Acceptance is not acceptance, but acceptance!

This journal recently published a paper by Vowles et al. 2020, entitled "Initial Evaluation of the Chronic Pain Acceptance Questionnaire – 2. The authors discuss the development of a two‐item measure to assess chronic pain acceptance. Items for this tool were derived from the Chronic Pain Acceptance Questionnaire (CPAQ) 20, and aim to map two key features of acceptance: (1) Activity Engagement, which entails participating in important or meaningful activities with continued pain and (2) Pain Willingness, which entails a willingness to experience pain without the need to reduce, avoid, or otherwise change it. The proposed development of the CPAQ‐2 is a further reduction of the CPAQ‐8, which was a reduction of the CPAQ‐20 and the CPAQ‐34 and fits within a broader effort to shorten questionnaires in order to increase their use in clinical settings. Furthermore, decreasing the number of items of a questionnaire reduces patient burden of (too) lengthy questionnaires and may boost the chance of recurring use to assess treatment efficacy. Yet, despite the many advantages of shortening questionnaires, caution is warranted as pruning items may not be without consequences.

A first consequence relates to a change in the content of the tool. Indeed, although an initial questionnaire may be developed to be content valid, subsequent item pruning, without taking content into account, frequently results in loss, narrowing or modification of the content and meaning of the questionnaire score. In this context, Lauwerier and colleagues (2015) indicated that this was not different for the CPAQ. Particularly, they indicated that although the original instrument captured all three core components of chronic pain acceptance (i.e. disengagement from pain control, pain willingness and engagement in important or meaningful activities despite pain), later reductions almost uniquely include items concerning engagement in important or meaningful activities despite pain and pain control, but lost the content related to disengagement from pain control, and pain willingness, key in the original definition(s) of chronic pain acceptance. Taken together, some concerns could be raised about the representativeness of the items included in the later versions of the CPAQ for the construct of ‘chronic pain acceptance’. Based upon the study of Lauwerier and colleagues (2015), reporting on both items of the CPAQ‐2, these concerns may also apply on the newly developed tool. Note additionally, that the item of the pain willingness scale of the CPAQ‐2 was found to measure pain control, rather than pain willingness. Although assessment of both constructs is valuable, they are not necessarily two sides of the same coin. Labelling subscales in line with the item content may be a step forward in this discussion.

The discussion about the content validity of the CPAQ‐2 may not come as a surprise as items were derived from the CPAQ‐8 using well‐validated pruning techniques, but without taking into account the importance of content validity. For ‘chronic pain acceptance’, one may even wonder whether it is possible to grasp its complexity using only 2 items, each tapping into one complex subdimension. Indeed, several scholars have argued that single item questionnaires may be particularly well‐suited for simple (uni‐dimensional) or concrete constructs that are well‐understood, but less for the assessment of complex constructs, such as ‘Chronic pain acceptance’ or even its subscales, which may need multi‐item measures to be representative for their respective construct definition. Based upon these findings, one may wonder whether acceptance before and after pruning of ‘irrelevant’ items over time is still the same.

The pruning of items may also impact upon the sensitivity for change. This change could go either way, decreasing or increasing sensitivity for change. Several studies have shown that eliminating items that poorly represent the construct which they intend to measure may increase sensitivity for change as including such items adds noise to the questionnaire. Otherwise, changing the content of a questionnaire by pruning items, some of which may capture unique features of each subscale, may adversely impact upon sensitivity for change. It may even be that differences exist of sensitivity for change between different versions depending upon the particular context and population, warranting future research to address this issue.

Concerns raised about content validity and change sensitivity (due to pruning items) are, however, not unique for the CPAQ‐2 (e.g. Crombez et al., 2020). Particularly, content validity has often been overlooked at the cost of other types of validity (e.g. construct validity) despite being a key feature of questionnaire development. This is surprising as several tools are at hand to assess content validity, such as the discriminant content validity method developed by Johnston and colleagues or cognitive interviewing techniques which ensure that items are relevant and representative for the content of the intended construct (see also Crombez et al., 2020). We believe that research and clinic would benefit from future questionnaire development where content validity and reduced questionnaire length go hand in hand. Aiming for short content valid questionnaires promises exciting times in the field of questionnaire development and validation.