MATHEMATICAL MODELING IN ARTIFICIAL INTELLIGENCE FOR

HEALTHCARE SYSTEMS

Abstract

Artificial intelligence (AI) is the imitation of human cognitive abilities. It’s changing the way
people think about medical services, thanks to increased access to medical information and
rapid advancements in examination techniques. The authors look at the current state of AI in
medical services and speculate on its future. Artificial intelligence can be used to analyses
various types of medical data (structured and unstructured). AI approaches for structured data
machine learning methods, such as the old-style support vector machine and neural network, as
well as modern deep learning, and regular language processing for unstructured data are
among the most well-known AI tactics. Finding and treatment suggestions, patient commitment
and adherence, and managerial exercises are all important classifications of utilizations.
Malignant growth, nervous system science, and cardiology, for example, are some of the major
disease areas that use AI devices.

Keywords: Artificial intelligence, treatment suggestions, human.

Introduction

Artificial intelligence (AI) is viewed as a key game changer for achieving the 17 Sustainable
Development Goals (SDGs), within the defined timeframe, to which India is a signatory.
Artificial intelligence (AI) which is a part of smart intelligent technologies has evolved
considerably since 1956 when this field came into existence. In current scenario, there is already
a large deployment in healthcare in high income countries to support and improve healthcare.
However, this is now rapidly gaining widespread attention even in low- and middle-income
countries (LMICs). The smartphones industry's explosive growth along with increasing
advancements in ancillary technologies such as mobile health (mHealth), electronic medical
records (EMR) and cloud computing, provide the environment required for the increasing uptake
of AI in healthcare in LMICs.

Developing countries such as India have been exhibiting rapid innovation in the field of
Artificial Intelligence (AI) applications in healthcare (Tayade and Karandikar, 2013). Healthcare
applications of AI are being used by a variety of stakeholders ranging from medical
professionals, healthcare professionals, healthcare management professionals, basic science and
life science researchers, global health professionals to even patients (Luxton, 2016; Tao, 2016).
However, in India, the opportunities and key considerations are larger and complex in nature
owing to vast geographic expanse and human capital. Diversity in terms of populations,
geography, literacy etc add up to the challenges.

The scientific community may have to adopt Artificial Intelligence (AI) to deliver high quality
services and achieve better outcomes through research. The application of AI has huge capability
to improve medical outcomes and quality of services being provided. This has high potential
especially in a country like India where there is no dearth of data owing to human capital and
high diversity. AI can be used for prediction, prevention, screening, diagnosis, treatment and
cure (Patel et al., 2009; Tao, 2016). Besides this it can also help in drug discovery, in chronic
condition management, health services delivery and detection of disease (Patel et al., 2009).

Prelude to Mathematical Modelling and the Scientific Method under uncertainty

Mathematics and physical science each had important effects on the development of the other.
Mathematics is starting to play a greater role in the development of the life and social sciences,
and these sciences are starting to influence the development of mathematics. This sort of
interaction is extremely important if the proper mathematical tools are going to be developed for
the various sciences. Some people feel that there is something deeper going on than simply an
interaction leading to the formulation of appropriate mathematical and physical concepts.

Moreover, Observation is the ultimate basis for our understanding of the world around us. But
observation alone gives information about particular events; it provides little help for dealing
with new situations. Out ability and aptitude to recognize similarities in different events, to distil
the important factors for a specific purpose, and to generalize our experience enables us to
operate effectively in new environments. The result of this is knowledge, an essential resource
for any intelligent agent.

Computational Intelligence in Healthcare

Health care is coming home. For a number of reasons, health care is increasingly occurring in
residential settings rather than in professional medical settings. This change in the locus of care
needs to be seen in context. In this first decade of the 21st century, great attention is being
devoted to world society ‘s needs for access to health care and health care delivery. To date,
however, there has been too little focus on the transition of care into the home. A wide range of
procedures and therapies are now performed far from any medical facility, often with no health
care professional on site, with highly variable results. Although each situation is unique, all are
dependent on the people involved the human factors.

Given converging trends of an aging population, an increasing human life span, medical
technology migrating into the home, and design features of the home in contrast to health care
institutions, it is important to gain an understanding of the most prevalent and serious threats to
safety, the quality of care experienced, and the well-being of care recipients and caregivers.
Developing human factors approach to health care in the home that can accommodate the
diversity, strengths, and limitations of humans, both as care recipients and caregivers, is critical
to addressing these threats effectively.

A number of factors are driving the migration of health care practice from professional facilities
to the home and, as a result, significantly increasing the numbers of people who must provide
health care in the home:

 The costs of providing health care at formal medical facilities are increasing. Advanced
medical technologies and procedures, as well as the training of medical professionals to
employ them, can be very expensive.

