Evaluating the sustainability of smart technology applications in healthcare after the COVID-19 pandemic: A hybridising subjective and objective fuzzy group decision-making approach with explainable artificial intelligence

Abstract

During the COVID-19 pandemic, some smart technology applications were more effective than had been expected, whereas some others did not achieve satisfactory performance. Consequently, whether smart technology applications in healthcare are sustainable is a question that warrants investigation. To address this question, a hybridising subjective and objective fuzzy group decision-making approach with explainable artificial intelligence was proposed in this study and then used to evaluate the sustainability of smart technology applications in healthcare. The contribution of this research is its subjective evaluation of the sustainability of smart technology applications followed by correction of the evaluation outcome on the basis of the applications’ objective performance during the COVID-19 pandemic. To this end, a fuzzy nonlinear programming model was formulated and optimised. In addition, the impact of several major global events that occurred during the pandemic on the sustainability of smart technology applications was considered. The proposed methodology was applied to evaluate the sustainability levels of eight smart technology applications in healthcare. According to the experimental results, three applications—namely healthcare apps, smartwatches, and remote temperature scanners—are expected to be highly sustainable in healthcare, whereas one application, namely smart clothing, is not.

Keywords

Smart technology healthcare sustainability explainable artificial intelligence fuzzy group decision-making

Introduction

Before the outbreak of the COVID-19 pandemic, smart technologies were being used extensively to improve people's quality of life.^1,2 However, since the onset of the pandemic, some smart technologies (e.g. remote temperature scanners) have still been frequently employed, whereas the use of others (e.g. wireless medical sensor networks) has been limited by their low effectiveness or efficiency.^3,4 Several studies have investigated the application of contact-tracing apps and sensors.^5–7 According to Khan et al.,⁸ robotics (including drones), artificial intelligence (AI), masks, and sensors were all relatively frequently used to minimise the spread of COVID-19, and according to Chen and Wang,⁹ robots, smartphone apps, wearable sensors and devices, and wireless medical sensor networks were considered the most effective smart technology applications during the pandemic. By contrast, the application of remote temperature scanners was deemed unfavourable. In summary,

Most studies conducted before or during the COVID-19 pandemic focused on the use of smart technology applications in healthcare^2,10;

State-of-the-art development has tended to focus on comparing the prevalence of various smart technology applications in healthcare during the COVID-19 pandemic and recommending the possible applications of such technologies^5,8; and

Few studies have discussed the reasons for this difference.^9,11

Thus, the following research gaps need to be addressed:

Smart technologies that will likely be widely used after the COVID-19 pandemic may be different from those widely used during the pandemic; this topic was discussed by Chen and Wang.⁹ However, the COVID-19 pandemic did not end as expected, resulting in a low reference value of previous analysis results. For example, Chen and Wang⁹ did not consider remote temperature scanners as the most suitable smart technology application, yet as COVID-19 remained prevalent in society, the cost-effectiveness of remote temperature scanners increased. Now, remote temperature scanners are used extensively worldwide.

Major global events that occurred during the COVID-19 pandemic—including the US–China trade war,¹² severe delays and congestion at ports and terminals due to COVID-19,¹³ a global shortage of semiconductor chips,¹⁴ inflation,¹⁵ and the Ukraine–Russia war¹⁶—have influenced the application of smart technologies in healthcare.¹⁷

Existing methods based on AI are not easy to understand or communicate.¹⁸

To fill these gaps, this study aimed to achieve the following goals:

To update information on the factors critical to the sustainability of smart technology applications in healthcare (i.e. the possibility that a smart technology application will continue to be used in healthcare after the COVID-19 pandemic);

To consider the impact of major global events on the sustainability of smart technology applications in healthcare; and

To enhance the explainability of AI methods by using explainable AI (XAI).

