=Paper=
{{Paper
|id=Vol-2768/paper8
|storemode=property
|title=Reducing the Psychological Burden of Isolated Oncological Patients by Means of Decision Trees
|pdfUrl=https://ceur-ws.org/Vol-2768/p8.pdf
|volume=Vol-2768
|authors=Samuele Russo,Salvatore Ivan Illari,Roberta Avanzato,Christian Napoli
}}
==Reducing the Psychological Burden of Isolated Oncological Patients by Means of Decision Trees==
https://ceur-ws.org/Vol-2768/p8.pdf
Reducing the Psychological Burden of Isolated
Oncological Patients by Means of Decision Trees
Samuele Russoa , Salvatore Ivan Illarib , Roberta Avanzatoc and Christian Napolid
a Sapienza University of Rome, Piazzale Aldo Moro 5, Roma, Italy
b Fondazione Istituto Oncologico del Mediterraneo, Via Penninazzo 11, Viagrande CT, Italy
c Department of Electrical, Electronic and Computer Engineering, University of Catania, Catania, CT, Italy
d Department of Computer, Control, and Management Engineering, Sapienza University of Rome, Via Ariosto 25, Roma, Italy
Abstract
This century has seen several outbreaks of epidemics caused by a common sub-family of coronaviruses such as the responsible
for COVID-19 outbreak. The most ominous variants have developed a peculiar viral mechanisms that allows the virus to
directly attack the pulmonary tissues often causing a set of dangerous symptoms. It made quite evident that we need a global
response to prepare health systems for future epidemics. Unfortunately, during such kind of diseases’ outbreaks a large
amount of time is required to the caregivers for sanitization and cleaning operations, therefore tampering with number and
duration of visits to patients, especially in oncology wards. Such patients are then left alone for a long time, it follows that
their perceived quality of service is greatly diminished, often determining ill-fated consequences also on the psychological
side, with significant fallbacks on the recovery possibilities and speed. In this paper we explore an algorithmic approach to
automatic communication interfaces that could enhance and enforce the perceived quality of care by the patients in in order to
reduce predisposing factors that could potentially tamper with the patient’s ability to recover, also preventing the occurrence
of precipitating factors that could lead a therapy to complete failure. The proposed interface could be used to connect the
patients with a psychological support when it is most needed, and, moreover, to connect them with their physicians and
families, and also to the outside world. In particular we aim to provide the psychological support that is actually excluded in
pandemics such as the COVID-19 emergency, mainly in order to enforce the healthcare and sanification protocols, due to its
potential unsafety related to the introduction of more personnel into the hospital.
Keywords
Psychology, Oncology, Medical Physics, hospitalization, decision support, decision trees.
1. Introduction matter of weeks. The healthcare system’s capacity to
respond has been under enormous pressure, to the point
This century has seen several outbreaks of epidemics that Intensive care specialists had been considering the
caused by a common sub-family of coronaviruses. The possibility to deny life-saving care to the sickest, giv-
most ominous variants have developed a peculiar viral ing priority to patients with better survival chances [2].
mechanism that makes use of the angiotensin-convert- While in several countries such a point of no return
ing enzyme 2 (ACE2). has been trespassed [3]. Such events made quite evi-
Such a mechanism allows the virus to directly attack dent that we need a global response to prepare health
the pulmonary tissues often causing a set of danger- systems for future epidemics. Official numbers of in-
ous symptoms that can be generalized as Severe Acute fected people during the COVID-19 virus outbreak have
Respiratory Syndromes (SARSs). Therefore such virus- been indicative of the spread of the infection, and of
es are characterized by extreme infectivity, rapid spread, the challenges that have been posed to Italian hospitals
and the concrete risk of developing pulmonary syn- and, in particular, intensive care facilities. The enor-
dromes that may require intensive care unit admis- mous demand for handling the COVID-19 outbreak
sion [1]. The spread of severe acute respiratory syn- challenged both the health care personnel and the med-
drome coronavirus 2 (SARS-CoV-2) has taken on pan- ical supply system. The COVID-19 emergency has ex-
demic proportions, affecting over 100 countries in a posed the fragility of many Health Care Systems around
the world. Two major critical factors have been re-
ICYRIME 2020: International Conference for Young Researchers in lated to the management of critical care units as well
Informatics, Mathematics, and Engineering, Online, July 09 2020 as of other wards hosting patients with immunolog-
" samuelerussoct@gmail.com (S. Russo);
salvatore.illari@fondazioneiom.it (S.I. Illari); ical deficiencies such as oncology. COVID-like dis-
roberta.avanzato@phd.unict.it (R. Avanzato); eases are generally transmitted by airborne pathogens
cnapoli@diag.uniroma1.it (C. Napoli) that grant a high contagion rate and rapidity. More-
© 2020 Copyright for this paper by its authors. Use permitted under Creative
over, such pathogens often tamper with the respira-
CEUR
Workshop
Commons License Attribution 4.0 International (CC BY 4.0).