 Hospitals are discharging patients, including premature infants, sooner into home care,
sometimes with complex care regimens.
  The world population is aging, and consequently the demand is growing for various
health services, particularly related to conditions associated with aging. At the same time,
people are focusing increasingly on overall wellness and quality of life, even into
advanced age.

 The prevalence of chronic conditions across the entire age spectrum is growing
(particularly conditions related to obesity, such as diabetes), and growing along with it is
the demand for health care. More people are living longer with increasingly complex
medical and social needs.

 Consumers want to be independent in their health management and are seeking more
home-based services.

In the recent years, all over the world the formulation quality of life‖ has become a popular
catchword in the political scene in view of promising people a better future. Analyzing this
catchword we find, besides expectations towards the standard of living a growing concern about
the human environment and about health.

Scope

The goal of health services is to protect and improve the health of individuals and populations. A
new health system for the 21st century proposed that the goals for health services should include
six critical elements which are patient safety, effectiveness, timeliness, patient centered,
efficiency and equity.

the quantity and quality of many health care interventions are improved through the results of
science, such as advanced through the medical model of health which focuses on the eradication
of illness through diagnosis and effective treatment. Many important advances have been made
through health research, including biomedical research and pharmaceutical research, which form
the basis for evidence-based medicine and evidence-based practice in health care delivery.

In addition, the results of health services research can lead to greater efficiency and equitable
delivery of health care interventions, as advanced through the social model of health and
disability, which emphasizes the societal changes that can be made to make population healthier.
Results from health services research often form the basis of evidence-based policy in health care
systems. Health services research is also aided by initiatives in the field of Artificial Intelligence
for the development of systems of health assessment that are clinically useful, timely, sensitive to
change, culturally sensitive, low burden, low cost, involving for the patient and built into
standard procedures.

Objective of the study

1. To examine the Efficient AI for Healthcare Systems using Mathematical Modelling.

2. To examine the Responsible Innovation Mode's Application of AI to Healthcare in India.

Review of literature
Josephine & Blanchet, Karl. (2019) Mathematical modelling has been a vital research tool for
exploring complex systems, most recently to aid understanding of health system functioning and
optimisation. System dynamics models (SDM) and agent-based models (ABM) are two popular
complementary methods, used to simulate macro- and micro-level health system behaviour. This
systematic review aims to collate, compare and summarise the application of both methods in
this field and to identify common healthcare settings and problems that have been modelled
using SDM and ABM. Methods: We searched MEDLINE, EMBASE, Cochrane Library,
MathSciNet, ACM Digital Library, HMIC, Econlit and Global Health databases to identify
literature for this review. We described papers meeting the inclusion criteria using descriptive
statistics and narrative synthesis, and made comparisons between the identified SDM and ABM
literature. Results: We identified 28 papers using SDM methods and 11 papers using ABM
methods, one of which used hybrid SDM-ABM to simulate health system behaviour.

Razavyan, Shabnam. (2023) Degeneration of the neurological system linked to cognitive deficits,
daily living exercise clutters, and behavioral disturbing impacts may define Alzheimer’s disease.
Alzheimer’s disease research conducted later in life focuses on describing ways for early
detection of dementia, a kind of mental disorder. To tailor our care to each patient, we utilized
visual cues to determine how they were feeling. We did this by outlining two approaches to
diagnosing a person’s mental health. Support vector machine is the first technique. Image
characteristics are extracted using a fractal model for classification in this method. With this
technique, the histogram of a picture is modeled after a Gaussian distribution. Classification was
performed with several support vector machines kernels, and the outcomes were compared. Step
two proposes using a deep convolutional neural network architecture to identify Alzheimer’s-
related mental disorders.

Szolovits, Peter & Atun, Rifat. (2018) Globally, health systems face multiple challenges: rising
burden of illness, multimorbidity and disability driven by ageing and epidemiological transition,
greater demand for health services, higher societal expectations and increasing health
expenditures. A further challenge relates to inefficiency, with poor productivity. These health
system challenges exist against a background of fiscal conservatism, with misplaced economic
austerity policies that are constraining investment in health systems. Fundamental transformation
of health systems is critical to overcome these challenges and to achieve universal health
coverage (UHC) by 2030. Machine learning, the most tangible manifestation of artificial
intelligence (AI) and the newest growth area in digital technology holds the promise of achieving
more with less, and could be the catalyst for such a transformation.

