Functional Analysis to Drive Research and Identify Regulation
Requirements: An Example with a Lithium Monitoring Device
K. Charrière
1a
, C. L. Azzopardi
2b
, M. Nicolier
1,3 c
, T. Lihoreau
1d
, F. Bellivier
4,5,6,7 e
,
E. Haffen
1,3,8 f
and B. Wacogne
1,2 g
1
Centre Hospitalier Universitaire de Besançon, Centre d’Investigation Clinique, INSERM CIC 1431, 25030,
Besançon Cedex, France
2
FEMTO-ST Institute, Univ. Bourgogne Franche-Comte, CNRS, 15B avenue des Montboucons, 25030 Besancon Cedex,
France
3
Department of Clinical Psychiatry, CHU de Besançon, EA 481 Neurosciences, University Bourgogne Franche-Comté,
25030, Besançon Cedex, France
4
AP-HP, GH Saint-Louis, Lariboisière, F. Widal, Department of Psychiatry and Addiction Medicine,
75475 Paris cedex 10, France
5
Inserm, U1144, Paris, F-75006, France
6
Université Paris Descartes, UMR-S 1144, Paris, F-75006, France
7
Université Paris Diderot, Sorbonne Paris Cité, UMR-S 1144, Paris, F-75013, France
8
FondaMental Foundation, Creteil, Hôpital Albert Chenevier, Pôle Psychiatrie, 40 rue de Mesly, 94000 Créteil, France
{emmanuel.haffen, bruno.wacogne}@univ-fcomte.fr, frank.bellivier@inserm.fr
Keywords: Functional Analysis, Medical Device, Lithium Monitoring.
Abstract: Medical device development is often understood as a linear process with design stages occurring sequentially.
First stages are usually performed in order to specify the future device definition through interviews/meetings
of the end-users, researchers and manufacturers. Because the medical device is original, these first stages
mainly involve end-users and researcher. However, regulation constraints and economic reality sometimes
makes manufacturers hesitant to base the industrial development on this initial basis. Functional analysis, well
known by manufacturers, is a method used to accurately define the final functions of a medical device. In this
conference, we estimate that the functional analysis can be put to profit in a more efficient way if researchers
and end-users get familiar with it prior to the interview/meeting stages. Although the results of such
knowledge democratisation is not demonstrated here, we present the function analysis conducted on a lithium
monitoring device according to this multidisciplinary approach. We also show that function analysis can be
used not only to drive research actions but also to identify regulation requirements.
1 INTRODUCTION
Research and development actions in technologies for
health are usually driven by discussions and
experience exchanges between practitioners,
researchers and industrial partners. Because the need
a
https://orcid.org/0000-0003-4542-8003
b
https://orcid.org/0000-0003-2147-2042
c
https://orcid.org/0000-0001-5135-5109
d
https://orcid.org/0000-0001-8417-6609
e
https://orcid.org/0000-0002-3660-6640
f
https://orcid.org/0000-0002-4091-518X
g
https://orcid.org/0000-0003-1490-5831
to be addressed involves innovations that have never
been studied before, first discussions are often led by
practitioners and researchers with an academic point
of view. However, in the case of medical devices
developments, the main goal is neither to increase
knowledge nor to invent new technologies. The goal
is to answer the need as quickly as possible while
300
Charrière, K., Azzopardi, C., Nicolier, M., Lihoreau, T., Bellivier, F., Haffen, E. and Wacogne, B.
Functional Analysis to Drive Research and Identify Regulation Requirements: An Example with a Lithium Monitoring Device.
DOI: 10.5220/0010382503000307
In Proceedings of the 14th Inter national Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2021) - Volume 1: BIODEVICES, pages 300-307
ISBN: 978-989-758-490-9
Copyright
c
2021 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
proposing a new medical device which not only
addresses the need but also meets the manufacturer
interests and end-users wishes.
