Chronically ailing or failing chronically: a typology of South African diners living with diabetes

ABSTRACT

Dining out serves both nourishment and entertainment needs. This study segments diabetics based on their dining out motivations, restaurant preferences, and satisfaction levels. The research in South Africa utilised an online survey on QuestionPro, resulting in 350 responses. The analysis took place in three stages: (a) socio-demographic profiling, (b) exploratory factor analyses (EFAs) on dining motivations, restaurant choice factors, and satisfaction, and (c) clustering to identify segments. Out of 350, only 239 complete responses were valid for cluster analysis. Four dining motivations were identified: celebration and socialisation, tourism and novelty, vitality and appreciation, gastronomy and enjoyment, and convenience and practicality. Restaurant choices hinged on value and relaxation and E-WOM and quality. Dining satisfaction (dining experience) emerged as a singular factor. The Two-step cluster analysis revealed two diabetic diner groups: Affluent, ageing, frequent, Caucasian, discontent Type 1’s and Middle-class, mature, infrequent, multi-racial, content Type 2’s. Recommendations for the hospitality sector include adaptability to evolving consumer demands, offering nutritional details, minimising food wastage, catering to health-conscious patrons, enhancing point-of-sale systems for meal customisations, and staff training for diabetic-friendly requests. The research improves understanding of diabetic dining behaviour in South Africa, unveiling relationships between variables defining two diabetic diner segments.

KEYWORDS:

• Market segmentation
• people living with diabetes
• dining behaviour
• hospitality management
• health belief model
• integrated behavioural model

Introduction

Wellness tourism has inconsistent meanings depending on cultural and philosophical perspectives (Voigt, 2013). The term is often used interchangeably with health tourism (Smith & Puczkó, 2015), medical tourism (Sandberg, 2017), thermal tourism (Pereira et al., 2023), spa tourism (Mihók & Marčeková, 2022) and well-being tourism (Smith & Diekmann, 2017). Conversely, wellness tourism broadly includes consideration of mental, physical, emotional and spiritual health and well-being (Majeed & Gon Kim, 2023). It should be noted that wellness, well-being and health perpetuate beyond the absence of illness, disease or infirmity (World Health Organization, 2024). Moreover, health is also impacted by various socioeconomic, environmental and biological factors (WHO, 2024). Considering these factors, the impact of food on wellness is clear. Food and access to nutritious food are linked to socioeconomic status, where low socioeconomic status (SES) can be a marker of a more “obesogenic” food environment (Kininmonth et al., 2020). As an agricultural product, food has a bi-directional impact on the environment. Food production contributes to greenhouse gas emissions (around 25%) and, subsequently, climate change, while climate change (droughts and floods) impacts food production (Mrówczyńska-Kamińska et al., 2021).

Biologically, food and water (nutrition) are each human’s most important physiological needs (Maslow, 1943) and are considered a universal human right (Article 25; United Nations General Assembly, 1948). Understanding the intersection between wellness, well-being, health, and food is paramount in creating just and equal societies and addressing specific sustainable development goals (SDGs). According to Tomičić (2023), food security is essential for human well-being and the overarching goal of poverty eradication. Ensuring food security involves providing all individuals with adequate, safe, and nutritious food that meets their dietary needs and preferences. This encompasses the availability of food and its accessibility and utilisation.

Unique intersections between food, health, well-being and wellness require individual investigations, especially regarding at-risk people. According to the WHO (2023), non-communicable diseases (NCDs), also known as chronic diseases, can be classified depending on the associated risk factors. Risk factors include behavioural (tobacco use, excessive salt/sodium use, alcohol abuse, insufficient physical activity), metabolic (raised blood pressure, overweight/obesity, hyperglycaemia [raised blood glucose], hyperlipidaemia [raised levels of fat in blood]), and environmental risks (pollution, radiation, noise, land use patterns, work environment, climate change). The main types of NCD include cardiovascular (heart attacks and stroke), cancers, chronic respiratory diseases (chronic obstructive pulmonary disease and asthma) and diabetes (WHO, 2023). The impact on health, wellness and well-being also disproportionately affects people in low- and middle-income countries, where more than three-quarters of global NCD deaths (31.4 million) occur (WHO, 2023).

In this exploratory study, we investigate a critical health concern within South Africa, an upper-middle-income country currently facing significant challenges with obesity and diabetes. Diabetes stands as a principal health issue, being the leading underlying natural cause of death among women and the second highest for the entire population (Sifunda et al., 2023). These health challenges are alarming due to their immediate impact on the population’s well-being and a substantial burden on the country’s healthcare system and economy.

Acknowledging the severity of this crisis, the South African government and various health organisations have initiated several strategies to mitigate the effects of these non-communicable diseases (NCDs). Notable among these is the National Strategic Plan for the Prevention and Control of NCDs (NSP-NCDs), launched by the National Department of Health (NDoH) in 2022. This strategic plan aligns with the Sustainable Development Goal (SDG) target 3.4, focusing on reducing premature mortality from NCDs through prevention and treatment and promoting mental health and well-being (NDoH, 2022). One of the primary objectives of this strategic plan is to promote healthy nutrition across all stages of life, an initiative that directly addresses the growing concerns over diet-related diseases like obesity and diabetes.

In parallel with governmental efforts, campaigns encouraging healthier dietary choices have been deployed, with entities such as UNICEF South Africa (2022) leading initiatives that educate and motivate the public towards adopting more nutritious eating habits. Collectively, these campaigns and policies aim to curb the rising tide of NCDs by fostering an environment where healthy choices are accessible and encouraged.

Our research focuses on recognising a gap in the literature concerning how people living with diabetes (PLWD) in South Africa manage their dietary needs, particularly in the context of dining out. Given the critical link between diet and nutrition and the management of diabetes, understanding the dining preferences, motivations, and needs of South Africans living with diabetes is essential. Therefore, our study aims to segment diners living with diabetes based on their profiles, dining motivations, dietary needs and preferences, and preferred restaurant attributes. This segmentation provides valuable insights for the hospitality industry, enabling it to tailor its offerings to better cater to this significant segment of the population. By doing so, the sector can play a crucial role in supporting public health objectives, particularly promoting healthy nutrition as outlined in the NSP-NCDs and overall wellness.

This research is pivotal for several reasons. It not only fills an existing gap in the academic literature by shedding light on the dining preferences and needs of South African diabetics but also underscores the importance of providing nutritional information to facilitate informed dining decisions. Such information is crucial for individuals managing diabetes, enabling them to make choices that align with their health requirements. Through this study, we contribute to the broader discourse on public health and nutrition, particularly in the context of NCDs, and offer practical insights that can help enhance the quality of life for PLWD.

Literature review

The diabetes pandemic

Insulin is pivotal in moderating blood glucose levels (Van den Berg & Webber, 2019). Diabetes, a chronic non-communicable disease (NCD), presents in three prevalent forms. As delineated by Mukhtar et al. (2020), Type 1 diabetes, also known as autoimmune diabetes mellitus, emerges suddenly due to insufficient insulin production by the pancreas. Conversely, Type 2 diabetes manifests gradually, often correlated with obesity, stemming from diminished insulin production or its ineffective metabolic utilisation. This progression can range from pre-diabetes to full-fledged Type 2, contingent on blood glucose levels (NIDDK, 2017). Additionally, gestational diabetes, marked by elevated blood glucose levels, emerges during pregnancy. The heightened hormone production and physiological changes like weight gain during pregnancy can induce insulin resistance, impacting up to 10% of pregnancies annually (CDC, 2022). Hyperglycemia, characterised by elevated blood glucose, results from unregulated diabetes, inflicting profound harm over time, particularly to the nervous system and circulatory structures (WHO, 2021). Alarmingly, in 2019, diabetes directly led to 1.5 million fatalities worldwide. An estimated 9.3% of the global population grapples with this NCD, with a concerning half remaining undiagnosed, precipitating detrimental health outcomes (Saeedi et al., 2019).

