A Soft Computing Approach to Ensuring Data Integrity in IoT-Enabled Healthcare Using Hesitant Fuzzy Sets

Abstract

The Internet of Medical Things (IoMT) is the latest advancement in the Internet of Things (IoT). Researchers are increasingly drawn to its vast potential applications in secure healthcare systems. The growing use of internet-connected medical device sensors has significantly transformed healthcare, necessitating the development of robust methodologies to assess their integrity. As access to computer networks continues to expand, these sensors have become vulnerable to a wide range of security threats, thereby compromising their integrity. To prevent such lapses, it is essential to understand the complexities of the operational environment and to systematically identify technical vulnerabilities. This paper proposes a unified hesitant fuzzy-based healthcare system for assessing IoMT sensor integrity. The approach integrates the hesitant fuzzy Analytic Network Process (ANP) and the hesitant fuzzy Technique for Order Preference by Similarity to the Ideal Solution (TOPSIS). In this study, a hesitant fuzzy ANP is employed to construct a comprehensive network that illustrates the interrelationships among various integrity criteria. This network incorporates expert input and accounts for inherent uncertainties. The research also offers sensitivity analysis and comparative evaluations to show that the suggested method can analyse many medical device sensors. The unified hesitant fuzzy-based healthcare system presented here offers a systematic and valuable tool for informed decision-making in healthcare. It strengthens both the integrity and security of healthcare systems amid the rapidly evolving landscape of medical technology. Healthcare stakeholders and beyond can significantly benefit from adopting this integrated fuzzy-based approach as they navigate the challenges of modern healthcare.

1. Introduction

Evaluating and ensuring the integrity of IoMT sensors is of paramount importance [1]. The healthcare, pharmaceutical, and IoMT sensor industries, both locally and globally, face unique vulnerabilities due to the immense value of personal healthcare data and historical reluctance to adopt robust security measures. Recognizing this critical need, there is a notable increase in cybersecurity spending projected for the IoMT sensor sector, from $869 million in 2020 to an estimated $1.2 billion by 2025, representing a 7.3% compound annual growth rate (CAGR) [2]. Nevertheless, this expenditure still accounts for just around 11.3% of the total health cyber-security spending and a mere 0.6% of the anticipated global security spending of $198 billion for 2025. This underscores the pressing importance of comprehensive evaluations to ensure the integrity of IoMT sensors.

The serious repercussions of cybersecurity breaches serve to further emphasize the urgency of assessing IoMT sensor integrity. Hackers exploit personal healthcare data for fraudulent activities, including manipulating IoMT histories, creating fake insurance claims, or illegally trading prescription medications. According to the US Department of Health and Human Services (HHS) Office of Civil Rights, there were healthcare data breaches that affected more than 41 million people in the US alone in 2021. Investigations continued in 2022, with cases impacting over 22.5 million individuals in the US, marking a 4.6% increase compared to the previous year [1,2]. In 2022, a significant breach at Shields Health Care Group exposed personal information, addresses, diagnoses, and other sensitive IoMT data from individuals in the USA, Saudi Arabia, India, and other countries [1,2]. Given these challenges, rigorous evaluations of IoMT sensor integrity are not merely a necessity but an ethical and regulatory imperative to safeguard patient information and well-being.

Medical world has witnessed rapid advancements in IoMT technology, leading to a healthcare landscape heavily reliant on various IoMT sensors for diagnostics, treatment, and patient care [3,4]. These sensors encompass a wide spectrum, from sophisticated imaging equipment to implantable sensors and portable monitors, all designed to enhance healthcare outcomes. However, the proliferation of IoMT sensors has brought about complex challenges related to their integrity, including factors such as functionality, safety, integrity, and compliance with regulatory standards. Ensuring the integrity of these sensors is crucial to maintaining the trust of patients and healthcare professionals and mitigating potential risks associated with sensor failures. This study is dedicated to the healthcare sector within the Saudi Arabian region, with a particular focus on evaluating the integrity of IoMT sensors in this context. Figure 1 illustrates the process of evaluating the integrity of IoMT sensors.

According to Figure 1, the literature review process will take place first, discussing issues and gaps in the proposed area. Furthermore, the criteria for integrity will be discussed in this process. In the next stage, the evaluation of integrity using the Hesitant Fuzzy Analytic Network Technique (HF ANP) will be conducted. The rating of alternatives using TOPSIS and sensitivity analysis will be carried out in the final stage.

The critical importance of IoMT sensor integrity cannot be overstated, as lapses in this regard can have dire consequences for patient health and safety [5,6]. Despite the existence of various assessment methods and standards, the multifaceted nature of integrity and the uncertainties present in real-world healthcare environments demand a more comprehensive and systematic approach.

To address this need, my research delves into an integrated, unified hesitant fuzzy-based healthcare system. This system combines the power of the HF ANP and the Hesitant Fuzzy Technique for Order of Preference by Similarity to Ideal Solution (HF TOPSIS) to comprehensively analyze the integrity of IoMT sensors [7,8]. Traditional ways of judging the quality of IoMT sensors do not always work well because they do not take into account how factors interact with each other, how important each criterion is, and the uncertainty that comes with real-world situations [7,8,9].