This study began by conducting a survey of the relevant literature and identifying factors critical to the sustainability of smart technology applications in healthcare. Subsequently, to evaluate the sustainability of smart technology applications, a hybridising subjective and objective fuzzy group decision-making approach was proposed in which the sustainability of a smart technology application in healthcare is evaluated both subjectively and objectively. Specifically, some smart technology applications can be evaluated only objectively, whereas others can also be evaluated subjectively on the basis of their performance during the COVID-19 pandemic. The hybridising subjective and objective fuzzy group decision-making approach is a fuzzy analytic hierarchy process (FAHP) method. Other fuzzy group decision-making methods—such as the fuzzy (ordered) weighted average, fuzzy Delphi method, fuzzy technique for order preference by similarity to the ideal solution, fuzzy measuring attractiveness through a categorical-based evaluation technique, fuzzy preference ranking organisation method for enriched evaluation, fuzzy elimination and choice expressing reality method, and fuzzy TODIM method—are also applicable^19–21; however, unless the costs of decisions can be quantified and compared, determining which fuzzy group decision-making method is the optimal method is difficult. Nevertheless, among FAHP methods, the proposed hybridising subjective and objective fuzzy group decision-making approach achieves high precision by deriving the exact values of fuzzy priorities, whereas other FAHP methods, such as the fuzzy geometric mean (FGM²²) and fuzzy extent analysis (FEA²³), only approximate the values of these priorities. Furthermore, several XAI techniques were applied to enhance the explainability of the hybridising subjective and objective fuzzy group decision-making approach. XAI techniques can be divided into two categories: those that enhance the practicality of an existing AI technology by explaining its execution process or results, and those that improve the effectiveness of existing AI technologies by incorporating easy-to-interpret visual features.^24,25 The XAI techniques applied in this study fell within the first category.

The remainder of this paper is organised as follows: “Literature review” section presents a literature review. “Key factors affecting the sustainability of smart technology applications in healthcare” section identifies the key factors affecting the sustainability of smart technology applications in healthcare. “Methodology” section introduces the hybridising subjective and objective fuzzy group decision-making approach. “Case study” section details the application of the proposed methodology to evaluate the sustainability of eight smart technology applications in healthcare. Finally, “conclusion” section concludes this paper and suggests directions for future research.

Literature review

Kakderi et al.²⁶ highlighted the reality that the adoption of smart technologies during the COVID-19 pandemic was far less extensive than had been expected, despite government policies aimed at transitioning to smart cities. This reality has called into question whether the application of smart technologies in healthcare, especially at the individual level, is definitely sustainable and motivated the present assessment of the sustainability of smart technologies applied in healthcare.

As discussed by Waheed and Shafi,³ in theory, the application of smart technologies can slow the spread of COVID-19. However, the practical benefits of smart technology applications appear to be far fewer in number than researchers had expected. In the face of a widespread pandemic, some traditional disposal methods are more effective, efficient, and convenient. Therefore, when cost efficiency is a priority, people may abandon the pursuit of expensive smart technologies. Some major global events that occurred during the COVID-19 pandemic, such as the aforementioned shortage of semiconductor chips and rising inflation, have increased people's economic burden, and this new reality will likely reduce people's willingness to use smart technologies in healthcare. This topic has yet to be explored in the literature, and thus the present study intended to analyse it.

Chen and Wang⁹ listed some characteristics of smart technology applications in healthcare after the COVID-19 pandemic, namely allowing users to regain their freedom of movement and travel,^7,26 meeting practical needs, having high effectiveness, the ability to be continually improved through software developments, and being cost-effective. In that study, restoring people's freedom of movement was considered the most crucial characteristic. Today, however, freedom of movement is no longer restricted in most regions. Therefore, other characteristics will likely start to become more important, and the present study intends to determine which characteristics these are.

Monitoring of body temperature has served as a common measure to control the spread of COVID-19.²⁷ For this reason, remote temperature scanners are popular, although they are more expensive and larger than wearable smart devices, such as smartwatches, smart bracelets, and smart clothing. In addition, remote temperature scanners can monitor many people visiting a public space at once, whereas one wearable smart device can monitor only one individual. Therefore, the market for such wearable smart devices may not expand as expected after the COVID-19 pandemic. In summary, the sustainability of certain applications in healthcare is questionable. Clarifying this question was one of the motivations for this study.

