Problematic Issues Around Behaviour Change Theory and Practice

The literature review leads to the conclusion that behaviour change frameworks have systematically tried to become more effective in changing behaviour by focusing on a continued concern for grounding proposals in sound evidence and state-of-the-art theory. However, throughout this process, key areas of attention with significant ethical and practical implications have been dismissed. This has contributed to make many of current frameworks less participatory and more focused on the views of experts than on the perspectives of the people whose behaviours are being targeted. But the

literature on behaviour change proclaims that stakeholder engagement is fundamental, thus pointing to a clear contradiction.

The Dominating Focus on Agency and Individual Factors

The work of Michael Marmot is fundamental to understanding the determinants of health. In his analysis, behavioural components appear as a core part of how health conditions and outcomes can be understood, both at the individual and at the population level. But his stance is highly critical of the predominant, individualistic model of behaviour that assumes people as all-powerful, rational beings whose informed choices determine their own health status. Marmot (2016) highlights the fundamental conflict between a perspective that assumes an empowered, rational individual capable of changing his/hers circumstances simply by making good choices, with a perspective whereby change and behaviour are largely dependent on external factors that comprise the social, political and economic organisation of the societies people live in.

Spotswood & Marsh (2016) share a perspective critical of a focus on individuals as

‘doers of behaviour’. In their view much attention is given to what people do within a given social structure without really questioning the structure itself – i.e. without trying to understand how and to what extent the circumstances where people live determine or direct individual behaviour; what they believe is lacking is a ‘culturalist’ perspective of behaviour (Spotswood & Marsh, 2016). Halpern et al. (2004), similarly, advocate for

21 See section 4.1.2 of Chapter 04.

43 a middle-ground perspective where responsibilities over behavioural outcomes are shared between individuals, community and state. They propose a model that, sensitive to contextual conditions, ultimately attempts to navigate questions about the

fundamental rights of people, the rights and responsibilities deemed to be ‘good’ by society (as to clarify what are the desirable behaviours to be targeted), who is/are

‘causally responsible’ for the outcomes, and who is ‘best placed to act’ (Halpern et al., 2004).

The Overstated Value of Evidence from RCTs

Another aspect inherited from the scientific tradition whence behaviour change emerged is the notion that RCTs provide the most reliable and desirable type of evidence to inform intervention design and, consequently, policymaking. The ramifications of this contentious issue have been recently summarised under the suggestive title ‘the evidence conundrum’ (Spotswood & Marsh, 2016). These

ramifications include the extent to which evidence of various types are taken seriously and embraced within behaviour change research and practice, and the relative

disadvantage of qualitative methodologies in light of a biased quantitative research ethos. Moreover, how, first and foremost, evidence emerging from RCTs are even appropriate in the context of transdisciplinary projects (frequently mixed-methods or exclusively qualitative) that look at ‘complex social problems in which value positions can shape both the questions asked by researchers and the conceptual and empirical frameworks deployed’ (Spotswood & Marsh, 2016).

A complementary argument brought by Trenchard-Mabere (2016) concerns how a more context-based, systems approach to intervention design questions the underlying assumptions about the appropriateness of a blunt quantitative approach. She argues that RCTs assume inaccurate presuppositions regarding complex social phenomena, including that: the relationships of cause and effect are linear; the environmental and the contextual factors can somehow be ‘controlled’; and the effects of an intervention can be predicted (as opposed to be cocreated and learnt from).

Being subservient to an ideology where RCTs occupy, single-handedly, the higher levels of the scientific hierarchy has additional implications that deeply affect research

funding, limit research design, and may consequently dismiss future qualitative, social research as something that cannot contribute to understanding and changing of behaviour within healthcare (Spotswood & Marsh, 2016). The disparity of scientific

44 paradigms that operate under distinct, and often at times, conflicting premises

concerning theories and methods has been noted as a detriment to design practice within quality improvement. Design, being predominantly qualitative and abductive, could have an a priori disadvantage over other dominant disciplines, approaches and methodologies to improving healthcare practice and service provision (Robert &

Macdonald, 2017)²².