Existing research in this field tends to focus on individual aspects of integrity, lacking a holistic approach. Consequently, there exists a gap in the literature that necessitates a comprehensive and adaptable methodology to assess IoMT sensor integrity effectively. Furthermore, the contribution of the work is as follows:

• Introduces and empirically validates a unified hesitant fuzzy-based healthcare system using HF ANP and HF TOPSIS to analyze and evaluate IoMT sensor integrity.
• Aims to fill the literature gap by proposing a method that considers the multifaceted nature of integrity, incorporates expert insights, and accommodates uncertainties, enhancing decision-making in healthcare settings.
• Highlights security and integrity issues in IoMT sensors and provides a solution to overcome these challenges.
• Demonstrates the effectiveness of the integrated HF ANP- HF TOPSIS technique in real-world situations, testing the reliability of various IoMT sensors.
• Offers valuable insights for healthcare stakeholders in the region by considering complex relationships among integrity criteria, quantifying their relative importance with fuzziness, and providing a systematic decision-making framework. Contributes to enhancing the integrity, security, and overall quality of healthcare systems.

2. Ensuring Integrity of IoMT Sensors

A recent survey found that 94% of healthcare organizations have been victims of a cyberattack. This finding indicates that the healthcare industry is becoming an increasingly attractive target for cybercriminals [3]. Implanted IoMT sensors are connected to patients’ bodies and placed inside them to record sensitive bodily activities [4,5]. The information is collected by the sensor and then sent to relevant IoMT professionals and laboratories over the network. Low-power sensor-based devices are the most susceptible to breaches because of their connection to the network. Hence, the integrity of IoMT sensors is of paramount importance in healthcare systems worldwide. IoMT sensors encompass a wide range of equipment and technologies, from simple thermometers to complex imaging machines and life-support systems. Ensuring the integrity of these sensors is critical to patient safety, accurate diagnoses, and effective treatments.

The IoMT Sensors and Products market is poised for significant growth in the coming years. By 2023, the value added to this market is anticipated to reach a substantial US$0.50 billion [6]. This projection demonstrates a steady upward trajectory, with an expected compound annual growth rate (CAGR) of 2.66% between 2023 and 2030. This indicates a promising landscape for businesses operating in this sector, encouraging investments and innovation to meet the increasing demand for IoMT sensors and products. Additionally, by 2023, the IoMT Sensors and Products market is expected to grow substantially, reaching US$1.48 billion. This growth represents a strong annual increase of 6.52% over the five-year period, as illustrated in Figure 2. Due to improvements in healthcare infrastructure and a growing focus on healthcare quality, this increase in production highlights the sector’s growing importance in the national economy.

The market is expected to experience an expansion in the number of enterprises as well, with an estimated 0.35 thousand businesses operating within the IoMT Sensors and Products sector in 2023 [10,11,12]. This expansion is forecasted to continue with a Compound Annual Growth Rate (CAGR) of 5.15% from 2023 to 2030. The spread of companies in this field is likely to foster competition and innovation, ultimately benefiting consumers and healthcare providers alike. These positive trends signify a prosperous and evolving sector that holds considerable promise for both investors and the healthcare sector. On the other hand, the security and integrity of IoMT sensors are becoming critical for public health and the safety of their data. It is crucial for stakeholders to monitor these developments closely and adapt their strategies to capitalize on the opportunities presented by this dynamic market.

Integrity in the context of IoMT (Internet of Medical Things) sensors refers to the ability to prevent unauthorized modifications to data during its collection, transmission, and storage. As illustrated in

Table 1. Core factors and sub-factors of integrity in the context of IoMT sensors.

S. No.	Core Factor	Sub-Factor	Description
1	Functional Integrity (F1)	Accuracy (F11)	Precision in measurements and dosages to ensure reliable functionality and effectiveness.
		Measurement Accuracy (F111)	Precision in measuring vital signs such as blood pressure and temperature.
		Dosage Accuracy (F112)	Ensures correct drug dosages for medication delivery.
2	Reliability (F2)	Failure Rate (F21)	Rate at which the sensor experiences malfunctions or breakdowns during operation.
		Dependability (F22)	Ability to consistently perform without unexpected failures, ensuring patient safety.
3	Safety (F3)	Patient Safety (F31)	Focus on features and mechanisms preventing harm to patients, including error prevention and alarms.
		Operator Safety (F32)	Ensuring the safety of healthcare professionals while operating the sensor, protection from hazards, and design.
4	Functionality (F4)	Core Functions (F41)	Primary functions that the IoMT sensor is designed to perform accurately and reliably.
		Additional Features (F42)	Supplementary functions enhancing utility without compromising core functionality.

, ensuring data integrity is critical, as it guarantees that the data transmitted via wireless networks reaches its intended destination unaltered. The broadcast nature of wireless networks makes them vulnerable to attacks, where unauthorized entities could gain access to and manipulate patient data, potentially leading to life-threatening consequences. Therefore, it is essential to develop robust methods for maintaining data integrity to prevent alterations during transmission due to malicious attacks. Additionally, ensuring that data remains unaltered while stored on IoMT servers is crucial for safeguarding the completeness and security of the information. Assessing the core factors of integrity for IoMT sensors is vital for several reasons, particularly in healthcare settings.

• The factors were identified through a two-step approach:
  • Literature review: We systematically examined prior studies on IoMT sensor integrity, hesitant fuzzy decision-making, and healthcare cybersecurity. Key references include [6,7,8,9,10], which discuss confidentiality, availability, fault tolerance, and related integrity criteria.
  • Expert consultation: To ensure practical relevance, we consulted three domain experts (two from healthcare technology and one from information security), who validated and refined the initial factor list.

