=Paper=
{{Paper
|id=Vol-3939/paper1
|storemode=property
|title=Modeling Biomedical Burdens in Basic Formal Ontology
|pdfUrl=https://ceur-ws.org/Vol-3939/paper1.pdf
|volume=Vol-3939
|authors=Paul Smart,Nic Fair,Michael Boniface
|dblpUrl=https://dblp.org/rec/conf/icbo/SmartFB24
}}
==Modeling Biomedical Burdens in Basic Formal Ontology==
https://ceur-ws.org/Vol-3939/paper1.pdf
Modeling Biomedical Burdens in Basic Formal
Ontology
Paul Smart1,*,† , Nic Fair1,† and Michael Boniface1,†
1
Electronics and Computer Science, University of Southampton, Highfield, Southampton, SO17 1BJ, UK
Abstract
The term “biomedical burden” refers to a number of types of burden that have been discussed in the
medical literature. This includes the burdens that stem from the presence of disease (e.g., disease burden
and symptom burden), as well as the burden associated with healthcare interventions (e.g., treatment
burden and side-effect burden). The present paper describes an ontology of biomedical burdens that
uses Basic Formal Ontology (BFO) as an upper-level ontology. The ontology provides support for the
representation of burden-related cognitions, the classification of burden types, and the representation of
burdensome situations. The ontology also captures our current theoretical understanding of burden,
thereby serving as the basis for future analytic and definitional work.
Keywords
Basic Formal Ontology, Medical Ontology, Treatment Burden, Symptom Burden
1. Introduction
The term “biomedical burden” refers to a number of types of burden that have been discussed in
the medical literature. These include the burdens associated with disease and illness (e.g., disease
burden, symptom burden), as well as the burdens associated with healthcare interventions (e.g.,
treatment burden, side-effect burden).
While the study of burden has long been a feature of medical research—with studies dating
back to at least the 1940s [1]—recent years have seen a renewed interest in burden. This is partly
due to shifts in the demographic and epidemiological landscape, which has meant that more
people are living with chronic, life-limiting conditions. Other changes relate to the provision of
healthcare, with patients expected to play a greater role in the prevention, management, and
treatment of disease. As May et al. [2] note:
Healthcare services increasingly seek to position patients and their supporters
as accountable for this work. In turn, this shift in accountability involves adding
the burden of treatment to the burden of symptoms, as patients experience new
and growing demands to organize and co-ordinate their own care, to comply with
complex treatment and self-monitoring regimens, and to meet a whole range of
expectations of personal motivation, expertise and self-care. (p. 2)
15th International Conference on Biological and Biomedical Ontology, July 17–19 2024, Enschede, The Netherlands
*
Corresponding author.
†
These authors contributed equally.
$ ps02v@ecs.soton.ac.uk (P. Smart); N.S.Fair@soton.ac.uk (N. Fair); M.J.Boniface@soton.ac.uk (M. Boniface)
0000-0001-9989-5307 (P. Smart); 0000-0003-1566-4689 (N. Fair); 0000-0002-9281-6095 (M. Boniface)
© 2024 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR
ceur-ws.org
Workshop ISSN 1613-0073
Proceedings
Despite the interest in burden, there has been little attempt to subject burden-related concepts
to analytic scrutiny. This has complicated efforts to arrive at a common understanding of burden,
leading to confusion about the meaning of burden-related terms and the absence of univocal
definitions [see 3, 4].
Paralleling the lack of theoretical attention, there is, at present, little support for burden-
related terms within medical ontologies. The Ontology Lookup Service (OLS) search engine, for
example, yields no results for the terms “symptom burden” or “treatment burden.”1 This lack of
support is problematic, for it complicates efforts to assimilate burden-related data into wider
nexus of biomedical datasets and biomedical applications.