CEUR Workshop Proceedings (CEUR-WS.org)
Proceedings
http://ceur-ws.org
ISSN 1613-0073
tory system causing various lung-related comorbidity.
These affections can also evolve in acute respiratory healthcare and sanification protocols, due to its poten-
syndromes, with variable or uncertain outcomes, such tial unsafety related to the introduction of more per-
as severe pneumonia, that commonly require hospital- sonnel into the hospital.
ization in intensive care. The paper is organized as follows. After this brief
Unfortunately, during such kind of diseases’ out- introduction, in the following Section 2 we discuss the
breaks a large amount of time is required to the care- related works and compare our contribution to the ex-
givers for sanitization and cleaning operations, there- isting literature. In Section 3 we describe the system,
fore tampering with number and duration of visits to its purpose and aim, while we deepen into the algo-
patients. Such patients are then left alone for a long rithm and topology in Section 4. Finally in Section 5
time, it follows that their perceived quality of service we will report the simulation results and draw our con-
is greatly diminished, often determining ill-fated con- clusions.
sequences also on the psychological side, with signifi-
cant fallbacks on the recovery possibilities and speed.
Most hospitals could not maintain their routine work 2. Related Works
due to the disaster-related new procedures. In facts
In literature the quality of services of a healthcare sys-
medical professionals caring for patients with highly
tem is defined as consistently delighting the patient by
infectious diseases such as COVID-19 are at high risk
providing efficacious, effective and efficient healthcare
of contracting such infections. All medical person-
services according to the latest clinical guidelines and
nel involved in the management of potentially infected
standards, which meet the patient’s needs and satisfies
patients must adhere to airborne precautions, hand hy-
providers [4]. Healthcare quality definitions common
giene, and donning of personal protective equipment.
to all stakeholders involve offering effective care that
All aerosol-generating procedures should be done in
contributes to the patient well-being and satisfaction.
an airborne infection isolation room. Double-gloving,
As shown in [5] the perceived health service quality is
as a standard practice at our unit, might provide ex-
an important determinant for health service satisfac-
tra protection and minimize spreading via fomite con-
tion and behavioral intentions.
tamination to the surrounding equipment after intu-
A recent study [6] reported that in general the hos-
bation. All these necessary safety measures come with
pitalized patients, while often lacking the education
an elevated cost, not only on the financial side but also
and knowledge regarding isolation, feels that it im-
on the amount of time and energy required to enforce
proves their care.
such practices, as well as in term of quality of care re-
On the other hand [7] shows that contact isolation
duction for the patients, that are often to be left alone
is associated with adverse effects in patients and lead
for the major part of the day.
to psychological and physical problems, and that hos-
In this paper we explore an algorithmic approach
pitalised patients placed under isolation often showed
to automatic communication interfaces that could en-
a negative impact on their mental well-being and be-
hance and enforce the perceived quality of care by the
haviour, including higher scores for depression, anxi-
patients in in order to reduce predisposing factors that
ety and anger [8].
could potentially tamper with the patient’s ability to
Moreover, as showin in [9], isolated patients are vis-
recover, also preventing the occurrence of precipitat-
ited fewer times than non-isolated patients, moreover
ing factors that could lead a therapy to complete fail-
such isolated patients generally benefit of a shorter
ure. Tumors represent a nefarious event of high im-
time span with their physicians. Because of the signifi-
portance. In fact, cancer always represents, for the
cantly lower contact time observed, particularly among
patient and for his family but also for the treatment
the most severely ill of floor patients, a reexamina-
system, an overwhelming existential test. This test
tion of the risk-benefit ratio of this infection control
concerns all aspects of life: the relationship with one’s
method has been proposed. In facts the attending physi-
body, the meaning given to suffering, illness, death, as
cians are about half as likely to examine patients in
well as family, social and professional relationships.