The assessment of core factors of integrity for IoMT sensors is of paramount importance due to its impact on patient safety, regulatory compliance, data privacy, and cost-effectiveness. Firstly, patient safety is a primary concern, as any malfunction or inaccuracy in IoMT sensors could result in incorrect diagnoses or treatment plans, posing serious risks to patients. Secondly, regulatory compliance necessitates the evaluation of sensor integrity to meet specific standards and certifications, ensuring their safety and effectiveness. Thirdly, data privacy and security are critical, given the digitization of healthcare. Thorough assessment of security measures, including encryption and authentication, helps prevent data breaches and unauthorized access, safeguarding patient confidentiality. Lastly, regular assessment of IoMT sensors enhances cost-effectiveness by prolonging their lifespan and reducing the need for frequent replacements. This proactive approach also mitigates potential vulnerabilities, contributing to the overall efficiency and security of healthcare systems. Therefore, a comprehensive assessment of IoMT sensor integrity is essential for improving healthcare quality and building trust in these devices.

3. Methodology Followed

The Hesitant Fuzzy ANP-TOPSIS (HF-ANP-TOPSIS) methodology is a robust hybrid technique designed to address multi-criteria decision-making (MCDM) problems in environments characterized by uncertainty and imprecision [13,14,15]. By integrating the Analytic Network Process (ANP) with the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) under hesitant fuzzy conditions, this approach enables decision-makers to evaluate and rank alternatives while accounting for the hesitation inherent in human judgment [16,17,18,19]. The reason for using the proposed methodology is as follows:

• Hesitant Fuzzy Sets (HFS) enable decision-makers to articulate several potential membership values for a criterion, rather than necessitating a singular choice.
• Contrary to AHP, which presumes the independence of criteria, the Analytic Network Process (ANP) incorporates feedback and interdependencies among decision elements.
• TOPSIS is a recognised method that evaluates alternatives by assessing their closeness to the ideal solution and their distance from the least favourable option.
• This integration demonstrates greater robustness, flexibility, and alignment with real-world complexity compared to the use of any single technique in isolation.

The step-by-step procedure of the HF-ANP-TOPSIS technique is illustrated in Figure 3, with a detailed description of each phase of the methodology provided in the following sections.

Step 1: Define the Problem and Decision Criteria

The decision problem is first identified, along with the relevant criteria and alternatives. Let the set of criteria be $\left\{{\mathrm{C}\mathrm{r}}_1,{\mathrm{C}\mathrm{r}}_2,\dots .,{\mathrm{C}\mathrm{r}}_{\mathrm{m}}\right\}$ and the set of alternatives be $\mathrm{A}\mathrm{t}=\left\{{\mathrm{A}\mathrm{t}}_1,{\mathrm{A}\mathrm{t}}_2,\dots ..{\mathrm{A}\mathrm{t}}_{\mathrm{n}}\right\}$. Each criterion represents an evaluation dimension (e.g., cost, reliability, efficiency), while alternatives represent the options being compared.

Step 2: Construct the Decision Matrix

The Decision Matrix (DM) contains the evaluation of each alternative with respect to each criterion. In the hesitant fuzzy environment, the element ${\mathrm{d}\mathrm{m}}_{\mathrm{i}\mathrm{j}}$ is expressed as a Hesitant Fuzzy Set (HFS):

$${\mathrm{d}\mathrm{m}}_{\mathrm{i}\mathrm{j}}=\left\{{\mathrm{h}\mathrm{f}}_{\mathrm{i}\mathrm{j}}^1,{\mathrm{h}\mathrm{f}}_{\mathrm{i}\mathrm{j}}^2,\dots ..,{\mathrm{h}\mathrm{f}}_{\mathrm{i}\mathrm{j}}^{\mathrm{k}}\right\}$$

(1)

where ${\mathrm{h}\mathrm{f}}_{\mathrm{i}\mathrm{j}}^{\mathrm{k}}$ is the k^th membership value for alternative ${\mathrm{A}\mathrm{t}}_{\mathrm{i}}$ under criterion ${\mathrm{C}\mathrm{r}}_{\mathrm{j}}$. This representation captures uncertainty when multiple possible membership degrees exist.

Step 3: Construct the Pairwise Comparison Matrix

Experts provide judgments on the relative importance of criteria using hesitant fuzzy linguistic terms (e.g., “equally important,” “moderately important”) [5]. These judgments form a PCM:

$${\mathrm{P}\mathrm{C}\mathrm{M}}_{\mathrm{c}\mathrm{r}\mathrm{i}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{a}}=\left(\begin{array}{cccc}1& {\tilde{\mathrm{a}}}_{12}& \dots & {\tilde{\mathrm{a}}}_{1\mathrm{m}}\\ {\tilde{\mathrm{a}}}_{21}& 1& \dots & {\tilde{\mathrm{a}}}_{2\mathrm{m}}\\ ⋮& ⋮& \ddots & ⋮\\ {\tilde{\mathrm{a}}}_{\mathrm{m}1}& {\tilde{\mathrm{a}}}_{\mathrm{m}2}& \dots & 1\end{array}\right)$$

(2)

where ${\tilde{\mathrm{a}}}_{\mathrm{i}\mathrm{j}}$ is a hesitant fuzzy number denoting the importance of ${\mathrm{C}\mathrm{r}}_{\mathrm{i}}$ over ${\mathrm{C}\mathrm{r}}_{\mathrm{j}}$ and m is the number of criteria.

Step 4: Calculate the Weights of Criteria Using ANP

(a) Normalization of Weights

The hesitant fuzzy weights for each criterion are computed by normalizing the pairwise comparison matrix:

$${\tilde{\mathrm{w}}}_{\mathrm{j}}={\displaystyle \frac{{\sum }_{\mathrm{i}=1}^{\mathrm{m}}{\tilde{\mathrm{a}}}_{\mathrm{i}\mathrm{j}}}{\mathrm{m}}, \mathrm{j} = 1, 2,\dots ,\mathrm{m}}$$

(3)

Here, $\mathrm{m}$ (earlier referred to as “PCM”) is the number of criteria, not the matrix itself.