Nurmaganbetova, Mugulsum. (2023) One of the primaries focuses of the Republic of

Kazakhstan concerning sustainable and stable improvements in the well-being of its population
is the advancement of the healthcare sector. A mathematical model for an automated medical
diagnostics system integrates machine learning algorithms, statistical models, and decision trees
to analyze patient data and facilitate accurate diagnoses. This model enables healthcare
professionals to enhance the efficiency and reliability of medical diagnostics by leveraging
advanced computational techniques. In diagnosing diseases, the differential diagnosis between
primary biliary cirrhosis of the liver and active hepatitis with cholestatic syndrome poses a
challenge due to their overlapping symptoms. Both conditions exhibit jaundice, pruritus, fatigue,
and hepatomegaly. However, distinguishing features based on clinical and laboratory
observations can aid in accurate differentiation between the two states. These distinguishing
features can be incorporated by developing a mathematical model for diagnosing diseases,
enabling precise identification and guiding appropriate treatment strategies. Predicting the
progression of diseases is a crucial aspect of healthcare, enabling personalized interventions and
improved patient outcomes.

Research methodology

Definition and Classification of Mathematical Modeling

A mathematical model is a tool for describing real-world or complex systems in mathematical

terms. It converts real-world problems into mathematical problems and formalizes them in such a
way that they can be analyzed and solved mathematically, thereby drawing quantitative or
qualitative conclusions about practical problems. Table 1 presents the basic concepts used in the
mathematical modeling process.

Table 1. Basic concepts in mathematical modeling process

Modeling Process Basic Concepts

The process of translating practical problems into mathematical

language. This includes identifying key variables, establishing
Problem Abstraction
mathematical descriptions of the problem, and setting goals for the
problem.

A mathematical model is an abstract representation of practical

problems. It usually includes components such as variables, equations,
Mathematical Model
constraints, and initial conditions. Mathematical models can be
deterministic or probabilistic and can be continuous or discrete.

Modeling methods are techniques and strategies used to construct

mathematical models. Common modeling methods include
Modeling Methods mathematical analysis, optimization theory, probability and statistics,
calculus, and differential equations. The choice of modeling method
depends on the nature and requirements of the problem.

Model validation is the process of confirming whether a mathematical

model can accurately reflect actual problems. This can be achieved by
Model Validation comparing the predicted results of the model with actual data. If the
model’s predictions match well with actual observations, then the
effectiveness of the model is verified.

Model solving is the process of analyzing and calculating established

mathematical models using mathematical tools and techniques. This
Model Solution
can involve methods such as symbolic computation, numerical
computation, and optimization algorithms to obtain the solution or
optimal solution of the problem.
 Model evaluation is the process of evaluating the quality of model
results and solutions. This includes evaluating the feasibility, stability,
Model Evaluation accuracy, and practicality of the solution. The purpose of model
evaluation is to determine the effectiveness and reliability of the model
in solving practical problems.

Mathematical models can be represented using different mathematical methods and techniques,
such as calculus, probability theory, and linear algebra, from the fields of computer science and
statistics. In engineering computing, a mathematical model is a mathematical expression or set of
equations that describes and analyzes an engineering problem and can be used to predict and
optimize the performance and efficiency of an engineering system.

With all the above-stated features, this system can prove to be a major advancement in the
current system used.

Figure 1 Types of machine learning models

Search strategy and information sources

The literature on ABM and SDM of health systems has not been confined to a single research
discipline, making it necessary to widen the systematic review to capture peer-reviewed articles
found in mathematical, computing, medicine and health databases. Accordingly, we searched
MEDLINE, EMBASE, Cochrane Library, MathSciNet, ACM Digital Library, HMIC, Econlit
and Global Health databases for literature. The search of health system literature was narrowed
to identify articles that were concerned with modelling facility-based healthcare, services and
related healthcare financing agreements which had been excluded or were not the focus of
previous reviews.

Data analysis

Accuracy, Re-call or sensitivity, Specificity, F1-score, and geometric mean sensitivity and
specificity (Gmeans) are used in this analysis to determine the effects of the classification. A
system ability to discover new patients is measured by its Recall and accuracy rate. Each class
overall performance, including classification accuracy and recall, is measured using the F1. This
section defines these standards.

All search results were uploaded to Mendeley reference software where duplicate entries were
removed. The remaining records were screened using their titles and abstracts, removing entries
based on eligibility criteria given in Table 1. Post-abstract review, the full text of remaining
articles was screened. Papers retained in final stage of screening were scrutinized, with data
imported to Excel based on the following categories; publication date, geographical and
healthcare setting modelled, purpose of research in addition to any policies or interventions
tested, rationale for modelling method and software platform, validation and limitations of
model.