The prevalence of diabetes is country-specific, and the International Diabetes Federation (IDF, 2021a) estimates that 537 million adults live with diabetes worldwide. The IDF also projected growth for each region from 2021 to 2045, with an average global increase of 46%. Increases per region are displayed in Table 1, with Sub-Saharan Africa (SSA) projected to increase substantially (134%).

Table 1. Number of people with diabetes worldwide and per IDF Region in 2021–2045.

Year	World	North America and the Caribbean	Europe	Western Pacific	South and Central America	Sub-Saharan Africa	Middle East and North Africa	South-East Asia
2021	537	51	61	206	32	24	73	90
2030	643	57	67	238	40	33	95	113
2045	783	63	69	260	49	55	136	152
Increase (%)	46%	24%	13%	27%	50%	134%	87%	68%

*Number of people in millions.

Source: Adapted from IDF (2021b).

About half (58%) of the IDF SSA region countries lack high-quality in-country data sources on diabetes (IDF, 2021a). One in 22 people in SSA have diabetes; however, 54% are undiagnosed – the highest among IDF regions. The SSA region also has the second lowest diabetes-related expenditure, representing 1% of global spending. In South Africa, diabetes prevalence is estimated at 11.3% (IDF, 2023). According to Statistics South Africa (StatsSA, 2021), the top six underlying natural causes of death in 2018 include tuberculosis (6%), diabetes (5.9%), cerebrovascular disease (5.1%), heart disease (5.1%), HIV (4.8%), and hypertensive disease (4.5%). Diabetes-related complications often impact large and small blood vessels, affecting the heart, eyes, kidneys, and nerves (WHO, 2021). In South Africa, a state-of-the-art diabetes centre at Groote Schuur Hospital was opened in 2021, the first of its kind in Africa (Western Cape Government, 2021). This indicates that public healthcare access in many developing countries in Africa is lacking and exacerbated by the fact that Africa is home to 70% of the least developed countries (LDCs) globally (United Nations Department of Economic and Social Affairs [UNDESA], 2021). According to the Global Obesity Observatory (GOO, 2022), South Africa has a national obesity risk score of 8 out of 10.

Theoretical underpinnings

Three foundational approaches, theories or models illuminate the research, offering insights into the dining habits of people living with diabetes (PLWD).

Firstly, the market segmentation approach postulates that segmenting a broad market into more distinct, homogenous sub-markets can be advantageous (Dickson & Ginter, 1987; Smith, 1956). These segments can be differentiated based on consumer needs, attributes, and actions (Dolnicar, 2014). Variables for segmentation can be categorised into socio-demographics, geographics, psychographic motives, and behavioural inclinations (Dolnicar, 2020; Dolnicar et al., 2018). This study intends to utilise a comprehensive segmentation methodology, incorporating all these bases, to profile diabetic diners in South Africa.

Secondly, the Health Belief Model (HBM) postulates that individuals are more inclined to modify their behaviour given a credible reason (Becker, 1974; Hochbaum, 1958; Kirscht, 1974; Rosenstock, 1960, 1974). A more comprehensive HBM introduces elements like individual perceptions, external factors, and potential actions (Aalto & Uutela, 1997). The model revolves around two principal constructs: behavioural beliefs, which gauge the pros and cons of a specific behaviour (Jin et al., 2017), and perceived risk, which assesses potential vulnerabilities and consequences if a behaviour is not adopted (Becker, 1974). For a PLWD dining out, these elements can emerge as the juxtaposition between meal selections, with socio-psychological variables complicating the decision. The challenge arises from balancing the pleasure of certain food choices with the potential health risks, especially when dining in a social context.

Lastly, the Integrated Behavioural Model (IBM) is pivotal in elucidating and anticipating behaviours (Fishbein, 2009). It amalgamates aspects of the Social Cognitive Theory (Bandura, 1986), the Theory of Reasoned Action (TRA) (Fishbein & Ajzen, 1975), and the Theory of Planned Behaviour (TPB) (Ajzen, 1991). Within IBM, intention is seen as the foremost behavioural predictor, emphasising personal habits, capabilities, and situational variables (Fishbein & Ajzen, 2011). IBM also accentuates the influence of societal settings, physical environments, and public policies on health behaviours (Sallis et al., 2008). Thus, when considering PLWD’s dining choices, several factors like knowledge, environmental limitations, and habitual choices play a role in their decision-making (Yusof et al., 2018).

The seminal work on general and diabetic diners’ preferences

Food is a universal connector in a diverse cultural setting (Ngcakani, 2021). While eating meets physiological needs as identified by Maslow in 1943 (Navy, 2020), dining out, which incorporates entertainment (Silverberg, 2019), dominates the food service sector, contributing to 55% of food expenditure in the USA in 2021 (USDA, 2022). Full-service and fast-food outlets have the majority market share (USDA, 2022), underscoring the food and beverage (F&B) sector’s significance in global economies (Hazari, 2021). Dining out has become a common leisure activity, with consumers facing numerous restaurant choices and intensifying competition (Cleave, 2020; Harrington et al., 2011; Jones, 2018). To thrive, restaurateurs must discern what drives consumers’ dining choices (Chua et al., 2020). Food preferences vary globally based on sensory attributes, health and wellness considerations, and price (Jones, 2018). Consumers weigh several elements, from food quality to menu diversity and price points (Baiomy et al., 2019; Ismail et al., 2016; Keshavarz et al., 2016; Park et al., 2021; Peters & Remaud, 2020; Zhong & Moon, 2020).

Apart from exceptional food and service, restaurant ambience, décor, and location significantly influence dining choices (Chang et al., 2011; Dhurup et al., 2013; Duncan et al., 2015; Jemaiyo Imamai et al., 2019; Matusitz & Breen, 2009; Pezenka & Weismayer, 2020). The rising trend in health-conscious eating, spurred by events like the COVID-19 pandemic, demands that restaurants cater to varied dietary requirements, especially among patrons with non-communicable diseases (NCDs) like diabetes (O’Meara et al., 2022; WHO, 2022). The Nielsen Group (2016) notes a significant gap in meeting these needs, with only 45% of those with special diets feeling adequately catered for in establishments.

The World Business Council for Sustainable Development’s (WBCSD, 2018) report suggests that future consumption patterns will pivot towards health and wellness-driven and desire-influenced decisions. Key health and wellness-driven choices include opting for less processed foods, reduced sugar consumption, ethically sourced products, flexitarian and plant-based diets, portion control, using food as a natural remedy, and catering to specific dietary needs such as gluten-free diets. Consequently, the hospitality and restaurant sectors should seize the opportunities presented by these evolving consumption trends. McKay and Subramoney (2017) highlight that the rising fast-food intake in South Africa correlates with increasing obesity and conditions like Type 2 diabetes. They recommend probing into perceptions of healthier food options and cooking methods, especially among general diners.

PLWD often grapple with hidden sugars in various sweet and savoury dishes. Product labels typically reveal sugar content, especially if ingredients list sugar, syrups, or compounds ending in “ose” such as sucrose or glucose (John Hopkins Medicine, 2022). Additionally, nutritional labels provide comprehensive dietary data, such as caloric content, fat, cholesterol, and carbohydrates. Specific health concerns like high blood pressure might also dictate dietary choices. For instance, those with hypertension typically opt for low-sodium dishes (Juraschek et al., 2021), while those with cardiac issues avoid high fat and cholesterol (Zinöcker et al., 2021).

