taxonomy of approaches to developing …...health. these interventions have multiple interacting...

REVIEW Open Access

Taxonomy of approaches to developinginterventions to improve health: asystematic methods overviewAlicia O’Cathain1* , Liz Croot1, Katie Sworn1, Edward Duncan2, Nikki Rousseau2, Katrina Turner3, Lucy Yardley3 andPat Hoddinott2

Abstract

Background: Interventions need to be developed prior to the feasibility and piloting phase of a study. There are avariety of published approaches to developing interventions, programmes or innovations to improve health.Identifying different types of approach, and synthesising the range of actions taken within this endeavour, caninform future intervention development.

Methods: This study is a systematic methods overview of approaches to intervention development. Approacheswere considered for inclusion if they described how to develop or adapt an intervention in a book, website orjournal article published after 2007, or were cited in a primary research study reporting the development of aspecific intervention published in 2015 or 2016. Approaches were read, a taxonomy of approaches was developedand the range of actions taken across different approaches were synthesised.

Results: Eight categories of approach to intervention development were identified. (1) Partnership, where peoplewho will use the intervention participate equally with the research team in decision-making about the interventionthroughout the development process. (2) Target population-centred, where the intervention is based on the viewsand actions of the people who will use it. (3) Evidence and theory-based, where the intervention is based onpublished research evidence and existing theories. (4) Implementation-based, where the intervention is developedwith attention to ensuring it will be used in the real world. (5) Efficiency-based, where components of anintervention are tested using experimental designs to select components which will optimise efficiency. (6) Steppedor phased, where interventions are developed with an emphasis on following a systematic set of processes. (7)Intervention-specific, where an approach is constructed for a specific type of intervention. (8) Combination, whereexisting approaches to intervention development are formally combined. The actions from approaches in all eightcategories were synthesised to identify 18 actions to consider when developing interventions.

Conclusions: This overview of approaches to intervention development can help researchers to understand thevariety of existing approaches, and to understand the range of possible actions involved in interventiondevelopment, prior to assessing feasibility or piloting the intervention. Findings from this overview will contributeto future guidance on intervention development.

Trial registration: PROSPERO CRD42017080553.

Keywords: Intervention development, Review, Methodology, Guidance, Health

* Correspondence: [email protected] Care Research Unit, Health Services Research, School of Health andRelated Research (ScHARR), University of Sheffield, Regent Court, 30 RegentStreet, Sheffield S1 4DA, UKFull list of author information is available at the end of the article

© The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

O’Cathain et al. Pilot and Feasibility Studies (2019) 5:41 https://doi.org/10.1186/s40814-019-0425-6

http://crossmark.crossref.org/dialog/?doi=10.1186/s40814-019-0425-6&domain=pdf

http://orcid.org/0000-0003-4033-506X

https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=80553

mailto:[email protected]

http://creativecommons.org/licenses/by/4.0/

http://creativecommons.org/publicdomain/zero/1.0/

BackgroundPolicy makers, health professionals, patient groups, thepublic, designers and researchers develop interventions,programmes or innovations to improve health. It is im-portant that the intervention development process maxi-mises the chances that an intervention will be effectiveand sustainable. Unless it does, there is a risk of researchwaste [1], where expensive evaluations are undertaken offlawed interventions that turn out not to be feasible, ac-ceptable or effective in subsequent feasibility studies orfully powered evaluations [2].In recent years, researchers have published journal ar-

ticles, websites and books on how to develop interven-tions. This international endeavour, proposing ways ofdeveloping interventions that others can follow, could bedescribed as the production of guides, guidance, meth-odology or frameworks. In this article, the umbrella term‘approaches’ is used. These approaches are distinct frompublications describing the development of a specificintervention. Approaches that show how to develop in-terventions are useful for those new to intervention de-velopment. They offer an opportunity for researchcommunities to refine and improve those approaches forfuture use.There are a variety of approaches to intervention de-

velopment and it is timely to bring these together andsynthesise them to understand the range of actions avail-able. Previous reviews of intervention development havefocused on identifying approaches used in the specificcontext of behaviour change in implementation science[3], optimisation in terms of making final modificationsto interventions prior to formal evaluation [4], the use oftheory in intervention development for a single condi-tion [5] and ways of adapting interventions for ethnicminority communities [6].Complex interventions are widely used to improve

health. These interventions have multiple interactingcomponents, target multiple groups or levels of an or-ganisation and attempt to affect multiple outcomes [7].The United Kingdom Medical Research Council (MRC)is widely cited for its guidance on developing and evalu-ating complex interventions [7], describing the fourphases of developing, feasibility/piloting, evaluation andimplementation. One part of the guidance has been ex-tended recently to offer more detail on process evalu-ation [8]. Guidance on the feasibility/piloting phase iscurrently being extended, following recent publication ofa systematic review of existing guidance for this phase[9]. Some researchers have considered enhancements tothe development phase of the MRC guidance for thespecific field of nursing studies [10]. The MRC hasfunded a study to produce guidance on intervention de-velopment: ‘IdentifyiNg and assessing different ap-proaches to DEveloping compleX interventions (the

INDEX study). As part of the INDEX study [11], a sys-tematic review of approaches to intervention develop-ment was undertaken to identify the range ofapproaches available, and to synthesise the actionswithin these approaches, in order to help researchers todevelop complex interventions and to inform futureguidance on intervention development.

MethodsSystematic methods overviewSystematic methods overviews are reviews of themethods literature [12–14]. Guidance has been pub-lished to help researchers to undertake systematicmethods overviews [14]. This guidance was followed toundertake a systematic methods overview of differentapproaches to developing complex interventions. Ex-haustive searching and inclusion of all relevant literatureassociated with systematic reviews of primary research isnot necessary because learning and arguments aboutmethodology and methods are repeated frequently in theliterature. Instead, there is an emphasis on broad search-ing to identify the range of relevant literature, and ondata saturation of learning and arguments [12, 13]. Theprotocol is available, registered at PROSPEROCRD42017080553.

The aim of the overviewThe aim of this overview was to identify a broad rangeof approaches to intervention development. The em-phasis was on recently produced or recently used ap-proaches, because of the rapid development of this field,with newer approaches building explicitly on older ap-proaches. The objectives were to construct a taxonomyof approaches to help future developers think about theapproach they might take, and to synthesise the actionswithin each approach to identify the full range of actionsdevelopers can consider.

Definitions used in this overviewInterventionA health intervention is an effort, activity or combin-ation of programme elements designed to improvehealth status. This overview focuses on complex inter-ventions that include a number of components whichmay act both independently and inter-dependently. Thisincludes policy innovations such as introducing a newhealth service or public health policy nationally (e.g.smoking ban in public places). It does not include thedevelopment of medicines and any invasive interven-tions (e.g. pills, procedures, devices). Complex interven-tions to improve health or health care outcomes can bedelivered in many settings including health care facil-ities, schools, local communities or national populations.They can be delivered by a range of individuals including

O’Cathain et al. Pilot and Feasibility Studies (2019) 5:41 of 27

health care, social care and public health practitioners,as well as professionals working outside of the healthcare sector, such as teachers, charity workers and peers.

Intervention developmentCraig et al. [15] proposed the development phase to bethe period when the ‘intervention must be developed tothe point where it can reasonably be expected to have aworthwhile effect’. (p. 9). The start and end points of thedevelopment phase are not always clear. There may beoverlap between the development phase and the subse-quent phase of feasibility and piloting, because some ex-ploration of feasibility is often part of the interventiondevelopment process [16]. A helpful indicator of the endof the development phase is the production of a docu-ment or manual describing the intervention and how itshould be delivered [16]. There may also be overlap be-tween the intensive development phase and a longerperiod of preparation prior to intervention development,when a team undertakes a series of studies over a num-ber of years before the point of formally developing anintervention. This may involve assessment of the evi-dence base, including reviewing the effectiveness ofexisting interventions, and/or qualitative research withstakeholders. Alternatively, these studies may be under-taken as part of the intensive intervention developmentphase. This overview focuses on the intensive develop-ment phase, recognising that the start and end of thisphase may be hard to define.

Refinement, optimisation, modification and adaptationDuring the development process, the initial version ofthe intervention may be repeatedly refined by makingimprovements based on early assessment of feasibilityand acceptability. This process continues throughout theformal feasibility/pilot testing and evaluation phases (seeFig. 1). Indeed, some researchers see intervention devel-opment as a long-term ongoing endeavour which laststhroughout the full evaluation and implementationphases [17]. Early refinement, during the developmentphase, is included in this overview but later refinement,during or after the formal pilot phase of an evaluation, isexcluded.Intervention optimisation is the process of improving

the efficiency of an intervention. Different componentsare assessed to identify those affecting intermediate out-comes, so that only effective components are included inthe intervention to be fully evaluated. A recent review ofoptimisation of complex health interventions prior to afull trial has been published [4]. Approaches to optimisa-tion are therefore not a focus of this overview unlesstheir authors frame them specifically as approaches tointervention development.

Sometimes researchers take interventions that havebeen shown to be effective at the evaluation phase, andperhaps implemented in the real world, and adapt themfor a new sub-population, health condition or context(Fig. 1); for example, an existing effective interventionmight be adapted for ethnic minority communities [6].Such adaptation may involve a formal developmentphase, so approaches to adaptation are included in thisoverview if they are framed by their authors as interven-tion development.

An ‘approach’ to intervention development‘Approach’ refers to the whole process of interventiondevelopment documented in a book, website or journalarticle where authors explicitly offer a guide to undertak-ing intervention development. Approaches may providedifferent amounts of detail about how to develop anintervention. All are included in this overview regardlessof the amount of detail offered.

Development versus designSometimes researchers use the terms ‘development’ and‘design’ interchangeably. In this overview, the term ‘de-velopment’ is used for the whole process of interventiondevelopment and the term ‘design’ is reserved for a pointin the development process where developers make de-cisions about the intervention content, format anddelivery.

SearchThe focus of systematic methods overviews can in-clude the literature describing or critiquing methodsor methodology, and the methods sections of primaryresearch papers [14]. The focus of this overview is re-cent literature documenting how to develop an inter-vention. Approaches were considered for inclusion ifthey describe how to develop an intervention in abook, website or journal articles published after 2007,or are cited in a primary research study reporting thedevelopment of a specific intervention published in2015 or 2016.In systematic methods overviews, the search strategy

should be transparent and broad rather than exhaustive[13, 14]. The process started with a primary search ofthe databases Medline, CINAHL, PsycINFO, ASSIA andERIC from January 2015 to December 2016 using thesingle search term ‘intervention development’. Thesehealth, social science and education databases were se-lected because they include research on complex inter-ventions with health outcomes. Title and abstractscreening, followed by full text search, identified journalarticles reporting primary research of the developmentof specific interventions. The methods sections of thesearticles were read by KS and AOC in order to identify


any intervention development approaches that wereused and referenced. The most up-to-date version ofjournal articles, books or websites referenced in these ar-ticles were obtained for data extraction. This search alsoidentified articles describing approaches, and systematicreviews of approaches, to intervention development. Be-cause the search term ‘intervention development’ wassimple and potentially limited the breadth of approachesidentified, a check was undertaken with a second searchusing a broader set of search terms to see if this yieldeda broader set of approaches. A set of diverse terms asso-ciated with intervention development were searched inthe same databases in the same time period: complexbehavioural intervention, develop, design, phase I, ex-ploratory, refine and translate. This identified 808

records. AOC and KS conducted a title and abstractscreen on a sample of records from this search: the first100 records and 1 in every 8 records. Allowing for over-lap, this identified 189 records. The full texts of 26 metthe inclusion criteria and did not identify further ap-proaches to intervention development. Both search strat-egies are detailed in Additional file 1.Searching should go beyond standard bibliographic da-

tabases because methodologies/methods are described inbooks as well as in journal articles [14]. The formalsearch described above was supplemented by a search inGoogle Scholar using the terms ‘intervention develop-ment’, ‘complex intervention development’, ‘interventionoptimisation’, ‘complex intervention pre-clinical’, ‘inter-vention adaptation’ and ‘intervention modification’.

