Specifying how intervention content is communicated: Development of a Style of Delivery Ontology

Background: Investigating and enhancing the effectiveness of behaviour change interventions requires detailed and consistent specification of all aspects of interventions. We need to understand not only their content, that is the specific techniques, but also the source, mode, schedule, and style in which this content is delivered. Delivery style refers to the manner by which content is communicated to intervention participants. This paper reports the development of an ontology for specifying the style of delivery of interventions that depend on communication. This forms part of the Behaviour Change Intervention Ontology, which aims to cover all aspects of behaviour change intervention scenarios. Methods: The Style of Delivery Ontology was developed following methods for ontology development used in the Human Behaviour-Change Project, with seven key steps: 1) defining the scope of the ontology, 2) identifying key entities and developing their preliminary definitions by reviewing 100 behaviour change intervention evaluation reports and existing classification systems, 3) refining the ontology by piloting the ontology through annotations of 100 reports, 4) stakeholder review by eight behavioural science and public health experts, 5) inter-rater reliability testing through annotating 100 reports using the ontology, 6) specifying ontological relationships between entities, and 7) disseminating and maintaining the ontology. Results: The resulting ontology is a five-level hierarchical structure comprising 145 unique entities relevant to style of delivery. Key areas include communication processes, communication styles, and attributes of objects used in communication processes. Inter-rater reliability for annotating intervention evaluation reports was α=0.77 (good) for those familiar with the ontology and α=0.62 (acceptable) for those unfamiliar with it. Conclusions: The Style of Delivery Ontology can be used for both annotating and describing behaviour change interventions in a consistent and coherent manner, thereby improving evidence comparison, synthesis, replication, and implementation of effective interventions.


Introduction
Behaviour change interventions vary considerably in the manner in which they are delivered.Specifying how an intervention is delivered is essential to understanding its effectiveness, its mechanisms of action, and reasons for variation in effectiveness across contexts (Dombrowski et al., 2016).To facilitate better replication and implementation of effective behaviour change interventions, we need to report them clearly and with a sufficiently granular level of specification (Michie et al., 2020;Michie et al., 2023).
The delivery of an intervention can include the following components (Michie et al., 2021): 1) mode of delivery, i.e. the medium through which an intervention is provided, for example printed materials (Dombrowski et al., 2016;Marques et al., 2021), 2) intervention source, i.e. who delivers the intervention, for example a health professional (Norris et al., 2021), 3) intervention schedule, i.e. the temporal organisation of how the intervention is provided, for example, the duration of parts of the intervention, and 4) style of delivery, i.e. the manner in which the intervention content, such as behaviour change techniques (Corker et al., 2023;Marques et al., 2023;Michie et al., 2013), is delivered, for example, cold and distant versus warm and accepting communication style (Dombrowski et al., 2016;Hagger & Hardcastle, 2014).We consider style of delivery in this paper.
The impact of intervention content on behaviour is likely to vary depending on how it is delivered to participants, either enhancing or undermining its effectiveness.For example, the behaviour change technique 'inform about health consequences' could be presented in an authoritarian formal manner, or in an informal conversational manner.Additionally, certain groups of people may prefer particular delivery styles due to differences in social or cultural norms or other factors.This likely contributes to differences in effectiveness of interventions.For interventions that depend on communication, style of delivery includes the type of language used and the nature of the interactions between two or more participants (Hagger & Hardcastle, 2014).A communication process can involve a variety of participants, such as humans interacting synchronously or asynchronously, a human talking to an artificial intelligence-based conversational agent (i.e., chatbots), or two computer systems exchanging information.
Style of delivery plays an important role in interventions targeting behaviours from a variety of domains including health, sustainability, and education.For example, motivational interviewing (MI) interventions use multiple techniques to increase motivation for behaviour change which are delivered using a client-centred interpersonal style (Miller & Rollnick, 2012).MI aims to elicit the client's own reasons for change, which requires the intervention source to adopt non-pressuring language and use a reflective, respectful, and empathic style (Hardcastle et al., 2017).In the medical field, extensive work has been conducted on training healthcare professionals to strengthen their communication skills to enhance patient outcomes.For example, the Roter Interaction Analysis System (Roter & Larson, 2002) provides a method to describe communication between healthcare professionals and patients.Studies using this method have found large variations in the communication profiles of healthcare professionals and that these variations may affect patient outcomes such as adherence to treatment (Pires & Cavaco, 2014).
Reporting guidelines aim to improve the quality of intervention reporting in terms of specificity, consistency, and comprehensiveness.However, reporting of delivery style components has tended to be neglected.Widely used existing guidelines such as the Template for Intervention Description and Replication (TIDieR) guidance (Hoffmann et al., 2014), and the CONsolidated Standards of Reporting Trials statement (CONSORT) (Schulz et al., 2010) and its extension for social and psychology interventions (CONSORT-SPI) (Montgomery et al., 2018) do not specify reporting the style in which intervention content is communicated.An attempt has been made to classify the delivery style of MI interventions (e.g.engaging techniques such as open-ended questions; focusing techniques such as permission to provide information and advice) (Hardcastle et al., 2017) however, there is no broader categorisation system.More work is needed to extensively specify style of delivery, in other words, the manner in which intervention content is communicated.
An ontology is one such method for specifying and classifying aspects of interventions (see Table 1 for glossary of bold, italicised ontology-related terms).Ontologies are standardised representational frameworks (i.e.classification systems) providing a set of terms for the consistent description of data and information across disciplinary boundaries (Arp et al., 2015).Ontologies represent information in the form of clearly expressed and unambiguous entities, labels and definitions corresponding to these entities, and the relationships between the entities (Arp et al., 2015;Hastings, 2017;Larsen et al., 2017).This provides formal specification and a 'controlled vocabulary' for a given discipline that can be understood across disciplines and topic domains (National Academies of Sciences, Engineering, and Medicine, 2022).Entities are also given unique identifiers, allowing them to be computer-readable.This can support the application of artificial intelligence and machine learning-based approaches in data extraction, evidence synthesis, and outcome prediction (Hastings, 2017).A further advantage of ontologies is that they are dynamic; they can and should be maintained and updated in response to new evidence emerging in the relevant discipline.
No ontology currently exists to specify and classify the full detail of style of delivery.A comprehensive Behaviour Change Intervention Ontology (BCIO) is being developed as part of the Human Behaviour-Change Project (Michie et al., 2017;Michie et al., 2020).The BCIO characterises behaviour change interventions, their contexts, and their evaluations.A benefit of the BCIO, compared to other