contact isolation compared with patients not in con-
The proposed interface could be used to connect the
tact isolation [10].
patients with a psychological support when it is most
Similarly, other studies have pointed out the con-
needed, and, moreover, to connect them with their physi-
cern that isolation may negatively affect not only the
cians and families, and also to the outside world. In
perceived quality of service but also the patients’ men-
particular we aim to provide the psychological sup-
tal health [11, 12], with a substantial increase in anx-
port that is actually excluded in pandemics such as the
iety and stress-related disorders [13, 14]. Finally [15]
COVID-19 emergency, mainly in order to enforce the
47
Figure 1: A schematic representation of the developed system’s purpose and application.
shows that isolation precautions are associated with vision to protect patients with cancer remain critical.
adverse effects which may result in poorer hospital Therefore the implementation of a psychological sup-
outcomes, a longer hospitalization, an higher cost of port within such units appear natural, as well as agree-
care, as well as an higher rate of readmission to hos- able. In fact, in [23] it has been shown that during the
pital within a month. The spread of COVID-19 is of COVID-19 outbreak, using online multimedia psycho-
particular concern in this vulnerable population, given educational intervention on perceived stress and re-
the fatality rate and the potentially increased severity silience of patients hospitalized in quarantine had a
of the disease course [16]. For this reason, as stated beneficial impact on the before-mentioned undesirable
in [17], a multitude of precautionary steps are imple- psychological effects.
mented by hospitals, departments of radiation oncol- Differently from other fields of medicine, psychol-
ogy to provide uninterrupted radiation treatment for ogy does not base its protocol on drugs and prescrip-
most patients with cancer amid the current COVID-19 tion, and neither on standard surgical procedures [24,
pandemic. The main care services in several countries 25], on the contrary it build the intervention around
has implemented clinical psychology units to cope with the patients needs starting from standardized proto-
the COVID-19 emergency outbreak. The unit’s main cols. While standardization comes with a price, since
goal has been to support and protect health care pro- it results in a lack of customization for the developed
fessionals, relatives of hospitalized patients, and pa- therapy, it also presents great advantages in terms of
tients themselves from further psychological distress. comparability and results testing among different pa-
Details and insights are shared [18]. tients.
Among such measures an extensive application of Moreover trough standardization the caregivers are
isolation protocols has been applied in onclology and guided in making decisions regarding the more appro-
radiotherapy units [19].Patients with cancer are known priate therapeutic plan for a specific conditions, while
to be at an increased risk for community-acquired res- the medical practices can be rationalized improving,
piratory viruses, such as influenza, because of their in the end, the general outcome for the therapy at full
frequently observed immunocompromised state [20]. advantage of the patient’s well being. Other fields of
Unfortunately for cancer patients the psychologi- medicine can rely on very effective clinical prediction
cal burden of isolation is heavier with respect to iso- rules in order to reduce the uncertainty inherent the
lated general medicine patients [21]. As pointed out medical practice by defining how to use clinical find-
in [22], radiation oncology clinics have always func- ings to make predictions [26]. Finally, it must be said
tioned as an interdisciplinary team of support staff, that in certain cases it is uttermost difficult to draw
nurses, therapists, dosimetrists, physicists, and physi- methodology-proof clinical practice guidelines due to
cians, all aiming to help patients with cancer. Heading the extreme statistical and subjective variability of the
into the fight with COVID-19, that team nature and matter at hand [27].
48
3. Purpose of the developed and radiation producing machines) used in a radiation
oncology program, enforce the radiation safety pro-
system gram (possibly shared with an institution’s radiation
Finally relevant focus should highlight the situation safety officer), focuses on the physical aspects of pa-
of patients with cancer may have compromised im- tients’ treatments and interacts with the medical com-
munity due to their malignancy and/or treatment, and munity.
may be at elevated risk of severe COVID-19. Com- The main objective of the treatment of the cancer
munity transmission of COVID-19 could overwhelm patient must be to improve the quality of life and to
health care services, compromising delivery of cancer limit the risk of psychopathological consequences such
care. This interim consensus guidance provides advice as to affect the future life of the patient and his family.