Fig. 1 PRISMA 2009 flow diagram: search for primary studies only


Different terms to those used in the searches of data-bases were used deliberately to facilitate broad searching.Finally, the authors of this overview drew on their exist-ing reference libraries because use of personal know-ledge is also important in reviews of complex evidence[18].

Inclusion and exclusion criteriaApproaches were considered for inclusion if they pre-sented a guide to developing an intervention that hadbeen produced or updated since 2007 or used in primaryresearch published 2015–2016. A purposeful approachto selection of literature is advocated in systematicmethods overviews, with use of maximum variationsampling. The inclusion of literature stops when new is-sues no longer emerge. As approaches were identified,members of the team (AOC, LC, KS) inductively devel-oped a taxonomy of approaches. Data saturation wasconsidered and further approaches not included if theteam considered that saturation was reached. For ex-ample, not all approaches to developing digital interven-tions were included because they repeated the actionsalready identified within that category of approach. At-tention was also paid to diversity of context within eachcategory and across all categories in the emerging tax-onomy. This process was led by AOC with team discus-sions with LC and KS.

Data extractionFor each approach, AOC read the article, website orbook and extracted the rationale stated by authors forthe approach, the context for which the approach wasconstructed, the key actions undertaken, the methodsused to deliver each action and the strengths and limita-tions of the approach. These strengths and limitationswere identified by the authors of the approach, the au-thors of other approaches included in the overview orthe research team (indicated by ‘INDEX’).

Quality appraisalQuality appraisal is a challenge in these overviews [14].There was no formal assessment of the quality of the ap-proaches to intervention development because assess-ment criteria do not exist.

AnalysisThe constant comparative method has been recom-mended for synthesising within systematic methodsoverviews [14]. Most overviews are aggregative in termsof bringing together different concepts, rather than in-terpretative in terms of developing new concepts [14]. Aconstant comparative aggregative approach was under-taken within three concurrent steps:

(i) AOC extracted data on rationale, context, keyactions, methods and strengths and limitations tosummarise each approach within a table.

(ii) AOC developed a taxonomy of approaches. AOCgrouped approaches together based on the statedrationales for each approach because these conveythe intentions of the authors. These rationales wereextracted from statements made by the authorswhen introducing their approach. The categories ofapproach, and the individual approaches includedwithin them, were discussed and refined by AOC,LC and KS until an agreed taxonomy wasproduced.

(iii)AOC listed the actions from each approach,grouped similar actions and brought these togetherto identify a comprehensive set of actions from allthe categories of approach, including the methodsthat could be used at each action. AOC, LC and KSdiscussed these actions until agreement wasreached.

ResultsApproaches identified and includedAs previously stated, the intention was to undertake abroad rather than an exhaustive search, where more in-formal searches are as important as formal searches.PRISMA flow charts are devised to display exhaustivesearches within standard systematic reviews. For thissystematic methods overview, a PRISMA flow chart isdisplayed for the search using the term ‘intervention de-velopment’ of primary studies reporting intervention de-velopment in 2015–2016 (Fig. 1).

Taxonomy of approachesEight categories of approach to intervention develop-ment were identified, distinguished by the rationalesstated by the authors of these approaches (Table 1). Thereview team identified the following labels for theseeight categories based on the language used by authorsof approaches.

1. Partnership intervention development. Threeapproaches were included, addressing a range ofcontexts (quality improvement, social innovationand radical innovation). Partnership approachesincluded co-production with equal participation indecision-making of the research team and thepeople whom the intervention aimed to help, anduser-driven development. Due to similarities withapproaches already included, a partnership ap-proach for implementation science was not in-cluded [19]. Primary research studies reported usingcommunity-based participatory research but articlesor books describing how to use this approach for


intervention development could not be located so itwas not included here.

2. Target population-centred intervention develop-ment. Three approaches were included, address-ing a wide range of contexts (health caredelivery, technology, behaviour change and self-management).

3. Evidence and theory-based intervention develop-ment. Six approaches were included, addressing arange of contexts (complex interventions in healthand health care, public health, social policy, behav-iour change and quality improvement). Some ofthese approaches also proposed a rationale of beingsystematic (see category 6) but were included in this

Table 1 Taxonomy of approaches to intervention development

Category INDEX team definition Defined approach Source

1. Partnership The people for whom the intervention aims tohelp are involved in decision-making about theintervention throughout the development process,having at least equal decision-making powers withmembers of the research team

Co-production, co-creation,co-design, co-operative design

Voorberg et al. 2015 [38]Bessant and Maher 2009 [39]Spencer et al. 2013 [40]

User-driven Kushniruk and Nøhr 2016 [25]

Experience-based co-design(EBCD) and accelerated EBCD

Robert et al. 2013 [41]Locock et al. 2014 [42]

2. Target population-centred

Interventions are based on the views andactions of the people who will use the intervention

Person-based Yardley et al. 2015 [17]

User-centred Erwin and Krishnan 2016 [44]Erwin and Krishnan 2016 [45]Erwin 2015 [43–45]

Human-centred design Norman 2013 [33]

3. Theory and evidence-based

Interventions are based on combiningpublished research evidence and formaltheories (e.g. psychological or organisationaltheories) or theories specific to the intervention

MRC Framework for developing andevaluating complex interventions

MRC Guidance [7, 15]

Behaviour change wheel (BCW) Michie et al. 2014 [26]

Intervention mapping (IM) Bartholomew Eldredge et al.2016 [27]

Matrix Assisting Practitioner’sIntervention Planning Tool (MAP-IT)

Hansen et al. 2017 [32]

Normalisation process theory (NPT)a Murray et al. 2010 [46]

Theoretical domains framework(TDF)

French et al. 2012 [47]

4. Implementation-based Interventions are developed with attention toensuring the intervention will be used in thereal world if effective

Reach, Effectiveness, Adoption,Implementation, Maintenance(RE-AIM)

RE-AIM.org [48]

5. Efficiency based Components of an intervention are testedusing experimental designs to determineactive components and make interventionsmore efficient

Multiphase optimization strategy(MOST)

Collins et al. [49]

Multi-level and fractional factorialexperiments

Chakraborty 2009 [50]Dziak et al. 2012 [50, 51]

Micro-randomisation trials Klasnja et al. 2015 [52]

6. Stepped or phasedbased

Interventions are developed throughemphasis on a systematic overview ofprocesses involved in intervention development

Six essential Steps for QualityIntervention Development (6SQUID)

Wight et al. 2015 [28]

Five actions model Fraser and Galinsky 2010 [29]Fraser et al. 2009 [24, 29]

Obesity-Related BehaviouralIntervention Trials (ORBIT)

Czajkowski et al. 2015 [34]

7. Intervention-specific An intervention development approach isconstructed for a specific type of intervention

Digital (e.g. Integrate, Design, Assessand Share (IDEAS))

Mummah et al. 2016 [30]Horvarth et al. 2016 [30, 53]

Patient decision support or aids Elwyn et al. 2011 [31]Coulter et al. 2013 [31, 54]

Group interventions Hoddinott et al. 2010 [55]

8. Combination Existing approaches tointervention developmentare combined

Participatory Action Researchbased on theories of BehaviourChange and Persuasive Technology(PAR-BCP)

Janols and Lindgren 2017 [56]

aCould be considered under implementation based approaches to intervention development because the theory is about implementation


http://re-aim.org

category because they emphasised the role of evi-dence and theory within their rationale. There werea large number of these. Some approaches identifiedin the primary research study search were not in-cluded here due to data saturation [20, 21].

4. Implementation-based intervention development.One approach was included, in the context ofhealth behaviour interventions.

5. Efficiency-based intervention development. Threeapproaches were included, although they were notindependent of each other. Two of the approachesdefined different ways of optimising components forparts in the first approach in this category.

6. Stepped or phased-based intervention development.Three approaches were included, addressing a rangeof contexts (public health, social policy and clinicalpractice).

7. Intervention-specific development. Five approacheswere included in three intervention groups—digitalbehaviour change interventions, patient decisionaids and group interventions. Other approacheswere identified but not included here because ofdata saturation, because they were outside the timerange of post 2007 and had not been used in theprimary research studies published 2015–2016 [22],or because they offered recommendations for thefuture rather than current guidance [23].

8. Combination approach to interventiondevelopment. There was one approach includedhere in the context of behaviour change.

The stated rationale, context, key actions and strengthsand limitations of each included approach are describedin Table 2 and these characteristics are consideredbelow.

ContextsThe approaches included were produced for a widerange of contexts and sometimes for multiple contexts.They addressed the contexts of behaviour change (11approaches), public health and health promotion (9),digital health (6), complex interventions (5), quality orservice improvement (3), clinical research (2), social pol-icy or innovation (2) and others (Table 2).

Key actionsThe actions from all the included approaches were syn-thesised to identify a total of 18 actions. These actionsare displayed in seven domains of intervention develop-ment. Although some authors describe intervention de-velopment as a broadly sequential process, in that someactions are usually undertaken prior to others [24], au-thors also emphasise that intervention development is acyclical or iterative rather than linear process (for

example, action four may generate understanding thattakes developers back to action two), or there may berepetition within a single action until the developers areready to move on to the next action [15, 17, 25–31].Some actions may be undertaken concurrently. There-fore, the actions are presented within domains, withsome attention to broad sequencing of domains and ac-tions, to facilitate understanding of the process of inter-vention development. The seven domains are presentedin three tables (Tables 3, 4 and 5) and described below.Authors of some approaches state that intervention

development is a cumulative or progressive process, ar-guing that it is necessary to spend time on the early ac-tions and get this right because later ones depend onthese [26].

Conception and planningThe first two domains of intervention development areConception and Planning (Table 3). Although the Con-ception domain has only one action, it emphasises theimportance of being transparent about where the ideafor an intervention has originated. The Planning domainhas seven actions, from deciding who will be involved inthe development process and how, through to consider-ing whether it is worthwhile designing an intervention(Table 3). Authors of some approaches identify the needto spend resources on the Planning domain to get theintervention right (e.g. [32]).

Designing and creatingThe next two domains of intervention development areDesigning an intervention and then Creating it (Table 4).The Creating domain can be an integral action withinthe Designing domain in some approaches. Two ways ofaddressing both of these domains have been proposed:one prioritises working with stakeholders, particularlythe target population, and the other focuses on theory.Constructing a rough prototype early in these domainsis key to some approaches. These domains require cre-ativity [24, 33, 34], with use of a multidisciplinary teamproposed as a way to maximise idea generation andinnovation [30, 34].