Term
Definition Source

Annotation
Process of coding, or tagging, selected parts of documents or other resources to identify the presence of ontology entities.

Annotation guidance manual
Written guidance on how to identify and tag pieces of text from intervention evaluation reports with specific codes relating to entities in the ontology, using for example EPPI-Reviewer software.

Inter-rater reliability
Statistical assessment of similarity and dissimilarity of coding between two or more coders.If inter-rater reliability is high this suggests that ontology entity definitions and labels are being interpreted similarly by the coders.

Interoperability
Two systems are interoperable to the extent that the data in each system can be used by the other system.Note: An ontology is interoperable with another ontology if it can be used together with the other ontology. http://www.obofoundry.org/principles/fp-010collaboration.html

Issue tracker
An online log for problems identified by users accessing and using an ontology.

Open Biological and Biomedical Ontology (OBO) Foundry
A collective of ontology developers that are committed to collaboration and adherence to shared principles.The mission of the OBO Foundry is to develop a family of interoperable ontologies that are both logically wellformed and scientifically accurate.

Ontology
A standardised representational framework providing a set of entities for the consistent description (or "annotation" or "tagging") of data and information across disciplinary and research community boundaries.

Parent class
A class within an ontology that is hierarchically related to one or more child classes (subclasses) such that all members of the child class are also members of the parent class, and all properties of the parent class are also properties of the child class.

Process
Something that takes place over time.Arp et al. (2015) Term Definition Source

Relationship
The manner in which two entities are connected or linked.A string of characters that unambiguously identifies an ontology or an individual entity within an ontology.

Versioning
Ontologies that have been released are expected to change over time as they are developed and refined, leading to a series of different files.Consumers of ontologies must be able to specify exactly which ontology files they used to encode their data or build their applications and be able to retrieve unaltered copies of those files in perpetuity.Versioning is one of the OBO Foundry principles. http://www.obofoundry.org/principles/fp-004-versioning.html

Web Ontology Language (OWL)
A formal language for describing ontologies.It provides methods to model entities of "things", how they relate to each other and the properties they have.OWL is designed to be interpreted by computer programs and is extensively used in the Semantic Web where rich knowledge about web documents and the relationships between them are represented using OWL syntax.
https://www.w3.org/TR/owl2-quick-reference/ classification systems such as taxonomies, is that aspects of a behaviour change intervention scenario can be linked, such as their content, delivery, mechanisms of action, target behaviours, engagement, and contextual features such as target population and setting (Michie et al., 2021).The BCIO consists of an upper level with 42 entities, one of which is behaviour change intervention style of delivery, specified as 'an attribute of a behaviour change intervention delivery that encompasses the characteristics of how a behaviour change intervention is communicated' (Michie et al., 2021).Figure 1 shows how style of delivery fits within the upper level of the BCIO.
To develop the Style of Delivery Ontology, we used the ontology development methodology specified for the Human Behaviour-Change Project (Wright et al., 2020).This method was designed to meet principles of good practice in ontology development, and is a tried and tested approach for developing the lower-level ontologies of the BCIO.

Aim
This study aimed to develop a Style of Delivery Ontology that can be reliably used to identify and specify the style in which behaviour change intervention content is delivered.

Ethics
Ethical approval was granted by University College London Research Ethics Committee (CEHP/2020/579).Informed participant consent was obtained in the expert stakeholder review step via an online Qualtrics survey.