However, carbohydrates, which metabolise into glucose post-consumption, pose significant challenges, primarily because of their ubiquity in foods (Volek et al., 2021). Common food groupings for carbohydrates include starchy items (like bread and rice) and sugar-containing foods, including fruits and sweets (Diabetes UK, 2022). The lack of standardised restaurant menu labelling worldwide often renders dining out a guessing exercise. To regulate insulin doses, the American Diabetes Association (ADA, 2021) suggests administering one insulin unit for every 10-15 g of carbohydrates, varying based on individual insulin sensitivity. Food’s glycaemic index (GI) is also vital for carbohydrate count (Jenkins et al., 1981). The GI gauges a food’s effect on blood sugar levels, ranking them between 0-100, with glucose equated to 100. Lower GI foods gradually increase blood sugar, whereas high GI foods, often processed and lacking in fibre, can lead to hyperglycaemia (Wolfram, 2016). Harvard Medical School (2021) defines foods with a GI of 55 or below as low, those between 56 and 69 as moderate, and those 70 or above as high.

A study by Rincón-Gallardo Patiño et al. (2020) mapped menu labelling policies from 2000 to 2020, revealing mandatory policies in select countries and a complete absence in certain regions, like Africa. The UK instituted mandatory calorie labelling in April 2022 (UK Government, 2022). The variance in global policies often emerges from national health department regulations, with implementations influencing the tourism and hospitality industries. Frequent dining out correlates with elevated risks of obesity, diabetes, and other chronic diseases (Du et al., 2021). Therefore, access to nutritious dining options translates to health equity, underpinning modern democratic principles (Weiler et al., 2015; UN, 2021). The concept of well-being as a primary part of social justice needs to be reinvestigated, and the functional role-players (like the hospitality sector) should be readdressed (Hood, 2018). Much of the existing literature approaches this from a health science perspective, sidelining the social sciences perspective. For instance, Table 2 illustrates dining out’s association with obesity and Type 2 diabetes risks (Dhinagaran et al., 2021), noting that sedentary lifestyles also contribute. While the hospitality sector cannot dictate dining frequencies, offering healthier choices can provide a competitive advantage. Potential strategies include nutritional education, endorsing plant-based diets, and cutting back on animal proteins and fried foods (Han et al., 2017; Parackal, 2017).

Table 2. Previous research on dining out and diabetes.

Author(s)	Domain	Diabetes Population	Results
Sarkar et al. (2018)	Health Sciences	Type 2 (UK)	The density (proximity/location) of ready-to-eat food environments within a 1 km catchment area was associated with higher odds of Type 2 diabetes for participants with the highest exposure to restaurants and cafeterias.
Choi et al. (2019)	Health Sciences	Type 1 and 2 (South Korea)	The higher the frequency of dining out, the higher the distribution of participants who reported undertaking light physical activity, being current smokers, having a graduate school education level, and having a higher monthly income. When selecting a dining-out menu, taste was reported to be the most relevant factor among all the participants, and health considerations decreased with the frequency of dining out.
Lee (2017) translated from Korean	Health Sciences	Labelling (South Korea)	From the point of view of diabetic patients, who have a limited range of meal options compared to the general public, it is inevitable to be cautious about eating out. However, living on only home-cooked meals in modern society is almost impossible, so learning and practicing healthy eating is essential. Therefore, the nutrition labelling system is suggested and applied to processed food, catering, and restaurants.
Parackal (2017)	Health Sciences	Type 2 (South Asian Diaspora)	South Asians have a four to fivefold higher risk for Type 2 diabetes mellitus (T2DM) than other Asian migrant groups. A significant shift in meal patterns with frequent dining out and eating fast foods, traditional festival foods and Western desserts and snacks were typical among South Asians. Consumption of potatoes, dairy, oil, meat and fish increased, and beans, lentils, fruits and vegetables decreased post-migration. “Animal protein” and “fried snacks, sweets and high-fat dairy” were associated with greater insulin resistance and lower HDL cholesterol. A “mixed” dietary pattern was associated with obesity and hypertension, and a “western” dietary pattern was associated with an overall risk for Metabolic Syndrome. A 70% increase in the odds of diabetes per standard deviation in a gram of protein intake was also observed.
Kelly et al. (2020)	Health Sciences	Diabetic status (USA)	Diabetes status associated with meal preparation, food procurement, and spending behaviours differs for males and females. All adults, but men with diabetes and pre-diabetes, spent less money dining out than those without diabetes. These findings support the importance of addressing the behavioural context of food choices to help support a healthier diet and lifestyle for all adults.
Han et al. (2017)	Health Sciences	Gestational diabetes (South Korea)	The major obstructive factors in dietary management were “eating more than planned when dining out” and “finding the appropriate menu when dining out”. In developing the management program, the preferred nutrition information contents were “nutritional information of food”, ““recommended food by major nutrients”, “the relationship between blood glucose and food”, and “tips” on menu selection when eating out”.
Joseph et al. (2017)	Health Sciences	Type 1 (India)	The study highlights the need for intense and continuous nutritional education to increase awareness about the individual nutrients in the diet of Type 1 diabetes mellitus patients. There is a need to enhance the availability of cost-effective, easy-to-prepare, and nutrient-rich foods, especially when dining out. The importance of meal-dependent insulin adjustments should be reinforced at every hospital visit.
Dhinagaran et al. (2021)	Social Sciences	Type 2 (Singapore)	Barriers to healthy living include frequent dining out, high stress levels, lack of work-life balance, and lack of free time to engage in physical activity. In contrast, discipline, preplanning, and routine enabled a healthy lifestyle.
Conway et al. (2018)	Health Sciences	Unreported diabetes (USA)	The incidence of diabetes was observed to be twice as high in individuals of African descent compared to their Caucasian counterparts with a Body Mass Index (BMI) within the normal range. This disparity in diabetes incidence diminished with an increase in BMI, culminating in the estimated 5-year probability of developing diabetes nearing 20% for both groups at higher BMI levels. Additionally, diabetes risk was found to correlate with lower socio-economic status, a correlation that was notably more pronounced among Caucasians. Factors such as current tobacco use, as well as reduced adherence to healthy eating habits and physical activity levels, were also associated with an increased risk of diabetes. Despite these variables, a high BMI emerged as the most significant risk factor for diabetes in both African descent and Caucasian populations.

Source: Compiled from the literature.

There is limited research on the profile and needs of diners with special dietary requirements. Most South African dining preferences research (Table 3) focuses on the general population and shows they are motivated by socialisation, relaxation, escape, and convenience (Kleinhans et al., 2019; Kruger & Saayman, 2016). The most critical attributes for restaurant choice are food quality, value/price, service quality, and ambience. However, there is still a need for more conclusive research on menu variety, especially for healthier options (Han et al., 2017; Joseph et al., 2017).

Table 3. Previous research on South African dining preferences.

Author(s)	Study results
Cant et al. (2014)	Consumers are satisfied with the availability of healthier foods in fast-food establishments, but the lack of variety and nutrition remains problematic.
Kruger et al. (2015)	Dining factors were investigated in a national park context, revealing that quality, variety, value, and natural ambiance enhance dining experiences.
Kruger and Saayman (2016)	Motives for dining out among black South Africans revealed status, socialisation, lifestyle and escape, and gastronomy enjoyment. Tipping decision-making factors reveal that mood and ambiance, waiter competence and service, ethnicity, and restaurant setting impact tipping.
Kleinhans et al. (2019)	Dining motivation includes socialisation, relaxation, and convenience. Dining experience dimensions include service, social, food, layout, restaurant health, food health, atmosphere and accessibility.
Tlapana and Sandlana (2021)	Food quality, price, hygiene, and service quality influence consumers’ restaurant choices.