Refining, documenting and planning for future evaluationThe final three domains are Refining, Documenting andPlanning for Future Evaluation (see Table 5). The Refiningdomain starts by testing early versions of the interventionon a small sample, and asking whether it merits morerigorous and costly testing [34] before moving to testingon a diverse sample to improve external validity [17, 27].Some approaches consider this domain to be part of theprocess of optimising the intervention and propose theuse of mixed methods research as necessary [17, 27].Other approaches propose quantitative experimental


Table 2 Description of different approaches to intervention developmentCategory Approach Rationale Context specified

by authorsSteps, activities oractions specified byauthorsa

Strengths specified byauthors of approach,authors of otherapproaches and theoverview team INDEX(source in brackets)

Limitations

1. Partnership Co-creation,co-production,co-design [38]

Active involvement ofend users in variousstages of the productionprocess produces moreeffective and efficientservices with higher usersatisfaction [38, 40]A key issue is an equalrelationship between theend users (and theirfamilies andcommunities) andprofessionals, withshared decision-making[40]. It can also be seenas ‘user-led innovation’[39]. Requires a shift inpower from professionalsto community or endusers [40]Co-creation producessustainable competitiveadvantage [39]Customises solutions tospecific contexts [39]Delivers servicesappropriate to the needsof patients and advancesequality [40]

Qualityimprovement inhealth and socialcareSocial innovation inpublic sectorservicesRadicalinnovation—asopposed toincremental—inhealth services

Six steps:1. Identify and build aninitial team includingend users and peopleimportant to the service,developing inclusivecommunicationprocesses. Use joint andequal involvement ofstaff, patients,researchers, peopleleading improvement,and design professionals2. Define and shareassets—knowledge,experience, skills andabilities, influence andconnections. Understandthe current problemthrough non-participantobservation, patient in-terviews, log books, films,local press, use of cam-eras, workshops, storytell-ing, etc.3. Co-create the visionby listening to all voices4. Co-design the solutionusing qualitative re-search, rapid ethnog-raphy and prototyping.Use tools to generatecreative thinking. Openup a range of potentialsolutions as described inuser and human centredapproaches below.5. Build the solutionpossibly using smallaction groups who canuse their relevantexpertise. Make use ofprototyping methods.6. Measure outcomestogether and plan this asan integral part of theprocess

There are examples ofchanges made toservices based on thisapproach [25] andreductions in the cost ofhealth care provision[40]Studies what people dorather than what theysay they do [39]A detailed guide isavailable [40]

Attention has not beenpaid to the outcomes ofco-creation [38]Quantitative methodsneed to be usedbecause qualitativeevaluation of co-creationis dominant [38][38] is a systematicreview of the use of co-creation discussing dif-ferent levels of involve-ment of end usersincluding co-design (thedevelopers lead theprocess in partnershipwith the end users), co-implementation (endusers implement a ser-vice with formal serviceproviders) and initiation(end users develop andimplement innovation).They offer insights intothe process rather than atool-kit (INDEX)

User-driven [25] A participatory approachgoes beyond user-centred design, withusers as active partici-pants in generating de-sign ideas and decision-making. In co-operativedesign, users and de-signers work together tocome up with a designand further refinements.In user-driven design,the users lead the cre-ative thinking and thedesigners facilitate theprocess.End user involvement iscritical to the adoptionof information systemsbecause it increasesfunctionality and thequality of the systemInvolvement empowersusers

Information systemsin health

Proposes three levels ofparticipation in design:user-centred (see nextgroup in this table), co-operative (see co-production earlier) anduser-driven.Important activitiesinclude:1. Establish co-operationbetween users anddesigners2. Gain insight intocurrent problems andneeds and generatingvisions for futuresolutions. This mayinvolve ‘design games’to free minds andcreativity.3. Continual and iterativeinput from end users4. Develop prototypesand undertake usability

Can be low cost andrapid and thus increasedissemination of newdesigns [25]Shown to be successfulat improving futureprototypes andpreventing theintroduction of systemsthat fail [25]

How, when and whereto engage users remainsopen to question [25]Ensuring the usersinvolved arerepresentative of thetarget population ischallenging [25]Reaching consensuswhen there are differingvoices is challenging [25]Difficult for clinical staffto give time for designbut there are ways ofworking rapidly toalleviate this [25]The ‘interventions’ arenot necessarily intendedfor evaluation in an RCTbut may be usedimmediately in the realworld (INDEX)


Table 2 Description of different approaches to intervention development (Continued)Category Approach Rationale Context specified



Limitations

testing of them in reallife environments or asimulation of this toidentify interactions withwider users and activitiesaffecting use.5. Bring the users whowere observed using theprototypes into furtherdesign meetings foractive participation inrefinement of prototype

Experiencedbased co-design(EBCD) andacceleratedexperiencebased co-design(AEBCD) [41, 42]

Need in-depth patientexperience (narrative) totake action and makeimprovements toservicesPatient accountsgenerate priorities andsolutions that serviceproviders may not thinkofPatient narratives canhelp patients and staffreflect on how toimprove services andestablish an emotionalconnection betweenstaff and patientsPatients as equalpartners in co-designcan generateimprovementAEBCD is more feasiblethan EBCD in thecomplex cash-strappedreal world and offers arigorous and effectiveapproach to qualityimprovement

Serviceimprovementspecific to a singleservice in a singlesetting

Core ‘strands’ are:-Participatory actionresearch-User centred-Reflective practice-NarrativeThere are six steps intwo phases:Phase 1 Discovery1. Project managementestablished2. Local staff areinterviewed about theirexperiences.3. Local patients areinterviewed aboutexperiences to producea ‘trigger film/video’ toprompt discussionamongst patients andstaff aboutimprovements needed.In AEBCD, the film isbased on a nationalarchive rather thangathering local patientexperiences. Patients andcarers are invited to viewthe video and identifypriorities.Phase 2 co-design wherefamily, patients and staffare equal partners insmall working groups4. The priorities of staffand patients, and thevideo, are considered bypatients, carers and staffin a workshop meetingto identify priorities forimprovement.5. Small co-designgroups established toimplementimprovements.6. Small groups re-convene to celebrateand review progress.

Draws on rigorousnarrative-based researchwith a broad sample ofpatients rather than anarrow group of people[42]Active partnershipbetween patients andstaff and focus ontangible resultsproduces results [42]Evaluation shown to besuccessful at producingimprovements in thetarget service and inwider aspects of thehospital [42]Hospitals commitinvestment to doing thisagain so they see it assuccessful [42]Online training toolkit isavailable (INDEX)

Discovery phase is timeconsuming so notpractical in real world ofhealth care. Therefore,AEBCD preferable [42]Some patients found thevideo more negativethan their ownexperiences; there was aheavy workload for localfacilitators but theyobtained wider benefitssuch as capacity building[42]Useful for local serviceimprovement ratherthan developing ageneralizableintervention (INDEX)

2. Targetpopulationbased

Person-basedapproach [17, 57]

Enhances acceptabilityand feasibility of anintervention at earlystages of developmentand evaluationSystematic investigationIn depth understandingof users leads tointerventions that aremore relevant,persuasive, accessible

Digital health-related behaviourchange interven-tions and illnessmanagement inter-ventions becausepeople use e-healthindependentlyHas also been usedoutside digitalinterventions for

Uses mixed methodsresearch and iterativequalitative studies toinvestigate beliefs, needs,attitudes and context oftarget populationTwo elements: First, adevelopmental processusing qualitativeresearch with a diversesample of target

Systematic way ofgaining in depthunderstanding of users’perspectives to makethe intervention morerelevant and engaging[17, 57]Shown to be successfulbecause interventionshave been effective inRCTs [17, 57]

Iterative approach maybe hard to respond toquickly in practice [17,57]





Limitations

and engagingComplements theory-based and evidence-based designMatches fundamentaldesign to needs andgoals of users

self-managementBehaviour changeinterventionsEarly stages ofdevelopment andevaluation

population. Goes beyondacceptability, usabilityand satisfaction tounderstand the psychosocial context of the userso can makeintervention relevant tothem. Second, identify‘guiding principles’ toguide interventiondevelopment. Theseelements are used atfour stages of theprocess:1. At the planning stageundertake synthesis ofqualitative studies orqualitative research toprioritise what isimportant or identifynew components of anintervention2. At the design stageidentify the interventionobjectives and featuresof the interventionrequired to deliver them3. When the prototype isavailable, evaluateacceptability andfeasibility4. Implement in real lifesetting to further modifyintervention

Advantage over co-design is that people arebasing views on actualuse of the intervention[17, 57]Different from user-centred approach usedin computer-based re-search because looks be-yond usability andtechnical issues [17, 57]Reasonable amount ofdetail given, withexamples (INDEX)

User-centreddesign [43]

Making delivery moreefficient and equitableby putting people at thecentre of any problem todevelop solutions thatbetter fit their everydaylives, activities andcontextMust designinterventions to fit users’needs and context tofacilitate translation ofevidence into the realworldMay need newapproaches to addresscomplexity

Innovation inorganisationsImproving healthcare delivery

Early and continuousstakeholderengagement, includinghaving stakeholders aspart of research team toundertake contextualinquiry. Three phases:Phase I Defining designrequirements: Use of roleplay and observation toidentify issues ratherthan only qualitativeinterviews; developprototypes to getspecific views on theinterventionPhase II Develop aprototype and refine initerative interviews: e.g.rank priority of concepts;converse withstakeholders to improvefitPhase III Evaluatestakeholder preferences:e.g. compare withalternatives and getquantitative feedback,card sorting ofstatements to obtainviews

Multi-stakeholder driven[44, 45]Focuses on what usersand practitioners actuallydo, not simply on whatthey say they do [44, 45]Shifts focus fromcontent of interventionto delivery in context sohelps to overcomebarriers toimplementation in thereal world [44, 45]Uses prototypes to getspecific rather thangeneric feedback [44,45]Focus is on utility, fitand engagement of keyusers of the intervention[44, 45]

Although there is a bookas well as journal articles,more details could begiven about how toachieve each action(INDEX)

Human-centreddesign [33]

Study people and taketheir needs and interestsinto account so thattechnology andappliances meet theneeds of people

Design of machines,appliances,technology foreveryday useNot health

Four activities areproposed, workingwithin a multidisciplinaryteam:1. Observing—Philosophyof early focus on

The focus on thestarting point of theprocess, and not closingdown questioning andideas too early areimportant actions not

Working within time,budget and otherconstraints [33]





Limitations

including that it isenjoyable and useable

observing the target usersand tasks rather thanasking users what theywant. Good designers donot start by trying tosolve the proposedproblem but by trying tounderstand what the realissues are.2. Ideation—Consider awide range of potentialsolutions and be creative3. Prototyping—buildquick rough prototypesto continue tounderstand the problem4. Testing andundertaking rapid testingof ideas/prototypes withthe target population inreal circumstances andmodifying approachafter each iterationThroughout, considerwider issues such as thecost of the object orstigmaAttached to using it

articulated well in otherapproaches (INDEX)

3. Theory andevidence based

MRC Frameworkfor developingand evaluatinginterventions[7, 15]

Spending timedeveloping interventionssystematically based onevidence and theoryproduces interventionswhich have a reasonablechance of having aworthwhile effect

Complexinterventions inhealth care, publichealth and socialpolicy

Three functions:1. Identifying theevidence base2. Identifying/developingtheory3. Modelling process andoutcomesQuestions are alsoidentified for researchersto ask themselves, suchas ‘Have you used thistheory systematically todevelop theintervention?’ and ‘Canyou describe theintervention fully, so thatit can be implementedproperly for thepurposes of yourevaluation, andreplicated by others?’