The Relative Disregard for Stakeholder Participation

The historical application and evolution of healthcare behaviour change interventions at the public health level has clearly contributed to a notion that stakeholder

participation should (or could) not be a major concern. The rationale for this stance, although apparently paradoxical, is grounded on the premise that the State needs to act on behalf of the population, and should favour public interests over individual benefits or preferences. In doing so, governments afford themselves the authority to ‘act for the public good’, often at times by restricting personal freedom (Maryon-Davis, 2016), which has evident ethical implications. Among the ethical consequences of this

perspective is that any technology or intervention based on such utilitarian grounds is likely to have unintended, negative impacts over ‘our understanding of the meaning of the human condition, letting go of any pretensions to human autonomy, integrity, dignity, and moral responsibility’ (Buchanan, 2006).

Bypassing individual choice may also be another form of denying active participation.

Saghai (2013) addressed this aspect when problematising the concept of behavioural

‘nudges’. This author recollects some fundamental ‘conditions for the preservation of freedom of choice’ which helps safeguard the autonomy of individuals (and, by extension, their participation in the determination of their own behaviour). These conditions include preserving the opportunity of making no-choice or opting-out, as well as to consciously and easily resist choices that are externally presented as default, as ideal, desirable or optimal. Making choices in the face of a predetermined set of options is arguably not the most empowering or most comprehensive way of

participating, but it certainly constitutes an important aspect to be addressed within a more traditional perspective of the behaviour change field.

22 This aspect will be explored more thoroughly in the last section of Chapter 06.

45 Overlooking the human dimensions of any intervention, at the level of the individual person, may be justifiable by practical arguments – such as the technical and logistic difficulties of involving massive amounts of people in decision-making processes – but it proves problematic when considered from an ethical perspective. That is because people are largely affected by value-based arguments not only rational explanations founded solely on scientific models of prediction aiming at controlling undesirable behaviour (Buchanan, 2006). That said, some level of personal restriction or coercion seems to be tolerated in public health interventions where the collective benefits clearly outweigh the individual costs or inconveniences (Tengland, 2012).

Again, as discussed in previous sections of this chapter, it appears that trying to transpose behaviour change approaches from the public policy/public health scale to the scale of healthcare practice and service provision poses problems of an

epistemological and methodological nature. What individuals may be willing to accept as citizens differs from what they are willing to give up as professionals or as members of a community of practice. Welch is an author that upholds this position. Despite his focus being on practice as a social phenomenon – the aggregated practices of many individuals along with knowledge, skills and the contextual circumstances surrounding actions and behaviours – Welch (2016) emphasises the importance of understanding individuals’ participation (or contributions) as an active aspect of how practice is perceived and performed. Within this perspective, stakeholder participation becomes fundamental, less so from a stance of assuring fundamental individual rights, but more so in terms of how specific human aspects (from psychological and behavioural issues to human factors issues related to usability, interface and design) are represented in understanding and designing practice.

The relative disregard for stakeholder participation in behaviour change can be further observed in two indirect ways: 1) the superficial way in which participation is reported and discussed in standard templates for describing behaviour change intervention design and implementation; and 2) the almost complete absence of reference to the participatory design (and largely the user-centred design) literature in behaviour change publications.

Poor Reporting and Discussion of Participation in Behaviour Change Interventions’ Checklists In the recent years, various templates or checklists for developing and reporting behaviour change interventions have been devised as a means to obtain greater

46 scientific rigour, and to try and facilitate comparisons and criticism. However, issues concerning participation are only superficially addressed within these tools. In this present section, three checklists are briefly analysed and compared to make evident the lack of a deeper discussion regarding how and with what purposes participants are (or could be) involved in behaviour change intervention design and implementation. The selected checklists are: the Template for Intervention Description and Replication, TIDieR, (Hoffmann et al., 2014); the checklist to improve reporting of group-based behaviour-change interventions (Börek et al., 2015); and the Single-Case Reporting Guideline in Behavioural Interventions, SCRIBE, (Tate et al., 2016).