Design
The Style of Delivery Ontology was developed in an iterative process of seven steps (Wright et al, 2020).Labels and definitions of each entity were written to be consistent with principles and guidelines for writing ontological definitions (Michie et al., 2019;Seppälä et al., 2017).
Step 1 -Development of the scope and definition of the Style of Delivery Ontology A preliminary definition for 'style of delivery' and the scope of the ontology was developed by reviewing dictionaries and key papers (e.g., Dombrowski et al., 2016;Hagger & Hardcastle, 2014).
Step 2 -Identifying key entities and developing the preliminary Style of Delivery Ontology A preliminary prototype version of the ontology was developed using both a bottom-up and top-down approach.
In the bottom-up approach, 100 published reports of behaviour change interventions were reviewed to develop a preliminary list of terms relevant to style of delivery.This number of reports was chosen to generate a wide preliminary range of entities for inclusion in the ontology (Wright et al., 2020).These reports were part of a large dataset of behaviour change intervention reports partially annotated for behaviour change techniques, mechanisms of action, and modes of delivery, covering a range of health behaviours (Carey et al., 2019;Michie et al., 2015).In the top-down approach, existing ontologies were searched using the list of terms created in the bottom-up approach, via specialist ontology search engines such as the Ontology Lookup Service.Terms from other ontologies were reused if the ontology used Basic Formal Ontology (BFO) as an upper level (Arp et al., 2015;Smith & Grenon, 2004).If a relevant class (note that the terms 'entity' and 'class' can be used interchangeably to refer to entities represented in an ontology.For simplicity, we use the term 'entity' throughout) was found that was not from a BFO-based ontology, a new entity was created based on that entity but made to fit with the BFO upper level.The previous entity was cross-referenced in the new entity.
Step 3 -Refining the ontology through literature annotation, discussion and revision The preliminary ontology was revised by the research team based on annotations of published, peer-reviewed intervention evaluation reports.Using EPPI-Reviewer 4 software (Thomas et al., 2010), pairs of researchers independently annotated 100 intervention reports using the preliminary Style of Delivery Ontology.This number of papers was used as previous work has found that no additional entities are likely to become apparent for inclusion in the ontology at this point (Wright et al., 2020).An open alternative to this software used for annotation is PDFAnno (Shindo et al., 2018).
Reports to annotate were identified from several systematic reviews on behavioural health interventions (Dambha-Miller et al., 2017;Kelley et al., 2014;Makin et al., 2021).These reviews were identified by using a search criterion informed by the terms identified in the bottom-up approach of entity identification.Titles and abstracts were screened from these reviews.Papers were included for annotation if: i) the outcome was a behaviour and the intervention description mentioned some aspect related to style of delivery or ii) the outcome was whether someone could be trained to use a particular style of delivery (list of papers used in the development of the ontology available on Open Science Framework).
The corpus that was used to guide the development of other ontologies in the Behaviour Change Intervention Ontology (e.g., Norris et al., 2021) was not used as early exploratory work found that relatively few of these reports contained information related to delivery style.
An annotation guidance manual was created to provide information on how and when to annotate each style of delivery entity.This promoted standardisation of the procedure.Discrepancies in annotations between annotators were discussed and the ontology structure, definitions and annotation guidance were revised accordingly.
Step 4 -Expert stakeholder review A total of 25 expert stakeholders were invited to give feedback on the Style of Delivery Ontology resulting from Step 3. The experts came from an existing list of behavioural science and public health experts who had provided feedback on previous projects at the UCL Centre for Behaviour Change, and stakeholders who expressed interest in being involved in the Human Behaviour-Change Project stakeholder initiatives in response to invitations on Twitter and the Project newsletter.Experts were purposively selected to ensure geographic diversity of participants.
Invitations to participate were sent via email, with feedback collected through an online questionnaire using Qualtrics software.The task was designed to take no longer than 45 minutes to complete.Alongside the link to the questionnaire, experts were emailed a copy of the full ontology as a spreadsheet and in a diagram in case they found these useful when completing the task.Experts were not reimbursed for their participation.
The task presented participants with the labels and definitions of the entities in the Style of Delivery Ontology.The task asked experts to: 1. Judge whether the labels or definitions of any entities needed to be refined.If so, the experts were asked to suggest changes to these entities 2. Suggest whether any entities were missing from the Style of Delivery Ontology.If so, the experts were asked to suggest labels and definitions from these additional entities 3. Suggest whether any entities needed to be re-organised in the ontology's hierarchy 4. Provide any other or general comments on the ontology.
Step 5 -Inter-rater reliability of annotations using the Style of Delivery Ontology Assessment of inter-rater reliability of the annotations by two researchers leading the development of the ontology was conducted using 50 papers.This number of papers was selected as it would give a 10-15% margin of error around the estimated percentage agreement between annotators (Gwet, 2014;Wright et al., 2020).Inter-rater reliability was also assessed for annotations of a separate batch of 50 papers by two behaviour change experts unfamiliar with the ontology but with experience in annotating behaviour change intervention reports.
To identify appropriate papers to annotate, we conducted a search using Web of Science and PubMed.We used the following search terms: "communication style" or "style of delivery" or "communication skills", combined with "intervention".We also screened individual studies included in relevant systematic reviews in the Cochrane Consumers and Communication Group library.Studies had to include one or more of the three style of delivery-related search terms in the title/abstract to be included for full-text screening.The inclusion criteria for full-text screening was: i) quantitative studies, and ii) an experimental/intervention design (i.e., not observational).The style of delivery could be part of an intervention or assessed as the outcome of an intervention (e.g., in studies of interventions to enhance healthcare professionals' communication skills).We selected the first 100 papers that met these criteria.
Inter-rater reliability was assessed using Krippendorff's Alpha (Hayes & Krippendorff, 2007) calculated using version 1.0.0 of the Automation Inter-Rater Reliability script developed by the Human Behaviour-Change Project team (Finnerty & Moore, 2020), incorporating the python script Krippendorrff 0.3.2.The research team made additional changes to the ontology based on the issues arising from inter-rater reliability testing.
Step 6 -Specifying relationships between Style of Delivery Ontology entities We established relationships between entities to formally capture the knowledge present in the ontology.Relationships were specified in line with Basic Formal Ontology principles described in Arp et al. (2015) and the Relation Ontology (Smith et al., 2005).The suitability of common relationships from these ontologies were assessed, including the basic hierarchical relationship 'is_a' which holds between entities where one entity is a parent class of another entity.
Step 7 -Disseminating and maintaining the Style of Delivery Ontology The Style of Delivery Ontology was initially developed as a table of entities, with separate rows for each entity annotated with a primary label, definition, comment, synonyms, examples, and relationships.When the Style of Delivery Ontology was at a stable level of development for initial release, it was converted into Web Ontology Language (OWL) (Antoniou & Harmelen, 2004) format, which is a standard representation format for ontologies widely used across domains.This allows it to be viewed and visualised using ontology software such as Protégé and to be compatible with other ontologies.The conversion to OWL used the ROBOT ontology toolkit library (Jackson et al., 2019), which provides a facility to create well-formatted ontologies from templates.
A ROBOT template is a comma-separated values (CSV) file that can be prepared easily in common spreadsheet software and annotated with instructions for translation from spreadsheet columns to OWL language and metadata attributes.Within the input template spreadsheet, separate columns represent the entity ID (e.g.BCIO:044000), name, definition, comment, relationship with other entities, examples, and synonyms.
This OWL version of the Style of Delivery Ontology was then stored on the project GitHub repository, an online platform for sharing and versioning resources.GitHub has an issue tracker which allows feedback to be submitted by members of the community and addressed in subsequent releases.The Style of Delivery Ontology was also made available as part of the growing Behaviour Change Intervention Ontology within the Ontology Lookup Service (Côté et al., 2006;Jupp et al., 2015) and the Behavioural and Social Science Ontology (BSSO) Foundry, a repository and community of practice for ontologies in the behavioural and social science domains.In addition, when all components of the Behaviour Change Intervention Ontology have been developed, it will be submitted to the Open Biological and Biomedical (OBO) Foundry (Smith et al., 2007).

