Conceptualisation of Food Addiction for development of a system to guide practice

Food addiction is a complex construct with mixed support in the academic community and debate surrounds its inclusion as a clinical diagnostic condition. Non-inclusion of food addiction as a clinical diagnostic category poses a risk as it limits access to clinical assistance for people in situations where a diagnosis, and subsequent support that can be accessed following diagnosis, is the only affordable means of treatment. With this in mind, further investigation of the evidence surrounding support of the construct is warranted. No diagnostic criteria or measurement instruments currently exist to measure food addiction as a behavioural concept. Without a proposed set of diagnostic criteria which have been shown to validly measure food addiction using a behavioural model, it is unlikely that it will be accepted into a diagnostic manual as a diagnostic clinical condition. Similarly, without an instrument to measure food addiction using a behavioural framework, the evidence required to support a behavioural model for food addiction cannot be established. To date, food addiction has been investigated using a primarily substance addiction model with the Yale Food Addiction Scale (YFAS) (Gearhardt, Corbin, and Brownell, 2009). An updated version of the YFAS has been developed based on the DSM-5 criteria for substance use disorder (Gearhardt et al., 2016).

While research has been conducted exploring food addiction using a primarily substance addiction model, no measurement tool has previously been examined based on a behavioural addiction model using proposed behavioural addiction criteria. This program of research explores the structure and utility of a behavioural scale to measure food addiction based on a behavioural model of food addiction and is presented in three stages.

In Stage 1, a behavioural scale for food addiction was developed based on Griffiths’ (1996) model of addiction, and Koob’s (2008) model of the addictive process. This understanding of how addiction occurs, and the criteria that can be applied to addiction, was used in the development of a scale to assess food addiction when conceptualised as a behavioural addiction. The behavioural food addiction scale was piloted and a Cronbach’s α of .93 was observed among a sample of 217 Australian adults aged 18 to 69 years (M = 35.89, SD = 12.78) recruited via Facebook, Twitter and email using a snowball sampling procedure. This stage resulted in the development of proposed diagnostic criteria for food addiction from a behavioural perspective.

In Stage 2, data measuring a range of factors associated with food addiction including, BMI, Binge Eating Disorder, impulsivity, emotional eating and other addictions was examined. Twenty Australian adults aged 20 to 55 years (M = 38.1, SD = 7.55) were recruited at university open days and residential schools. During this stage correlates of food addiction were investigated and the extent to which food addiction (as measured by the behavioural food addiction scale) was influenced by factors such as Binge Eating Disorder, impulsivity, emotional eating and other addictions were explored.

In Stage 3, the acceptance of Behavioural Food Addiction Scale, and the usability of the scale, was explored with four clinicians working with people demonstrating problem eating behaviour. In this stage, support for the newly developed scale was established among clinicians working with people who demonstrated problem eating behaviours. Additionally, this support was illustrated both in terms of support for use of the Behavioural Food Addiction Scale and with support for usability of the scale.

Taken together this program of research provides a strong basis for viewing food addiction as a behavioural addiction, as opposed to a substance addiction as it is currently commonly presented. the developed behavioural addiction scale achieved a Cronbach’s α of .93 in the pilot study and small positive corelations were discovered between diagnosis and three eating situations; eating when watching television (τ = 0.15, n = 216, p = .02 (2-tailed)), eating when alone (τ = .14, n = 216, p = .031 (2-tailed)), and eating when upset (τ = .13, n = 216, p = .046 (2-tailed)). This indicates that the scale validly measures the construct of food addiction from a behavioural perspective, supporting continued use and further investigation. The results provide a solid basis for continuing work to encourage inclusion of food addiction as a behavioural addiction in diagnostic manuals. The research also demonstrates preliminary support for the scale to be utilised in clinical contexts to assist those with problem eating behaviours. This utilisation includes identifying problem food, identifying settings in which problems occur and establishing whether food addiction may be a contributing factor in an effort to direct treatment.

Further research is necessary with larger sample sizes to establish estimates of behavioural food addiction among specific populations. This should include a larger, representative Australian adult population as well as clinical populations (e.g. participants with diagnosed binge eating disorder, overweight and obese populations) to establish how generalisable these results are to the Australian population as a whole and specific clinical populations. There is also scope for adaptation of the scale to suit specific populations such as adolescents or children. Further research is needed to assess the utility of the existing scale among these populations and develop adaptations of the BFAS specific to these populations. The results demonstrated here provide support for food addiction as a behavioural addiction and develop the concept of food addiction as a clinical diagnostic category.