User driven medical device development is now
widely applied in different collaborative formats likes
Living Labs (Korman, 2016) or Hacking Health
organizations (Chowdhury, 2012). These approaches
have been more or less conceptualized a while ago. In
a well-organized and well-documented paper
(Money, 2011), it is recall that user-centred usability
engineering methods have already been proposed to
improve the Medical Device Design and
Development (MDDD) (Gosbee, 2002). Money and
co-workers describe the MDDD using 4 main
development stages, each of them involving distinct
actions. They can be summarized as follows. Stage 1
identifies the user’s need and involves enquiries and
interviews, both dealing with contextual and usability
issues. This forms the basis for the initial concept
which is further defined during stage 2 which uses
reduced focus groups deepening the initial concept
ideas. This produces requirements documents leading
to stage 3, fully oriented to the device manufacturing.
The device can then be evaluated during stage 4 in
order to check whether or not the users’ requirements
are satisfied. We assume that prototyping actions are
present between stages 2 and 3 or completely
included in stage 3.
In (Money, 2011), what manufacturers think of
the user-driven approach is analysed through
interviews. Instead of trying to rewrite their
conclusion, we simply reproduce the main points they
report. The findings reveal that despite standards
agencies and academic literature offering strong
support for the employment formal methods,
manufacturers are still hesitant due to a range of
factors including: perceived barriers to obtaining
ethical approval; the speed at which such activity
may be carried out; the belief that there is no need
given the ‘all-knowing’ nature of senior health care
staff and clinical champions; a belief that effective
results are achievable by consulting a minimal
number of champions. Furthermore, less senior
health care practitioners and patients were rarely
seen as being able to provide valuable input into the
process”.
More generally, what is mentioned above also
applies to researchers who often misunderstand both
the manufacturer economic constraints and
patients/practitioners possibilities. Indeed, some
technical or scientific solutions can be incompatible
with a practical use or they can lead to an over-
expensive device. In these 2 cases other and more
realistic solutions must be explored, otherwise,
conditions to be safe, efficient and to match with the
local regulation will be difficult.
Indeed, since the very beginning of the MDDD,
discussions between end-users, researchers and
manufacturers must be somehow oriented/guided
towards realistic solutions. For this a conceptual tool
must be employed and this tool should ideally be
known (even partially) by all the stakeholders.
Functional analysis is a technology design tool
already well known by manufacturers but very rarely
by researchers and the reduced number of end-users
interviewed during the above mentioned stage 2. We
believe that functional analysis should be
pedagogically presented to end-users and researchers
prior to any enquiry. This pedagogical effort should
be clearly codified to allow beginners quickly
understanding main issues of this method and how it
can be used to drive the research and development
actions and to identify the corresponding regulation
issues. In this way interviews conducted to set-up the
initial prototype design can be focused on functional
analysis key points and the “expert” users can be
efficiently involved in stage 3 of the MDDD. At the
end, we think that what was called “end-users’
wishes” at the beginning of this introduction will
become “end-users’ highlights”, that researchers will
directly focus on realistically translatable solutions
and that manufacturers will be less hesitant to these
data into account.
In this conference, we do not present a
retrospective study on how this pedagogic effort is
put to profit to enhance MDDD, but we describe the
functional analysis which has been conducted in the
frame of the H2020 R-LiNK project (grant agreement
n°754907) according to this multidisciplinary
concept. The goal is to present how this analysis
accounts for end-users’ requirements, how it drives
the research actions to be privileged and how it allows
identifying regulation requirements. In the next
section, we rapidly presents the main goals of the R-
LiNK project. The functional analysis is described in
section 3. A short discussion is then proposed in
section 4 before some conclusive remarks are given.
2 THE R-LiNK PROJECT
The consortium of this European project led by Pr.
Franck Bellivier (INSERM UMR-S1144) is
composed of 22 European partners including research
institutes, hospitals, clinical investigation centers and
companies.