for clinicians managing patients with cancer during Social support therefore represents a constitutive ele-
the pandemic. [28]. Among the experts that take care ment of the treatment of the cancer patient and falls
of the patient with cancer a peculiar figure is consti- within the responsibility of each therapeutic figure.
tuted by the medical physicist: a specialist who ap- The adaptation to the disease and to the treatments
plies the principles and methods of both physics and depends largely on the quality of the relational ap-
medicine, focusing on the areas of prevention, diagno- proach of the treating team, which is the author above
sis, and treatment, as well as ensuring quality services all through the control of the side effects of the thera-
and prevention of risks to the patients, and members pies, the control of pain, anxiety and depressive symp-
of the public in general. Unfortunately the work of toms. This is possible through an individualized care
the MP, as well as the other oncology team members, of the patient, through information on the various as-
has been tremendously affected by the COVID-19 out- pects of the pathology as well as through the evalua-
break. In fact the MP shares the responsibility to plan tion of his needs, his possibilities of choice, his family
the radiotherapy and radiosurgery intervention also and social situation.
for patients with potentially compromised immunity The psychological and relational dimension repre-
system. sents an element of particular importance in oncology.
As it will be shown in the following, the psychologi- In fact, the carers must from time to time be able to tol-
cal and emotional status of the patients it is paramount erate and contain the emotional and affective reactions
to determine the therapeutic outcome, and, often, this of patients and their families on a daily basis, develop-
is strongly affected by the isolation protocols that de- ing a particular sensitivity with respect to the percep-
prive the person of human contact and relationships. tion of signs of discomfort and the inherent limits in
Therefore in this work we explore the development the possibility of adaptation of the patient himself to
of a decision three for oncology patients deployed from the disease.
the collaboration of different figures such as computer Following the dramatic COVID-19 pandemic, the ac-
scientists, psychologists and radiation oncology physi- tivities of the operators in oncological radiotherapy
cist. The first responsibility of the radiation oncology department have been extensively remodelled, so as
physicist is to the patient, trying to assure the best pos- to ensure greater safety for the entire staff operating
sible treatment given the state of technology and the in the facility. First of by applying social distancing,
skills of the other members of the radiation oncology equipping the staff equipped with personal protective
department. A radiation oncology physicist brings a equipment, installing sanitizing gel dispensers in ev-
unique perspective to the clinical team in a radiation ery hallway and waiting room, but also determining a
oncology program: he shows his abilities as a scientist maximum limit of two people at the same time in the
who trained in physics, including radiological physics, same room.
and also in clinical, basic medical, and radiobiologi- Even in the waiting room of an oncology depart-
cal sciences. The physicist performs an important role ment, patients are subjected to limitations, in order to
working along with the radiation oncologist, the ra- maintain the correct social distancing. Their relatives
diotherapy technologist and others, to assure the ac- have to wait upstairs, thus avoiding further gather-
curate delivery of all aspects of a treatment prescrip- ings, and sometime they are not admitted in the ward.
tion. In radiation oncology, physicists have the pri- The effect of these necessary limitations is to increase
mary responsibility for the following for planning the the isolation effect on the oncological patient.
resource allocation with radiation oncologists, admin- Therefore, while the patient follows a cure protocol,
istrators, and technologists, takes care of the physical he must also be helped, with the same accuracy, by
aspects of all radiation sources (radioactive materials means of a parallel protocol that takes care also of the
solitude experienced by the person. In the following
49
Has benign paroxysmal positional vertigo, and vestibular
company neuritis. In [40] the authors present the results of a
prospective, cross-sectional study involving patients
with acute headache and demonstrate that their best
NO
PRIORITY
Is bedside decision rule identified all cases of subarach-
autonomous
noid hemorrhage among emergency department pa-
tients presenting with new, isolated headaches.
NO Good
PRIORITY mobility In [41] uses fuzzy decision-making rules adapted to
classification problems by using the methodology of
exploratory analysis followed by unification of partic-
LAST VISIT LAST VISIT ular decision rules into fuzzy groups. On the other
< 12 h <8h
hand clinical decision rule must be based on evidences,
when no evidence-based guideline exists, i.e. due to
the extreme variability of a disease, then a consensus-
NO LOW LOW depression
PRIORITY PRIORITY PRIORITY or anxiety based clinical practice guideline is the best option [42].