In the present paper, we aim to address these twin concerns by describing the Biomedical
Burden Ontology (BBO)—a theoretically-grounded ontology of biomedical burdens.2 To support
interoperability with other medical ontologies, the BBO is implemented in the Web Ontology
Language (OWL) and uses Basic Formal Ontology (BFO) [see 5] as an upper-level ontology. The
BBO also relies on a number of ontologies that form part of the Open Biological and Biomedical
Ontology (OBO) Foundry. These include the Ontology for General Medical Science (OGMS)
[6], the Mental Functioning Ontology (MFO) [7], the Emotion Ontology (MFOEM) [8], and the
Information Artifact Ontology (IAO) [9].
The bulk of the paper is devoted to a description of the BBO. This is presented in Section 3.
In Section 2, we describe the theoretical grounding of the BBO, reporting the results of an initial
analysis of the burden concept. Throughout the paper, we rely on typographic conventions
to distinguish between ontology elements. OWL classes are rendered LIKE THIS, OWL
properties likeThis, and OWL individuals :LIKE THIS. Property instances are rendered
in italic font, likeThis.
2. Understanding Burden
One of the main challenges in developing an ontology of biomedical burdens is that burden-
related concepts have seldom been at the forefront of analytic attention. This makes it difficult
to answer questions that are relevant to the ontology development effort. What, for example,
is a burden? Are burdens members of a common ontological kind (such as the category of
continuants), or are they of different kinds? What, moreover, of the nature of burdensomeness?
What is it that makes some things burdensome, while others are not burdensome?
A useful starting point is to consider a dictionary definition of burden. According to Merriam-
Webster, the term “burden” can refer to multiple things.3 Firstly, it may refer to a heavy load, as
in the case of someone who is forced to carry a heavy backpack. Call this the load-based view of
burden. Secondly, a burden may refer to a certain duty or responsibility, as when one reluctantly
agrees to water a neighbor’s plants while they are on vacation. Call this the obligation-based
view of burden. Finally, a burden may refer to something oppressive or worrisome, as when an
individual is burdened by a persistent sense of guilt or regret. Call this the emotion-based view
of burden.
1
See https://www.ebi.ac.uk/ols4/ [Accessed: 09 April 2024].
2
The current version of the BBO is available via GitHub: https://github.com/ps02v/BBO.
3
See https://www.merriam-webster.com/dictionary/burden [Accessed: 16 March 2024].
At first sight, these views suggest that burdens are unlikely to belong to a common ontological
kind, such as the category of continuants or occurrents. Under the load-based view, for example,
the thing that is deemed to be a burden is a physical load of some sort, such as a heavy backpack.
Given that backpacks are material entities (and thus independent continuants), the load-based
view suggests that some burdens can qualify as independent continuants.
The emotion-based view yields a different result. In this case, the thing that is deemed to be
burdensome is a particular sort of experience, such as a feeling of guilt or regret. This suggests
that some burdens are best understood as occurrent or processual entities, at least according
to contemporary BFO-conformant ontologies, such as the MFO and MFOEM. Much the same
could be said of the obligation-based view. If, for example, one is burdened because one has an
obligation to water a neighbor’s plants, then it is hard to resist the conclusion that processes
(specifically, plant-watering processes) must have something to do with this.4
The problem, then, is that different views yield different conclusions regarding the ontic
nature of burdens. The load-based view suggests that burdens are best understood as continuant
entities, while the emotion- and obligation-based views suggest that burdens are best understood
as occurrent entities. Having said this, it should be noted that occurrent entities are not entirely
absent from cases that fall under load-based view. Consider, again, the backpack case. While the
backpack is readily understood to be a burden (or the thing that is burdensome), it cannot be the
mere presence of the backpack that establishes its status as a burden. What makes the backpack
burdensome is not so much the fact that it exists; it is more the fact that one is obliged to carry
it. In short, the backpack is burdensome because one must enter into some sort of relationship
with the backpack. This relationship is established courtesy of one’s participation in a process,
specifically, a carrying process. This is important, for it highlights a point of commonality
between the load-based, obligation-based, and emotion-based views: In all these cases, we
encounter state-of-affairs in which some individual is obliged to participate in a process of some
sort. In the backpack case, that process is a carrying process; in the plant-watering case, it is a
watering process, and, in the guilt-ridden case, it is an emotional process.