Results
Step 1 -Development of the scope and definition of the Style of Delivery Ontology Our starting definition of style of delivery was "An attribute of intervention delivery that encompasses the characteristics of how intervention content is communicated." Step 2 -Identifying key entities and developing the preliminary Style of Delivery Ontology The preliminary version of the Style of Delivery Ontology comprised 101 entities arranged into two groups: 'communication process' and 'communication style'.Communication processes are specific, directly observable communication activities (e.g.asking open-ended questions) while communication styles reflect the manner in which intervention content is communicated that is not directly observable without inference (e.g.empathic communication style).Communication processes and styles are related, in that the use of a particular set of communication processes contributes to a given communication style.They both relate to style of delivery when used as part of an intervention content communication process.
Step 3 -Refining the ontology through literature annotation, discussion and revision Annotating intervention reports led to numerous changes made to the Style of Delivery Ontology.A total of 30 entities were added, the definitions of four entities were revised, and the parent class of one entity was changed.Review by the wider ontology development group and an expert in MI led to a further nine entities being added, three removed, and changes made to the labels or definitions of 12.For example, the entity 'positivity' (defined as "A communication style characterised by finding optimistic aspects of the topic") was removed because it was not sufficiently differentiable from the entity 'communication style conveying hopefulness' (defined as "A communication style which imparts a feeling of optimism that a person will attain a desired outcome").At the end of this phase, the ontology contained 137 entities.
Step 4 -Expert stakeholder review Of the 25 experts contacted, 10 (40%) replied saying they were willing to participate in the stakeholder review of the ontology and eight (32%) completed the task.These participants were from institutions based in the following countries: Australia (n = 3), England (n = 2), Canada (n = 1), Greece (n = 1), and Brazil (n = 1).All of the experts' suggestions and how they were addressed within the ontology development process are reported in online Extended data.
As a result of expert feedback, two entities were added and three removed.For example, the entity 'communication using non-word vocalisations' was removed because experts had difficulty differentiating it from the entity 'communication using interjections to signal attentiveness'.The labels of eight entities were revised.Updates were made to the definitions of 25 entities, with eight including a change of parent class.Many of the changes to definitions were in response to experts' comments or queries about whether a given entity could apply to interventions where the content was delivered through digital or other means, rather than through an interpersonal interaction.We added an entity labelled 'communication', defined as "A process involving the exchange of information between two or more participants" to contrast with the entity now labelled 'human communication behaviour' and defined as "interpersonal behaviour that involves the transmission of information between two or more human beings".Given the intention for the ontology to be broadly applicable, definitions were rewritten where appropriate to remove mention of "person".Instead, we refer to "participants" in a communication process.Where we needed to distinguish the actions of the participant initiating or leading a particular communication process or using a certain communication style from the participant who was the target of the communication, we referred to the leading participant as the "initiator" or "instigator" and the communication target as the "recipient".At the end of the expert stakeholder feedback review, the revised version of the ontology had 136 entities.
Step 5 -Inter-rater reliability of annotations using the Style of Delivery Ontology Inter-rater reliability from the 50 papers annotated by the two researchers familiar with the ontology was 'good' (α = 0.77; https://osf.io/dvcy6)(Hayes & Krippendorrff, 2007).Inter-rater reliability from the 50 papers annotated by the two researchers unfamiliar with the ontology was 'acceptable' (α = 0.62; https://osf.io/2b874).Final revisions were made to the ontology based on issues raised by the annotators.The following changes were made: 12 entities were added, five entities were removed, and the label of one entity was updated.For example, the two entities 'communication style conveying compassion' (defined as "A communication style which conveys an awareness of the feelings and behaviours of another, coupled with valuing the other's welfare") and 'communication style conveying sensitivity' (defined as "A communication style which empathically responds to the thoughts and feelings of the recipient") were removed because annotators had difficulty distinguishing them from the entity 'empathic communication style' (defined as "A communication style which conveys an insightful awareness of the feelings and behaviours of the recipient").Instead, we added 'conveying sensitivity' and 'conveying compassion' as synonyms of the entity 'empathic communication style'.
Step 6 -Specifying relationships between Style of Delivery Ontology entities The primary relationship used in the ontology is the 'is_a' relation which captures hierarchical classification.This relationship is specified for every entity in the ontology, that is, every entity is assigned a single classification parent within the ontology.In addition, the parent class 'communication process attribute' is related to the overarching communication process using an 'is process attribute of' relationship.This relationship is inherited down the hierarchy, thus, 'intervention style of delivery' and all its subclasses can also be inferred to be process attributes of communication processes.
Step  (Michie et al., 2017).This is the overarching vision of the Human Behaviour-Change Project, and the role of ontologies is to provide organising structure to guide the application of artificial intelligence and machine learning approaches to extract and synthesise evidence, and make inferences and predictions from that evidence to generate new understanding.
Ontologies should be maintained and updated in response to new evidence about entities and their relationships (Arp et al., 2015).To support ontology interoperability and limit duplication of work, ontology developers should collaborate with others where possible (He et al., 2018).
As with other ontologies produced as part of the Human Behaviour-Change Project (Michie et al., 2020), the Style of Delivery Ontology will be refined through application and feedback from users via GitHub.In the future, updated versions of the Style of Delivery Ontology will be released via GitHub so we recommend prospective users of the ontology check there to ensure they have the most recent version of the ontology.
The Style of Delivery Ontology is applicable to a wide range of intervention content communication processes.
In addition to contributing to the Behaviour Change Intervention Ontology (Michie et al., 2020), the Style of Delivery Ontology provides a stand-alone classification system for describing and reporting how intervention content is communicated with potential applications including environmental, educational, sports psychology, healthcare, and psychotherapy interventions.Such a classification system is relevant to training healthcare and other professionals to deliver intervention content in particular ways to their patients and clients by allowing for a detailed specification of the ways in which content might be communicated.