The main objective of R-LiNK is to identify the
eligibility criteria for treatment with lithium in bipolar
Functional Analysis to Drive Research and Identify Regulation Requirements: An Example with a Lithium Monitoring Device
301
disorder type 1 (BD1) patients in terms of response,
safety and tolerability. Research actions conducted to
find lithium response biomarkers involve
multidisciplinary research fields like “omics”
investigations, blood analyses, nuclear magnetic
resonance imaging and activity assessment. The
functional analysis presented in this paper is related
to treatment adherence as explained below.
2.1 Origin of the Project
Bipolar disorders (BD) are prevalent mental disorders
and a leading cause of suicide. Bipolar disorders are
lifelong lasting, with an episodic course of the illness
in most cases. Mood stabilizers are the mainstay of
treatment of BD and lithium is the gold standard
(Sani, 2017).
Indeed, a substantial minority of individuals
remain asymptomatic for years on lithium (about
20%) but most show only partial response and up to
one third do not respond (Burgess, 2001).
Furthermore, in current clinical practice, lithium
exposure is poorly controlled using laboratory tests.
First, it is usually verified only once or twice a year.
Second, adherence to chronic treatment is known to
be poor. Meanwhile, the prescription of lithium
remains delicate. For the treatment to be effective,
serum concentrations 10-14 h after the last dose taken
must reach 0.5 to 1.0 mEq/L. If plasmatic
concentration exceeds 1.2 mEq/L, toxic effects are
likely to occur (Amdisen, 1967; Baldessarini, 2013;
Bauer, 2016; Tondo, 2019).
Therefore, there is an important medical need to
first, provide the patients/practitioners with a simple
tool which could allow patients to become actor of
her/his treatment, hence improving the adherence to
treatment and second, to increase the frequency with
which the lithium level can be assessed in a non-
invasive manner. To this end, the device is intended
to be used at home by patients, so it’s a class C in vitro
medical device, according the European regulation
(EU 2017/746). The idea is not to replace laboratory
lithium dosing techniques but to create a simple
home-based lithium level indicator so that the patient
can check if her/his lithium level is below, within or
above the therapeutic window. A rapid analysis of the
home-based usability possibilities led us to consider
the detection of lithium in saliva. Also, because it will
be used at home by patients or by non-specifically
trained medical staff, the device must be carefully
designed and a complete functional analysis must be
performed.
3 FUNCTIONAL STATEMENT OF
THE NEED
Functional specifications documents are well
codified, even if different methodologies can be used
to write them. They can be separated in three parts:
need analysis
functional analysis
technical specification
Before and during constitution of functional
specifications, studies are carried out to better
understand the needs of potential customers.
3.1 Analysis of the Need
Analysis of the need is an essential phase because it
dictates the direction of the future work. The needs
and the objectives should be clearly identified and
formalized.
To do this, different tools can be used; one of them
is the APTE® method (see for example (APTE, no
date)). This method starts from the expression of a
need, without considering any technical solutions. It
constitutes the first phase of design leading to the
edition of the functional specifications. Three
questions have to be considered.
To whom is the “product” useful?
What does the “product” acts on?
What is the purpose of the “product”?
3.1.1 Expression and Characterization of
the Need for R-LiNK Monitoring
Device
In our project, the “product” is the monitoring device.
To whom is theproduct useful? The R-LiNK
monitoring device is useful to patients suffering from
bipolar disorder and the medical staff. What does the
“product” acts on? It acts on patients’ saliva to assess
lithium levels. What is the purpose of the “product”?
It aims at improving treatment adherence and to avoid
relapse by self-monitoring and self-management.
Furthermore, it aims at minimizing (avoid if possible)
the risk of lithium overdose. This is illustrated in
figure 1.
This first expression of the need is of course
essential but not sufficient. Indeed, the need will be
correctly addressed only if the expected performances
of the device are obtained. Therefore, the need has to
be clearly characterized specifying qualification and
quantification to estimate measurable quantities.
For the R-LiNK monitoring device, the need will
be satisfied if the device allows lithium detection
ClinMed 2021 - Special Session on Dealing with the Change in European Regulations for Medical Devices
302
between 0.5 and 4 mM, with 0.2 mM accuracy and
improve treatment adherence, among others.
