those experiencing food insecurity. Further work is required to test the 78-item comprehensive household food security tool in various settings and populations for reliability and validity. Subsequently, the tool could be used to examine the relationship and increase understanding between risk factors, components of the food insecurity experience, and health outcomes. With an increased understanding of the issue, practitioners, policymakers, and governments will be better placed to identify and implement the required solutions.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of Deakin University (2020-038 4 May 2020) and the University of Nebraska (0642-20-EX).

Informed Consent Statement:
Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The comprehensive household food security tool can be obtained from the corresponding author upon email request.

Conflicts of Interest:
The authors declare no conflict of interest. Added new response option "Permanently staying with family or friends" (R) Changed homeowner response options to "Homeowner no mortgage" or "Homeowner with mortgage" (R)

Appendix A
What is your main source of income?
Wages or salary Any government pension or allowance Self funded retirement Nil or negative Don't know Any other regular source-please answer 14a Interview 5: Add in an additional question 'are you the main income earner in the home' Interview 7: Answer for self or partner?
Changed question item wording to include household so the question now reads "What is your main source of household income"? (Q) Changed wording in this question to include the text "members of your household who you share finances with" (Q) Added text "before tax is taken out" to question (Q) Added text "child support" into question (Q)     Added in a question prior to this; "Do you follow a vegetarian or vegan diet?" If they select no then they are asked to answer this question, if they select partly (pescatarian) then they are asked to skip this question but answer other questions on fish, if they select yes then they can skip the question (A) Have included the word "frozen" in the question item so the question now reads; "Over the last month, how often did you eat fresh or frozen meat (including beef, veal, chicken, lamb, pork)?" (Q) Over the last month, how often did you eat take away or fast foods (such as burgers, chips, pizza, Indian)? Include foods eaten at the restaurant or at home (e.g., Uber eats, take away) Never Less than once per month 1-3 times per month 1 time per week 2 times per week 3-4 times per week 5-6 times per week 1 time per day 2 times per day 3 or more times per day