This insight serves as the basis for a view of burden we call the Atlassian view.5 According
to the Atlassian view, burdens are best understood as situations or states-of-affairs in which
some entity (e.g., a human individual) is obliged to participate in a process. This notion of
obligatory participation is central to the Atlassian view. According to the Atlassian view, what
makes something burdensome is the fact that an individual is obliged to participate in a process.
This might be because an individual has no control over the process, as is often the case with
symptom processes.6 Alternatively, there might be a substantial cost associated with one’s
refusal to participate in a process, as when one is obliged to participate in a particular course of
treatment as the a means of ridding oneself of a disease.7
4
An alternative reading of the obligation-based view stems from Arp et al. [5, p. 103]. They suggest that obligations
are best understood as relational qualities. This would place burdens (under the obligation-based view) in the
category of specifically-dependent continuants.
5
The Atlassian view is named after the Titan, Atlas, who was condemned (obliged) to hold up the heavens after the
Titanomachy.
6
Pain, for example, is something that happens to us; we do not choose to participate in a pain-related process, nor is
it easy to suspend our participation in pain-related processes whenever we wish to do so.
7
For more on the notion of participation, and the different types of participatory relationship included in BFO, see
Smith and Grenon [10].
Table 1
Biomedical burdens according to the Atlassian view. Different types of burden are distinguished
according to the nature of a process in which a human individual is obliged to participate.
Burden Type Process Type Comments
Disease Burden Pathological Bodily Pathological bodily processes correspond (ei-
Process ther wholly or in part) to the realization of
diseases. Diseases are represented as disposi-
tions in BFO.
Symptom Burden Symptom Process A process that often occurs as part of a disease
course (e.g., pain, fatigue, nausea, or anxiety).
Treatment Burden Treatment Process A process that is performed as part of a treat-
ment regimen or treatment plan (e.g., the pro-
cess of injecting insulin).
Side-Effect Burden Side Effect A process that occurs as the consequence of
(i.e., as the effect of) a treatment process.
One virtue of the Atlassian view is that it is readily applicable to many of the burdens
discussed in the medical literature. In respect of treatment burden, for example, the burden
arises as the result of an individual’s obligatory participation in processes that are intended to
treat an ongoing illness or disease (i.e., treatment processes). The same applies to other types
of biomedical burden, such as symptom burden, side-effect burden, and disease burden (see
Table 1). What is common to these burdens, we suggest, is that a given individual (e.g., a patient)
is obliged to participate in a certain sort of process. It is then the nature of this process (e.g., a
treatment process) that serves as the basis for the distinction between burden types.
Notwithstanding the merits of the Atlassian view, there are reasons to think that it cannot
be the whole story of burden. In particular, while obligatory participation may be a common
(perhaps universal) feature of burden, it cannot be sufficient for burden. Consider that one may
be obliged to participate in an eating process, but this does not mean that the eating process
(or one’s participation in the eating process) is thereby rendered burdensome. In order for
the eating process to qualify as burdensome, something more must be added. The missing
ingredient, it seems, is something to do with the ‘hedonic tone’ of the process in which one is
obliged to participate. If one wants to eat, and the food is palatable, then there is no reason to
think the eating process will be burdensome. If, by contrast, one is forced to eat despite the fact
that one is full, then the eating process is apt to acquire more burdensome ‘flavor’.
In addition to obligatory participation, then, the apparent ‘goodness’ or ‘badness’ of a process
looks to be important to our understanding of burden. In particular, we suggest that an
individual’s obligatory participation in a process only counts as burdensome if the process is
non-preferred, by which we mean, the individual would much prefer not to be participant in the
process in which they are obliged to participate. Ideally, of course, the appeal to non-preferred
processes would benefit from an account that explicates the distinction between preferred
and non-preferred processes. While we have explored this issue in recent research, a detailed
discussion of this issue would take us too far afield.8 For present purposes, then, we will simply
8
In brief, our approach relies on predictive processing (aka. active inference) accounts of motivated and goal-directed
acknowledge that the distinction between preferred and non-preferred processes is likely to be
important to our theoretical understanding of burden.