This latter is often used for psychological treatments
planning, sometime also along with more orthodox clin-
Memory MEDIUM ical decision rules.
Loss PRIORITY It follows that physicians, therapists, psychologists,
and caregivers in general could obtain great advan-
tages from specific support systems in order to be in-
LAST VISIT MEDIUM formed of the existing decision making rules. When
<4h PRIORITY
such rules are not available the implementation of de-
cision threes could be of great advantage.
MEDIUM HIGH A decision tree is a decision support tool that uses
PRIORITY PRIORITY a tree-like model of decisions and their possible con-
sequences, including chance event outcomes, resource
Figure 2: The topology, attributes and nodes of the imple-
costs, and utility. It is one way to display an algorithm
mented decision tree.
that only contains conditional control statements. The
decision tree consists of three types of nodes: deci-
sions, chances and endings.
we will describe how a decision tree can take respon-
Decision trees are commonly used in operations re-
sibility for the latter.
search and operations management. One advantage of
decision threes is the possibility to linearize them into
4. The implemented decision decision rules. From a decision tree it is possible to ex-
trapolate a chain of decision that are basically driven
tree by the comparison of measurements at a constant time.
If such measurements are coming from a set of obser-
Decision making rules have been adopted since many vation regarding the psychological state of an isolated
years and with different purposes. E. g. in [29] de- hospitalized patient, then the decision tree can be used
scision making rules have been developed as a guide to understand when it is needed a psychological help
for hospitalization of patients presenting community- to improve his mental status.
acquired pneumonia, while in [30, 31, 32] decision mak- Decision Tree is a classification algorithm that de-
ing rules are adopted to define when x-rays are needed cides whether a specific value should be accepted or
in acute ankle injuries. In facts such a support tool is rejected, and it provides with the set of the IF-THEN
often used for trauma treatments and when diagnosti- rules for transforming present state to future state [43].
cal imagery is involved [33, 34]. The tree structure is used to represent decision tree
There are many works in literature about the extrac- in which variant types of the nodes are connected by
tion and formulation of decision making rules. In [35] the branches where the topmost node is called as root
decision making rules have been extracted by means node and the leaves are called decision node [44, 45,
of a decision tree [36, 37, 38, 39] for the diagnostic 46].
workup of patients with Meniere’s disease, vestibu- In our implemented model (see Figure 2) we aimed
lar schwannoma, traumatic vertigo, sudden deafness, to discriminate whether or not a patient should urge a
50
visit by a physician, not only regarding the therapeu- Table 1
tic routine, but also in order to decrease the patient’s The table shows the simulated results obtained by the im-
psychological burden. In our model we used simple plemented decision support system in terms of elapsed time
observable variables that could be easily recorded dur- between visits. The last column shows the relative variation
ing the patient’s hospitalization period also by means which represent a beneficial reduction of time intervals for
of scarce automation. These variable take into account the most needful classes of patients.
whether or not the patient has a company, both in Standard Simulated
terms of a related, a visitor, or a conscious and interac- Average Average Δ Δ%
tive roommate, as well as the degree of autonomy and Time Time
mobility of the patient jointly with his mental status Patients with
∼ 10 h ∼ 12 h +2 h +20 %
(with particular focus on depression and anxiety). Fi- company
nally a special attention is given to patients with mem- Autonomous
∼ 10 h ∼8h −2 h −20 %
pateints
ory loss or mental impairment. In our approach the
Depressed
data can be collected automatically and stored in the ∼ 10 h ∼6h −4 h −40 %
or anxious
form of an input vector
Mentally
∼6h ∼3h −3 h −50 %
Impaired
𝑥⃗ = (𝑥𝑖 )𝑖=1∶𝑁 = (𝑥1 , 𝑥2 , ..., 𝑥𝑁 −1 , 𝑥𝑁 ) (1)
in order to feed the training algorithm of our deci-
sion stream. Given a set of known samples, the de- then benefit of a positive fallback on their mental sta-
cision three has can be trained with a C4.5 algorithm tus which also improves their remission, therefore re-
[47], using the Kullback–Leibler divergence [48] to ducing hospitalization and relieving also the general
measure the homogeneity of the target variable within burden for the healthcare service, with a positive feed-
the subsets. In this manner For a value 𝑣𝑘 taken by the back loop that should exponentially benefit the care-
attribute 𝑥𝑘 of the input vector 𝑥⃗ ∈ 𝑋 , given a related giving system.
