4. Extreme demand avoidance questionnaire

The ‘Extreme Demand Avoidance Questionnaire’ (EDA-Q) is a measure designed to quantify traits of extreme/‘pathological’ demand avoidance in children aged 5-17, on the basis of parent or teacher-report. The questionnaire was developed to ensure that PDA traits were being measured in a consistent way for research purposes. The English version of the EDA-Q can be downloaded with scoring information hereThe full research study describing its preliminary validation can be downloaded here

Background & preliminary validation

Questionnaire items were designed with input from ten clinicians working in a variety of settings in the UK who had experience of children with this profile. A preliminary validation study was conducted, in which 326 parents of children aged 5-17 participated. Parents were recruited through schools and online web-forums/ web-groups, including those specialising in PDA or ASD. As well as asking parents to complete the EDA-Q, we asked what (if any) diagnoses their child had received, and used this information, plus information from the Strengths and Difficulties Questionnaire, to allocate participants into sub-groups.

The six sub-groups we identified were as follows: (1) typically developing children (N=102), (2) children with ASD without disruptive behaviour (N=36), (3) children with ASD with disruptive behaviour (N=48), (4) children for whom PDA was suspected by parents (irrespective of other diagnoses) (N=67), (5) children who had, according to parents, been identified as having PDA by a health professional, irrespective of other diagnoses (N=50), and (6) disruptive behaviour or behavioural problems without suspected/identified ASD or PDA (N=23).

When scores were compared between the two PDA groups (diagnosed and suspected) and the other groups on the Strengths and Difficulties Questionnaire total difficulties scale, there were no differences between the PDA groups and those with ASD and disruptive behaviour. However, score on the EDA-Q was significantly higher in PDA than in all comparison groups, including ASD with disruptive behaviour (see Figure 1).


Figure 1: Scores across groups. Boxes represent the lower and upper quartiles, with the horizontal line in the middle reflecting the median value. Whiskers indicate the range of scores, and outliers are identified as crosses. The six participant groups are TD= typically developing; ASD DB-= ASD without disruptive behaviour; ASD DB+= ASD with disruptive behaviour; PDA suspected: PDA suspected by parents; PDA identified: PDA reported to have been identified by a health professional; DB/Behaviour: Disruptive behaviour or behaviour problems.

We ran analyses to identify a cut-off for the measure to best differentiate PDA from comparison groups. This cut-off would identify individuals at elevated risk of having a profile consistent with PDA, but should not be considered diagnostic. A full assessment is required to examine the nature and quality of particular features to inform clinical decision making.

Different risk cut-offs were required for older and younger age groups. A cut-off score of 50 was appropriate for ages 5-11 years and 45 for 12-17 years. Statistics that assess how well the measure discriminated across groups suggested that it was effective in doing this (sensitivity = .80 and specificity = .85). 

However, there are certain limitations. Because we recruited participants from specialised sources, the likelihood of ‘false positives’ is lower in our sample compared to what it would be in the general population, where individuals with extreme/‘pathological’ demand avoidance are less common. In addition, this study included relatively few children with disruptive behaviour disorders without ASD ('emotional and behavioural difficulties’), who exhibit some features of extreme/‘pathological’ demand avoidance, in particular non-compliance. Finally, group allocation was based on information reported by parents, rather than comprehensive clinical assessments. As such, further validation of this measure is required. 

To conclude, the EDA-Q may be useful in quantifying behavioural features of extreme/‘pathological’ demand avoidance on the basis of informant report. The cut-offs from the preliminary validation study may be useful as a guide to the reported severity of these behavioural features, but are not diagnostic. 

Usage of the EDA-Q & translations:


To date, the EDA-Q has to date been used to quantify features of extreme/'pathological' demand avoidance by several different research groups.


A German translation of the measure is currently available, and can be sent upon request. If you would like to translate the measure into other languages (or already have a translated version), please get in touch with me directly. 


Other methods to quantify extreme/'pathological' demand avoidance:


Recent work has identified items from within the Diagnostic Interview for Social and Communication Disorders (DISCO) that are both common in extreme/'pathological' demand avoidance and specific to it as compared to the rest of the autism spectrum. The publication describing the identification of these DISCO items can be downloaded here