Putting all this together, yields the following working definition of burden: A burden is a
situation (O) in which a subject (S) is obliged to participate in a process (P) and P is a non-
preferred process, meaning that S would much prefer they were not a participant in P.
We can now make a distinction between what we call the object of burden and the subject of
burden. The object of burden is the situation in which an individual is obliged to participate in a
(non-preferred) process (i.e., O in the above definition). The subject of burden is the entity that
is being burdened (i.e., S in the above definition). For present purposes, we will assume that
the subject of burden is always a human individual (as opposed to a collective entity, such as a
nation state). Accordingly, when we talk of treatment burden, what we mean is that a human
individual is obliged to participate in one or more treatment processes (e.g., a kidney dialysis
process). Likewise, when we talk of symptom burden, what we mean is that a human individual
is obliged to participate in one or more symptom processes (e.g., a chronic pain process or
chronic fatigue process).
Having summarized the effort to advance our theoretical understanding of burden, our
attention now turns to the BBO. In the next section, we describe how the various elements of
the Atlassian view of burden are accommodated by the BBO.
3. Representing Burden
The BBO is implemented as an extension of existing ontologies, some of which are included in
the OBO Foundry. Aside from BFO, the BBO incorporates entities from OGMS [6], MFO [7],
MFOEM [8], and the Cognitive Process Ontology (CPO) [12, 13]. The BBO also relies on entities
from the IAO [9], as well as the more recent information modeling extensions provided by the
Common Core Ontologies (CCO) suite [14].
At the heart of the BBO is the BURDEN PROCESS class. This is represented as a subtype of
AFFECTIVE PROCESS, which is one of the classes included in the MFOEM. Different types
of biomedical burden (e.g., TREATMENT BURDEN and SYMPTOM BURDEN) are represented
as subtypes of the BURDEN PROCESS class (see Figure 1).
As with the modeling of other affective phenomena, such as emotions [8], we assume that
additional processes occur as part of burden processes. These include the likes of physiological
processes (e.g., those associated with a stress response), cognitive processes (e.g., appraisal
processes), and experiential processes (e.g., the subjective feeling of burden). While these
processes need not occur at the same time, we assume they occur as part of burden processes,
which is to say that they all occur (if they occur) within the TEMPORAL REGION occupied
by the burden process.
Although it is not explicitly represented in the BBO, we allow for the possibility that burden
processes may be linked to phenomena that are typically studied as part of research into
behavior [e.g., 11]. In particular, we suggest that preferred processes are tied to neurally-realized predictions
that govern the choice between action policies. In the BBO, these predictions are represented via a CONATIVE
REPRESENTATION class (see Figure 3). Non-preferred processes are then the processes that disrupt an individual’s
capacity to bring about the state-of-affairs denoted by conative representations.
Occurrent
Health Care
Symptom Process
Process
Pathological Treatment
Bodily Process
Bodily Process Process
Appraisal
Mental Process
Process
Affective Process
Subjective
Emotion Process Burden Process
Affective Feeling
Treatment Feeling of
Fear Symptom Burden
Burden Burden
Side-Effect
Disease Burden
Burden
Figure 1: Different types of biomedical burden (e.g., treatment burden) are represented as subtypes
of the BURDEN PROCESS class. Colored boxes indicate classes that are specific to the BBO, while
non-colored boxes indicate classes sourced from other ontologies, such as BFO, OGMS, MFO, and
MFOEM.
biomedical burdens. In respect of treatment burden, for example, there is considerable interest
in the relationship between treatment burden and medical non-adherence [e.g., 15]. Indeed, non-
adherence is one of the most widely cited consequences of treatment burden [e.g., 16]. Figure 2
illustrates one way of representing this relationship in the BBO. In this case, a BURDEN
PROCESS causes a GAIN OF REALIZABLE ENTITY, which represents the gain of a
particular tendency, namely, TREATMENT NON-ADHERENCE TENDENCY.9 [Note that the
nodal elements (or boxes) in Figure 2 correspond to OWL individuals, not OWL classes.]