training set 𝑆𝑘 , and the conditional entropy as
|𝑆 | References
𝐻 (𝑋 |𝑣𝑘 ) = ∑ 𝑘 ⋅ 𝐻 (𝑆𝑘 ) (2)
|𝑋 |
[1] A. Remuzzi, G. Remuzzi, Covid-19 and italy:
the expected information gain is the change in in- what next?, The Lancet (2020).
formation entropy from a prior state, mediated by the [2] C. Mannelli, Whose life to save? scarce resources
a pirori Shannon entropy 𝐻 (𝑋 ), to a state that takes allocation in the covid-19 outbreak, Journal of
some information, mediated by the conditional entropy. Medical Ethics 46 (2020) 364–366.
Therefore it is possible to compute the information [3] R. D. Truog, C. Mitchell, G. Q. Daley, The
gain as toughest triage—allocating ventilators in a pan-
demic, New England Journal of Medicine 382
Γ(𝑋 , 𝑣𝑘 ) = 𝐻 (𝑋 ) − 𝐻 (𝑋 |𝑣𝑘 ) (3) (2020) 1973–1975.
therefore obtaining a good measure for deciding the [4] A. M. Mosadeghrad, Healthcare service quality:
relevance of each attribute in our recursive partition- towards a broad definition, International journal
ing. of health care quality assurance (2013).
[5] K. Hadwich, D. Georgi, S. Tuzovic, J. Büttner,
M. Bruhn, Perceived quality of e-health ser-
5. Results and conclusions vices, International Journal of Pharmaceutical
and Healthcare Marketing (2010).
In our approach we used a modified version of the C4.5 [6] L. B. Gasink, K. Singer, N. O. Fishman, W. C.
algorithm, introducing time and causality, in order to Holmes, M. G. Weiner, W. B. Bilker, E. Lauten-
manage the visiting time of the caregivers in an oncol- bach, Contact isolation for infection control in
ogy ward. I our simulations (see Table 1) the results hospitalized patients: is patient satisfaction af-
has showed an enhanced and improved time distri- fected?, Infection Control & Hospital Epidemi-
bution and time-consumption efficiency, with a short- ology 29 (2008) 275–278.
ened isolation time for the most needful classes of pa- [7] C. Abad, A. Fearday, N. Safdar, Adverse effects
tients. It is possible to state that the patients should of isolation in hospitalised patients: a systematic
51
review, Journal of hospital infection 76 (2010) 97– 1239–1242.
102. [17] Y.-L. Chen, F.-M. Hsu, C. J. Tsai, J. C.-H. Cheng,
[8] G. Lo Sciuto, S. Russo, C. Napoli, A cloud-based Efforts to reduce the impacts of covid-19 out-
flexible solution for psychometric tests valida- break on radiation oncology in taiwan, Advances
tion, administration and evaluation, in: Proceed- in Radiation Oncology (2020).
ings of SYSTEM 2019, volume 2468, CEUR-WS, [18] E. Cao di San Marco, J. Menichetti, E. Vegni,
2019, pp. 16–21. Covid-19 emergency in the hospital: How the
[9] H. L. Evans, M. M. Shaffer, M. G. Hughes, R. L. clinical psychology unit is responding., Psycho-
Smith, T. W. Chong, D. P. Raymond, S. J. Pelletier, logical Trauma: Theory, Research, Practice, and
T. L. Pruett, R. G. Sawyer, Contact isolation in Policy (2020).
surgical patients: a barrier to care?, Surgery 134 [19] M. S. Ning, M. F. McAleer, M. D. Jeter, B. D. Min-
(2003) 180–188. sky, R. A. Ghafar, I. J. Robinson, P. L. Nitsch, D. J.
[10] S. Saint, L. A. Higgins, B. K. Nallamothu, Zaebst, S. E. Todd, J. Nguyen, et al., Mitigating
C. Chenoweth, Do physicians examine patients the impact of covid-19 on oncology: Clinical and
in contact isolation less frequently? a brief re- operational lessons from a prospective radiation
port, American journal of infection control 31 oncology cohort tested for covid-19, Radiother-
(2003) 354–356. apy and Oncology (2020).