Figure 1: R-LiNK monitoring device: expression of the
need.
3.1.2 Validation of the Need
Here, the questions to be answered are defined as
follows. What is the origin of the need or why is the
device wanted? What can change the need or even
makes the need disappear?
The need of detecting lithium with a simple
monitoring device comes from practitioner facing
some clinical issues.
First, as described above, 30-55% of individuals
selected for treatment with lithium will not have the
predicted outcome. One possible cause to the non-
efficacy of the treatment is the non-adherence to
treatment. Second, lithium can be toxic and the
therapeutic window is limited, between 0.5 and 1.2
mM in plasma with toxic effects above the maximum
value.
These two issues will not disappear but
technological or pharmaceutical developments may
change clinical practices. One major evolution has
been identified: emergence of new medicine for
bipolar disorder leading to stopping the use of lithium
therapy.
Taking these elements into account, it is therefore
unlikely that the need for a new lithium monitoring
device disappears completely.
3.2 Establishment of Service Functions
and Constraints
Since the need has been validated, the functional
analysis can continue and a list of functions can be
established. First, it’s important to better understand
what a function is. Looking normative definition
issued from the French Association for
Standardization (AFNOR), a function is an action of
a product or one of its constituents expressed
exclusively in terms of purpose.
Therefore, it is required to disregard technological
solutions. In order to help defining these functions, an
interactions diagram of the device with its
environment can be drawn. The diagram defines the
limits of the device and specifies the life situation
being studied. This is presented in figure 2. In this
figure, bubbles represent the physical elements of the
environment that have impact on the device. Lines
characterize functions linking the system to its
environment.
Figure 2: R-LiNK monitoring device interaction diagram
for the “normal use” life situation.
Two types of functions exist: the main function
and constraint functions. The main function reflects
actions performed by the device, the constraint
functions reflect an adaptation of the device to its
environment. Therefore, the environment has to be
defined for a specified life situation. In our R-LiNK
example, the diagram represents the life situation
“normal use”.
Environment of the R-LiNK monitoring device
consists of:
users
biological sample
user’s place
European standards (IVD MD 2017/746)
European market
Functions are defined as follows.
Main function:
MF1: the device indicates the level of lithium in
the biological sample.
Functional Analysis to Drive Research and Identify Regulation Requirements: An Example with a Lithium Monitoring Device
303
Constraint functions:
CF1: the device is suitable for all users
CF2: the device is suitable for home use
CF3: the device respects the regulations in force
CF4: the device integrates the European market
at a reasonable cost
Note that the market is not a physical element, but
cost is really a constraint to take in account at the
beginning of the development. That’s why we chose
to include this notion in our functional analysis of the
need.
3.2.1 Justifications of Functions
To ensure the function is really necessary, it’s
possible to ask, as for the need: what is the purpose,
the origin and the probability to disappear for each
function?
For the R-LiNK device, the purpose of MF1 is to
improve treatment adherence by self-monitoring and
to minimize or avoid the risk of lithium overdose or
under dosing. New medicine for bipolar disorder
leading to the stopping lithium therapy can make the
function disappear. Eliminating the function is
unlikely. The function is validated.
The purpose of CF1 is to meet the expectations of
user comfort whatever the patient's physical condition
is because: the patient can use the product easily and
quickly so they can monitor themselves without
demotivation, the patient is looking for effective
products that provide him with a certain comfort of
use. Eliminating the function is unlikely. The
function is validated.
The purpose of CF2 is to meet the expectations of
user comfort whatever the place of use because
patients have to regularly monitor themselves at home
or at their place of holiday, during their work travels.
Eliminating the function is unlikely. The function is
validated.
The purpose of CF3 is to be able to commercialize
the device and to assure the user that the product has
been approved and that he can therefore use. The
products sold must be certified by standards of quality
and safety. Eliminating the function is unlikely. The
function is validated.