Appraisal processes are of central importance to burden processes, for they are the means by
which individuals appraise certain situations as being burdensome. In the MFOEM, appraisal
processes are represented by the APPRAISAL PROCESS class, which is a type of MENTAL
9
See Jansen [17], for more on the ontological characterization of tendencies and their relation to other types of
realizable entity (e.g., dispositions).
participatesIn
:Treatment :Gain of
:Human Being
participatesIn Burden isCauseOf Realizable Entity
:Treatment Non-
inheresIn Adherence Tendency affects
Figure 2: Using the BBO to represent the relationship between treatment burden and non-adherence.
Specifically Dependent
Occurrent
Continuant
Representation Mental Quality Process
Mental
Bodily Process
Representation
Cognitive Conative
Mental Process
Representation Representation
Affective Appraisal
Appraisal
Representation isOutputOf Process
Appraisal of Burden 1 Appraisal of 1 Appraisal of
Object Burden Burdensomeness
Appraisal of Appraisal of Appraisal of Side-
Symptom Burden Treatment Burden Effect Burden
Figure 3: Burden appraisals are represented as subtypes of the APPRAISAL class. Each APPRAISAL
OF BURDEN consists of an APPRAISAL OF BURDEN OBJECT and an APPRAISAL OF
BURDENSOMENESS. [Dashed lines represent indirect taxonomic relationships.]
PROCESS (see Figure 1). The output of an APPRAISAL PROCESS is an APPRAISAL OF
BURDEN, which is represented as a type of COGNITIVE REPRESENTATION (see Figure 3).
The APPRAISAL OF BURDEN is a composite representation, which consists of at least two
further representational entities. These are the APPRAISAL OF BURDEN OBJECT and the
APPRAISAL OF BURDENSOMENESS (see Figure 3). The former entity (i.e., APPRAISAL
OF BURDEN OBJECT) represents what we earlier referred to as the object of burden (see
Section 2). That is to say, it represents the situation or (state-of-affairs) that an individual deems
to be burdensome (e.g., a situation in which the individual is obliged to carry a heavy backpack).
The latter entity (i.e., APPRAISAL OF BURDENSOMENESS) represents the reason why
the object of burden is deemed to be burdensome. In the BBO, this is understood in terms of
the impact or effect of one’s (obligatory) participation in a process. Specifically, situations are
appraised as burdensome when they threaten to disrupt (or interfere with) an agent’s capacity to
bring about the state-of-affairs denoted by instances of the CONATIVE REPRESENTATION
class (see Figure 3 and Footnote 8). Due to space constraints, we will refrain from further
discussion of the CONATIVE REPRESENTATION class and its link to negative impacts. For
present purposes, we will limit our attention to the APPRAISAL OF BURDEN OBJECT
class.