(b) Form the Weighted Supermatrix

The normalized weights are organized into a supermatrix, which captures interdependencies among criteria and alternatives:

$$\mathrm{W}\mathrm{t}=\left[\begin{array}{cc}{\mathrm{W}\mathrm{t}}_{\mathrm{C}}& {\mathrm{W}\mathrm{t}}_{\mathrm{A}}\end{array}\right]$$

(4)

where ${\mathrm{W}\mathrm{t}}_{\mathrm{C}}$ is influence among criteria and ${\mathrm{W}\mathrm{t}}_{\mathrm{A}}$ influence of alternatives with respect to criteria.

Dimension of $\mathrm{W}\mathrm{t}$: If there are $\mathrm{m}$ criteria and $\mathrm{n}$ alternatives, the supermatrix has dimension $\left(\mathrm{m}+\mathrm{n}\right)\ast (\mathrm{m}+\mathrm{n})$. For example: For 2 criteria and 2 alternatives, $\mathrm{W}\mathrm{t}$ is a 4 × 4 matrix.

(c) Convergence to Limit Supermatrix

The weighted supermatrix is raised to successive powers until convergence is reached:

$${\mathrm{W}\mathrm{t}}_{\mathrm{l}\mathrm{i}\mathrm{m}\mathrm{i}\mathrm{t}}=\underset{\mathrm{k}\to \mathrm{\infty }}{\lim }{\mathrm{W}\mathrm{t}}^{\mathrm{k}}$$

(5)

(d) Convergence Criteria

The process is stopped when the difference between two successive matrices satisfies:

$${\mathrm{W}\mathrm{t}}_{\left(\mathrm{k}+1\right)}-{\mathrm{W}\mathrm{t}}_{\left(\mathrm{k}\right)}<{10}^{-4}$$

(6)

This ensures stable priority weights for criteria and alternatives.

Step 5: Construct the Weighted Hesitant Fuzzy Decision Matrix

The decision matrix is weighted by the final criteria weights obtained from ANP:

$${\tilde{\mathrm{v}}}_{\mathrm{i}\mathrm{j}}={\mathrm{w}\mathrm{t}}_{\mathrm{j}}.{\mathrm{d}\mathrm{m}}_{\mathrm{i}\mathrm{j}},\mathrm{f}\mathrm{o}\mathrm{r} \mathrm{a}\mathrm{l}\mathrm{l} \mathrm{i},\mathrm{j}$$

(7)

where ${\tilde{\mathrm{v}}}_{\mathrm{i}\mathrm{j}}$ is the weighted hesitant fuzzy evaluation of alternative ${\mathrm{A}\mathrm{t}}_{\mathrm{i}}$ under criterion Crj and ${\mathrm{d}\mathrm{m}}_{\mathrm{i}\mathrm{j}}$ is the evaluation of ${\mathrm{A}\mathrm{t}}_{\mathrm{i}}$ with respect to. C_rj.

Step 6: Determine the Positive Ideal Solution and Negative Ideal Solution

The Positive Ideal Solution (PIS) and Negative Ideal Solution (NIS) are determined for each criterion:

$${\widetilde{\mathrm{A}\mathrm{p}}}^{+}=\left\{\underset{\mathrm{i}}{\max } {\tilde{\mathrm{v}}}_{\mathrm{i}\mathrm{j}}\right\},{\widetilde{\mathrm{A}\mathrm{n}}}^{-}=\left\{\underset{\mathrm{i}}{\min } {\tilde{\mathrm{v}}}_{\mathrm{i}\mathrm{j}}\right\}$$

(8)

PIS represents the most favorable values, while NIS represents the least favorable.

Step 7: Calculate the Separation Measures

Calculate the separation measures of each alternative from the PIS and NIS:

Separation distance from PIS:

$${\mathrm{S}\mathrm{p}}_{\mathrm{i}}^{+}=\sqrt{{\sum }_{\mathrm{j}=1}^{\mathrm{m}}{\left({\tilde{\mathrm{v}}}_{\mathrm{i}\mathrm{j}}-{\widetilde{\mathrm{A}\mathrm{t}}}_{\mathrm{j}}^{+}\right)}^2}$$

(9)

Separation distance from NIS:

$${\mathrm{S}\mathrm{p}}_{\mathrm{i}}^{\_}=\sqrt{{\sum }_{\mathrm{j}=1}^{\mathrm{m}}{\left({\tilde{\mathrm{v}}}_{\mathrm{i}\mathrm{j}}-{\widetilde{\mathrm{A}\mathrm{t}}}_{\mathrm{j}}^{\_}\right)}^2}$$

(10)

${\mathrm{S}\mathrm{p}}_{\mathrm{i}}^{+}$ is separation distance of alternative ${\mathrm{A}\mathrm{t}}_{\mathrm{i}}$ from PIS and ${\mathrm{S}\mathrm{p}}_{\mathrm{i}}^{\_}$ is Separation distance of alternative ${\mathrm{A}\mathrm{t}}_{\mathrm{i}}$ from NIS.

Step 8: Calculate the Relative Closeness to the Ideal Solution

The Relative Closeness (${\mathrm{R}\mathrm{C}}_{\mathrm{i}}$) of each alternative is defined as:

$${\mathrm{R}\mathrm{C}}_{\mathrm{i}}={\displaystyle \frac{{\mathrm{S}\mathrm{p}}_{\mathrm{i}}^{-}}{{\mathrm{S}\mathrm{p}}_{\mathrm{i}}^{+}+{\mathrm{S}\mathrm{p}}_{\mathrm{i}}^{-}}}$$

(11)

where ${\mathrm{R}\mathrm{C}}_{\mathrm{i}}$ ranges from 0 to 1 and Higher ${\mathrm{R}\mathrm{C}}_{\mathrm{i}}$ values indicate alternatives closer to the ideal solution. Further, Here, ${\mathrm{C}}_{\mathrm{i}}$ refers to the closeness coefficient, while ${\mathrm{R}\mathrm{C}}_{\mathrm{i}}$ denotes the relative closeness. In this work, both represent the same measure, and the unified notation ${\mathrm{R}\mathrm{C}}_{\mathrm{i}}$ is adopted.