Not prescriptive [7]Well cited and used ingrant proposals [58]Used by manyresearchers in primaryresearch (INDEX)

Little detail [28, 47],INDEXIssues were underintense developmentand debate at time ofwriting guidance [7]Lacks attention tocomplexity science [58]

BehaviourChange Wheel(also actionby action approach)[26]

Comprehensive andsystematic approach,encouraging designersto consider the fullrange of optionsthrough systematicevaluation of theory andevidence

Behaviour changeinterventions inhealth and can beused in othersettings

Eights steps in threestages:1. Understanding thebehaviouri. Define the problem inbehavioural termsii. Select the behavioursyou are trying to changeiii. Specify the targetbehaviour, i.e. whoneeds to do whatdifferently and wheniv. Identify what willbring about the desiredbehaviour change usingCOM-B or TheoreticalDomains Framework2. Identify interventionoptions that will bringabout changei. Identify interventionfunctions

As well as aidingintervention design itimproves evaluation andtheory development byhelping to understandwhy interventions havefailed or how they haveworked [26]Explicitly draws attentionto the different levels atwhich an interventionmay need to work [26]Clear and detailedexplanation of eachaction with multipleexamples ([32], INDEX)Well known [32]Popular in that used bymany researchers inprimary research (INDEX)

Acknowledges thatjudgements are requiredwhere there is noevidence but does notsay who should beinvolved in making thesejudgements e.g.stakeholder groups(INDEX)Although reference ismade to working withstakeholders, theemphasis is onbehaviour change(INDEX)Needs more emphasison the target populationbeing involved inprocess [56]Requires substantialknowledge ofpsychological processes





Limitations

ii. Identify policycategories3. Identify content andimplementation optionsi. Identify behaviourchange techniques fromlist of 93, e.g. goalsettingii. Identify mode ofdelivery

[32]

Interventionmapping [27]

A systematic andthorough approachusing theory andevidence will produce aneffective intervention

Health promotionPublic healthComplex problems

Addresses planning,implementation andevaluation.6 steps:1. Undertake a needsassessment to develop alogic model of theproblem2. Produce a logic modelof the change processthat leads to outcomes3. Design the scope,sequence, methods andpractical applications ofthe program4. Produce the programincluding the materials5. Plan implementationand maintenance of theprogram6. Develop an evaluationplanUsing with acommunity-based par-ticipatory approach mayhelp external validity

Extremely rigorous andelaborate approach tointerventiondevelopment ([28],INDEX)Used by manyresearchers [32] andcites a long list ofpublished interventionsdeveloped with thisapproach (INDEX—seep34–38 of book)Addressesenvironmental as well aspersonal factorsaffecting the problem[32]

Highly technical,prescriptive, can requireyears to implement, anddifficult to operationalise[28]Does not cover the fullrange of interventionoptions available [26]So comprehensive that itrequires time resourcesthat make it unfeasiblefor use by manydevelopers [32]

Matrix AssistingPractitioner’sInterventionPlanning Tool(MAP-IT) [32]

Making the use oftheoretical knowledgeand empirical evidenceeasy can helppractitioners to developeffective interventions atlow cost

Health promotionBehaviour changecomplex healthinterventions

A matrix is determinedby a small group ofexpert researchersfocused on a specificbehaviour change for aspecific age group, e.g.promoting physicalactivity in older adults.The experts create amatrix of personal andenvironmentalmechanisms thatpromote positivebehaviour, relevanttheories and functions ofan intervention thatcould address eachmechanism. This matrixcan then be used bypractitioners to developa theory-driven andevidence-basedintervention

It undertakes one part ofinterventiondevelopment forbehaviour change sothat developers do nothave to understandpsychological theory indepth (INDEX)Links scientific researchwith practical real worldapplications [32]Offers a feasible and lowcost approach forpractitioners developinginterventions [32]Synthesises concepts inother well-known ap-proaches [32]

One matrix is presentedhere. Matrices need tobe produced for otherconditions/risk factors ina variety of age groups[32]It is insufficient becauseit does not take thecontext in which theintervention will be usedinto account [32]It facilitates one part ofintervention mappingrather than offering a fullapproach to interventiondevelopment (INDEX)

NormalisationProcess Theory(NPT) [46]

Using theory aboutnormalisinginterventions in routinepractice can helpdevelop and evaluateinterventions that will beimplemented in the realworld if found to beeffective

Complexinterventions inhealth and healthcare

The components of thetheory can help to1. Describe the contextin which the proposedintervention will beimplemented2. Define theintervention usingliterature reviews,observation, interviewsand surveys

Focuses on widersystem issues andinteractions betweendifferent groups of staffand patients, addressingboth individual andorganisational levelfactors [46]Addresses a neglectedaspect of interventiondevelopment (INDEX)

Focuses on one aspectof interventiondevelopment (INDEX)No detail about how todevelop interventions(INDEX)





Limitations

TheoreticalDomainsFramework(TDF) [47]

Using a theoreticalframework in asystematic way todevelop an interventionwill help to makehypothesisedmechanisms of changeexplicit and changeclinical practice

ComplexinterventionsClinical behaviourchangeImplementationinterventions to getevidence intopracticeQualityimprovement

A four-step systematicmethod based on guid-ing questions:1. Who needs to dowhat, differently?2. Which barriers andenablers need to beaddressed (using atheoretical framework)?3. Which componentscould overcomemodifiable barriers andenhance enablers?4. How can behaviourchange be measuredand understood?

A conceptual aid andnot a rigid prescription[47]Uses theory, evidenceand mixed methodsresearch [47]Using a broadly basedtheoretical frameworkfor behaviour change isbetter than using asingle theory [47]

Requires considerabletime and resources butspending this time andresource may be a goodinvestment [47]No detail about to howto undertake each action(INDEX)

4.Implementation-based

Reach,Effectiveness,Adoption,Implementation,Maintenance [48]

To encourageintervention plannersand other stakeholdersto pay more attention toexternal validity toimprove the sustainableadoption andimplementation ofeffective interventionsTo help planinterventions andimprove their chances ofworking in ‘real-world’settings.To facilitate translation ofresearch to practice

Health behaviourinterventions

The RE-AIM PlanningTool [48] is a series ofquestions which serve asa checklist for key issuesto consider when plan-ning an intervention. Thequestions are within fivegroups:1. Planning to improvereach to the targetpopulation2. Planning foreffectiveness3. Planning to improveadoption by target staff,settings, or institutions4. Planning to improveimplementation5. Planning to improvemaintenance ofintervention effects inindividuals and settingsover time

The approach has beenused to evaluate andreport a wide range ofinterventions [48]The emphasis ondeveloping interventionsthat will be used in thereal world if effective iscomplementary to someexisting approaches tointerventiondevelopment (INDEX)

RE-AIM [48] wasoriginally developed as aframework for consistentreporting of researchresults and then as aframework for evaluatinginterventions. As such,there is little detail abouthow to developinterventions (INDEX)

5. Efficiency-based

MultiphaseOptimizationStrategy (MOST)[49]

Conceptually rooted inengineering, MOSTemphasises efficiencyand careful managementof resources to moveintervention scienceforward systematicallyRandomisedexperimental approachesto optimisation leads tomore potentinterventions

Multicomponentbehaviouralinterventions inpublic health

There are three phases:1. Preparation:information from sourcessuch as behaviouraltheory, scientificliterature and secondaryanalyses of existing datais used to form the basisof a theoretical model.2. Optimisation:randomised experimentto test the effectivenessof different components.Fractional factorialexperiments (see below)sequential multiple-assignment randomisedtrials (SMARTs) or micro-randomised trials (seebelow) may be usedhere.3. Evaluation: standardRCT.A continuous cycle ofoptimisation andevaluation can occur

A number of projectsusing MOST have beenfunded by nationalfunding agencies [49]

Focuses on a narrowaspect of interventiondevelopment, occurringafter the components ofthe intervention havebeen assembled ordesigned (INDEX)

Multi-level andfractional factorialexperiments [50, 51]

Simultaneous screeningof candidatecomponents of anintervention to test foractive components offers

Multi componentinterventions withbehavioural,delivery orimplementation

Conduct a ‘screeningexperiment’ todetermine whichcomponents go forwardto experimental

Superior to mediationalanalyses from first RCTfollowed by second RCT[50]

Lack of statistical powerto do this at thedevelopment phase(INDEX)Focuses on a narrow





Limitations

an efficient way ofoptimising interventions

factors and wherethere is clustering

evaluation. Starts with anumber of potentialcomponents andremoves the least activeones. Uses fractionalfactorial design to screenout inactive componentsrather than evaluate theutility of a combinationof components over asingle component.Focuses on main effectsand a few anticipatedtwo-way interactions

aspect of interventiondevelopment, occurringafter the components ofthe intervention havebeen assembled ordesigned (INDEX)

Micro-randomisedtrials [52]

Delivering the rightinterventioncomponents at the righttimes and locations canoptimise support tochange individuals’health behaviours

‘Just in timeadaptiveinterventions’(mobile healthtechnologies)Behaviour change

Multiple components arerandomised at differentdecision points for anindividual. An individualmay be randomisedhundreds of times overweeks or months.Intermediate outcomescan be measured ratherthan primary outcomes

Only suitable for sometypes of interventionwhere participants areprompted to dosomething, whereevents are common andwhere measurement ofintermediate outcome islow burden [52]Focuses on a narrowaspect of interventiondevelopment, occurringafter the components ofthe intervention havebeen assembled ordesigned (INDEX)

6. Stepped/phased

Six essentialActions forQualityInterventionDevelopment(6SQuID) [28]

To guide researchersPractical, logical,evidence basedapproach to maximiseeffectiveness ofinterventionsTo reduce waste ofpublic money by notevaluating uselessinterventions

Public health butauthors say widerrelevance

1. Define andunderstand problem andits causes2. Identify which causalor contextual factors aremodifiable, and whichhave the greatest scopefor change3. Identify how to bringabout change (themechanisms of action)4. Identify how to deliverthe mechanisms ofchange5. Test and refine theintervention on a smallscale6. Collect enoughinformation abouteffectiveness to proceedto full evaluation

Systematic, logical andevidenced to maximiselikely effectiveness [28]Practical guidancewhere none exists [28]Attention to both earlyand later stages of thedevelopment process(INDEX)Based on experience ofdevelopment andevaluation ofinterventions (INDEX)

Offers an overview ratherthan detail (INDEX)Although authorsrecommend taking someof the actions withinvolvement fromstakeholders, and usingqualitative research atlater stages, littleattention needed toinvolvement of thosereceiving and deliveringthe intervention (INDEX)

Five actionmodel ininterventionresearch fordesigning anddevelopinginterventions[24, 29]

A systematic process ofdeveloping a manualleads to interventionsthat change practiceA detailed manual allowreplication of effectiveinterventions

Social workSocial and publichealth programsBased ondeveloping andtestinginterventions inchild development

The focus is on creatingthe intervention andthen refining it duringevaluation There are fivesteps:1. Develop both problemtheory and programtheory: specify theproblem, the rationalefor the intervention andthe theory of change2. Design interventionmaterials to articulatestrategies for changingmalleable mediators.Develop first draft ofmanual specifying theformat of manual(content, order of

Specifies link betweenthe problem theory andthe intervention content[29]Specifies process ofdeveloping treatmentmanuals [29]

The five actions coverevaluation as well asdevelopment so there isnot as much detailabout the developmentstage as in otherapproaches (INDEX)Although practitionersare considered early inthe process, the targetpopulation is consideredlate in the process ofdevelopment (INDEX)





Limitations

content and whodelivers it). Revisions andadaptations to themanual occurthroughout the furtheractions.3. Refine and confirmprogram components inefficacy tests. Submitmanual for review byrelevant stakeholdersincluding targetpopulation and thosedelivering theintervention. Undertakemixed methodsfeasibility testing.4. Test effectiveness in avarietyof practice settings5. Disseminate programfindings and materials

Obesity-RelatedBehaviouralIntervention Trials(ORBIT) [34]

A systematic, progressiveframework for translatingbasic behaviouralscience into treatmentsthat address clinicalproblems in a way thatstrengthens thetreatments andencourages rigorousevaluation

ClinicalBehaviouraltreatments forpreventing andtreating chronicdiseases

Flexible and progressiveprocess making use ofiterative refinement andoptimisation. The fivesteps are:1. Identification of asignificant clinicalquestion2. Phase 1a Design:Develop a hypothesisedpathway from behaviourtreatment to a solutionfor the clinical problem3. Phase 1b Refine:Optimise content anddelivery of anintervention, and tailorto sub-groups4. Phase IIa Proof ofconcept: Whentreatment manual isavailable, undertakestudy on small numbersto see if it merits morerigorous and costlytesting5. Phase IIb Pilot testing:Look for benefitsachieved over and abovea control group orconsider the feasibility ofa full evaluation

Clinically relevant anduses language fromdrug development toappeal to medicalstakeholders [34]Constructed for use witha broad number ofchronic diseases ratherthan a single category ofdisease [34]Details milestonesneeded at the end ofone phase prior tomoving on to nextphase (INDEX)

Takes a similar approachto MRC Guidance byusing the phases of drugtrials in an iterativephased approach. Onlyfocuses on the firstphases of drug trials andalthough there is moredetail aboutdevelopment than theMRC guidance, there isstill a lack of detailcompared with otherapproaches (INDEX)

7. Intervention-specific

Digital: IDEAS(Integrate, Design,Assess, and Share)Framework fordigital interventionsfor behaviourchange [48]