As with other types of appraisal (e.g., the appraisal of pleasantness), the APPRAISAL
OF BURDEN is intended to represent the appraisal-related cognitions that individuals have
about certain situations. This notion of ‘aboutness’ is what is sometimes referred to as
the content of a representation. In the BBO, the content of burden appraisals is given by
a particular type of INFORMATION CONTENT ENTITY, called REPRESENTATIONAL
INFORMATION CONTENT ENTITY (RICE). This class is taken from the CCO suite, specif-
ically, the Information Entity Ontology [see 14]. Instances of the REPRESENTATIONAL
INFORMATION CONTENT ENTITY class are used to connect burden appraisals to partic-
ular situations or states-of-affairs. In the case of the APPRAISAL OF BURDEN OBJECT
class, this situation is what we understand to be the object of burden; i.e., a situation in
which an individual is obliged to participate in a particular process. In the BBO, the ob-
ject of burden is represented by instances of the SITUATION class, which is asserted as
a subtype of the (top-level) BFO ENTITY class.10 The ‘aboutness’ of burden appraisals is
then established via the use of two relations: concretizes (which connects a burden ap-
praisal to a REPRESENTATIONAL INFORMATION CONTENT ENTITY) and isAbout
(which connects a REPRESENTATIONAL INFORMATION CONTENT ENTITY to a
SITUATION). The elements of a SITUATION are then specified via a custom BBO relation,
namely, hasSituationPart. In the case of appraisals relating to the object of burden
(i.e., APPRAISAL OF BURDEN OBJECT), we assume that the relevant SITUATION
will consist of at least one PROCESS. We then rely on participatory relationships (e.g.,
participatesIn, agentIn, isAffectedBy, etc.) to represent the involvement of
continuant entities in the PROCESS.11 One such (continuant) entity is the individual that
is obliged to participate in the PROCESS. This individual is what we earlier referred to as
the subject of burden (see Section 2). If the subject of burden is the individual entertaining
burden-related cognitions (i.e., the same individual who is appraising a situation as burdensome),
then we have a form of self-referential cognition—i.e., a state-of-affairs in which the individual
appraises themselves as being the subject of burden.
10
In the current version of the BBO, a situation is defined as “. . . a Portion of Reality that refers to a configuration of
entities.”
11
Note that processes can have multiple participants. The process of carrying a backpack, for example, would have
both the individual who is tasked with carrying the backpack and the backpack itself as participants in the carrying
process.
:Affective participatesIn :Feeling of
Representation Burden
hasOccurrentPart
inheresIn
participatesIn :Treatment
Burden
hasOccurrentPart
bearerOf :Appraisal of isOutputOf :Appraisal
:Human Being
Burden Process
hasContinuantPart hasContinuantPart
:Appraisal of
:Appraisal of Burden Object
Burdensomeness
bearerOf concretizes concretizes
:RICE :RICE
:Patient Role
isAbout isAbout
:Situation :Situation
hasSituationPart hasSituationPart
:Treatment
:Negative Effect
participatesIn Process isCauseOf
Figure 4: Using the BBO to represent a case of treatment burden. Blue arrows indicate relations from
the Modal Relation Ontology (see main text for details).
An example of this representational scheme is depicted in Figure 4. In particular, Figure 4
shows how the BBO can be used to model a case of treatment burden. The subject of burden
in this figure is represented by :HUMAN BEING. :HUMAN BEING is the bearerOf a :PATIENT
ROLE, which establishes the status of :HUMAN BEING as a patient. In addition to :PATIENT
ROLE, :HUMAN BEING is also the bearerOf an :APPRAISAL OF BURDEN, which is produced
by :APPRAISAL PROCESS.12 The :APPRAISAL OF BURDEN OBJECT concretizes a :RICE,
which isAbout a :SITUATION. Within this :SITUATION, we find a :TREATMENT PROCESS
that involves :HUMAN BEING as a participant. Figure 4 thus depicts a state-of-affairs in which
a patient (i.e., an individual who is the bearer of a patient role) appraises a situation as being
burdensome, and this situation is one that features the patient as a participant in a treatment
process. This establishes :HUMAN BEING as the subject of (treatment) burden, while the object
of burden is the :SITUATION that is being appraised as burdensome.13
12
For reasons of simplicity, we have represented burden appraisals as inhering in a human individual. Readers
should, however, note that the more correct characterization is one of burden appraisals (qua mental qualities)
inhering in some part of the cognitive system of a human individual. See Limbaugh et al. [13], for more on this.
13
Figure 4 also shows how the subjective experience of burden (or the feeling of burden) is represented within the
BBO. Here, the BBO follows the strategy adopted by the MFOEM. Specifically, the feeling of burden is represented
by the FEELING OF BURDEN class, which specializes the MFOEM SUBJECTIVE AFFECTIVE FEELING
(a) Appraisal of Treatment Burden (b) Treatment Burden
Figure 5: The characterization of APPRAISAL OF TREATMENT BURDEN and TREATMENT
BURDEN in the Protégé ontology editor.