Step 9: Rank the Alternatives

Alternatives are ranked based on their ${\mathrm{R}\mathrm{C}}_{\mathrm{i}}$ values. The alternative with the highest ${\mathrm{R}\mathrm{C}}_{\mathrm{i}}$ is considered the most preferable.

This methodology provides a systematic approach for decision-making in environments where uncertainty and hesitation are present. It combines the strengths of ANP in handling interdependencies among criteria and TOPSIS in ranking alternatives based on their distance from ideal solutions.

4. Numerical Analysis and Results

Evaluating the integrity of IoMT sensors is a critical process that demands a significant investment in terms of time and resources. This meticulous undertaking requires careful consideration, especially given the growing demand for affordable IoMT sensors, sometimes prioritizing affordability over ensuring the integrity and protection of patients’ personal information. It is crucial to understand that even the smallest flaws in the design of IoMT sensors could be exploited by hackers, posing a serious risk to patients’ health. While international authorities periodically update guidelines for securing IoMT sensors, there is a compelling need to establish a standardized and highly precise approach for assessing the integrity of these sensors.

In response to this need, I have developed an integrated mechanism for evaluating the integrity of IoMT sensors using the HF ANP TOPSIS technique. Figure 3 present the integrated procedure in a step-by-step format. The first step involves collecting standard data, identifying the criteria, and selecting the methodology. For this assessment, I have chosen the HF ANP TOPSIS technique. Following methodology selection, I design the network structure of criteria based on the chosen methodology. In my case, the HF ANP TOPSIS technique is employed. The network structure is divided into levels 1, 2, and 3 for a comprehensive assessment.

To assess the integrity of an IoMT sensor, I have followed methodology and Figure 3 to identify and design the network of factors. According to Table 1 and Figure 3, the constructed network has three levels: at level 1, the network has one group of factors, including Functional Integrity (F1), Reliability (F2), Safety (F3), and Functionality (F4). At level 2, the network has four groups of factors, including Accuracy (F11), Failure Rate (F21), Dependability (F22), Patient Safety (F31), Operator Safety (F32), Core Functions (F41), Additional Features (F42), and level 3 has one group of factors, including Measurement Accuracy (F111), Dosage Accuracy (F112).

To gather the necessary data for assessment, I have designed a questionnaire and collected data from 578 experts in the relevant field, each with more than 5 years of experience in health informatics and cybersecurity. Subsequently, the data is filtered to include only the required inputs. Two types of data are included: subjective data used to calculate criteria weights and objective data based on alternatives.

After data collection and filtering, I have constructed pairwise comparison matrices for the factors at level 1 using Table 1, Figure 3, and Equations (1) and (2). Next, Equations (3) and (4) are used to calculate the local-level weights. Specifically, Equation (3) is used to calculate the normalized hesitant fuzzy weights, Equation (4) is used to form the supermatrix using the normalized weights, and Equations (5) and (6) is used to calculate the limit supermatrix by raising the weighted supermatrix to a sufficiently large power until it converges. This process is repeated for the factors at level 2.

Table 2. Pair-wise comparisons matrixes of the groups.

Characteristic A/Characteristic B	Hesitant Fuzzy Pair-Wise Comparisons Matrixes	Defuzzified Pair-Wise Comparisons Matrixes
Functional Integrity (F1)/Reliability (F2)	0.9450, 1.2445, 1.4568, 1.6454	1.2457
Functional Integrity (F1)/Safety (F3)	0.2545, 1.1125, 1.5654, 1.8300	1.3145
Functional Integrity (F1)/Functionality (F4)	0.4758, 0.4478, 0.5265, 1.3653	1.1674
Reliability (F2)/Safety (F3)	0.0255, 0.0654, 0.0958, 1.2874	0.4575
Reliability (F2)/Functionality (F4)	0.4452, 1.4521, 1.5698, 1.8989	0.7156
Safety (F3)/Functionality (F4)	0.4547, 0.4525, 1.4475, 1.9653	1.1745
Accuracy (F11)/Accuracy (F11)	0.3563, 0.4789, 0.6445, 1.0658	0.4975
Failure Rate (F21)/Dependability (F22)	0.0592, 0.5849, 0.9006, 1.5356	1.1445
Patient Safety (F31)/Operator Safety (F32)	1.1500, 1.4400, 1.7000, 2.0056	0.8974
Core Functions (F41)/Additional Features (F42)	0.3568, 0.4456, 0.8897, 1.1653	0.6975
Measurement Accuracy (F111)/Dosage Accuracy (F112)	0.2356, 0.2889, 0.3678, 0.6325	0.3745

and Figure 4 display the pairwise comparison matrices for all three levels.

To ensure clarity in the computation of weights, this paper employs the alpha-cut method, which has now been explicitly defined in the revised manuscript. The alpha-cut method is a widely used technique in fuzzy set theory that transforms a fuzzy set into a crisp interval at a specified confidence level (α). By selecting a particular α value, the fuzzy set is reduced to an interval containing all elements whose membership degree is greater than or equal to α. This transformation facilitates easier handling of uncertain or imprecise information, thereby simplifying the computational process while preserving decision-making reliability. In the context of this study, the alpha-cut method was used for normalizing the weights, ensuring that hesitant fuzzy information could be consistently incorporated into the pairwise comparison matrices.