Guiding interventiondevelopment using thebest combination ofapproaches helps todeliver effective digitalinterventions that canchange behaviourNeed a combination ofbehavioural theory anduser-centred designthinking to develop ef-fective interventions.These must be evaluatedand disseminated tomaximise benefit

DigitalBehaviour change

Covers development andevaluation. Ten phases infour stages1. Integrate insights fromusers and theoryi. Empathise with targetusersii. Specify targetbehaviouriii. Ground in behaviouraltheory2. Design iteratively andrapidly with usersiv. Ideateimplementationstrategiesv. Produce prototype

Offers action by actionguide about combiningbehaviour theory anddesign thinking [48]Strikes a balancebetween offeringsufficient detail withoutbeing overly prescriptive[48]

Less experienced usersmay find it difficult toapply [48]There may bedisagreements amongstteam members that arechallenging to manage[48]





Limitations

vi. Obtain user feedbackvii. Create a product3. Assess rigorouslyviii. Pilot test to assesspotential efficacy andusabilityix. Evaluate in RCT4. Sharex. Share intervention andresults

Digital—practicaladvice forinternet-basedhealthinterventions[53]

Concrete examples fromexperience of digitalinterventiondevelopment cancomplement bestpractices guidance

Online healthinterventionsPublic Health

Based on the views ofresearchers andpractitioners:1. Hire the right researchteam, e.g. includecomputer scienceexperts2. Know the needs ofthe target population2. Plan the processbefore engaging a webdesigner3. Recognise thatdifferent stakeholdershave different values andlanguage e.g. researchersand web designers4. Develop a detailedcontract5. Document alldecisions6.Use a contentmanagement system7. Allow extra time fortesting and refining

Based on views ofresearchers withexperience and offerscomplementaryknowledge ofinterventiondevelopment to existingpublished sources [53]

The focus is largely onhow to work withcommercial webdesigners in the contextof a digital intervention(INDEX)

Web-baseddecision supporttools for patients[31]

A clear projectmanagement andeditorial process willhelp to balance differentpriorities of variety ofstakeholders [31]Need close consultationwith target users anditerative developmentprocess to developaccessible and usefulintervention [31]

Decision aidsavailable in web-based versions

A process map fordeveloping decision aidsaddressing two areas:First, contentspecification bycombining scientificevidence and patientperspectives. Secondcreative design to tailorit to specific audiencesby consideringpresentation ofinformation, help forpatients to assess howthey feel about futureevents and allowpatients to formulate apreferenceFive groups areestablished: a projectmanagement group of3–4 people to drive theprocess; an advisorygroup of 6–10stakeholders who advisebut do not have editorialrights; a virtual scientificreference group ofexperts to reviewevidence synthesis andthe evolving tool; atechnical productiongroup which will createand host the website;and stakeholderconsultations with a

Use of creative designand consultation as wellas scientific evidence[31]Close liaison with targetusers [31]Iterative method ofrefinement [31]

Time consuming [31]One action dependenton earlier actions so canbe delays [31]Can be disagreementsbetween experts, andbetween healthprofessionals andpatients [31]





Limitations

series of prototypesincluding patientsundergoing the decisionand practitioners whointeract with patientsOverlapping steps are:1. Identify patients’needs using qualitativeresearch2. Evidence synthesis3. Consensus onevidence4. Construct storyboard5. Undertake sandpittesting with experts6. Undertake usabilitytesting7. Undertake field testingwith real patients

Patient decisionaids [54]

Systematic andtransparent process ofdevelopment allowsusers to check validityand reduce chance ofcausing harm andincrease chance ofbenefit. Explicit thatthere is no hardevidence to support thisrationale

Decision support Based on a review ofdifferent approaches todeveloping decision aids,core features commonto all are:1. Scoping and design2. Development of aprototype3. Iterative ‘alpha’ testingby patients, cliniciansand other stakeholdersinvolved in thedevelopment4. Iterative ‘beta’ testingin real-life contexts withpatients and cliniciansnot involved in thedevelopment5. Production of finalversionThe process is overseenby a multi-stakeholdergroup

More comprehensivethan previous guides[54]

Uncertainty remainsabout how best toaddress the individualelements of the guide[54]Lack of detail about howto undertake differentactions (INDEX)

Group interventions[55]

More systematicapproach to designinginterventions

Healthimprovementinterventions orbehaviour changeinterventionsoccurring in agroup setting inpublic health andprimary care

Interventions arecomplex adaptive socialprocesses withinteractions between thegroup leader,participants, and thewider community andenvironment. Whendesigning themconsider:1. What the interventionis and the quantitydelivered2. How someonebecomes a groupmember3. The social andbehavioural theories thatinform the intervention4. How the groupinfluences members’attitudes, beliefs andbehaviours. Existingtheories may inform this,e.g. social support theory5. The intendedoutcomes6. Who should be thetarget population

Fills a gap in theevidence base [55]Can be used inconjunction withanother approach whendelivery to groups isrequired (INDEX)

Framework also coversevaluation so there is alack of detail aboutdevelopment (INDEX)Details issues to thinkabout rather than howto develop theintervention (INDEX)


designs for optimising the intervention. A number of ap-proaches recommend iterative processes, as new insightsemerge and changes are made to the intervention.The Documenting domain involves writing a set of in-

structions so others can use the intervention. The com-pilation of this document or manual is likely to startearly in the process and be finalised towards the end ofthe development process. The document or manual, likethe intervention, is likely to undergo multiple iterationsduring the development process and be further refinedafter any formal pilot or evaluation. The description ofthe intervention in any document can follow guidelinesfor reporting interventions: why, what, who provides,how, where, when and how much, and tailoring [35].Some approaches recommend planning for an even-

tual evaluation from the beginning of the developmentprocess [24, 27].

Methods and activities used in the development processAuthors of different approaches propose using differentresearch methods and activities to deliver each action(Tables 3, 4 and 5). Methods and activities include usingnon-participant observation to understand the problemand the context in which the intervention will operate,playing games to generate ideas for the content of theintervention and using ‘think aloud’ methods to

understand the usability, acceptability or feasibility ofearly versions of the intervention.

Strengths and limitations of approachesThe strengths of each approach, reported largely by itsauthors, included those related to processes (that theyare systematic, practical, clear or detailed) and outcomes(that they have been used a number of times, or there isevidence that they have produced effective interven-tions). Authors did not formally compare their approachwith others. Limitations, again reported largely by au-thors, included that some approaches lacked detail, weretime consuming to undertake, required expertise or didnot address how to deal with conflicting opinions fromdiverse stakeholders. These strengths and limitations arefurther detailed below.

ComprehensivenessSome of the approaches were more comprehensive thanothers in addressing a higher proportion of the 18 ac-tions (Table 6). More comprehensive approaches in-cluded Intervention Mapping and the Behaviour ChangeWheel. Some offered more detail about how to under-take specific actions because they were published asbooks or included a lot of examples. These includedIntervention Mapping, the Behaviour Change Wheel and




Limitations

8. Combination Participatory ActionResearch processbased on theorieson BehaviourChange andPersuasivetechnology (PAR-BCP) [56]

Aids the integration oftheories into aparticipatory actionresearch design processbecause behaviour ishard to change

Behaviour changesystems for healthpromotion (possiblyin digital health)

Combines theory fromtwo fields (behaviourchange and persuasivetechnology) with aparticipatory actionresearch methodology. Achecklist includes1. Understand anddefine the behaviour totarget2. Understand the targetgroup’s experiences andattitudes towards thebehaviour andintervention3. Consider ease of useof intervention4. Understand what kindof proactive feedback isneeded to changebehaviour5. Understand how tovisualise progress6. Explore what aboutthe patient-health pro-fessional relationshipbuilds trust7. Describe how socialinteractions can promotebehaviour change8. Evaluate prototypes

Brings together twocategories of approachto interventiondevelopment:partnership and theory-based (INDEX)

No detail on how toundertake actions(INDEX)Although the label‘participatory actionresearch’ is used, someexamples describe atarget populationcentred approach(INDEX)

aThese actions are summaries and readers are advised that source documents should be read to understand the detail of each approach


Table 3 Synthesis of actions in conception and planning (based on all approaches in taxonomy)

Domain Action Methods

1. Conception 1. Identify that there is aproblem in need of a newintervention [28, 29, 33, 34]

Authors of stepped or phased approaches to intervention developmentstart by describing how a problem has been identified. The existence ofa problem may be identified from published evidence synthesis, clinicalpractice, political strategy or needs assessment [28]. Alternatively,researchers or practitioners may have worked in a field for many yearsand identified the need for a new intervention [29]. In a clinical setting,the clinical significance of the problem, and the ability to make aclinically significant difference, is identified as the driver for selection ofproblems in need of a new intervention [34].

2. Planning 2. Establish a group or set ofgroups to guide the developmentprocess, thinking about engagementof relevant stakeholders such as thepublic, patients, practitioners and policymakers [27, 31, 34, 39–41, 53, 54]

Authors of a range of categories of intervention development explicitlyconsider the number, membership and role of groups that need to beestablished and run throughout the whole development process. Someauthors recommend that a group is established that has ‘editorial rights’(that is, makes final decisions about the intervention) and other groupsare established that may deliver any technical expertise needed or offeradvice and expertise for decision-making [31]. The ‘editorial rights’group—sometimes called ‘the development team’—includes the devel-opers and, in some approaches, includes members of the target popula-tion at which the intervention is aimed and practitioners likely to deliverit. Authors of user-centred approaches recommend including a variety ofdisciplines and expertise in this development team to generateinnovation [33, 43]. Authors of stepped or phased approaches also rec-ommend diverse membership to facilitate the development process, e.g.include people with computer science skills when designing digital inter-ventions [34, 53].In partnership approaches, the development team includes a diverserange of stakeholders, particularly members of the target population,who are equal partners with other team members, that is, have editorialrights [39, 40]. Those leading the intervention development will makeefforts to encourage engagement of members of the target population,especially of hard-to-reach groups, develop inclusive communication pro-cesses for the group, and consider the assets (knowledge, experience,skills and abilities, influence and connections) available within the group[40]. This focus on bringing a variety of stakeholders together, and collab-orative working with the target population and those who will deliverthe intervention, is not unique to partnership approaches. Authors ofsome theory and evidence based approaches value this, working with a‘planning group’ throughout the process, and seeking consensus afteropen discussion of diverse views [27]. Membership of these groups maychange over time as the intervention, its target population, and who willdeliver it become clear [24]. However, a unique aspect of partnership ap-proaches is that members of the target population have decision-makingrights throughout the development process.

3. Understand the problems or issuesto be addressed

Different authors address this action in different ways (see below). Forpartnership and target population-based approaches the focus is on in-depth understanding of the target population and the context in whichthe intervention will be delivered. For theory and evidence-based ap-proaches this understanding is gained from theory and published re-search. Some approaches include both of these strategies but may placedifferent weights on them. There are five sub-actions (i)-(v).

(i) Understand the experiences, perspectives andpsycho-social context of the potential target populationThe target population may be clients, patients, staff ora combination of these. This can involve identifyingthe priorities and needs of the potential target population,what matters most to people rather than what is thematter with them, why people behave as they do andunderstanding the lived experience of the potentialtarget population [17, 25, 27, 30, 33, 34, 39–42, 44,45, 53, 54, 56]

Some authors highlight this as the first action in the process and onethat shapes the whole process [30, 33]. It is central to partnership andtarget population centred approaches where understanding the livedexperiences and needs of the target population is the basis of theintervention. Secondary and primary qualitative research isrecommended: synthesis of qualitative research; iterative qualitativeresearch using diverse samples and open questions to explore people’sexperiences and needs; use of patients’ narratives or archives of patientexperiences and observation; consultation with stakeholders; and use ofpatient and public involvement [17, 41, 42]. Use of observation or‘shadowing’ patients and families is recommended as well as obtainingthe views of the target population because people may not be able toarticulate the problem fully [33, 40]. Theory and evidence-based ap-proaches, and stepped or phased approaches, also make use of qualita-tive research with the target population, including observation [27, 34].