Source: Compiled from the literature.

Examining the experiences of diabetic diners in South Africa, this study sheds light on the challenges faced when dining out by focusing on dining motivations, dietary needs and preferences, and restaurant attributes. The lack of research on the preferences and motivations of diabetic diners, particularly from a tourism and hospitality perspective, highlights a gap in the literature that this study aims to address. Identifying this population’s demands will enable the hospitality industry to offer diabetes-friendly and healthier choices for all diners.

Methodology

This exploratory research followed a quantitative research approach using an online questionnaire.

Development of the questionnaire

Developing the survey instrument represents a pioneering approach within the context of dining research among individuals living with diabetes in South Africa. This instrument was meticulously adapted from existing research focusing on the general motivations, preferences, behaviours, and attributes associated with dining in South Africa (Kleinhans et al., 2019; Kruger & Saayman, 2016; Saayman, 2014). This adaptation process was not merely an effort to repurpose existing tools but a significant endeavour to integrate and highlight elements critically relevant to understanding the unique dining motivations and behaviours of individuals managing diabetes.

Given the exploratory nature of the research, it was imperative to ensure that the survey instrument could capture the nuanced dining experiences of the target demographic. To this end, many sections of the questionnaire were developed through a collaborative effort involving the authors and the co-founder of the Sweet Life Diabetes Community. This partnership was instrumental in ensuring that the survey addressed general dining experiences and delved into aspects specific to individuals living with diabetes, such as dietary needs and preferences, and the impact of these on dining out behaviours and restaurant choice.

The survey consisted of five carefully designed sections. Section A collected essential socio-demographic details, enabling us to understand the diverse backgrounds of the respondents, including their specific type of diabetes. Sections B and C explored dining habits and motivations, respectively, offering insights into how frequently individuals dine out, their preferred dining settings, and the underlying reasons for these preferences. Section D focused on dietary preferences, a critical area for diabetics, capturing information on specific dietary needs, food allergies, and intolerances. Lastly, Section E assessed the criteria for restaurant selection and satisfaction with dining experiences, including perspectives on nutritional labelling. These sections were carefully crafted to ensure a comprehensive understanding of the dining landscape experienced by South Africans living with diabetes.

The collaborative effort in developing the questionnaire was pivotal in addressing the gap in the literature, aiming to shed light on the dining preferences and needs from a perspective that had not been extensively explored before. This approach underscores the novelty of the research, highlighting the importance of co-creating research tools with community stakeholders who possess intimate knowledge of the population of interest. Such collaboration ensures that the instrument is not only relevant and sensitive to the specific needs of the community but also enriches the research findings with nuanced insights that might otherwise be overlooked.

By emphasising the exploratory nature of this study and the collaborative effort in developing the survey sections, we acknowledge the innovative approach taken to understand the dining preferences and needs of South African diners living with diabetes. This methodological choice reflects a significant step forward in tailoring research instruments to meet the specific needs of underrepresented populations in academic inquiry, particularly in the context of dietary management and lifestyle choices among individuals living with chronic conditions.

Data collection method

In addressing the representativeness of our sample for the diabetic population, we must acknowledge certain limitations inherent to our methodology. The survey, designed digitally via QuestionPro (Research Analytics, 2022), employed a combination of purposive and snowball sampling techniques. Initially, we distributed the survey through an existing database, adopting a purposive sampling strategy. This was supplemented by a snowball sampling approach, leveraging social media campaigns to reach potential participants who are diners living with diabetes. Our collaboration with the Sweet Life Diabetes Community, a critical access point, enabled the dissemination of our survey through their significant online presence, including 34,000 social media followers and 12,000 email subscribers.

Despite these efforts, the recruitment of participants yielded 350 responses, slightly below the ideal sample size of 381, as suggested by Krejcie and Morgan’s (1970) formula for our target audience size of approximately 46,000. Only fully completed surveys were considered for analysis, resulting in 239 valid responses. When considering Cochran’s (1963) guidelines, which recommend a sample size of 204 for a population of 50,000 with a ±7% precision level, our valid response count meets this criterion, suggesting an adequate sample for analysis within the defined parameters.

However, it is critical to discuss the limitations related to the general population of individuals living with diabetes. The sampling methods, particularly the reliance on social media campaigns and an existing database for participant recruitment, may not fully capture the diversity within the broader diabetic population. This approach potentially introduces a selection bias, favouring individuals more engaged with diabetes-focused online communities or those with better access to digital resources.

Furthermore, using social media campaigns as a recruitment tool may bias outcomes. Such campaigns will likely attract participants who are already somewhat proactive about managing their diabetes through digital engagement, possibly differing significantly in behaviours, attitudes, or awareness from those in the wider diabetic community who might not be as active online.

Ethical considerations

Sweet Life granted permission and distributed the survey link across their database and social media channels. The study aimed to engage adults aged 18 and above who were in a position to give informed consent. A snowball sampling approach was adopted for recruitment, encouraging individuals to participate willingly and share the survey link, ensuring no obligations or repercussions for choosing not to participate. Alongside the survey was a letter detailing the research objectives, the method for giving informed consent, estimated completion time, copyright details, and the ethical clearance identifier (NWU-00594-22-A4) the faculty’s research ethics committee granted. Guarantees were made regarding the protection and anonymisation of collected data; findings would be published in aggregate, omitting individual data points. The study was deemed to carry minimal risk to participants. While the questionnaire did ask for personal data points such as yearly earnings and ethnic background, in compliance with South Africa’s Protection of Personal Information Act (POPIA), such inquiries were not obligatory. If they chose, participants could discontinue the survey at any stage. To alleviate online privacy concerns, the Respondent Anonymity Assurance (RAA) feature was enabled in QuestionPro, ensuring e-mail and IP addresses were stripped from the dataset. Once collated, the data was transferred to Microsoft Excel and securely housed on an encrypted client-server (Nextcloud), with access restricted solely to the authors.

Data analysis

We analysed the exported data (Microsoft Excel) using Statistical Package for Social Sciences (SPSS) version 28.0 (IBM, 2023). The analysis had three phases:

1. We analysed the profile of the respondents using descriptive analysis.

2. We conducted three exploratory factor analyses (EFAs) on dining motivation, restaurant choice attributes, and dining experience satisfaction.

3. We performed a two-step cluster analysis to segment the respondents based on the categorical and continuous variables measured in the questionnaire.

Results

Profile of respondents

Table 4 shows the profile of the respondents. Eighty-two percent (82%) of the respondents were people living with diabetes (PLWD), while 18% were carers for a PLWD. The majority of the PLWD was diagnosed with Type 1 diabetes (58%), followed by Type 2 (38%) and pre-diabetes (8%). Most respondents were female (84%), while only 16% were male. The average age of the respondents was 46 years, with the largest group being between 35 and 44 years old (25%). Carers likely make up the 3% of respondents under 18 and are likely parents or guardians of children, most likely mothers based on gender representation. Regarding ethnicity, most respondents identified as White (Caucasian) (68%), followed by Coloured (14%), Black African (10%), and Indian/Asian (8%). We included ethnicity in the analysis since some ethnic groups are more susceptible to developing diabetes due to biological, cultural, economic, and political factors (Bancks et al., 2021; Goff, 2019; Sifunda et al., 2023; Spanakis & Golden, 2013). English (57%) was respondents’ most common home language, followed by Afrikaans (35%). The Other category (8%) included the remaining official languages in South Africa, including isiZulu and isiXhosa at 2% each. The majority of respondents came from the Gauteng Province (36%), followed by the Western Cape (33%) and KwaZulu-Natal (9%). In terms of household income, the largest group (33%) earned up to R216,200.00 ($12,717) annually, followed by 21% who earned up to R337,800.00 ($19,870). Thirteen percent (13%) of respondents earned up to R1,656,600.00 ($97,447) annually.