Figure 2 Flow-chart for systematic review of SDMs

The remaining eight papers explored factors leading to undesirable emergency care system
behaviour, simulating hospital waste management systems and predicting future waste
generation, estimating future demand for cardiac care, exploring the impact of patient admission
on health professionals stress level in an integrated care system, and variation in physician
decision-making.

Table 2 Eligibility criteria for review

 Criteria Inclusion Exclusion

Studies that describe the Poster presentations, conference abstracts, review

Type of development and papers (reference list reviewed), commentaries,
study/model presentation of SDM or debate papers, papers that describe the qualitative
ABM or hybrid model. data used to inform a later developed model.

Facility-based healthcare
Papers that primarily describe a
or related
Setting disease/transmission model or delivery of non-
policies/financing
facility-based healthcare
arrangements

Healthcare setting and purpose of research

The healthcare settings modelled in the SDM, ABM and hybrid simulation papers are presented
in Figure 3. Healthcare settings modelled using SDM included systems that were concerned with
delivering emergency or acute care (11/28), elderly or long-term care services (LTC) (12/28) and
hospital waste management (4/28). Twenty of the SDM papers selected in this review assessed
the impact of health policy or interventions on the modelled system. Common policy targets
included finding robust methods to relieve stretched healthcare services, ward occupancy and
patient length of stay, reducing the time to patient admission, targeting undesirable patient health
outcomes, optimizing performance-based incentive health system policies and reducing the total
cost of care.

Figure 3 The health system sector locations modelled

Coronavirus belongs to a family of viruses with positive-sense (+) RNA, which have the
capability of infecting the host by inducing the host with symptoms of cold and flu in its mild
stage and severe respiratory ailments and multiorgan failure in its lethal stage. This virus can
infect humans, and several cases of pets getting infected (Figure 4) have also been reported in
different parts of the world. Some countries have a history of underreporting the disease, and it
acts as a catalyst in the spread of infection. Lack of infrastructure, proper testing techniques, and
high population may be the reasons for the spread of this deadly virus throughout the countries,
continents, and subcontinents.

Figure 4 The SDM-simulated healthcare environments

Minimal requirements of novel malaria interventions to achieve a defined health goal

For the five aforementioned malaria interventions, we explored their optimal profiles for a broad
set of target reduction levels, creating landscapes of intervention profiles according to their
minimal characteristics across various transmission settings. These landscapes provide a broad
and comprehensive overview of the intervention potential capabilities and limitations in
achieving a desired health goal. For example, as opposed to an anti-infective monoclonal
antibody which requires high efficacy and duration to achieve large PfPR reduction in only a
limited number of settings, attractive targeted sugar baits that kill mosquitoes achieve a wider
range of target PfPR reductions in high-transmission settings as well. Similarly, while in settings
with lower transmission, anti-infective and transmission-blocking vaccines had comparable
requirements in achieving similar PfPR reduction targets, anti-infective vaccines showed a
higher potential and reached additional targets in high-transmission, endemic settings (show in
figure 5).
 Figure 5 Estimated optimal intervention and delivery profiles (TPPs)

Result and Discussion

In this study, we introduced a new modeling and machine-learning framework that for the first
time enables quantitative differentiation between operational, setting, and intervention
parameters as determinants of intervention impact, using detailed simulation models of disease.
Our framework can be used for any disease where a valid model of disease progression or natural
history of disease is available. We provided mathematical tools for efficiently and quantitatively
defining the minimum profiles of malaria interventions as well as delivery approaches required
to reach a desired health goal. Furthermore, our methodology provides a means to refine the
identified optimal efficacy and duration characteristics as additional information becomes
available.

As a result, we can apply fully-detailed disease models to direct the design of novel interventions
and understand how intervention-specific, epidemiological and systems factors jointly contribute
to impact and thus inform TPP guidance. Most immediately, the approach is highly relevant to
define successful interventions against emerging diseases such as SARS-CoV-2, and to support
efficient, fast development of operational strategies. As uncertainties in disease progression and
epidemiology can be incorporated in our approach, it also provides a way to systematically sort
through large complex landscapes of unknowns and thus refine properties of interventions or
clinical trials as more knowledge is available.

Our approach constitutes a powerful tool to help address the challenges of current malaria
strategies and develop new interventions to progress towards malaria elimination. While
currently promising interventions such as insecticide treated nets, seasonal malaria
chemoprevention (SMC) and intermittent preventative treatment (IPT) have been very successful
at reducing malaria incidence and saving lives, their improved burden reduction and future
success is currently challenged by limited adherence, resource and time constraints to increase
coverage and usage in underserved populations, as well as resistance. Furthermore, for settings
where SMC has not been implemented or not recommended (for example in East Africa or in
perennial settings), there remains a gap in available interventions to protect vulnerable
populations who experience the highest burden of malaria. Similarly, for settings with outdoor
biting mosquitoes, the development and rollout of novel vector control interventions is needed.