(ii) Assess the causes of the problems Authors of a range of approaches recommend the use of the evidence


Table 3 Synthesis of actions in conception and planning (based on all approaches in taxonomy) (Continued)


This will include the determinants of these causes,influences on the problems, the size of problemsand who will benefit most and least from anyintervention [24, 27–29, 49]

base through literature or systematic reviews [24, 29, 34]. Alternatives aredrawing a logic model of the problem or model of causal pathways [27,28] and creative approaches, such as group discussions, as a way ofdeveloping questions for research evidence reviews [27].

(iii) Describe and understand the wider context ofthe target population and the context in which theintervention will be implementedConsider context at different levels: macro, meso,micro. Consider this context throughout the process[7, 17, 26, 27, 33, 44–46, 53, 56]

This sub-action can be undertaken as part of the earlier sub-actions (i)and (ii) but some approaches emphasise the importance of understand-ing context and so it is described as a separate action here. Bartholomewspecifies the contexts of population, setting and community [27]. Again,the use of qualitative research, particularly observation, is recommended.The observation may be of service delivery where the intervention willoccur [41, 42] or of the target population in their real life context [33, 43–45]. Conducting an asset assessment, that is, determining the strengthsof the community in which an intervention will take place is useful for ahealth promotion intervention [27]. Some theories can help to under-stand important aspects of context for implementation of the interven-tion in the real world [46].

(iv) Identify evidence of effectiveness of interventionsfor these problems, or for similar interventions oncedecisions have been made about the intervention type,so do not reinvent the wheel.Understand why previous interventions failed socan learn from this [7, 17, 31]

A range of approaches recommend systematic reviews of quantitativeevidence of effectiveness of interventions to identify what has worked,and qualitative evidence to understand why interventions have workedor not [7, 17, 31].

(v) Understand wider stakeholders’ perspectives of theproblems and issues [24, 28, 29, 39, 40, 59]

Authors of partnership and stepped/phased approaches recommendworking with wider stakeholders such as policy makers, communityleaders or service providers to clarify and understand the problems. Thiscan involve using research methods to obtain their views, meetings tofacilitate communication, or equal partnership with stakeholders usingactivities to encourage active engagement in the context of partnershipapproaches. Wider stakeholders may already be fully engaged withinpartnership approaches or because they are members of groupsestablished in Action 2.

4. Make a decision about the specific problem orproblems that an intervention will address, andthe aims or goals for the intervention. This mayinvolve defining the behaviours to target [27, 56]

If a list of problems has been identified then decisions will need to bemade about which to prioritise and focus on [27, 56].

5. Identify possible ways of making changes toaddress the problems.This involves identifying what needsto change, how to bring about thischange and what might need to changeat individual, interpersonal, organisational,community or societal levels [7, 17, 26, 27,29, 30, 34, 48, 55]

This action is addressed differently depending on the category ofapproach, and aim and context, of the intervention. Interventions aimingto address behaviour change in public health specify this action in detail,recommending the creation of a ‘logic model for change’ showingmechanisms of change and causal relationships between theory andevidence-based change methods [27, 28]. The emphasis is on drawingon existing theory or theories, and the research evidence base, to link de-terminants of a problem and the objectives of the intervention [27]. Iden-tifying a variety of theories rather than a single theory, including theoriesrelevant to later parts of the development process, e.g. implementationtheory, is recommended [27] at this action.Other approaches offer less detail about how to do this but suggestdrawing a ‘conceptual map’ [26] or point out that it should be influencedby the earlier qualitative research with stakeholders, including the targetpopulation and those who will deliver intervention [30]. Qualitativeresearch can be used to ask why people would make any proposedchanges, how change should occur and barrier and facilitators tochange.

6. Specify who will change, how and when.Selections may depend on considerationof the likely impact of the change, howeasy it is to change, how influentialit is for the problem being addressed,and how easy it is to measure [26, 27, 47]

Authors of theory and evidence-based approaches detail this action,recommending using the combination of a theory or theoretical frame-work with data from multiple sources such as interviews, focus groups,questionnaires, direct observation, review of relevant documents, litera-ture and involvement of stakeholders such as staff or patients [26]. Theremay be a long list of issues to change and these will need to be priori-tised at this action [26].

7. Consider real-world issues about cost anddelivery of any intervention at this early stageto reduce the risk of implementation failure ata later stage [7, 24, 27, 29, 33, 44–46, 48]

Understanding the context (see Action 3.iii above) can help here. Authorsrecommend considering wider issues such as the cost of an interventionor the stigma attached to using it [33] or how it fits with currentexpectations of a professional group that would deliver it [24]. This is a


the Person-Based Approach. There appeared to be a ten-sion between offering detailed description and being tooprescriptive.

Utility or successSome authors reported how their approach had beenused to develop interventions that were found to be ef-fective. Such approaches included Experience BasedCo-design, Intervention Mapping, the Behaviour ChangeWheel and Person Based Approach.

Resource and expertiseSome authors explicitly addressed concerns about theresource needed to develop an intervention. Some hadundertaken a part of one approach that required expert-ise and constructed a new approach to reduce the costand knowledge required (MAP-IT). Some proposed thattheir approach could be completed with limited re-sources by reducing the length of time or resource spenton each action (Intervention Mapping). One proposedthat the time and resources allocated should be influ-enced by the importance of the problem being addressed(Behaviour Change Wheel).

Relationship between approachesMost approaches were proposed as independent ways ofdeveloping interventions. Some were proposed as com-plementary to existing approaches (Person-Based Ap-proach), as explicit combinations of existing approaches(PAR-BCP) or as ways of facilitating existing approaches(MAP-IT). Some included the same actions but were in

different categories of the taxonomy because of theweight placed on those actions in the stated rationale.

DiscussionSummary of findingsThis is the most comprehensive review and synthesis ofdiverse approaches to intervention development under-taken to date. Eight categories of approach were identi-fied and descriptive summaries of approaches withineach category offered. There is considerable overlap be-tween categories in terms of the actions required, al-though there are differences in the methods that authorsof different approaches use to address each action. Someapproaches are more comprehensive than others interms of addressing a wide range of actions, and someoffer more detailed accounts than others to help re-searchers develop their own interventions. Interventiondevelopment is a rapidly developing field and recent ad-ditions have been proposed as complementary to exist-ing approaches [17] or as enhancements [10].

How to use the taxonomy and synthesis of actionsThis overview presents a broad range of approaches andactions common to intervention development, to helpdevelopers to select the approach or actions to suit theircontext, values and needs. Intervention developersshould consider the following questions when decidinghow to develop an intervention:

1. What is the intention of the intervention? e.g.changing behaviour

Table 3 Synthesis of actions in conception and planning (based on all approaches in taxonomy) (Continued)


key action for implementation-based approaches. The authors of [48] rec-ommend consideration of the barriers to reaching the target population,how the intervention will function for different sub-groups, what percent-age of organisations would be willing to adopt the intervention whentested and ability to overcome any barriers [48]. This thinking and plan-ning occurs early in the process and can involve formative research withwider stakeholders. Authors of a range of other approaches propose thatimplementation is considered at this early stage [24, 29], with the use oftheory to facilitate understanding of this [46] and the need to keep im-plementation issues in mind throughout later processes [27]. Issues otherthan effectiveness and cost-effectiveness that are related to implementa-tion can be considered: affordability, practicability, acceptability, safetyand equity [26]. Authors of partnership approaches recommend bringingstaff and patients together to increase engagement and improve imple-mentation [42]. Authors of stepped or phased approaches recommendthat developers have to understand the real world of practice so theycan develop not only effective interventions but interventions that practi-tioners adopt [24].

8. Consider whether it is worthwhile continuingwith the process of developing an intervention [7, 48]

The cost of delivering an intervention may outweigh the benefits it canpotentially achieve. This issue is addressed in economic modellingundertaken alongside RCTs but can also be considered at the planningstep by modelling processes and outcomes to determine if it is worthdeveloping an intervention [7]. Alternatively, if solutions to barriers tofuture implementation in the real world (see Action 7) cannot be foundthen it might not be worth developing an intervention [48].


Table 4 Actions within designing and creating (based on all approaches in taxonomy)


3. Designing 9. Generate ideas about solutions, andcomponents and features of anintervention [7, 17, 25–27, 30, 32, 33, 39,40, 42]

Ways of generating ideas for the intervention differ based on the category of approachto intervention development:Work with stakeholders creativelyPartnership and target population-centred approaches recommendbringing together a number of groups (e.g. patients, service providersand product designers) to generate diverse ideas for solutions fromdifferent perspectives. This is the central tenet of a co-design approachwhere patients are equal partners in the whole process rather than simplyhaving their views sought [39–41]. Authors of partnership approachespropose that listening to all voices is important, that processes to ensurethat this is undertaken in a meaningful way may be needed [40] and thatactive engagement of diverse groups of stakeholders is ongoing throughoutthe whole process [25, 42, 44, 45]. Encouraging all members of the developmentteam to interact directly with members the target population can guide thedevelopment of solutions that are more relevant and acceptable to the targetpopulation [30]. Methods to engage stakeholders may involve the use ofgames/exercises/tasks to promote creativity [25, 30, 33, 40] and the iterativeuse of prototypes (see step 4).

Target population involvement in intervention development at this designdomain is essential for authors of a range of approaches [7, 17, 25, 27, 30, 31, 42]with a proposal to make this involvement short and creative for busy people [25].Starting with divergent thinking and moving to convergent thinking is proposedas a way of maximising the potential to identify the most powerful solutions [30].

Use theoryTheory and evidence-based approaches to intervention development recommendmapping behavioural determinants to behaviour change techniques. This is a keyfocus of the Behaviour Change Wheel, where lists of behaviour change techniquesare given so that developers can identify intervention functions such as educationor persuasion that can address the selected behaviours [26]. This is also a core actionin Intervention Mapping with the construction of matrices to facilitate this action [27].Matrices have been constructed by a group of experts for a specific behaviour changefor a specific age group so that developers who are not experts in psychological theorycan undertake this action [32]. Authors of some theory-based approaches advocate creativethinking as well as use of theory at this action [27]. A variety of theories rather than a singletheory may be considered because one theory cannot explain everything of relevance [27]

10. Re-visit decisions about where tointerveneThis can involve consideration of thedifferent levels at which to intervene, andthe wider system in which theintervention will operate [7, 17, 24, 26–29,34, 54–56]

Consideration of where to intervene starts earlier at Action 4 but at this point finaldecisions need to be made. The authors of some approaches propose that this will requireseveral team meetings but they are not always clear about who should be involved inthese meetings. The ‘planning group’ [27] or ‘editorial group’ [31] may do this. Decisionsare made about:• the scope of the intervention• the target population (this may be narrower or broader than at the Conception andPlanning steps)

• levels at which the intervention is aimed: individual, community or population• key features of the intervention (which may be components of other interventions)• the components that will address the change required• the amount of exposure needed to obtain effect

11. Make decisions about the content,format and delivery of the intervention[24, 26–28, 30, 31, 47, 48, 54, 55]

Ideas generated in Action 9 are prioritised for inclusion in the intervention. Decision-making can be guided by the involvement of stakeholders, and theory and evidence in-cluding theories on what motivates people to engage in processes as well as produce out-comes, and use of taxonomies of modes of delivering interventions and evidence ofeffectiveness of these modes. Spencer [40] recommends using small ‘action groups’ ofstakeholders who can use their relevant expertise to build the solution to the problemidentified. Feasibility, budget and time constraints can inform choices [27]. Authors of onlyone approach recommend a formal consensus exercise for decisions in the specific casewhere published research evidence is summarised within the intervention [31]. Issues iden-tified for consideration here, based on a range of different approaches, include:• what will be delivered (content)• who will deliver it• where it will be delivered• how it will be delivered (format)• how many times it will be delivered• the point in any treatment or illness trajectory it will be delivered• the order in which different parts of the intervention will be delivered• the time period over which it will be delivered• interactions between components


2. What is the context of the intervention? e.g. publichealth

3. What values inform the intervention development?e.g. working in partnership with the targetpopulation

4. What skills and experience does the team bring?5. Which approaches have resulted in interventions

shown to be effective?6. What resources are available for the intervention

development?