[11] K. B. Kirkland, J. M. Weinstein, Adverse effects [20] K. A. Thom, M. Kleinberg, M.-C. Roghmann, In-
of contact isolation, The Lancet 354 (1999) 1177– fection prevention in the cancer center, Clinical
1178. infectious diseases 57 (2013) 579–585.
[12] B. Guilley-Lerondeau, C. Bourigault, A.-C. G. des [21] M. Powazek, J. R. Goff, J. Schyving, M. A. Paul-
Buttes, G. Birgand, D. Lepelletier, Adverse ef- son, Emotional reactions of children to isolation
fects of isolation: a prospective matched cohort in a cancer hospital, The Journal of Pediatrics 92
study including 90 direct interviews of hospital- (1978) 834–837.
ized patients in a french university hospital, Eu- [22] A. Rivera, N. Ohri, E. Thomas, R. Miller, M. A.
ropean Journal of Clinical Microbiology & Infec- Knoll, The impact of covid-19 on radiation on-
tious Diseases 36 (2017) 75–80. cology clinics and cancer patients in the us, Ad-
[13] G. Capizzi, C. Napoli, S. Russo, M. Wozniak, Less- vances in Radiation Oncology (2020).
ening stress and anxiety-related behaviors by [23] M. Shaygan, Z. Yazdani, A. Valibeigi, The ef-
means of ai-driven drones for aromatherapy, in: fect of online multimedia psychoeducational in-
Proceedings of the 6th Italian Workshop on Ar- terventions on the perceived stress and resilience
tificial Intelligence and Robotics co-located with of hospitalized patients with covid-19: a quasi-
the XVIII International Conference of the Ital- experimental study (2020).
ian Association for Artificial Intelligence (AIxIA [24] S. Russo, C. Napoli, A comprehensive solution
2019), volume 2594, CEUR-WS, 2019, pp. 7–12. for psychological treatment and therapeutic path
[14] G. Catalano, S. H. Houston, M. C. Catalano, A. S. planning based on knowledge base and expertise
Butera, S. M. Jennings, S. M. Hakala, S. L. Bur- sharing, in: "Proceedings of ICYRIME 2019", vol-
rows, M. G. Hickey, C. V. Duss, D. N. Skelton, ume 2472, CEUR-WS, 2019, pp. 41–47.
et al., Anxiety and depression in hospitalized pa- [25] S. I. Illari, S. Russo, R. Avanzato, C. Napoli, A
tients in resistant organism isolation., Southern cloud-oriented architecture for the remote as-
medical journal 96 (2003) 141–146. sessment and follow-up of hospitalized patients,
[15] K. Tran, C. Bell, N. Stall, G. Tomlinson, in: Symposium for Young Scientists in Technol-
A. McGeer, A. Morris, M. Gardam, H. B. Abrams, ogy, Engineering and Mathematics, volume 2694,
The effect of hospital isolation precautions on CEUR-WS, 2020.
patient outcomes and cost of care: a multi-site, [26] B. M. Reilly, A. T. Evans, Translating clinical re-
retrospective, propensity score-matched cohort search into clinical practice: impact of using pre-
study, Journal of general internal medicine 32 diction rules to make decisions, Annals of inter-
(2017) 262–268. nal medicine 144 (2006) 201–209.
[16] Z. Wu, J. M. McGoogan, Characteristics of and [27] T. M. Shaneyfelt, M. F. Mayo-Smith, J. Roth-
important lessons from the coronavirus disease wangl, Are guidelines following guidelines?:
2019 (covid-19) outbreak in china: summary of a The methodological quality of clinical practice
report of 72 314 cases from the chinese center for guidelines in the peer-reviewed medical litera-
disease control and prevention, Jama 323 (2020) ture, Jama 281 (1999) 1900–1905.