Table 4. Profile of the respondents.

Variable	Result	Variable	Result
RESPONDENT TYPE		GENDER
People living with diabetes (PLWD)	82%	Male	16%
Carer for PLWD	18%	Female	84%
DIABETES TYPE		ETHNICITY ^b
Type 1	58%	Black African	10%
Type 2	38%	Coloured	14%
Pre-diabetes	8%	Indian/Asian	8%
		White (Caucasian)	68%
ANNUAL INCOME ^a		HOME LANGUAGE
< $12,717	33%	English	57%
$12,718 to $19,870	21%	Afrikaans	35%
$19,871 to $27,500	11%	IsiZulu	2%
$27,501 to $36,094	10%	IsiXhosa	2%
$36,095 to $46,011	9%	Other (Ndebele, Pedi, Sotho, Swati, Tsonga, Tswana, Venda)	4%
$46,012 to $97,447	13%
$97,448>	3%
AGE CATEGORIES		PROVINCE OF ORIGIN
Under 18	3%	Gauteng	36%
18–25	7%	Western Cape	33%
26–34	17%	KwaZulu-Natal	9%
35–44	25%	Eastern Cape	5%
45–54	17%	Mpumalanga	5%
55–64	17%	North West	5%
Above 65	14%	Free State	3%
Average age	46 years	Limpopo	2%
		Northern Cape	1%
		Namibia ^c	1%

^a Exchange rate (1USD = 17ZAR); ^bPer SA Government Classification; ^cOutside RSA border.

Exploratory factor analysis (EFA) results

Exploratory factor analyses (EFAs) were conducted on variables related to dining motivations, preferences for restaurant attributes, and satisfaction levels with the dining experience. While Confirmatory Factor Analysis (CFA) is valuable for testing hypothesised models in studies where constructs are well-defined and understood, EFA is more suitable for this study because it allows for identifying underlying factor structures without imposing a predetermined model. This approach is particularly beneficial in exploratory research or when dealing with underexplored contexts, as in the case of the present research. To ensure the sample was adequate and that the data could be factorised, the Kaiser-Meyer-Olkin (KMO) measure was utilised, with values exceeding 0.7 indicating adequacy (Kaiser, 1960). Bartlett’s sphericity test was also applied and considered significant at p < 0.05 (Dziuban & Shirkey, 1974). Dimension reduction employed the principal component extraction approach, setting criteria at an Eigenvalue (EV) greater than 1. To enhance clarity in interpreting the resulting factor structures, an Oblimin rotation coupled with Kaiser normalisation was employed. Factor reliability was established through Cronbach’s alpha (α), with values surpassing 0.6, denoting reliability (Cronbach, 1951). These values aligned with recommended averages for inter-item correlations ranging from 0.15 to 0.55 (Cohen, 1988). Lastly, only items exhibiting a factor loading coefficient above 0.40 were retained. The constructs were created based on the EFA results.

As shown in Table 5, five motivational factors were extracted. Diabetic diners mainly dine out for celebration and socialisation (x̄ = 3.98), followed by tourism and novelty-related reasons (x̄ = 3.29). Vitality and appreciation (x̄ = 3.21) and gastronomy and enjoyment (x̄ = 3.15) were also important reasons for dining out, while convenience and practicality (x̄ = 3.07) were less important compared to the other motives.

Table 5. EFA on Motives of diabetic diners.

Motivational statements	Factors (N = 286)
	Vitality and appreciation	Convenience and practicality	Gastronomy and enjoyment	Tourism and novelty	Celebration and socialisation
For my health and well-being	0.81
To experience and appreciate local South African cuisine/produce/techniques (e.g. Cape Malay Bobotie / Karoo lamb)	0.72
To try out available specials/promotions	0.54
To support local business owners	0.46
Because it is more convenient (not having to cook)		0.82
Because it is loadshedding (scheduled power cuts)		0.72
Because it is more affordable than cooking for myself		0.54
To relax after a long day		0.44
Dining out is a reward/treat for “good behaviour”		0.42
I consider myself to be a “foodie”			0.72
Because I enjoy dining out			0.71
It is a habit of mine			0.70
Because I am travelling /on holiday				0.64
For business-related reasons, i.e. lunch meetings				0.57
To try out new food				0.56
To eat dishes that I cannot easily make at home (e.g. sushi)				0.54
To celebrate special occasions					0.82
To socialise with friends and family					0.81
EV	4.47	1.75	1.54	1.24	1.08
Var. (%)	24.83	9.73	8.56	6.88	5.97
Cum. Var. (%)	24.83	34.56	43.12	50.00	55.97
α	0.69	0.67	0.66	0.56^a	0.65
Avg. inter-item correlation	0.35	0.29	0.40	0.25	0.48
Mean value (x̄)	3.21	3.07	3.15	3.29	3.98

* 5-point Likert scale: 1 = completely disagree to 5 = completely agree; [KMO = 0.80, Bartlett’s test of sphericity: x² (153) = 1153.01, p < 0.05].^a The α-value is close to the threshold of 0.6, but 0.5 can be considered moderately reliable (Hinton et al., 2014).

Two restaurant choice attributes were extracted (Table 6), with value and relaxation (x̄ = 3.60) regarded as the most crucial attribute when diabetic diners select a restaurant, followed by E-WOM and quality (x̄ = 3.48).

Table 6. EFA of restaurant choice attributes.

Restaurant choice attributes	E-WOM and quality	Value and relaxation
Positive online reviews	0.80
Online reservation ability	0.74
High-quality ingredients	0.63
Excellent service quality	0.56
Affordable prices and value-for-money		0.84
Convenient location		0.73
Relaxing atmosphere		0.58
EV	2.79	1.23
Var. (%)	39.90	17.61
Cum. Var. (%)	39.90	57.51
α	0.67	0.64
Avg. inter-item correlation	0.36	0.37
Mean value (x̄)	3.48	3.60

* 5-point Likert scale: 1 = not at all important to 5 = very important; [KMO = 0.72, Bartlett’s test of sphericity: x² (21) = 437.10, p < 0.05].

Regarding diabetic diners’ satisfaction when dining out, the low rating of the factor, dining experience (x̄ = 2.46) (Table 7), reveals that diners are dissatisfied to only moderately satisfied with their overall dining experiences, indicating an urgent need for restaurants to better cater to the needs of diners living with diabetes.

Table 7. EFA of dining experience satisfaction.

Dining experience satisfaction	Dining experience
Availability of nutritional information on request	0.88
The staff’s knowledge regarding diabetes meal requests	0.87
Availability of nutritional information on the menu	0.85
Variety of diabetes-friendly meal options	0.84
Availability of food safety and allergy information	0.83
Variety and availability of tasty, healthier options	0.82
The adaptability of meal requests (substitutions)	0.81
The adaptability of meal price (exclusions are removed from the bill)	0.78
The understanding of fellow diners	0.62
EV	5.98
Var. (%)	66.39
Cum. Var. (%)	66.39
Α	0.94
Avg. inter-item correlation	0.62
Mean value (x̄)	2.46

* 5-point Likert scale: 1 = very dissatisfied to 5 = very satisfied; [KMO = 0.92, Bartlett’s test of sphericity: x² (36) = 1840.35, p < 0.05].