Some of these issues can be considered by readingTable 2 where the different approaches are summarisedalong with their contexts, strengths and limitations, andTable 6 which shows which approaches address the dif-ferent actions reported in this overview.The synthesis of actions from all the approaches

shows the range of issues to consider. It may not bepossible to undertake all these actions, and indeednot all of them may be necessary. By laying out therange of actions in this overview, researchers canconsider the possibilities available to them and whichapproaches address the actions they value for theirspecific context (see Table 6).

Placing the findings in the context of other researchThe approaches identified in other more narrowly fo-cused reviews of intervention development are includedin this overview, for example MRC Guidance, Interven-tion Mapping and multiphase optimization strategy(MOST) [3, 4]. One of the earlier reviews identifiedfour tasks of intervention development which are in-cluded in the actions identified in this overview: identi-fying barriers, selecting intervention components, usingtheory and engaging end-users [3]. A review of the suc-cess of different types of interventions in diabetes inchildhood assessed the scientific rigour of the develop-ment processes of each type of intervention [36]. Thechecklist for scientific rigour was based on the contentof one approach to intervention development, andreporting guidelines. Items in this checklist appear inthe actions in this overview, e.g. existence of a manualto guide procedures, and explicit statement of theoret-ical basis of the intervention. A recent paper was pub-lished on enriching the development phase of the MRCFramework by including steps from seven other ap-proaches [10]. Again, the steps identified are includedin this overview: problem identification, systematicidentification of evidence, identification or development

Table 4 Actions within designing and creating (based on all approaches in taxonomy) (Continued)


• how users will be recruited• the resources available for delivery• how implementers will be trained• the potential for harm• the meaning of fidelity

12. Design an implementation plan,thinking about who will adopt theintervention and maintain it [27, 48]

Consideration is given to implementation at the Planning domain (Action 7 earlier) but thisaction relates to establishing a formal implementation plan once the content, format anddelivery of the intervention is known. Some authors recommend that this plan is based onthe formative research undertaken earlier to understand barriers to implementation [48].Others recommend basing this plan on a combination of theory and evidence fromimplementation science, and participation of stakeholders, to promote the use of theintervention in the real world [27]

4. Creating 13. Make prototypes or mock-ups of theintervention, where relevant [17, 25, 27,30, 31, 33, 39, 40, 53, 54, 56, 59]

This action starts in the Design domain, and indeed is seen as an essential action in theDesign domain by authors of some approaches. It is identified as a separate domain herebecause it is identified as such a key part of the process of intervention development bysome authors. Testing prototypes can help developers to make decisions about thecontent, format and delivery of the intervention. It also continues into the Refining domainwhere refinements are made to prototypes as feasibility and acceptability is assessed.Authors of approaches to digital interventions recommend creating an early prototype ofany physical intervention to get feedback from the target population using think aloud,usability testing, interviews or focus groups [17, 30, 57]. The prototype can be rough, e.g.paper copies of what a digital application could look like, and can be changed rapidly afterfeedback from stakeholders. These prototypes can generate further ideas for differentprototypes as well as refine a current prototype. In some user-centred and digital ap-proaches, authors recommend producing multiple rough and cheap prototypes at first andthen reducing these down to a single prototype after stakeholder feedback [30, 33, 59].They do this rather than focus on a single prototype too soon and call them ‘build to think’prototypes [59].If commercial designers are involved in creating prototypes, then differences in languageand values between academia and commercial designers will need to be discussed, acontract established and decisions defined [53]


Table 5 Actions in refining, documenting and planning for future evaluation (based on all approaches in taxonomy)


5. Refining 14. Test on small samples for feasibility andacceptability and make changes to the interventionif possible [17, 24, 27, 28, 30, 31, 34, 53, 54, 57]

Authors of a range of approaches recommend iterative testingand formative evaluation for this action. Some recommendqualitative research with those receiving and delivering theintervention. For example, think aloud interviews with the targetpopulation as they use the intervention, videos of people usingthe intervention [17, 30, 31, 57] or asking users to keep a diary ofissues arising when using the intervention to prompt theirmemory during interviews. For some approaches the prototypediscussed in the Creating domain is used to obtain specificviews rather than general views on the intervention [31, 44, 45]and the use of observation moves beyond people’s views. Theprocess can start small, for example with only the developmentteam commenting on the first prototype and then widening thesample to members of the target population [31]. Quantitativeas well as qualitative research is recommended, for example pre-test post-test comparison looking for changes in some inter-mediate outcomes for a small number of the target populationusing the intervention [34]. The potential for unintended conse-quences can also be considered [48].

15. Test on a more diverse population, moving awayfrom the single setting where early development of theintervention took place and seeking a more diverse sample.This can involve asking questions such as ‘is it working asintended?’, ‘is it achieving short term goals?’,‘is it having serious adverse effects’? [17, 24, 25, 28, 31, 34, 54, 57]

The iterative approach used in Action 14 continues here bymaking changes to the intervention and continuing to usemixed methods to check if changes are working as planned onmore diverse samples. Authors of a range of approachesrecommend using pre-test post-test design, n-of-1 trials and ob-servation or video to consider acceptability and early feasibility.They also recommend using real members of the target popula-tion in a real-life environment to identify interactions and rela-tionships between different service providers and patients toiteratively modify the intervention. Groups of wider stakeholderscan review the intervention as it iterates [24, 54].

16. Optimise the intervention for efficiency prior to full RCT[34, 49–52]

Some approaches consider Actions 14 and 15 to be part of theprocess of optimisation of the intervention through the use ofmixed methods. Case series can be used to consider issues suchas dose, patterns of use over time, and safety [34]. Efficiency-based approaches and some stepped/phase approaches take amore quantitative approach: fractional factorial designs can beused to identify active components, interactions between com-ponents of an intervention, the doses that lead to best out-comes and tailoring to sub-groups [34, 50]. A review ofoptimisation of interventions has been published [4].

6. Documenting 17. Document the intervention, describing the intervention soothers can use it and offer instructions on how to trainpractitioners delivering the intervention and on how toimplement the intervention [7, 17, 24, 26, 29, 34, 48, 57]

This document is sometimes called a manual. The manual iswritten by the developers. Authors of some approachesrecommend that it undergoes external review by stakeholders,including the target population and those delivering theintervention, to make sure it is feasible for use in the real world[24, 26, 29].

7. Planning forfutureevaluation

18. Develop the objectives of the outcome and processevaluations.This includes determining how outcomes and mediators ofchange can be measured, developing measures, specifyingevaluation design, planning recruitment and consideringfeasibility of a full RCT [24, 27, 29, 31, 40, 47, 48]

Authors of some approaches recommend planning for arandomised study or experimental design with controls formeasuring effectiveness [29], whereas others recognise thatthere may not always be the resources to evaluate with a RCT sothe intervention may be implemented in practice andmonitored [31].Authors of some approaches recommend involving stakeholderssuch as funders, implementers and the target population in thisaction [27], particularly getting agreement with key stakeholdersabout how to define and measure success [48]. Partnershipapproaches recommend that this is undertaken withstakeholders [40].


of theory, determination of needs, examination ofcurrent practice or context and modelling process andexpected outcomes.

Strengths and limitationsThis is the first time such a broad and detailed re-view of approaches to intervention development hasbeen undertaken. The overview is timely given thecurrent interest in undertaking good intervention de-velopment. There were five limitations. First, therewas a subjective aspect to the selection of approachesthat is common to all qualitatively oriented reviews.Team discussion was used to address subjectivity anda description has been given of how decisions weremade by authors of this overview. It is likely that themajority—but not all—of the included approacheswould appear in a similar overview undertaken by adifferent team. Second, as planned, this is not an ex-haustive list of approaches and readers will be able toname other approaches. Third, the taxonomy wasbased on the rationale provided by authors of ap-proaches. This rationale was not always explicit, andwas dependent on the report given in the websites,books or articles included; authors of approaches maynot necessarily agree with the interpretation of theirrationale. Fourth, this is a rapidly developing fieldand new approaches were published after the end offormal data extraction and were not included [10,

37]. Finally, a comparison of the success of each ap-proach has not been undertaken (for example, thenumber of times effective interventions have been de-veloped using each approach), although authors’ ownreports of success have been included. Further re-search could synthesise evidence of effectiveness ofdifferent categories of approach, specific approachesor actions within approaches.

Research gapsAnother important part of the picture is to understandhow interventions are actually developed in practice. A re-view of primary research reporting this activity is beingundertaken as part of the wider INDEX study. Theintention is to compare the approaches reported here withpractice. It is already apparent that some approaches areused in primary studies but documentation describing theapproach in the context of intervention developmentcould not be found for inclusion in this overview, in par-ticular community-based participatory research. The useof patient and public involvement within intervention de-velopment is also absent from the approaches describedhere. This may be more prominent in the review of pri-mary research studies in the INDEX study noted above.

ConclusionsThis overview of approaches to intervention develop-ment can help researchers to understand the variety of

Table 6 Actions undertaken in intervention development within each approach (black dot in a cell indicates that an action labelled1–18 described in Tables 3, 4 and 5 is recommended by an approach)


approaches available and the range of possible actionsinvolved in intervention development, before undertak-ing any feasibility or piloting phase. Findings from thisoverview will contribute to future guidance on interven-tion development.

Additional file

Additional file 1: Search strategy. (DOCX 21 kb)

Abbreviations6SQUID: Six essential Steps for Quality Intervention Development;BCW: Behaviour Change Wheel; EBCD: Experience-based co-design;IDEAS: Integrate, Design, Assess and Share; IM: Intervention mapping;INDEX: IdentifyiNg and assessing different approaches to DEvelopingcompleX interventions; MAP-IT: Matrix Assisting Practitioner’s InterventionPlanning Tool; MOST: Multiphase optimization strategy; MRC: MedicalResearch Council; NPT: Normalisation process theory; ORBIT: Obesity-RelatedBehavioural Intervention Trials; PAR-BCP: Participatory Action Research basedon theories of Behaviour Change and Persuasive Technology; RE-AIM: Reach,Effectiveness, Adoption, Implementation, Maintenance; TDF: Theoreticaldomains framework; UK: United Kingdom

AcknowledgementsThe authors thanks Vera Fibisan for help with obtaining source materials andreferencing the manuscript.

FundingMedical Research Council MR/N015339/1. The funders had no role in thissystematic methods overview.

Availability of data and materialsAll available data are reported within the manuscript in the tables.

Authors’ contributionsAOC led the overview, undertook data extraction and synthesis and wrotethe first draft of the paper. KS undertook the searches. KS and LCparticipated in discussions with AOC about data extraction and synthesis. Allauthors discussed the issues addressed in this overview and commented onmultiple drafts of the paper, including the final draft.

Authors’ informationThe authors have experience of intervention development. PH and LY havewritten about approaches to intervention development. PH, LY, LC and ED havedeveloped or adapted interventions. AOC and KT have been co-applicants onstudies where others have developed or adapted interventions.

Ethics approval and consent to participateNot applicable because it is a review of research methods.

Consent for publicationNot applicable.

Competing interestsTwo authors of this article have produced approaches to interventiondevelopment included in this overview (PH and LY).

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Author details1Medical Care Research Unit, Health Services Research, School of Health andRelated Research (ScHARR), University of Sheffield, Regent Court, 30 RegentStreet, Sheffield S1 4DA, UK. 2NMAHP Research Unit, University of Stirling,Stirling FK9 4NF, UK. 3Population Health Sciences, Canynge Hall, 39 WhatleyRoad, University of Bristol, Bristol BS8 2PS, UK.