Note that some of the relations depicted in Figure 4 are rendered in blue font (e.g., isAbout,
hasSituationPart, participatesIn, and so on). These relations are so-called modal re-
lations, which form part of the CCO, specifically, the Modal Relation Ontology [see 14]. The
use of these modal relations enables the BBO to refer to non-actual states-of-affairs, as when
an individual who is obliged to water their neighbor’s plants appraises themselves as being
burdened, even though they have not (as yet) begun to water the plants. In short, the use of
modal relations enables the BBO to refer to expected (i.e., future) states-of-affairs. At present, it
remains unclear whether all burden appraisals are about future states-of-affairs; nevertheless,
we suspect that modal relations will be a common feature of burden modeling efforts. This is
not to say that the past (or present) is irrelevant to our understanding of burden; it is merely to
note that burden appraisals are often tied to things that will happen in the future (i.e., beyond
the present moment).14
While one can explicitly denote burden processes as being of a particular type (as is the case
in Figure 4), it is also possible to rely on the semantic description of classes within the BBO to
support the automatic classification of burden instances via subsumption reasoning. Figure 5
shows the logical expressions that support subsumption reasoning in respect of treatment burden.
In this case, an APPRAISAL OF BURDEN is classified as an APPRAISAL OF TREATMENT
BURDEN based on the type of PROCESS that features as part of the SITUATION that is
being appraised as burdensome (see Figure 5a). This classification then serves as the basis for
classifying a (generic) BURDEN PROCESS as a particular instance of TREATMENT BURDEN
(see Figure 5b).
4. Conclusion
In recent years, the assessment and management of biomedical burdens has emerged as an
important focus area for both scientific research and clinical practice. As a means of supporting
this effort, the present paper describes an ontology of biomedical burdens—the BBO—that
uses BFO as an upper-level ontology. The ontology provides support for the representation
class (see Figure 1).
14
See Cassell [18] for a similar (future-oriented) approach to suffering.
of burden-related cognitions, the classification of burden types, and the representation of
burdensome situations. It also demonstrates the way in which recent information modeling
practices, specifically those associated with the CCO [14], can be applied to the modeling of
cognitive/affective phenomena.
Future work will seek to extend and refine the BBO. One of the aims for future work is to
extend the scope of the BBO, providing support for the representation of entities that feature as
part of research into burden (examples include health literacy, demographic factors, education
background, early life experiences, and so on). A second focus area for future work relates
to the representation of data emanating from scientific studies of burden. This includes the
data emanating from survey instruments such as the Patient Experience with Treatment and
Self-management (PETS) [19]. Finally, we aim to use the BBO to support the data-driven analysis
of burdensome health conditions, combining the BBO with ontologies that specialize in the
representation of diseases, symptoms, treatments, and other types of patient data.
Acknowledgments
This paper is independent research funded by the National Institute for Health Research Artificial
Intelligence for Multiple Long-Term Conditions (AIM) programme (Multidisciplinary Ecosystem
to study Lifecourse Determinants and Prevention of Early-onset Burdensome Multimorbidity
(MELD-B), award identifier NIHR203988). The views expressed in this publication are those of
the author(s) and not necessarily those of the NHS, the National Institute for Health Research
or the Department of Health and Social Care.
References
[1] J. Downes, Illness in the chronic disease family, American Journal of Public Health and
the Nations Health 32 (1942) 589–600.
[2] C. R. May, D. T. Eton, K. Boehmer, K. Gallacher, K. Hunt, S. MacDonald, F. S. Mair, C. M.
May, V. M. Montori, A. Richardson, Rethinking the patient: using Burden of Treatment
Theory to understand the changing dynamics of illness, BMC Health Services Research 14
(2014) 1–11.
[3] M. Bastawrous, Caregiver burden—A critical discussion, International Journal of Nursing
Studies 50 (2013) 431–441.