The goal of CF4 is to enable the use of the device
throughout the European Union at a reasonable cost
because all patients are concerned and reimbursement
rules are not identical for the entire European
community. Eliminating the function is unlikely. The
function is validated.
3.2.2 Functional Analysis and Usability
At this stage of the functional analysis, details can be
provided to really meet the end-users’ need. It is often
made by brainstorming or interview with the various
stakeholders. In our project, these investigations led
to the conclusions given below.
The device must measure lithium quickly (5
minutes if patients have to wait for results, more
if the result is recorded).
The device must to detect lithium in the
therapeutic window (ideally 0.5 mM to 5 mM,
with an accuracy of 0.2 mM in saliva).
The device must be non-invasive (use of saliva
instead of blood).
The device has to deliver the results in an
understandable manner.
The device must be easy to use, compact and
mobile.
Finally, the device must be easily stored after
use.
3.2.3 Characterization of Service Functions
and Constraints
Based on all these indications, the functions are then
defined with criteria, levels, and tests to be
performed. This is summarized in table 1.
Note that in table 1, we inserted a column titled
“flexibility”. This estimates how negotiable can be
the results expected in columns “criteria” and
“levels”. Flexibility is defined as follows:
F0 means zero flexibility, imperative level
F1 means low flexibility, little negotiable level
F2 means good flexibility, negotiable level
F3 means strong flexibility, very negotiable
level.
After this functional analysis of the need,
technological solutions can be envisaged. At this
stage, it is essential to think about the solutions
objectively and without restrictions. To do this, a
Functional Analysis System Technique (FAST) is
useful (FAST1, FAST2, FAST3, no date). Among
other methods, the use of FAST allows a highly
multidisciplinary consortium to speak the same
language. Structured analysis and design technique,
also termed SADT (Ross, 1985), may be used too, but
we think it is more adapted when the technical
solution is already chosen.
ClinMed 2021 - Special Session on Dealing with the Change in European Regulations for Medical Devices
304
Table 1: Definition of the functions.
Function Criteria / standard Levels Flexibility Tests
MF1
Monitor the level of salivary
lithium
Indicate the level of lithium
in saliva comprehensively
for the patient
0,5 mM to 5 mM
100 µL saliva
Without
contamination
Accuracy 0.2 mM
F1
Performance tests in the
laboratory on artificial samples
then on real calibrated samples.
Tests in real conditions of use
CF1 Suitable for all users
The patient easily uses the
device without errors
Pass the aptitude
to use tests
F0
Formative and summative
evaluation tests
CF2 Suitable for home use
The device works in the
patient's environment:
home, holidays
Pass the aptitude
to use tests
F0
Formative and summative
evaluation tests
CF3
Respect the regulations in
force
Complies with the new
European IVMD regulation
(auto-monitoring)
UE 2017/746 F0 Respect the regulation in force
CF4
Integrate the European
market at a reasonable cost
The cost is not a hindrance
for the patient or the
institutions
Consumable 10 €
Cartridge reader
300 €
F1
F2
Estimated costs by item,
including raw materials and an
estimate of manufacturing costs
FAST allows, among other things, organizing and
understanding relationship between functions. It
helps identifying missing or redundant functions. All
functions can be analyzed in the same way. The
method consists in asking, for each of them: why do
you do [Function], how do you do [Function] and
when you do [Function] is there another function that
occur together with or as a result of [Function]? This
can be formalized using the codified diagram
presented in figure 3.
Figure 3: General principle for establishing a FAST
diagram.
The FAST diagram corresponding to the device
developed in the frame of the R-LiNK project is
presented in figure 4. In this figure, the first column
(on the left hand side) corresponds to functions.
Identification of these functions should ideally take
place during the stage 1 of the MDDD described in
the introduction of this document. The identification
of the items presented in the second columns should
be made during stage 2. The third and fourth columns
should be filled during the end of stage 2 or the
beginning of stage 3 depending on where the
prototyping actions are performed. The idea of this
conference is not to go through each items depicted
in figure 4. However, we shortly describe how the
main function analysis lead to “usability” solutions.