As a general characteristic, all three checklists describe and discuss participation overwhelmingly in the context of sampling, demographics, recruitment and

randomisation. Both the checklists proposed by Börek et al. and by Tate et al. include either a section or subsection titled Participants, under which aspects such as group composition and size, group allocation, selection criteria, method of recruitment, and demographic characteristics (and clinical features) of participants are considered.

Within SCRIBE, the extent to which participants have completed sequences of

interventions or partaken in trials is also included in the Results section; this section of the checklist additionally looks into adverse events that might have occurred across the stages of the intervention. A topic concerning if and how informed consent was

obtained from participants is also present in the SCRIBE.

The predominant focus of the TIDieR checklist is on the rationale and mechanics of intervention implementation. Topics concerning why the intervention is taking place, what constitutes the materials, their format, and how and by whom they were

delivered are the principal parts of this template. Any allusion to participants (other than the experts delivering the intervention) is implicit or indirect, and the word

‘participant’ is only used once (in reference to the materials ‘provided to participants’).

What is observed is not surprising, given the epistemological tradition and the motivations behind these proposals. Within all these checklists, participants are perceived either as objects of research projects or as receivers of interventions

developed, designed and deployed to them by professional experts. That stance, which aligns with a positivist, ‘hard science’, quantitative research orientation, views

participants in a completely different light of that of participatory design – a field largely ignored by the behaviour change scientific community, as explored below.

47 Lack of Awareness of Participatory Design Literature

Among all the behaviour change publications consulted regarding the present literature review, there is an observed neglect concerning participatory design references. All discussion about the values, motivations and aspirations of PD epistemology, as well as the practical and operational aspects of PD methodology seem mostly absent from the broad behaviour change literature. That is not to say that participation, as an

encompassing concept is not discussed; but to affirm that the particular approach to participation brought by PD is not contemplated, aside from one or two timid exceptions.

Generally speaking, stakeholder participation in behaviour change interventions is understood in the context of a scientific research project – either guided by a more social sciences orientation, or by a more health sciences model. Within these perspectives, the participation of stakeholders is largely perceived as a necessary condition to obtain information from people that may constitute a subsample of a larger group of beneficiaries of a research initiative. As such, stakeholders are seen predominantly as informants or users, not as cocreators. Under this logic, the methodologies employed to engage with participants are based primarily on questionnaires and interviews, and, sometimes, observations and focus groups. The intent is to understand how stakeholders think and feel about things so that experts can design interventions for them.

Participatory design, on the contrary, assumes stakeholders as active, engaged members of the design team. Their remit may encompass everything from the definition of ‘first principles’ and objectives, to data collection, analysis and solution development. This view has, thus, philosophical and operational consequences, impacting on both how stakeholders are perceived, and what roles and activities they perform through the intervention process. Without referring to participatory design literature, behaviour change discussions have a limited perspective on participation at the epistemological and the methodological levels.

(To some extent) the two exceptions to this would be Intervention Mapping

(Bartholomew et al., 1998; Bartholomew Eldredge et al., 2016) and the Person-Based Approach (Yardley et al., 2015). Although not mentioning PD authors or publications, Intervention Mapping discusses the participation of stakeholders in the context of their involvement in intervention design, emphasising engagement from the diagnosis of behavioural issues through to piloting and implementation. The Person-Based

48 approach does allude to user-centred methods, such as ‘usability testing’ and its

development within human-computer interaction (HCI) (Yardley et al., 2015) which shares some common ground with the extended PD literature. Both these examples, however, seem to dialogue more closely with participatory action research than with participatory design.