Strengths and limitations
A strength of this study is the use of a standardised, tried and tested method for ontology development created within the Human Behaviour-Change Project (Wright et al., 2020) that reflects the OBO Foundry principles of good practice in ontology development (Norris et al., 2019;Smith et al., 2007).As a result, the ontology fares well against several of Vrandečić's (2009) ontology evaluation criteria.In particular, the ontology's completeness, or how well it covers the domain of interest, was tackled by preliminary scoping

person-centred intervention delivery [BCIO:044006]
intervention delivery An intervention delivery characterised by efforts by an intervention source to give a person the resources they need to manage their own life and make them an active participant in deciding how to move forward.

participant-led intervention delivery [BCIO:044007]
intervention delivery An intervention delivery process in which the participant takes the lead in determining the focus of the intervention.

source-led intervention delivery [BCIO:044008]
intervention delivery An intervention delivery process in which the intervention person source takes the lead in determining the focus of the intervention.

human communication behaviour [BCIO:036034]
inter-personal behaviour An interpersonal behaviour that involves the transmission of information.

linguistic communication behaviour [BCIO:050237] human communication behaviour
A human communication behaviour in which the information that is communicated is encoded in language.

non-linguistic communication behaviour [BCIO:050238] human communication behaviour
A human communication behaviour in which information is transmitted without being encoded in the meaning units of any language.

non-linguistic communication behaviour
A non-linguistic communication behaviour that involves expressing thoughts or feelings through bodily movement or posture.

non-linguistic communication behaviour using gesture [BCIO:036002]
non-linguistic communication behaviour using body language A non-linguistic communication behaviour using body language involving a deliberately chosen movement of the head, neck, torso, or limbs to express an idea or feeling.

Label [URI]
Parent A communication using language in a manner which does not demonstrate an opinion on a particular issue.

communication using non-pressurising language [BCIO:044052] communication using particular language
A communication using language chosen to avoid coercing or persuading the recipient to think, feel or behave in a certain manner.

communication using slang [BCIO:044053] communication using particular language
A communication using informal language shared by members of a particular community.

communication using specialist language [BCIO:044054] communication using particular language
A communication using words or expressions used by a profession or group that are difficult for others not in that profession or group to understand.

communication avoiding offensive language [BCIO:044055] communication using particular language
A communication using language chosen by the instigator to avoid using words or phrases that could be perceived as rude or insulting.
communication using selected words from a second language [BCIO:044056] communication using particular language A communication using occasional words chosen from a language spoken by both the instigator and the recipient that is not the main language.

summarising reflective communication [BCIO:044072]
reflective communication A reflective communication in which the instigator pulls together key points from a complex story told by the recipient in a succinct restatement.

double-sided reflective communication [BCIO:044073]
reflective communication A reflective communication in which the instigator aims to highlight both the positive and negative sides of an issue as previously expressed by the recipient.

metaphorusing reflective communication [BCIO:044074]
reflective communication A reflective communication in which the instigator replaces a word or phrase of the recipient's prior assertion with a figure of speech suggesting an analogy.

emotion-emphasising reflective communication [BCIO:044075]
reflective communication A reflective communication highlighting the emotional aspects of an issue as previously described by the recipient.

re-evaluation prompting reflective communication [BCIO:044076]
reflective communication A reflective communication in which the instigator restates the recipient's prior assertation in a less credible way to prompt reconsideration of the assertion.

relationship building [BCIO:044124] communication
A communication aimed at building or managing aspects of the interpersonal relationship between an initiator and a recipient rather than focusing on delivering communication content.

readability [BCIO:044130] attribute
An attribute of a text used in communication process, concerning the ease of understanding the wording or other qualitative aspects of the text that facilitate comprehension.