The MF1 function is analyzed as follows.
In order to assess the lithium level in the
biological sample it is necessary to put the
saliva in contact with the sensor’s reagents.
For this, the user must be able to collect the
saliva.
In order to collect the saliva, different
technical solutions can be used: use an existing
saliva sampling device (Salivette® type) or
spit in a recipient.
Here, we recall that the scientific or technological
solutions are listed in column 4 but the final choice is
not yet made. This is because, in the frame of our
functional analysis, the goal is not yet to select the
final solutions but to identify all of them as we already
mentioned above while writing that we should
disregards technological solutions.
Indeed, when the FAST diagram is finalized,
practical work can start. This is the prototyping phase
which should take place during stage 2 or 3 of the
MDDD method already mentioned. Choice of the
technical solutions or specialty areas where research
efforts should be put can now be finalized in
accordance with the end-users’ “highlights” and
regulation constraints. Because our device is intended
to be used in the European community, the regulation
documents listed in figure 4 correspond to a part a part
of required standards to comply with the European
Regulation.
Functional Analysis to Drive Research and Identify Regulation Requirements: An Example with a Lithium Monitoring Device
305
4 A BIT OF DISCUSSION
We have seen that the functional analysis allows
defining what a medical device should be in order to
meet requirements of end-users, researchers and
manufacturers in accordance with the regulation
constraints. The FAST diagram summarizes these
aspects. Including this diagram in the general MDDD
plane proposed elsewhere, we understand that the
functional analysis takes place at the beginning of the
development, mainly during stages 1 and 2.
In order to be efficiently conducted, the functional
analysis must involve all the stakeholders which
includes not only manufacturers who already know
this kind of analysis but also end-users and
researchers. For this, a pedagogic effort must be made
which is not in the scope of this conference.
However, if we only stick to this understanding,
there is a risk that the functional analysis is
considered as a background task which extends over
the beginning of the MDDD process. The
consequence could likely be that important
information highlighted during the functional
analysis are under-estimated during the prototyping
phase, if not simply forgotten. It is therefore crucial
that actors try to consider the development of a
medical device not like a process linear in time but as
a whole.
The linear perception of technological
developments is nowadays probably due to the
importance that the TRL scale has gained these last
decades. This “Readiness Technology Level” scale
was originally proposed by the NASA to improve
technology developments. However, the NASA
proposed a more general and less linear development
model with the so-called CML scale, namely the
“Concept Maturity Level” scale (Ziemer, 2013).
Readers interested in this method can refer to (CML1,
CML2, no date). The CML scale has recently been
adapted to the medical device development (Béjean,
2019). The development is now not only described
but also driven through 3 main dimensions: need,
science and technology, programmatic.
We think that the functional analysis can
efficiently be used to link the above mentioned 3-
dimensional development descriptions.
Figure 4: FAST diagram obtained for the home-based lithium monitoring device of the R-LiNK project.
ClinMed 2021 - Special Session on Dealing with the Change in European Regulations for Medical Devices
306
5 CONCLUSION
In this paper, we have presented how function
analysis can be used to drive research and
development actions and to identify regulation
constraints during the development of medical
devices. An example of functional analysis conducted
to design a home-based lithium monitoring device is
given. It is shown that function must be identified and
characterized and that technological solutions and
regulation constraints arise from this analysis. At this
stage, scientific or technological techniques used in
the final medical device are not yet chosen. They will
be chosen during the subsequent prototyping phase
according to the identified regulation constraints.
But, beyond the only description of a functional
analysis, we pointed out that an efficient design of a
medical device implies controlled discussion between
end-users, researchers and manufacturers. In order to
ensure these fruitful discussions and exchanges of
experience, a common innovation frame must be
adopted. The idea here is that functional analysis can
be this common frame to the condition that it is
pedagogically explained and presented to
stakeholders less familiar with it than manufacturers.
Functional analysis can then be regarded as a
common thread in the design and development
process.