Finally, Figure 3 illustrates the calculation of the weights of each factor in the designed network, while

Table 3. Global priority weights of the IoMT sensor integrity evaluation criteria derived using the hesitant fuzzy ANP method.

Characteristic	Symbols	Independent Weight of the Groups	Overall Weights Through Hierarchy	Percentage	Priority
Factors of Group 1 at Level 1
Functional Integrity	F1	0.301862	0.301862	30%	2
Reliability	F2	0.210246	0.210246	21%	3
Safety	F3	0.126984	0.126984	13%	4
Functionality	F4	0.360909	0.360909	36%	1
Factors of Groups 1, 2, 3, 4 at Level 2 and Group 1 at Level 3
Measurement Accuracy	F111	0.09465	0.11383	11%	4
Dosage Accuracy	F112	0.16514	0.19860	20%	2
Failure Rate	F21	0.11149	0.09339	9%	5
Dependability	F22	0.08809	0.07379	7%	7
Patient Safety	F31	0.15910	0.08049	8%	6
Operator Safety	F32	0.11663	0.05900	6%	8
Core Functions	F41	0.17097	0.24584	24%	1
Additional Features	F42	0.09393	0.13506	14%	3

presents both the local and global weights for each level, providing a comprehensive overview of the decision-making structure.

Subsequently, the author utilizes the HF TOPSIS method outlined in Steps (5)–(9) and Figure 3 to evaluate the performance of the alternatives. Ensuring the integrity of IoMT sensors is crucial for safeguarding patient safety and optimizing healthcare treatments. Real-time applications, or alternatives, of IoMT sensor integrity play a pivotal role in monitoring, maintaining, and enhancing the performance of these sensors [15,16,17,18,19].

In the present research work, we explore five alternatives or real-time applications, each addressing specific aspects of IoMT sensor integrity. Remote Patient Monitoring (RPM) involves the continuous monitoring of a patient’s vital signs, such as heart rate, blood pressure, and oxygen levels, utilizing IoMT sensors. The real-time assessment of these sensors ensures the accuracy and reliability of data, enabling healthcare providers to make timely decisions and interventions based on patient conditions.

Smart Infusion Pumps (SIP), crucial for delivering medications and fluids intravenously, benefit from real-time integrity monitoring. This ensures accurate drug dosages, consistent flow rates, and safe sensor operation, with alerts and alarms triggered in case of anomalies. Implantable IoMT Sensors (IMD), including pacemakers and insulin pumps, require continuous real-time integrity monitoring to ensure their proper functioning within the body. Monitoring encompasses assessing battery life, sensor status, and communication with external programming sensors.

In telemedicine and telehealth applications (TT), IoMT sensors play a vital role in remote diagnostics and treatment. Real-time integrity checks are crucial for confirming the accuracy of data transmitted from patients’ homes or remote locations. This ensures healthcare providers can make informed decisions based on trustworthy data. Operating Room Equipment (ORE), including surgical instruments and equipment like anesthesia machines and monitoring sensors, necessitates continuous integrity monitoring. Real-time assessments help maintain patient safety during surgery by ensuring the functionality and accuracy of these sensors.

These alternatives collectively contribute to the overall quality of healthcare delivery. By enhancing patient safety, improving the accuracy of diagnoses and treatments, and reducing the risk of adverse events, real-time monitoring of IoMT sensor integrity empowers healthcare professionals to make informed decisions and respond promptly to any anomalies or malfunctions. This, in turn, leads to better patient outcomes. Following Figure 3 and Equations (6) and (7), construct the weighted decision matrix by multiplying the hesitant fuzzy decision matrix by the criteria weights. Equations (7)–(10) are used to determine the Positive Ideal Solution (PIS) and the Negative Ideal Solution (NIS). Finally, Equation (11) is used to calculate the relative closeness Ci of each alternative to the ideal solutions.

Table 4. Final ranking of IoMT medical device sensor alternatives based on the unified hesitant fuzzy ANP–TOPSIS evaluation.

Factors at Final Level	RPM	SIP	IMD	TT	ORE
Measurement Accuracy	0.9193	0.9051	0.9193	0.9051	0.7336
	0.6493	0.6492	0.6493	0.6492	0.6578
	0.7642	0.9193	0.7642	0.7641	0.7571
Dosage Accuracy	0.8832	0.6493	0.9193	0.9051	0.7336
	0.9193	0.9193	0.6493	0.6492	0.6578
	0.6493	0.6493	0.7642	0.7641	0.7571
Failure Rate	0.6492	0.6493	0.6492	0.6578	0.9191
	0.9193	0.7642	0.6492	0.6493	0.6492
	0.6493	0.9193	0.9193	0.7642	0.7641
Dependability	0.9193	0.6493	0.6493	0.9193	0.9051
	0.6493	0.7642	0.9193	0.6493	0.6492
	0.6492	0.6493	0.6492	0.6493	0.6492
Patient Safety	0.9193	0.7642	0.9193	0.7642	0.7641
	0.6492	0.6493	0.6492	0.9193	0.9051
	0.9193	0.7642	0.7641	0.6493	0.6492
Operator Safety	0.6493	0.9193	0.6492	0.6493	0.6492
	0.9193	0.6493	0.9193	0.7642	0.7641
	0.6492	0.6493	0.6493	0.9193	0.9051
Core Functions	0.6492	0.6493	0.6492	0.6493	0.6492
	0.9193	0.7642	0.9193	0.7642	0.7641
	0.6493	0.9193	0.6493	0.9193	0.9051
Additional Features	0.9193	0.6493	0.9193	0.6493	0.6492
	0.6492	0.6493	0.6492	0.6493	0.6492
	0.9193	0.7642	0.9193	0.7642	0.7641

illustrates the computations and outcomes obtained.