Two-step cluster analysis results

A two-step cluster analysis was conducted on a dataset of 239 respondents, yielding a silhouette measure of cohesion and separation valued at 0.2. This surpasses the baseline threshold of 0.0, suggesting acceptable within-cluster and between-cluster distinctions (Norusˇis, 2009). Nonetheless, the size ratio was 1.13, falling short of the advised benchmark of 3. This analysis discerned two distinct clusters, identified by a reduced BIC value, minimal BIC change, and a specific distance measure. An assessment involving cross-tabulations, chi-square tests, and independent sample t-tests verified that the 19 integrated variables differed across these clusters. Consistent clusters emerged in the preliminary and validation results when the dataset was bifurcated for verification purposes. The relative significance of the variables in predicting cluster membership remained congruent across the final and validation analyses (as detailed in Table 8). Each variable’s role in the two-step cluster analysis is noted alongside its respective predictive importance score (Tkaczynski et al., 2010). Variables scoring between 0.8 and 1.0 play a pivotal role in the cluster formation. Those ranging from 0.2 to 0.7 have a moderate to high predictive relevance, while scores from 0.0 to 0.2, although significant, exert a lesser influence on the formation of clusters.

Table 8. Final cluster solution.

Identified segments		Segment 1: Affluent, ageing, frequent, Caucasian, and discontent Type 1’s		Segment 2: Middle-class, mature, occasional, multi-racial, and content Type 2’s
Segment size (N)		N = 127		N = 112
Segment (%)		53%		47%
Categorical variables		N	%	N	%
Type of diabetes (1)	Type 1	116	91%	28	25%
	Type 2	11	9%	84	75%
Gender (0.02)	Male	15	12%	20	18%
	Female	112	88%	92	82%
Home language (0.14)	Afrikaans	63	50%	24	21%
	English	63	50%	73	65%
	isiXhosa	0	0%	2	2%
	isiZulu	0	0%	2	2%
	Sesotho Leboa	0	0%	2	2%
	Sesotho	1	1%	1	1%
	Setswana	0	0%	3	3%
	Tshivenda	0	0%	2	2%
	Xitsonga	0	0%	1	1%
	Other	0	0%	2	2%
Ethnicity (0.29)	Black African	2	2%	17	15%
	Coloured	7	6%	24	21%
	Indian/Asian	6	5%	10	9%
	White	112	88%	61	54%
Province of residence (0.07)	Eastern Cape	3	2%	9	8%
	Free State	4	3%	4	4%
	Gauteng	53	42%	37	33%
	KwaZulu-Natal	3	2%	13	12%
	Limpopo	3	2%	3	3%
	Mpumalanga	4	3%	8	7%
	Northern Cape	2	2%	0	0%
	North-West	6	5%	3	3%
	Western Cape	49	39%	34	30%
	Outside RSA borders	0	0%	1	1%
Annual household income (0.45)	<$12,717	16	13%	59	53%
	$12,718 to $19,870	31	24%	22	20%
	$19,871 to $27,500	11	9%	16	14%
	$27,501 to $36,094	15	12%	6	5%
	$36,095 to $46,011	14	11%	5	4%
	$46,012 to $97,447	31	24%	4	4%
ZAR17 = 1US$	$97,448>	9	7%	0	0%
Dining frequency (0.33)	Never	2	2%	7	6%
	Once a week	28	22%	20	18%
	2–3 times a week	24	19%	10	9%
	More than 3 times a week	3	2%	2	2%
	Once a month	16	13%	55	49%
	2–3 times a month	48	38%	14	13%
	More than 3 times a month	6	5%	4	4%
Dietary intolerances or allergies (0.01)	Yes	107	84%	90	80%
	No	20	16%	22	20%
Other diabetic-related complications that impact dining preferences (0.29)	Yes	106	83%	57	51%
	No	21	17%	55	49%
To better manage blood glucose, support the introduction of food labelling and nutritional information in restaurants (0.02)	Yes	125	98%	108	96%
	No	2	2%	4	4%
Identified segments		Segment 1: Affluent, ageing, frequent, Caucasian, and discontent Type 1’s		Segment 2: Middle-class, mature, occasional, multi-racial, and content Type 2’s
Segment size (N)		N = 127		N = 112
Segment (%)		53%		47%
Continuous variables		M	SD	M	SD
Age (0.43)	Avg. age	40.50	11.24	52.76	15.25
Motives for dining out	Vitality and appreciation (0.18)	2.99	0.64	3.40	0.86
	Convenience and practicality (0.01)	3.08	0.75	3.03	0.72
	Gastronomy and enjoyment (0.01)	3.11	0.81	3.15	0.87
	Tourism and novelty (0.06)	3.39	0.66	3.39	0.66
	Celebration and socialisation (0.00)	3.39	0.66	3.98	0.68
Restaurant choice attributes	E-WOM and quality (0.06)	3.33	0.71	3.54	0.81
	Value and relaxation (0.04)	3.47	0.71	3.63	0.78
Satisfaction when dining out	Dining experience (0.29)	2.17	0.71	2.79	1.00

As delineated in Table 8, the specific type of diabetes diagnosed in the respondent (1) emerged as the paramount predictive variable, mainly shaping the distinctions between the clusters. Subsequent variables of significance include annual household income, age, frequency of dining out, ethnicity, the presence of other diabetes-related complications influencing dining choices, and overall satisfaction with the dining experience. The remaining variables exerted a diminished influence on cluster delineation. The results further delineate distinct typologies of South African diners living with diabetes.

Segment 1 is the largest segment, with 127 respondents. This segment is represented by 91% of people with Type 1 diabetes. Compared to Segment 2, this segment is younger (average age of 41) with a higher annual household income [$12,718 to $19,870 and $46,012 to $97,447 (24% respectively)]. This segment also tends to dine out more frequently [2–3 times a month (38%), once a week (22%) and 2–3 times a week (19%)]. Eighty-eight percent (88%) of respondents in this segment are White/Caucasian. Most respondents (83%) also have other diabetic-related complications that impact their dining preferences. Regarding their overall dining experience satisfaction, this segment rated the factor significantly lower (x̄ = 2.17), indicating they are most dissatisfied. Based on these characteristics, we labelled this segment the Affluent, ageing, frequent, Caucasian, and discontent Type 1s.

Segment 2 is the smaller segment with 112 respondents. This segment is represented by 75% of respondents with Type 2 diabetes. This segment is older (average age of 53), has a lower annual household income [< $12,717 (53%)], and has a good representation of the different ethnic groups in South Africa [White/Caucasian (54%), Coloured (21%) and Black African (15%)]. This segment does not dine out as frequently as Segment 1 [once a month (49%)]. However, 18% indicated they dine out once a week. A lower percentage (51%) indicated they have other diabetic-related complications impacting dining preferences. They are further moderately satisfied with their overall dining experience when dining out (x̄ = 2.79). This segment was therefore labelled the Middle-class, mature, occasional, multi-racial and content Type 2’s.

Looking at additional similarities and differences between the segments, both had more female (88% and 82%) and English-speaking (50% and 65%) respondents. Segment 1 had more Afrikaans-speaking respondents (50%), while Segment 2 had a greater representation of the other official languages in the country. Gauteng (42% and 33%) and the Western Cape (39% and 30%) had the highest representation in both segments. Both segments support introducing food labelling and nutritional information in restaurants (98% and 96%). Regarding the motives for dining out, Segment 2 mainly dines out for celebration and socialisation (x̄ = 3.98), followed by vitality and appreciation (x̄ = 3.40). Interestingly, Segment 1 dines out for tourism and novelty and celebration and socialisation (x̄ = 3.39 respectively) reasons. Regarding the restaurant attributes, both segments seek value and relaxation (x̄ = 3.47 and x̄ = 3.63, respectively), while E-WOM and quality (x̄ = 3.54) are more important to Segment 2.