This methodology is currently applied to the
lithium monitoring device developed in the frame of
the H2020 R-LiNK project and results of this
development will be available soon.
ACKNOWLEDGEMENTS
R-LiNK has received funding from The European
Union’s Horizon 2020 Research and Innovation
program Under Grant Agreement N° 754907.
REFERENCES
Amdisen, A., 1967. Serum Lithium Determinations for
Clinical Use, Scandinavian Journal of Clinical and
Laboratory Investigation, Vol. 20, pp. 104-108. DOI:
10.3109/00365516709076929
APTE, http://methode-apte.com/
Baldessarini, R. J., 2013. Chemotherapy in Psychiatry, third
edition, Springer, ISBN 978-1-4614-3710-9
Bauer, M., 2016. The Essential Guide to Lithium
Treatment, Springer, ISBN 978-3-319-31214-9
Béjean, M. and Siqueira, E., 2019. Organizing Medtech
Innovation with Concept Maturity Levels. Proceedings
of the 12th International Joint Conference on
Biomedical Engineering Systems and Technologies -
Volume 5: ClinMed, pp. 621-626. DOI:
10.5220/0007697006210626
Burgess, S., Geddes, et al., 2001. Lithium for maintenance
treatment of mood disorders. The Cochrane Database of
Systematic Reviews, Vol. 3, CD003013. doi.org/
10.1002/14651858.CD003013
Chowdhury, J. 2012. Hacking Health: Bottom-up
Innovation for Healthcare. Technology Innovation
Management Review, Vol. 2, pp. 31-35. doi.org/
10.22215/timreview/579
CML1, https://spacepolicyonline.com/images/stories/
JPL_re_Concept_Maturity_Level_Nomenclature.pdf
CML2, https://trs.jpl.nasa.gov/bitstream/handle/2014/444
57/13-4156_A1b.pdf?sequence=1
FAST1, https://fr.wikipedia.org/wiki/Function_analy
sis_system_technique
FAST2, https://www.slideshare.net/baggiojoset/fast-
method
FAST3, https://new.aitriz.org/documents/TRIZCON/
Proceedings/2011-06_FAST-Diagrams-The-Foundati
on-for-Creating-Effective-Function-Models.pdf
Gosbee J., 2002. Human factors engineering and patient
safety. Qual Saf Health Care. Vol. 11, pp.352-354. doi:
10.1136/qhc.11.4.352. PMID: 12468696; PMCID:
PMC1758019.
Korman, M., Weiss, P. L. et al., 2016. Living Labs:
overview of ecological approaches for health
promotion and rehabilitation, Disability and
Rehabilitation, Vol. 38, pp. 613-619, doi: 10.3109/09
638288.2015.1059494
Money, A.G., Barnett, J. et al., 2001. The role of the user
within the medical device design and development
process: medical device manufacturers’ perspectives.
BMC Medical Informatics and Decision Making, pp.
11:15. doi:10.1186/1472-6947-11-15
Ross, D.T., 1985. Applications and Extensions of SADT.
Computer, vol. 18, no. 4, pp. 25-34. doi:
10.1109/MC.1985.1662862.
Sani, G., Perugi, G. et al., 2017. Treatment of Bipolar
Disorder in a Lifetime Perspective: Is Lithium Still the
Best Choice?. Clin Drug Investig, Vol. 37, pp. 713–727
.doi.org/10.1007/s40261-017-0531-2
Tondo, L., Alda, M. et al, 2019. Clinical use of lithium salts:
guide for users and prescribers. Int. J. Bipolar Disord.,
Vol. 22, pp.16. doi: 10.1186/s40345-019-0151-2.
PMID: 31328245; PMCID: PMC6643006
Ziemer, J., Ervin, et al., 2013. Exploring Mission Concepts
with the JPL Innovation Foundry A-Team.
In AIAA Space 2013. San Diego, California.
doi.org/10.2514/6.2013-
Functional Analysis to Drive Research and Identify Regulation Requirements: An Example with a Lithium Monitoring Device
307