52
[28] R. Weinkove, Z. K. McQuilten, J. Adler, M. R. [41] N. Korenevskiy, Application of fuzzy logic
Agar, E. Blyth, A. C. Cheng, R. Conyers, G. M. for decision-making in medical expert systems,
Haeusler, C. Hardie, C. Jackson, et al., Managing Biomedical Engineering 49 (2015) 46–49.
haematology and oncology patients during the [42] T. Ashizawa, C. Gagnon, W. J. Groh, L. Gutmann,
covid-19 pandemic: interim consensus guidance, N. E. Johnson, G. Meola, R. Moxley, S. Pandya,
Medical Journal of Australia (2020). M. T. Rogers, E. Simpson, et al., Consensus-based
[29] M. J. Fine, T. E. Auble, D. M. Yealy, B. H. Hanusa, care recommendations for adults with myotonic
L. A. Weissfeld, D. E. Singer, C. M. Coley, T. J. dystrophy type 1, Neurology: Clinical Practice 8
Marrie, W. N. Kapoor, A prediction rule to iden- (2018) 507–520.
tify low-risk patients with community-acquired [43] R. R. Halde, A. Deshpande, A. Mahajan, Psychol-
pneumonia, New England journal of medicine ogy assisted prediction of academic performance
336 (1997) 243–250. using machine learning, in: 2016 IEEE Interna-
[30] I. G. Stiell, G. H. Greenberg, R. D. McKnight, R. C. tional Conference on Recent Trends in Electron-
Nair, I. McDowell, J. R. Worthington, A study ics, Information & Communication Technology
to develop clinical decision rules for the use of (RTEICT), IEEE, 2016, pp. 431–435.
radiography in acute ankle injuries, Annals of [44] I. Nincevic, M. Cukusic, Z. Garaca, Mining de-
emergency medicine 21 (1992) 384–390. mographic data with decision trees, in: The 33rd
[31] C. J. McDonald, J. M. Overhage, Guidelines you International Convention MIPRO, 2010.
can follow and can trust: an ideal and an example, [45] P. M. Arsad, N. Buniyamin, J.-L. Ab Manan,
Jama 271 (1994) 872–873. N. Hamzah, Proposed academic students’ perfor-
[32] J. H. Wasson, H. C. Sox, Clinical prediction rules: mance prediction model: A malaysian case study,
have they come of age?, Jama 275 (1996) 641–642. in: 2011 3rd International Congress on Engineer-
[33] C. Madden, D. B. Witzke, A. B. Sanders, J. Valente, ing Education (ICEED), IEEE, 2011, pp. 90–94.
M. Fritz, High-yield selection criteria for cranial [46] R. Chuchra, Use of data mining techniques for
computed tomography after acute trauma, Aca- the evaluation of student performance: a case
demic Emergency Medicine 2 (1995) 248–253. study, International Journal of Computer Science
[34] I. D. Graham, I. G. Stiell, A. Laupacis, A. M. and Management Research 1 (2012) 425–433.
O’Connor, G. A. Wells, Emergency physicians’ [47] J. R. Quinlan, C4. 5: programs for machine learn-
attitudes toward and use of clinical decision rules ing, Elsevier, 2014.
for radiography, Academic Emergency Medicine [48] T. Van Erven, P. Harremos, Rényi divergence and
5 (1998) 134–140. kullback-leibler divergence, IEEE Transactions
[35] E. Kentala, I. Pyykkö, K. Viikki, M. Juhola, Pro- on Information Theory 60 (2014) 3797–3820.
duction of diagnostic rules from a neurotologic
database with decision trees, Annals of Otology,
Rhinology & Laryngology 109 (2000) 170–176.
[36] W.-Y. Loh, Classification and regression trees,
Wiley Interdisciplinary Reviews: Data Mining
and Knowledge Discovery 1 (2011) 14–23.
[37] T. K. Ho, Random decision forests, in: Proceed-
ings of 3rd international conference on document
analysis and recognition, volume 1, IEEE, 1995,
pp. 278–282.
[38] S. Bernard, L. Heutte, S. Adam, On the selec-
tion of decision trees in random forests, in: 2009
International Joint Conference on Neural Net-
works, IEEE, 2009, pp. 302–307.
[39] J. Ali, R. Khan, N. Ahmad, I. Maqsood, Random
forests and decision trees, International Journal
of Computer Science Issues (IJCSI) 9 (2012) 272.
[40] D. E. Newman-Toker, J. A. Edlow, High-stakes di-
agnostic decision rules for serious disorders: the
ottawa subarachnoid hemorrhage rule, JAMA
310 (2013) 1237–1239.
53