Conclusion

The paper entitled mathematical modeling in artificial intelligence for healthcare systems is an
attempt to analyze the various decision-making problems faced by the health department to
reduce the human productivity loss. The thesis develops the various mathematical models based
on single and multi-objective linear programming model, relation and cluster analysis, by using
medical experts ‘knowledge, matrix and Euclidean distance in order to take the effective
decisions in the decision-making problems in health care process under uncertain context.
Throughout, the researcher has employed triangular number as a parameter to depict the
vagueness governing the decision-making problems faced by the health department in health care
process. Finally, it concludes that mathematical modeling making healthcare analysis enables to
predict the effective decisions of decision-making problems faced by the Healthcare
Organizations under uncertain environment to minimize the human productivity loss in an
effective and an efficient manner.

References

1. Adamo, J.M., Fuzzy decision trees, Fuzzy Sets and Systems, Vol. 4, pp. 207- 219, 1980.

2. Agarana, M.C., Optimization of health care pathways in covenant university health care
using linear programming model, Far east journal of applied mathematics, 2015, Vol. 91,
No.3, pp.215 – 228.

3. Asghari, G., et al., Designing Fuzzy Algorithms to Develop Healthy Dietary Pattern.
International Journal of Endocrinol Metab, Vol. 11, No.3, pp. 154-61, 2013.

4. Bector, C.R., and Suresh Chandra, Fuzzy Mathematical Programming and Fuzzy Matrix
Games, Springer Berlin Heidelberg, New York, 2005.

5. Bellman, R.E. and Zadeh, L.A., Decision making in a fuzzy environment, Management
Science, Vol. 17, pp. 141-164, 1970.

6. Boah, D. K., Nurse Scheduling at Navrongo War Memorial Hospital in Ghana Using
Linear Programming, Journal of Innovative Technology and Education, 2014, Vol.1, No.
1, pp. 25 – 33.

7. Bhatia, H.L., et al. (1977), ―A procedure for time minimizing transportation problem‖,
Indian Journal of Pure and Applied Mathematics, 8, pp. 920-929.

8. Buisson, J.C. and Garel, A., "Balancing meals using fuzzy arithmetic and heuristic search
algorithms," Fuzzy Systems, IEEE Transactions, vol.11, no.1, pp. 68- 78, 2003.
9. Buisson, J.C. and Garel, A., "Balancing meals using fuzzy arithmetic and heuristic search
algorithms," Fuzzy Systems, IEEE Transactions, vol.11, no.1, pp. 68- 78, 2003.

10. Chakraborty Ananya and Chakraborty M., (2010), Cost-time Minimization in a

Transportation Problem with Fuzzy Parameters: A Case Study, Journal of Transportation
Systems Engineering and Information Technology, 10(6), pp. 53−63.

11. Chang-Shing Lee, et al., A Type-2 Fuzzy Ontology and Its Application to Personal
Diabetic-Diet Recommendation, Fuzzy Systems, IEEE Transactions on, vol.18, no.2,
pp.374-395, 2010.

12. Crina Grosan et al., (2008), Multi Criteria Programming in Medical Diagnosis and
Treatment, Applied Soft Computing, 8(4), pp. 1407 – 1417.

13. Denton, B.T., Optimizing the Start time of Statin Therapy for Patients with Diabetes.
Med Decis Making, 2009, Vol. 29, No.3, pp.351 – 367.

14. Dubois, D. and Prade, H. Fuzzy numbers: an overview, in J.C. Bezdek (ed.) Analysis of
Fuzzy Information, Mathematics and Logic, CRC Press, Boca Raton, FL, 1997.

15. Dubois, D. and Prade, H., Ranking fuzzy numbers in the setting of possibility theory,
Information Sciences, Vol. 30, pp. 183-224, 1980.

16. Dumitrescu, D., Lazzzerini, L., and L.C. Jain, Fuzzy Sets and Their Application to
Clustering and Training, CRC Press LLC, Florida (USA), 2000.

17. Earnshaw, S.R., et. al., Optimal allocation of resources across four interventions for type
2 diabetes. Medical Decision Making, 2002, Vol. 22, No. 5, pp. 80 – 91.

18. Earnshaw, S.R., Hicks, K., Richter, A., Honeycutt, A., A linear programming model for
allocating HIV prevention funds with state agencies: A pilot study. Health Care
Management Science, 2007, Vol. 10, pp. 239–252.