source visual style [BCIO:044128]
attribute An attribute of an intervention person source concerning how they appear to others.

object layout [BCIO:044129]
object attribute An attribute of an object used in communication process, concerning how the object's components are arranged.

object visual style [BCIO:044127]
object attribute An attribute of an object used in a communication process, concerning how information sensed with the eyes is presented, which may include colours/ fonts/images/spacing/sizing.
of terms from 100 papers, and annotations of a further 100 papers describing a wide range of interventions.In the stakeholder review, experts were asked if they thought any terms were missing from the ontology.The ontology's accuracy, in terms of how well it accorded with experts' knowledge, was targeted by having the ontology developed by a team of researchers with considerable experience in behaviour change interventions and by subjecting the ontology to expert stakeholder review.The ontology's clarity, in terms of whether it communicated the intended meaning of the defined terms was targeted through literature annotation, where differences between annotators were considered to indicate problems with communicating meaning and the need for revisions.It was further examined through inter-rater reliability testing which indicates whether the communicated meaning of terms is sufficiently clear that annotators can agree on what constitutes an example of an entity.
A limitation of this study is that most intervention reports used in the ontology development process came from high income countries, potentially limiting the ontology's applicability to interventions in low-and middle-income countries.
While the ontology benefits from the incorporation of international expert stakeholder feedback in the ontology development process, only a third of invited experts provided feedback on the ontology.Feedback from a larger group of experts might have led to a greater range of suggested changes to the ontology.Further, our definition of 'expert' (i.e.those with a doctoral degree in behavioural science or a related discipline) may differ to other people's definitions, thus another group of experts from different disciplines or domains may have provided different feedback.
It is also important to acknowledge the limitations of poor reporting in intervention reports used to develop the ontology.
A key part of ontology development is to identify entities from published reports during the 'bottom-up' phase (i.e.step 2).Many reports did not describe style of delivery at all, or if they did, not in detail.This might limit the comprehensiveness of the Style of Delivery Ontology.However, it is expected that this will improve as the ontology is used more in the design and reporting of behaviour change interventions.Future applications of the Style of Delivery Ontology to a wider range of interventions and contexts will provide opportunities for it to be extended and improved.

Conclusions
The Style of Delivery Ontology provides a classification system that can be reliably used to specify the style with which intervention content is communicated.Its use by researchers and other stakeholders will contribute to improved consistency in reporting, facilitating easier replication and simplifying evidence synthesis.The ontology will increase understanding of how the impact of intervention content (e.g., behaviour change techniques) (Marques et al., 2023;Michie et al., 2013)

Bryan A Kutner
Albert Einstein College of Medicine, Bronx, USA The authors present their process for the development of an ontology intended to allow the characterization of all aspects of the style in which behavior change interventions are delivered.This is an ambitious endeavor, the manuscript is well-written and easy to follow, and the ontology will be a useful addition to behavioral science.I intend my suggestions below to strengthen an already very clearly written article, in particular to add details that allow readers to replicate the 7step process.
General Comments: A. Among the entities in the ontology, I've not seen anything that might capture that cultural affinity between deliverer and audience might be an important aspect of Style of Delivery that influences effectiveness.For example, when peers deliver an intervention there may be shared language, experience and/or communication styles that convey the relevance of behavior change differently than a would-be credible source who is credentialed but not as closely aligned in the style of how they communicate as someone who more directly shares cultural experiences with an audience.Would the authors consider this to be an aspect of Style of Delivery?
B. For dissemination, would the authors consider inclusion of the some of the coded excerpts that exemplify each entry, as the Human Behaviour Change Project has for the Behaviour Change Techniques Taxonomy v. 1?This would be helpful to understand how to operationalize the entities within the ontology and perhaps aid users in recognizing distinctions between entries.

Methods:
A.
Step 1 -Development of the scope and definition of the Style of Delivery Ontology.How did you select definitions and key papers, who reviewed these definitions and key papers, and how did you come to a final determination of the scope and definition of the Style of Delivery?These would be useful to share for those of us interested in replication.

B.
Step 2 -Identifying key entities and developing the preliminary Style of Delivery Ontology.For the bottom-down approach, how did you select 100 out of what look like more than 400 reports in the linked dataset repository?Perhaps useful as well to say "existing dataset" and to specify that these include reports through 2018 in case the linked site changes.For the top-up approach, I'd make the parenthetical here outside of the sentence itself ("If a relevant class (note that the terms 'entity' and 'class' can be used interchangeably to refer to entities represented in an ontology.For simplicity, we use the term 'entity' throughout) was...").Also who conducted the searches? C.
Step 3 -Refining the ontology through literature annotation, discussion and revision.How many pairs reviewed the 100 articles?Were discrepancies in annotations between annotators discussed just between annotators or within a larger research team (in other words, what if annotators could not come to consensus or their determinations were ambiguous)?D.
Step 4 -Expert stakeholder review.Who resolved and how did they resolve all of the feedback from reviewers?Knowing about this process would aid replication.

E.
Step 5 -Inter-rater reliability of annotations using the Style of Delivery Ontology.Do you want to specify from among the authors who the two pairs of raters were here?Why only quantitative and experimental/interventional design studies?It seems that qualitative studies might also be informative about style of delivery.Using searches across two publication databases, how did you determine the first 100 to select?Was this organized by date or by likelihood of matching the search criteria?Readers may also want to know more about the kinds of "additional changes" as examples ("The research team made additional changes to the ontology based on the issues arising from inter-rater reliability testing").