Table 5. Final ranking of alternatives.

S. No.	Applications	Closeness Coefficients
1	Remote Patient Monitoring (RPM)	0.652547
2	Smart Infusion Pumps (SIP)	0.563325
3	Implantable IoMT Sensors (IMD)	0.465878
4	Telemedicine and Telehealth (TT)	0.512265
5	Operating Room Equipment (ORE)	0.498755

and Figure 5 present the final results of the assessment of IoMT sensor integrity.

This study employed the HF TOPSIS method to assess and rank various competitive alternatives in the field of IoMT sensors. These alternatives included RPM, SIP, IMD, TT, and ORE. I evaluated these alternatives based on preference scores or relative proximity scores. The objective was to identify the most suitable application for addressing critical concerns related to the reliability of IoMT diagnoses and specific criteria relevant to the healthcare landscape. After conducting a thorough investigation, RPM emerged as the superior choice among these applications. RPM consistently outperformed the others in my assessment, making it the preferred solution for addressing the unique challenges and requirements of the healthcare system. Additionally, my use of hesitant fuzzy decision-making strategies highlighted RPM’s efficiency in sensors of patient health, which is crucial for informed decision-making in the Saudi healthcare sector.

In this present work, I conducted a sensitivity analysis to ensure the reliability of conclusions in the face of ambiguities and differences in the opinions of individuals considered experts. This study, renowned for its efficiency in validating data, was carried out meticulously throughout this research by repeatedly modifying variables. Given that this article identified eight components reflecting the final network level in Table 1, my research comprised a total of eight different test cases, labeled Test Cases 1 through 8. These test cases are designed with adjusting variables for this specific outcome. As I adjusted variables based on the findings, it became evident that alternative 1 (remote patient monitoring) consistently ranked as the best option. Figure 6 depicts the experimental outcomes compared to the initially used weightings.

Consistent with the initial weightings, Test 1 (representing remote patient monitoring) consistently achieved the highest degree of satisfaction, as indicated in Figure 6. The results of all eight experiments remained consistent. Despite facing various situations and alterations during the review process, these findings offer additional evidence that RPM remains the most reliable option for meeting critical healthcare requirements. Specifically, it underscores the significance of RPM in addressing challenges in the healthcare industry and enhancing patient care.

IoMT sensor manufacturers are progressively implementing software solutions and encryption techniques to secure data transfers and mitigate potential data breaches. This demonstrates a growing awareness of the critical role IoMT sensor integrity plays in safeguarding patient well-being and maintaining the quality of healthcare services. The implications of failing to uphold the integrity of IoMT sensors are profound, as patient safety and the efficacy of IoMT treatments rely heavily on the secure and proper functioning of these sensors.

Despite these challenges, the IoMT sensor security market remains dynamic and competitive. Industry leaders are actively innovating their products and establishing strategic partnerships to bolster their market presence. In a time when IoMT sensors are playing an increasingly pivotal role in patient care, ensuring the integrity and security of these sensors is a paramount priority. It not only enhances individual well-being but also bolsters the overall healthcare industry’s resilience.

Based on my research, I have proposed a novel approach to assess the security integrity of IoMT sensors. This strategy aims to improve the level of care that IoMT specialists offer. To validate my research findings, I conducted a comprehensive analysis, comparing the proposed approach with various Multi-Criteria Decision-Making (MCDM) methods in the relevant healthcare landscape. The goal was to conclusively determine whether the results of the achieved analysis align with the superiority of the proposed method.

In this study, I evaluated the performance of my proposed method by comparing it with MCDM methods such as HF ANP-VIekriterijumsko KOmpro-miso Rangiranje (VIKOR) and HF ANP-Elimination Et Choix Traduisant la Realité (ELECTRE) [5,6,9,12,13,14,15,16]. My findings, when using HF ANP-TOPSIS, HF ANP-VIKOR, and HF ANP-ELECTRE, consistently yielded results similar to those obtained with the HF ANP-TOPSIS technique. To illustrate these comparisons,

Table 6. Comparisons with other techniques.

S. No.	Applications	HF ANP-TOPSIS	HF ANP-VIKOR	HF AHP-TOPSIS	HF ANP-ELECTRE
1	RPM	0.652547	0.652547	0.644577	0.657854
2	SIP	0.563325	0.574855	0.574589	0.567859
3	IMD	0.465878	0.452458	0.456368	0.466532
4	TT	0.512265	0.545255	0.515478	0.514574
5	ORE	0.498755	0.485697	0.478553	0.497758

and Figure 7 provide visual representations of the differences between these methods.

Table 6 presents a comparative analysis of the Hesitant Fuzzy ANP-TOPSIS technique against other prominent MCDM methods, including HF ANP-VIKOR, HF ANP-TOPSIS, and HF ANP-ELECTRE, across various applications. The data reflect a consistent performance by HF ANP-TOPSIS, often yielding slightly higher or equivalent values compared to the other techniques.

Figure 7 shows that the rankings generated by the four HF-based methods are highly consistent, with only minor variations. This supports the robustness and stability of the unified hesitant fuzzy ANP–TOPSIS approach, as it produces results that align closely with other HF-based decision-making methods.

For instance, in the RPM application, HF ANP-TOPSIS achieves a score of 0.569547, slightly outperforming HF ANP-VIKOR (0.568547) and closely aligning with HF ANP-ELECTRE (0.569758). A similar pattern is observed in the SIP application, where HF ANP-TOPSIS registers a score of 0.497458, marginally higher than HF ANP-VIKOR (0.496525) and almost identical to HF ANP-ELECTRE (0.497587). In the IMD application, HF ANP-TOPSIS again demonstrates its effectiveness with a score of 0.394523, closely mirroring the results from the other methods.