Received: 24 September 2018 Accepted: 24 February 2019

References1. Macleod MR, Michie S, Roberts I, Dirnagl U, Chalmers I, Ioannidis JP, et al.

Biomedical research: increasing value, reducing waste. Lancet. 2014;383(9912):101–4.

2. Hawe P. Lessons from complex interventions to improve health. Annu RevPublic Health. 2015;18(36):307–23.

3. Colquhoun HL, Squires JE, Kolehmainen N, Fraser C, Grimshaw JM. Methodsfor designing interventions to change healthcare professionals' behaviour: asystematic review. Implement Sci. 2017;12(1):30.

4. Levati S, Campbell P, Frost R, Dougall N, Wells M, Donaldson C, et al.Optimisation of complex health interventions prior to a randomised controlledtrial: a scoping review of strategies used. Pilot Feasibility Stud. 2016;2:17.

5. Redfern J, McKevitt C, Wolfe CDA. Development of complex interventions instroke care: a systematic review. Stroke. 2006;37(9):2410–9.

6. Netto G, Bhopal R, Lederle N, Khatoon J, Jackson A. How can healthpromotion interventions be adapted for minority ethnic communities? Fiveprinciples for guiding the development of behavioural interventions. HealthPromot Int. 2010;25(2):248–57.

7. Craig P, Dieppe P, Macintyre S, Michie S, Nazareth I, Petticrew M.Developing and evaluating complex interventions: new guidance. London:Medical Research Council; 2008.

8. Moore GF, Audrey S, Barker M, Bond L, Bonell C, Hardeman W, et al. Processevaluation of complex interventions: Medical Research Council guidance.BMJ. 2015;350:h1258.

9. Hallingberg B, Turley R, Segrott J, Wight D, Craig P, Moore L, et al.Exploratory studies to decide whether and how to proceed with full-scaleevaluations of public health interventions: a systematic review of guidance.Pilot Feasibility Stud. 2018;4:104.

10. Bleijenberg N, JMdM-v G, JCA T, RGA E, Sinob CG, Heima N, et al. Increasingvalue and reducing waste by optimizing the development of complexinterventions: Enriching the development phase of the Medical ResearchCouncil (MRC) Framework. Int J Nurs Stud. 2018;79:86–93.

11. INDEX. INDEX Study (IdentifyiNg and assessing different approaches toDEveloping compleX interventions) [Available from: https://www.sheffield.ac.uk/scharr/sections/hsr/mcru/indexstudy.

12. Hutton JL, Ashcroft R. What does "systematic" mean for reviews ofmethods? In: Black N, Brazier J, Fitzpatrick R, Revees B, editors. Healthservices research methods: a guide to best practice. London: BMJBooks; 1998. p. 249–55.

13. Edwards SJL, Lilford RJ, Kaiuka S. Different categories of systematic review inhealth services research. In: Black N, Brazier J, Fitzpatrick R, Revees B, editors.Health services research methods: a guide to best practice. London: BMJBooks; 1998. p. 255–60.

14. Gentles SJ, Charles C, Nicholas DB, Ploeg J, McKibbon KA. Reviewing theresearch methods literature: principles and strategies illustrated by a systematicoverview of sampling in qualitative research. Syst Rev. 2016;5:172.

15. Craig P, Dieppe P, Macintyre S, Michie S, Nazareth I, Petticrew M.Developing and evaluating complex interventions: the new MedicalResearch Council guidance. BMJ. 2008;337:a1655.

16. Hoddinott P. A new era for intervention development studies. PilotFeasibility Stud. 2015;1:36.

17. Yardley L, Morrison L, Bradbury K, Muller I. The person-based approach tointervention development: application to digital health-related behaviorchange interventions. J Med Internet Res. 2015;17(1):e30.

18. Greenhalgh T, Peacock R. Effectiveness and efficiency of search methods insystematic reviews of complex evidence: audit of primary sources. BMJ.2005;331(7524):1064–5.

19. Curran GM, Mukherjee S, Allee E, Owen RR. A process for developing animplementation intervention: QUERI series. Implement Sci. 2008;3:17.

20. De Silva MJ, Breuer E, Lee L, Asher L, Chowdhary N, Lund C, et al. Theory ofchange: a theory-driven approach to enhance the Medical ResearchCouncil's framework for complex interventions. Trials. 2014;15:267.

21. Green L, Kreuter MK. Health program planning: an educational andecological approach. New York: McGraw Hill; 2005.

22. van Meijel B, Gamel C, van Swieten-Duijfjes B, Grypdonck M. Thedevelopment of evidence-based nursing interventions: methodologicalconsiderations. J Adv Nurs. 2004;48(1):84–92.


https://doi.org/10.1186/s40814-019-0425-6

https://www.sheffield.ac.uk/scharr/sections/hsr/mcru/indexstudy

https://www.sheffield.ac.uk/scharr/sections/hsr/mcru/indexstudy

23. Michie S, Yardley L, West R, Patric K, Greaves F. Developing and evaluatingdigital interventions to promote behavior change in health and health Care:recommendations resulting from an international workshop. J Med InternetRes. 2017;19(6):e232.

24. Fraser MW, Richman JM, Galinsky MJ, Day SH. Intervention research:developing social programs. New York: Oxford University Press; 2009.

25. Kushniruk A, Nøhr C. Participatory design, user involvement and health ITevaluation. Stud Health Technol Inform. 2016;222:139–51.

26. Michie S, Atkins L, West R. The behaviour change wheel: a guide todesigning interventions. London: Silverback Publishing; 2014.

27. Bartholomew Eldredge LK, Markham CM, Ruiter RA, Fernandez ME, Kok G,Parcel GS. Planning health promotion programs: an intervention mappingapproach, 4th edition. San Francisco: Jossey-Bass; 2016.

28. Wight D, Wimbush E, Jepson R, Doi L. Six steps in quality interventiondevelopment (6SQuID). J Epidemiol Community Health. 2015;70(5):520–5.

29. Fraser C, Galinsky MJ. Steps in intervention research: designing anddeveloping social programs. Res Soc Work Pract. 2010;20(5):459–66.

30. Mummah SA, Robinson TN, King AC, Gardner CD, Sutton S. IDEAS (integrate,design, assess, and share): a framework and toolkit of strategies for thedevelopment of more effective digital interventions to change healthbehavior. J Med Internet Res. 2016;18(12):e317.

31. Elwyn G, Kreuwel I, Duranda MA, Sivell S, Joseph-Williams N, Evans R, et al.How to develop web-based decision support interventions for patients: aprocess map. Patient Educ Couns. 2011;82:260–5.

32. Hansen S, Kanning M, Lauer R, Steinacker JM, Schlicht W. MAP-IT: a practicaltool for planning complex behavior modification interventions. HealthPromot Pract. 2017;18(5):696–705.

33. Norman DA. The design of everyday things, revised and expanded edition:MIT press; 2013.

34. Czajkowski SM, Powell LH, Adler N, Naar-King S, Reynolds KD, Hunter CM,et al. From ideas to efficacy: the ORBIT model for developing behavioraltreatments for chronic diseases. Health Psychol. 2015;34(10):971–82.

35. Hoffmann TC, Glasziou PP, Boutron I, Milne R, Perera R, Moher D, et al.Better reporting of interventions: template for intervention description andreplication (TIDieR) checklist and guide. BMJ. 2014;348:g1687.

36. Savage E, Farrell D, McManus V, Grey M. The science of interventiondevelopment for type 1 diabetes in childhood: systematic review. J AdvNurs. 2010;66(12):2604–19.

37. Hawkins J, Madden K, Fletcher A, Midgley L, Grant A, Cox G, et al.Development of a framework for the co-production and prototyping ofpublic health interventions. BMC Public Health. 2017;17(1):689.

38. Voorberg WH, Bekkers VJJM, Tummers LG. A systematic review of co-creation and co-production: embarking on the social innovation journey.Public Manag Rev. 2015;17(9):1333–57.

39. Bessant J, Maher L. Developing radical service innovations in healthcare—therole of design methods. Int J Innov Manag. 2009;13(4):555–68.

40. Spencer M, Dineen R, Phillips A. Co-producing services—co-creating health; 2013.41. Robert G. Participatory action research: using experience-based co-design

(EBCD) to improve health care services. In: Ziebland S, Coulter A, CalabreseJD, Locock L, editors. Understanding and using health experiences:improving patient care. Oxford: Oxford University Press; 2013. p. 138–49.

42. Locock L, Robert G, Boaz A, Vougioukalou S, Shuldham C, Fielden J, et al.Testing accelerated experience-based Co-design: a qualitative study of using anational archive of patient experience narrative interviews to promote rapidpatient-centred service improvement. Health Serv Deliv Res. 2014;2(4).

43. Erwin K. Communicating the new: methods to shape and accelerateinnovation. New Jersey: Wiley; 2013.

44. Erwin K, Krishnan JA. Redesigning healthcare to fit with people. BMJ. 2016;354:i4536.

45. Erwin K, Krishnan JA. Using design methods to provide the care that peoplewant and need. J Comparative Effective Res. 2016;5(1):13–5.

46. Murray E, Treweek S, Pope C, MacFarlane A, Ballini L, Dowrick C, et al.Normalisation process theory: a framework for developing, evaluating andimplementing complex interventions. BMC Med. 2010;8(1):1.

47. French SD, Green SE, O’Connor DA, McKenzie JE, Francis JJ, Michie S, et al.Developing theory-informed behaviour change interventions to implementevidence into practice: a systematic approach using the TheoreticalDomains Framework. Implement Sci. 2012;7(38)

48. Checklist for Study or Intervention Planning [Internet]. [cited 23.11.2017].Available from: http://re-aim.org.

49. Collins LM, Murphy SA, Strecher V. The multiphase optimization strategy(MOST) and the sequential multiple assignment randomized trial (SMART):new methods for more potent eHealth interventions. Am J Prev Med. 2007;32(5 Suppl):S112–8.

50. Chakraborty B, Collins LM, Strecher VJ, Murphy SA. DevelopingMulticomponent Interventions using Fractional Factorial Designs. Stat Med2009;28(21):2687–2708.

51. Dziak JJ, Nahum-Shani I, Collins LM. Multilevel factorial experiments fordeveloping behavioral interventions: power, sample size, and resourceconsiderations. Psychol Methods. 2012;17(2):153–75.

52. Klasnja P, Hekler EB, Shiffman S, Boruvka A, Almirall D, Tewari A, et al. Micro-randomized trials: an experimental design for developing just-in-timeadaptive interventions. Health Psychol. 2015;34S:1220–8.

53. Horvath KJ, Ecklund AM, Hunt SL, Nelson TF, Toomey TL. DevelopingInternet-based health interventions: a guide for public health researchersand practitioners. J Med Internet Res. 2015;17(1):e28.

54. Coulter A, Stilwell D, Kryworuchko J, Mullen PD, Ng CJ, Weijden T. Asystematic development process for patient decision aids. BMC Med InformDecis Mak. 2013; 13(Suppl 2):S2.

55. Hoddinott P, Allan K, Avenell A, Britten J. Group interventions to improvehealth outcomes: a framework for their design and delivery. BMC PublicHealth. 2010;10:800.

56. Janols R, Lindgren H. A method for co-designing theory-based behaviourchange systems for health promotion. Stud Health Technol Inform. 2017;235:368–72.

57. Yardley L, Ainsworth B, Arden-Close E, Muller I. The person-based approachto enhancing the acceptability and feasibility of interventions. PilotFeasibility Stud. 2015;1:37.

58. Craig P, Petticrew M. Developing and evaluating complex interventions:reflections on the 2008 MRC guidance. Int J Nurs Stud. 2013;50(5):585–7.

59. Erwin K. Communicating the new: methods to shape and accelerateinnovation. New Jersey: Wiley; 2015.


http://re-aim.org

taxonomy of approaches to developing …...health. these interventions have multiple interacting...

Documents