Results:
A.
Step 2 -Identifying key entities and developing the preliminary Style of Delivery Ontology.These definitions mirror the Motivational Interviewing Training Integrity (MITI) assessment tool that distinguishes between behavior counts and global aspects of the delivery of motivational interviewing.This may not have informed your own definitions; if it did, that lineage would be useful to share, given that you mention the involvement of one expert in motivational interviewing.

B.
Step 5 -Inter-rater reliability of annotations using the Style of Delivery Ontology.What was the experience of the two raters who were unfamiliar with the ontology in terms of their use of ontologies in general?This has implications for implementation of the ontology among 'lay' users.Related to the earlier comment on the MITI, my understanding is that inter-rater reliability for MITI counts is greater than for MITI global entities.Was this similar in your study?It would be useful for readers to see whether IRR varied among entities that fell within specific buckets (e.g., "communication processes", "non-linguistic communication behaviours", "communication styles" and "attributes of objects used in communication processes").In general, seeing which codes had more or less IRR would also help advance where clarity might be furthered within the ontology.Thank you for the invitation to review this article reporting the development of a style of intervention delivery ontology.This is an interesting paper which I enjoyed reading, and will be of great value to readers across a large number of disciplines.
Some minor suggestions are provided below: Intro/general rationale -For the sentence "Additionally, certain groups of people may prefer particular delivery styles due to differences in social or cultural norms or other factors."-An example or two might be helpful here (e.g.people with versus without a term health conditions, or people from higher or lower socio economic backgroundsthose are just examples).
○ Also in the proceeding sentences a citation or two could be provided with examples showing differences in intervention effectiveness according to delivery style.This is a key aim of the work and just a couple of citations would be helpful for added context, particularly for readers not familiar with this area.
○ I like the point "A communication process can involve a variety of participants, such as humans interacting synchronously or asynchronously, a human talking to an artificial intelligence-based conversational agent (i.e., chatbots), or two computer systems exchanging information".-I wonder if more explicit examples could be provided e.g. for humans interacting synchronously or asynchronously computer systems talking to each other.

○
Methods -it might be helpful to somewhere define what the term "reports" relates to and/or includes (assuming peer reviewed publications, as well as others?)-I think this includes papers and reviews as per the link (https://www.bct-taxonomy.com/interventions) but it might be good to explicitly define this in the text.

○
Might it be helpful to list all of the specialist ontology search engines used?(Step 2 in methods).

○
Good to see inter-rater reliability for the annotations -was it also done for the title and abstract screening stage?○ I think the advantages of converting into Web Ontology Language (OWL) could be a little more explicit -presumably one of the advantages is that it can potentially be integrated into IT and/or AI systems?
○ I like the iterative process of developing and presenting the different versions of the ontology and looking at the links (also great to provide these) "A communication style which imparts a feeling of optimism that a person will attain a desired outcome").At the end of this phase, the ontology contained 137 entities."It looks like these correspond to the different versions (as well as the labels i.e. version 0.2) -you could label the version in the text more explicitly to show the iterative process.E.g. this led to version 2… etc.

○
In step 7 could make it clear that the link is the final version.Might also be helpful to say it's free to access and use with no licensing restrictions (if that is the case).

○
For the sentence "Such a classification system is relevant to training healthcare and other professionals to deliver intervention content in particular ways to their patients and clients by allowing for a detailed specification of the ways in which content might be communicated."Can you provide an example or two of the "content" of interventions.

○
For the 100 papers that were used could you give a very brief description of the inclusion ○ criteria to highlight the strength and amount of evidence reviewed (e.g.peer reviewed research studies and reviews spanning x years across x countries, although I do note the mention of high income countries in the proceeding section so you may not feel this is relevant)."Health professionals" and "healthcare professionals" are used interchangeably -perhaps just stick with the latter for consistency.The authors provide a clear explanation of how they developed the ontology, supported by supplementary materials accessible via embedded links.In my opinion, the details provided, clear presentation, supporting tables and figures and additional documentation accessible to readers sufficiently support their description and conclusion, thereby validating the resulting ontology as being potentially useful for researchers interested in the area of behaviour change communication.
As a qualitative researcher, however, I am not able to comment on the statistical process used to assess inter-rater reliability (Krippendorff's Alpha, pg 7).Another point I would like to highlight is that the ontology separates communicative practices into discrete units, which is necessary for interventional study designs, while actual conversations, including those focused on behaviour change, are interactive and collaboratively accomplished.Moreover, in the ontology itself, some of the communicative practices listed refer to specific interactional practices (e.g.asking closedended questions), which can be discretely observed and measured, while others refer to communicative events which are jointly accomplished by two or more interactants (e.g.interrupting, negotiation, classroom style discussion).In this sense, the labelling of the ontology as Style of Delivery may be a little limiting as it conveys a one-sided discrete behaviour and the ontology also includes communicative practices which are not necessarily related to 'style' (e.g.interruptions).Thus, the authors may want to acknowledge the limitations of the ontology as far as the need to reduce the collaborative nature of human conversation into discrete describable/measurable actions to be focused on in interventional research.
In addition, interpretation of some items could be ambiguous, subjective or context-bound, particularly those related to communication style (interactive communication style, informal communication style, rushed communication style).In the long-term, therefore, it may be useful to work towards developing a supplementary database of example/transcripts from authentic communication to illustrate some of the items in the ontology.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound?Yes

Are sufficient details of methods and analysis provided to allow replication by others? Yes
If applicable, is the statistical analysis and its interpretation appropriate?