Discussion

Theoretical findings

This study aimed to segment South African diners living with diabetes into distinct groups based on their socio-demographic profiles, dining motivations, dietary needs and preferences, and preferred restaurant attributes. The two-step cluster analysis yielded two separate segments of diabetic diners, namely the Affluent, ageing, frequent, Caucasian, and discontent Type 1’s and the Middle-class, mature, occasional, multi-racial, and content Type 2’s. This typology represents a description of the dining behaviour of PLWD in the South African dining context. The type of diabetes the respondent had (Type 1 or Type 2) was found to be the most significant factor in differentiating between the two segments, followed by socio-demographic and dining behaviour characteristics, annual household income, age, dining frequency, ethnicity, and living with other diabetic-related complications that impact dining preferences and dining experience satisfaction. The findings contribute to the existing theoretical models that guided the research by illustrating the relationships between the variables that distinguished the two segments and by proposing an integrative model illustrated in Figure 1 for understanding diabetic dining behaviour, incorporating the multi-segmentation, the constructs of the Health Belief Model (HBM) and the Integrated Behavioural Model (IBM).

Figure 1. Proposed model of diabetic dining behaviour with behavioural determinants and contextual considerations.

The relationship between socio-demographic characteristics and dining behaviour is illustrated in Figure 1 and is a crucial aspect of understanding the latter. The study found that age, annual household income, and ethnicity were crucial determinants differentiating the two segments of people living with diabetes (PLWD). The first segment, referred to as Affluent, ageing, frequent, Caucasian, and discontent Type 1’s, was found to be younger, have a higher income, and primarily of Caucasian ethnicity. The second segment, Middle-class, mature, occasional, multi-racial, and content Type 2’s, was found to be older, have a lower income, and be more ethnically diverse. It should be noted that the study did not inquire about the time of diagnosis of diabetes and thus cannot verify the relationship between age and diabetes. However, it is well established that Type 2 diabetes is more common in older individuals and among ethnic groups such as African Americans, Hispanics/Latinos, American Indians, Asian Americans, and Pacific Islanders (NIDDKD, 2022).

Additionally, lower socioeconomic status has increased diabetes risk among Caucasians (Conway et al., 2018). The study’s results suggest that socio-demographic characteristics, such as dining frequency, play a role in dining behaviour and may be linked to individual perceptions.

As depicted in Figure 1, the type of diabetes can be considered a component of individual perceptions in the form of perceived risk, which is one of the key constructs of the Health Belief Model (HBM) (Becker, 1974; Champion & Skinner, 2008; Hochbaum, 1958; Kirscht, 1974; Rosenstock, 1960, 1974). The results indicated that most of the Affluent, ageing, frequent, Caucasian, and discontent Type 1’s (83%) and the Middle-class, mature, occasional, multi-racial and content Type 2’s (51%) also have other diabetic-related health complications. This could significantly impact these segments’ dining (restaurant and meal) choices (Juraschek et al., 2021; Zinöcker et al., 2021). Other health conditions, such as high cholesterol, can influence dining preferences. The results showed that having dietary restrictions and allergies was not a distinguishing factor between the two segments, as most of both segments reported suffering from these conditions. However, it is still an important variable to consider when understanding dining behaviour, particularly regarding health-related risks.

The Integrated Behavioural Model (IBM) is a continually refined theoretical framework that builds upon the Theory of Planned Behaviour (TPB) by considering the external and non-cognitive factors that influence the expression of health-related behaviours, such as overconsumption and eating unhealthy food. The IBM Model identifies three factors that support the Health Belief Model (HBM) and IBM as modifying factors (Aalto & Uutela, 1997) and environmental constraints (Fishbein & Ajzen, 2011). According to TPB, intentions play a crucial role in predicting behaviour; thus, habits are considered the first variable in Figure 1. The results indicate that dining frequency is a crucial factor distinguishing between the two segments of people living with diabetes. The Affluent, ageing, frequent, Caucasian, and discontent Type 1 segment dine out more often than the Middle-class, mature, occasional, multi-racial, and content Type 2 segment.

The attributes of restaurant choice (E-WOM and quality and value and relaxation) and motives for dining out (vitality and appreciation, convenience and practicality, gastronomy and enjoyment, tourism and novelty, and celebration and socialisation) had less impact on differentiating the two segments. Other studies have also recognised comfort, value, quality, and an enjoyable environment as significant restaurant attributes that diners look for (Alonso & O’Neill, 2010; Goodman-Smith et al., 2020; Kruger et al., 2015; Tlapana & Sandlana, 2021; Yi et al., 2018). Promotion and recommendation are essential factors for restaurant choice (Park et al., 2021), with E-WOM being a unique attribute identified in this research.

Regarding motives for dining out, socialisation and relaxation were noted by Kleinhans et al. (2019), Kruger and Saayman (2016), Choi and Zhao (2014), Azim et al. (2014), and Kanyan et al. (2016); convenience by Kleinhans et al. (2019); gastronomy enjoyment by Kruger and Saayman (2016); and healthy options in the form of vitality by Cant et al. (2014) and Kessler et al. (2020). The combination of motives in this research is unique, and identifying tourism and novelty as a dining-out motive is novel.

The findings concur with the observation made by Choi and Zhao (2014) that differences can be seen in the reasons and context of dining out. Both segments prioritise value and relaxation when choosing a restaurant. At the same time, the primary motivators for Middle-class, mature, occasional, multi-racial and content Type 2’s are celebration and socialisation, followed by vitality and appreciation. For Affluent, ageing, frequent, Caucasian, and discontent Type 1’s is tourism and novelty, followed by celebration and socialisation. The study by Choi et al. (2019) also revealed that health considerations become less important as the frequency of dining out increases among both Type 1 and Type 2 diabetics in South Korea. This trend seems to be mirrored by the lower motivation for vitality and appreciation among Affluent, ageing, frequent, Caucasian, and discontent Type 1’s.

The second construct, health value, is an important factor for individuals living with diabetes, as described in the Health Belief Model (Aalto & Uutela, 1997). Three aspects are crucial for this population: access to healthy menu options (Choi et al., 2019; Joseph et al., 2017), satisfaction with the dining experience (Cant et al., 2014; Kleinhans et al., 2019; Tlapana & Sandlana, 2021), and self-efficacy in managing glucose levels when dining out (Dhinagaran et al., 2021; Han et al., 2017; Kelly et al., 2020; Lee, 2017). The results showed that Affluent, ageing, frequent, Caucasian, and discontent Type 1 diabetics have a significantly lower overall dining experience satisfaction compared to Middle-class, mature, occasional, multi-racial and content Type 2 diabetics. This could be attributed to the fact that people with Type 1 diabetes, dependent on insulin, would generally have been diagnosed much earlier in life. This might indicate they have been dissatisfied with dining out for far longer. Additionally, self-efficacy may be hindered when choosing an unhealthy option from a restaurant menu, and the negative impact of missing out on a pleasurable dining experience with friends/family may also play a role (Gillibrand & Stevenson, 2006).