For statistically validating the results, we have performed an ANOVA test on the above Table 6 and F statistic value is 0.0116 and the p-value is 0.9982. Since the value of p is less than 0.05, we consider the null hypothesis rejected, and this confirms that HF-ANP TOPSIS slightly outperforms other methods.

The graphical representation in Figure 6 further illustrates these comparisons, highlighting the consistent performance of HF ANP-TOPSIS across different applications. This statistical analysis underscores the robustness and reliability of the HF ANP-TOPSIS technique, particularly in complex decision-making environments, where precision and accuracy are paramount.

5. Discussion

The following segment explores the advantages and limitations of employing this integrated hybrid hesitant fuzzy-based healthcare system for assessing IoMT sensor integrity. The integrated approach allows for a comprehensive assessment of IoMT sensor-based devices’ integrity. By considering multiple core factors and their interdependencies, it offers a holistic view of the sensor’s performance. HF ANP’s capacity to incorporate expert judgments through pairwise comparisons accommodates subjectivity and captures the nuanced understanding of domain experts. This ensures that assessments benefit from both data-driven and expert-driven insights. In the healthcare landscape, where outcomes are influenced by numerous variables and factors, the hesitant fuzzy logic employed in both HF ANP and HF TOPSIS can effectively capture and model uncertainties. HF ANP assigns relative weights to different integrity factors based on expert judgments, quantifying their importance.

This prioritization aids decision-makers in focusing on the most critical aspects of sensor integrity. HF TOPSIS ranks potential solutions, providing a clear order of preference based on their proximity to the ideal integrity state and deviation from the least desirable state, facilitating decision-making and resource allocation. Increased accuracy values directly improve the reliability of healthcare decision-making. For example, the accuracy of Respiratory Rate per Minute (RPM) readings affects the trustworthiness of data collected from patients. High-accuracy RPM measurements allow clinicians to detect abnormal breathing patterns more precisely, enabling timely interventions such as oxygen therapy or further diagnostics. Conversely, lower accuracy may lead to misclassification of a patient’s condition, resulting in delayed or inappropriate treatment. Therefore, improving sensor accuracy not only refines the decision-making process but also strengthens overall clinical outcomes. The empirical validation of this integrated technique across diverse IoMT sensors demonstrates its applicability and effectiveness in real-world healthcare settings, enhancing its credibility.

However, implementing HF ANP and HF TOPSIS necessitates a solid understanding of this hybrid hesitant fuzzy expert system. Training and expertise are essential to ensure the correct application of these methods, which can be time-consuming and resource-intensive. The effectiveness of this approach relies on the availability of accurate and comprehensive data, which can be challenging to obtain, especially for new or innovative IoMT sensor-based devices in the Saudi context. While expert judgments are valuable, they introduce an element of subjectivity into the assessment process, and variations in expert opinions can impact the results, requiring careful consideration and sensitivity analysis.

This work could be expanded to a number of different areas by future research. First, in order to give a more comprehensive evaluation of healthcare monitoring systems, it is possible that additional sensor modalities, including glucose, SpO2, and electrocardiograms might be integrated into the framework. Secondly, the generalisability and clinical reliability of the study would be improved through the implementation of a large-scale validation within a number of Saudi Arabian institutions and patient categories. Thirdly, the use of artificial intelligence (AI) and machine learning algorithms has the potential to further enhance weighing criteria and provide support for dynamic decision-making in real-time settings.

Last but not least, longitudinal studies could investigate the ways in which enhanced sensor prioritisation influences long-term patient outcomes, financial expenses associated with healthcare, and operational efficiency in medical facilities. In future studies, researchers will look at more factors, such as how well the sensors work with hospital information systems, how much energy they use for constant monitoring, how quickly they can respond to emergencies, and how long the sensors last. Adding these factors will make the suggested framework more complete and make it easier to use in a wider range of healthcare situations.

6. Conclusions

The integrated hybrid hesitant fuzzy-based healthcare system, utilizing HF ANP and HF TOPSIS, offers a systematic and robust approach to enhancing patient safety and ensuring the integrity of IoMT sensors. This methodology efficiently prioritises essential integrity parameters and offers thorough, data-driven insights on sensor performance by including expert judgements and compensating for uncertainties. Although these strategies are sophisticated and require a lot of resources to implement, they have considerable benefits and can be used in real-world healthcare settings.

This research proposed a unified hesitant fuzzy-based healthcare system that integrates HF ANP and HF TOPSIS to evaluate and ensure the integrity of IoMT sensors. The framework systematically addressed the complex interdependencies among integrity criteria while incorporating expert judgments and managing uncertainty. The findings highlighted Remote Patient Monitoring (RPM) as the most effective application for ensuring IoMT sensor integrity, consistently outperforming alternatives across sensitivity analyses and comparative evaluations. Furthermore, the comparative study against other MCDM techniques, such as HF ANP-VIKOR and HF ANP-ELECTRE, confirmed the robustness and reliability of the HF ANP-TOPSIS approach.

Future efforts should focus on simplifying implementation processes, developing user-friendly tools, and expanding applications to emerging healthcare technologies to make these advanced assessment methods more accessible to healthcare professionals across the globe. Acknowledging both the strengths and challenges of this approach is essential for its successful adoption and for advancing excellence in healthcare delivery.

By doing so, the proposed approach can significantly contribute to strengthening healthcare resilience, enhancing patient outcomes, and ensuring the secure integration of IoMT technologies in an increasingly digital healthcare landscape.