Not applicable
Are all the source data underlying the results available to ensure full reproducibility?Yes

Are the conclusions drawn adequately supported by the results? Yes
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: doctor-patient interaction, discourse analysis I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.
of communicating aimed at inducing or avoiding certain kinds of responses in others, or demonstrating certain characteristics of the initiator.that is an attribute of a BCI content communication process.
involves the use of declarative propositions based on what the communicator believes to be logical.discussion [BCIO:044019] communication A communication involving an extended exchange of utterances between an instigator and at least one other person.by the initiator in a manner typical of a teacher or lecturer.better understand what a person thinks or feels about a topic or an experience.communication exploring a person's thoughts about a topic [BCIO:050386] exploring communication An exploring communication focused on what a person thinks about a matter.involves suggesting adoption of certain beliefs, theories or lines of action by others.Interrupting [BCIO:044061] communicationA communication in which one person interjects before another has finished making their point.which at least one participant endeavours to not interject before another participant has finished making their point.the particulars of a person's life or a course of events in the style of telling a story.goal-oriented collaboration to arrive at an agreement or compromise on a matter of importance to the individuals involved.to provide hope and strength in times of anxiety, worry, confusion, grief, distress, or suffering.a person's previously expressed thoughts and feelings back to them.
follows a pre-written sequence of instructions that are interpreted or carried out by a person or by a program.marked by fostering the involvement of the recipient as an engaged actor in the activity to be undertaken rather than as a passive recipient.characterised by a lack of empathy or compassion for the recipient.characterised by a sense of the initiator and recipient working together, the initiator showing sensitivity to the others' needs, in order to obtain an outcome desired by all.which imparts a feeling that the instigator believes themselves to be higher in status than the recipient.communicationstyle conveying support for change [BCIO:0440894] communication styleA communication style which imparts one's belief that the recipient can change.that demonstrates that one is not assigning a positive or negative valence to another's views or behaviours.which aims to induce or urge the adoption of certain beliefs, theories or lines of action.characterised by a lack of tension or anxiety.suggesting an understanding of and regard for the feelings, wishes, or rights of others.involving the communication content having some constituent pre-planned parts but also being responsive to inputs.characterised by one participant having more control over a conversation than the other.an oral communication relating to the rhythm and pitch of what is said.

○
Is the work clearly and accurately presented and does it cite the current literature?Yes Is the study design appropriate and is the work technically sound?Yes Are sufficient details of methods and analysis provided to allow replication by others?Yes If applicable, is the statistical analysis and its interpretation appropriate?Yes Are all the source data underlying the results available to ensure full reproducibility?Yes Are the conclusions drawn adequately supported by the results?Yes Competing Interests: No competing interests were disclosed.Reviewer Expertise: Behaviour change I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.a larger effort to develop a Behaviour Change Ontology encompassing all elements of behaviour change scenarios.The research efforts focus on a much-needed area, where the numerous elements involved in behaviour change interventions as well as the diversity of understandings of what each element encompasses hinders meaningful discussion and comparison of the effectiveness of health behaviour change interventions.
7 -Disseminating and maintaining the Style of Delivery Ontology The developed ontology consisted of 145 entities relevant to style of delivery.A downloadable version of the Style of Delivery Ontology is available from GitHub and it can be browsed in the dedicated BCIOSearch tool and diagrammatically represented using the BCIOVisualise tool.What works, compared with what, for what behaviours, how well, for how long, with whom, in what setting, and why?" (Norris et al., 2020)Identifiers (URIs), labels and definitions for all entities are described in Table2.Further details (e.g.comments, synonyms, examples) are found in Extended data.The ontology is accompanied by an annotation guidance manual that provides guidance on how to annotate for these entities in behaviour change intervention.(Norrisetal., 2020), mechanisms of action(Schenk et al.,  2023), dose (in preparation), schedule of delivery (in preparation), engagement (in preparation), fidelity (in preparation), target behaviour (in preparation), and target population (in preparation).These ontologies are connected and can be used together to answer variants of the 'big question': "When it comes to behaviour change interventions:

Table 2 . URIs, labels and definitions for all Style of Delivery Ontology entities.
varies according to the style in which such content is delivered.The Style of Delivery Ontology provides a starting point upon which future research and translation efforts can build.Its development is intended to be an ongoing and collaborative process.

Table 1
unfamiliar with ontology development B. Entity: "It Includes" and is a comma intended here to distinguish a cognitive process as either cognitive or emotional?"the process and content of cognitive representations, and emotions" C. EPPI-Reviewer: Check grammar . Glossary:There are a few typos to correct. A. Basic Formal Ontology (BFO): entity->entities; please define "continuants" and "occurrents" for an audience

Is the work clearly and accurately presented and does it cite the current literature? Yes Is the study design appropriate and is the work technically sound? Yes Are sufficient details of methods and analysis provided to allow replication by others? Partly If applicable, is the statistical analysis and its interpretation appropriate? Yes Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes Competing Interests:
I attended a course on the Behaviour Change Wheel with Drs.Michie and West, hosted online by University College London from 12-16 July 2021.I also organized a 3-part series of seminars about mechanisms of action in behavioral science and HIV research hosted by the HIV Center for Clinical and Behavioral Studies at Columbia University and New York State Psychiatric Institute; Drs.Michie, Johnston, Hastings and West presented the first seminar on 21 September 2022.I have no other competing interests to disclose.

have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. Chris Keyworth Manchester
Centre for Health Psychology, Division of Psychology and Mental Health, University of Manchester, Manchester, UK