The third construct, social support, can enhance health value. Key aspects of this construct include understanding by dining companions, the restaurant’s ability to accommodate dietary needs and menu labelling. The first two aspects were evaluated as part of the dining experience factor, rated lower by Affluent, ageing, frequent, Caucasian, and discontent Type 1s. Although both segments rated the dining experience as only moderately satisfactory, this still indicates a lack of social support. However, menu labelling was not a significant predictor of segment differences. Both segments valued restaurant menu labelling and nutritional information for better glucose management. This finding aligns with previous research indicating the importance of menu labelling in restaurants (Han et al., 2017; Joseph et al., 2017; Lee, 2017), although it is currently unregulated in South Africa.

The market segmentation variables discussed in this research effectively distinguished the profile, needs, and preferences of people with diabetes in a dining context. These constructs explain the dining behaviour of PLWD, but they are interdependent and influence one another. The final construct, likelihood of action, part of the Health Belief Model (Aalto & Uutela, 1997), represents either the perceived benefits or costs of preventative actions taken by diners or actions taken by restaurants and the hospitality industry. The perceived benefits or costs of action are influenced by the constructs discussed previously, and the findings indicate the need for a comprehensive understanding of these constructs to address the needs of PLWD in a dining context effectively.

The dining behaviour of people with health-related complications is complex and can be influenced by restaurant options. Considering future trends prioritising health and desire-based choices (WBCSD, 2018), restaurants and the hospitality industry can gain a competitive advantage by incorporating healthier meal options. Figure 1 highlights the interdependence between constructs and dining behaviour, indicating the importance of understanding this relationship to cater to the needs of PLWD and other diners following health-related lifestyles. When faced with limited or unsatisfactory options, PLWD may risk their health, settle for the healthiest option available, hope the restaurant can accommodate their needs, or select a different one. Restaurants and the hospitality industry can respond by maintaining their current offerings or implementing healthier options and staff training to accommodate meal requests.

Practical implications aimed at the hospitality sector

The findings of this research have important implications for the hospitality sector and dining management. Service quality and customer satisfaction with service are traditionally considered key indicators of excellence in the hospitality sector and predictors of future behaviours such as customer loyalty. However, this study suggests that a particular segment of consumers, PLWD, are only moderately satisfied with current dining options, hindering long-term customer relationships. Given that PLWD’s overall health and well-being are closely tied to their food consumption and that they are frequent diners with characteristics similar to those of the general population, it raises concerns about social sustainability. As a result, the hospitality sector is responsible for ensuring equitable access to nutritious and healthy meal options for PLWD and all diners who prioritise their health and wellness.

Several approaches can be taken to address consumers’ changing needs and ensure inclusiveness in the hospitality sector. Firstly, the widespread implementation of menu labelling in restaurants that includes standardised portion size recommendations and comprehensive nutritional information. This will benefit people with specific health requirements and those interested in a healthy and balanced lifestyle. Additionally, reducing food waste must be a priority in the sector. Excessive food waste increases costs and contributes significantly to greenhouse gas emissions through energy and water use in food production and cultivation. To address this, the industry should set minimum standards for portion sizes, collaborating with professionals such as dieticians to promote a balanced and nutritious meal.

The second approach for ensuring inclusivity in the hospitality sector involves using advanced point of sale technology (POS). This technology should facilitate more precise and accurate ordering for consumers with dietary restrictions, such as PLWD. One common challenge PLWD faces when dining out is the need for substitutions in their meals, particularly in the case of carbohydrates, which often comprise large portions and are staples of many meals. Given that PLWD needs to reduce or avoid certain types of carbohydrates, offering healthy alternative options and accurately reflecting omitted items in the bill becomes a priority for the sector. For example, if a customer decides not to order fries, the POS system should reflect a decreased overall meal price.

Thirdly, the quality of the dining experience is crucial to diners’ satisfaction, particularly PLWD. The interaction between waitstaff and diners and the end product delivered by the kitchen plays a significant role in determining the dining experience. Hence, waitstaff must be knowledgeable, empathetic, and trained in catering to the needs of PLWD. This can be achieved through training that equips waitstaff with the skills to communicate requests accurately and prioritise them in the kitchen. Similarly, kitchen staff must be trained to understand and prioritise requests from PLWD, treating meal requests that cater to their dietary needs with caution as requests for food allergies. Training for managers, waitstaff, and kitchen staff is required, which can be conducted by professionals such as dieticians. If waitstaff or kitchen staff are not equipped to handle such requests, the restaurant manager and chef or kitchen manager should be held responsible for ensuring that PLWD’s meal requests are fulfilled to the best of their ability. It is also important to note that there is currently a lack of training and understanding of the potential negative consequences of not catering to PLWD’s needs, making it imperative for the sector to prioritise this aspect of service delivery.

Finally, health and wellness-orientated consumer segments are particularly interesting, as seen in the “shift” towards more health-related consumption. Therefore, it is essential to understand what these consumers require when dining out, regardless of location or occasion.

Conclusion

This study fills a void in existing literature by investigating the dining habits of individuals with diabetes in South Africa. While they share similar dining motivations with those not diagnosed with the condition, their available meal choices are constrained, often culminating in less-than-satisfactory dining experiences. The findings delineate two unique segments of diners with diabetes, each characterised by varied socio-demographics, dining tendencies, and preferences. Grounded in the Market Segmentation Theory, Health Belief Model, and Integrated Behavioural Model, the authors put forth a framework aimed at aiding scholars and industry professionals in deciphering the dining patterns of these individuals. The research underscores the pivotal role of educational efforts, adequate menu labelling, and comprehensive staff training within the restaurant and hospitality sectors to cater to individuals with diabetes and other patrons with specific dietary and health considerations.

Nonetheless, the study is not without its constraints; the gathered sample does not encapsulate the entirety of the diabetic population in South Africa, and the exploration remains confined to a demand-driven perspective. The sample was impacted by the use of gatekeepers to access potential respondents. The use of an online questionnaire, although very cost and time-effective, has limitations regarding access, especially in a developing country. Although a database and social media campaign were employed, the responses remained unrepresentative, and an alternative approach might include a qualitative approach with interviews. While the study provides valuable insights into the dining preferences and needs of South African diners living with diabetes, these findings must be contextualised within the limitations of the sampling strategy. The potential biases introduced using social media for participant recruitment highlight the need for caution when generalising these results to the entire diabetic population. Future research should consider more inclusive and varied recruitment strategies to ensure a broader representation of the diabetic community, thereby enhancing the generalisability of the findings.

Future endeavours should further probe the practicality of the introduced model and gauge the hospitality industry’s readiness for adaptive measures. Notably, this study did not evaluate the construct related to the probability of acting, signalling another gap for upcoming research endeavours.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability

The data will be made available on request.

Additional information

Funding

This research received no specific grant from funding agencies in the public, commercial, or not-for-profit sectors. This work is supported by the National Research Foundation (NRF). Any opinion, finding and conclusion or recommendation expressed in this material is that of the authors, and the NRF does not accept any liability in this regard.

Notes on contributors

Adam Viljoen

Adam Viljoen is an Associate Professor based at TREES (Tourism Research in Economics, Environs and Society), North-West University, Potchefstroom, South Africa. He obtained his PhD in 2018 and currently has a research focus on niche experiences within events, festivals and nature-based contexts. He is predominantly interested in expanding the culinary and food tourism understanding from an experience perspective. Adam is a Y1 rated researchers by the National Research Foundation (NRF).

Martinette Kruger

Martinette Kruger is based at TREES (Tourism Research in Economics, Environs and Society), North-West University, Potchefstroom, South Africa. She is a Professor in Tourism Management and devotes her career to better understanding the needs of the tourism market and how tourism can be facilitated in a developing country and multi-cultural society context. She is an established researcher with a C3 research rating from the National Research Foundation (NRF).