[4] A. Sav, M. A. King, J. A. Whitty, E. Kendall, S. S. McMillan, F. Kelly, B. Hunter, A. J. Wheeler,
Burden of treatment for chronic illness: a concept analysis and review of the literature,
Health Expectations 18 (2013) 312–324.
[5] R. Arp, B. Smith, A. D. Spear, Building Ontologies with Basic Formal Ontology, MIT Press,
Cambridge, Massachusetts, USA, 2015.
[6] R. H. Scheuermann, W. Ceusters, B. Smith, Toward an ontological treatment of disease
and diagnosis, Summit on Translational Bioinformatics 2009 (2009) 116–120.
[7] J. Hastings, W. Ceusters, M. Jensen, K. Mulligan, B. Smith, Representing mental functioning:
Ontologies for mental health and disease, in: 3rd International Conference on Biomedical
Ontology: Towards an Ontology of Mental Functioning Workshop, Graz, Austria, 2012, pp.
1–5.
[8] J. Hastings, W. Ceusters, B. Smith, K. Mulligan, The emotion ontology: Enabling interdisci-
plinary research in the affective sciences, in: M. Beigl, H. Christiansen, T. R. Roth-Berghofer,
A. Kofod-Petersen, K. R. Coventry, H. R. Schmidtke (Eds.), Modeling and Using Context,
volume 6967, Springer, Karlsruhe, Germany, 2011, pp. 119–123.
[9] B. Smith, W. Ceusters, Aboutness: Towards foundations for the Information Artifact
Ontology, in: F. M. Couto, J. Hastings (Eds.), 6th International Conference on Biomedical
Ontology, volume 1515, CEUR Workshop Proceedings, Lisbon Portugal, 2015, pp. 1–5.
[10] B. Smith, P. Grenon, The cornucopia of formal-ontological relations, Dialectica 58 (2004)
279–296.
[11] G. Pezzulo, F. Rigoli, K. J. Friston, Hierarchical active inference: A theory of motivated
control, Trends In Cognitive Sciences 22 (2018) 294–306.
[12] D. Limbaugh, D. Kasmier, W. Ceusters, B. Smith, Warranted diagnosis, in: A. D. Diehl,
W. D. Duncan, G. Sanso (Eds.), 10th International Conference on Biomedical Ontology,
volume 2931, CEUR Workshop Proceedings, New York, New York, USA, 2019, pp. 1–10.
[13] D. Limbaugh, J. Landgrebe, D. Kasmier, R. Rudnicki, J. Llinas, B. Smith, Ontology and
cognitive outcomes, Journal of Knowledge Structures and Systems 1 (2020) 3–22.
[14] CUBRC, Modeling Information with the Common Core Ontologies, Technical Report,
CUBRC, Inc., Buffalo, New York, USA, 2020.
[15] S. Demain, A.-C. Goncalves, C. Areia, R. Oliveira, A. J. Marcos, A. Marques, R. Parmar,
K. Hunt, Living with, managing and minimising treatment burden in long term conditions:
A systematic review of qualitative research, PLoS One 10 (2015) e0125457.
[16] D. T. Eton, R. T. Anderson, J. L. St Sauver, E. A. Rogers, M. Linzer, M. K. Lee, Longitudinal
trajectories of treatment burden: A prospective survey study of adults living with multiple
chronic conditions in the midwestern United States, Journal of Multimorbidity and
Comorbidity 12 (2022) 1–14.
[17] L. Jansen, Tendencies and other realizables in medical information sciences, The Monist
90 (2007) 534–554.
[18] E. J. Cassell, The nature of suffering and the goals of medicine, New England Journal of
Medicine 306 (1982) 639–645.
[19] D. T. Eton, K. J. Yost, J.-s. Lai, J. L. Ridgeway, J. S. Egginton, J. K. Rosedahl, M. Linzer, D. H.
Boehm, A. Thakur, S. Poplau, Development and validation of the Patient Experience with
Treatment and Self-management (PETS): A patient-reported measure of treatment burden,
Quality of Life Research 26 (2017) 489–503.