Key factors affecting the sustainability of smart technology applications in healthcare

The literature review shows that the sustainability of a smart technology application after the COVID-19 pandemic can be evaluated by considering the following aspects:

A smart technology application is sustainable if it can provide value-added services on the basis of vaccination information^28–30: Although the demand for vaccination information from travellers is now in decline,³¹ providing different services for travellers with unequal vaccination statuses can still minimise health risks.³²

A smart technology application is sustainable if it is cost-effective^6,10,33: The cost of a smart technology application is determined by its supply and demand, both of which are vulnerable to influence by global events.^34,35 Furthermore, the cost-effectiveness of smart technology applications cannot be directly assessed. For example, a remote temperature scanner can be used to monitor the body temperature of thousands of customers who visit a department store; therefore, the more customers who visit that store, the more cost-effective is the remote temperature scanner.

A smart technology application is sustainable if it can promote healthy mobility among the public^30,36: A smart technology application can reduce the risk of people becoming infected with a virus while they are on the move and remind people to take measures against such infection.

A smart technology application is sustainable if it is necessary or irreplaceable^36–38: During the COVID-19 pandemic, public body temperature monitoring using remote temperature scanners seemed to be more effective than having individuals use a smart bracelet to monitor their own body temperature.

A smart technology application is sustainable if it can be combined with other smart technology applications to achieve synergy^6,10: Remote body temperature scanners are often combined with noncontact alcohol sterilisers or automated camera devices (to track footprints).

A smart technology application is sustainable if it is easy to implement and maintain^36,38,39: Most smart devices can be operated intuitively. However, smart devices are difficult to repair when they malfunction. For example, Djuric⁴⁰ estimated that an Apple Watch smartwatch (original version) is only 50% repairable.

During the COVID-19 pandemic, several major global events had a considerable impact on the demand and supply of smart technologies; that such impacts are possible is not conducive to the sustainability of using smart technology applications, as illustrated in Figure 1.

Figure 1.

Factors affecting the sustainability of a smart technology application in healthcare.

Major global events can influence the application of smart technologies in healthcare in two ways: (1) by raising the threshold for user acceptance of the technology application in question and (2) by changing users’ preferences related to specific criteria, as illustrated in Figure 2. In addition, a global event can give rise to other such events; for example, both the US–China trade war and delays and congestion at ports and terminals (owing to the COVID-19 pandemic) led to a global shortage of semiconductor chips.

Figure 2.

Illustration of how global events give rise to other such events.

Methodology

Procedure

A hybridising subjective and objective fuzzy group decision-making approach with XAI was proposed in this study; it can be used to evaluate the sustainability of smart technology applications in healthcare. Fuzzy set theory has been used extensively to make practical management decisions related to healthcare, such as those related to the evaluation of medical waste disposal strategies and techniques,⁴¹ the prioritisation of lean-supply-chain management initiatives in healthcare service operations,⁴² and the assessment of patients’ health check results.⁴³

The proposed methodology comprises the following steps:

Step 1. Form a team of experts in related fields.

Step 2. Compare (or modify) the relative priorities on the basis of criteria in linguistic terms.⁴⁴

Step 3. If the fuzzy judgment matrix is consistent, proceed to the next step; otherwise, return to Step 2.

Step 4. Aggregate the fuzzy judgment matrixes of all the experts by using the FGM approach.⁴⁵

Step 5. Evaluate the sustainability of a smart technology application in healthcare subjectively and in linguistic terms⁴⁶ on the basis of its effectiveness during the COVID-19 pandemic.

Step 6. Solve a fuzzy nonlinear programming (FNLP) problem to derive the fuzzy priorities of criteria from the aggregation result.

Step 7. Evaluate the sustainability of a smart technology application in healthcare objectively on the basis of the derived fuzzy priorities.⁴⁷

As illustrated in Figure 3, the proposed methodology is novel for the following reasons:

The process of evaluating a methodology's sustainability is treated as a hybrid (supervised and unsupervised) optimisation problem, whereas in most existing methods, the problem in question is treated as an unsupervised fuzzy group decision-making process.^2,48,49

The sustainability of a smart technology application in healthcare is evaluated both subjectively and objectively.

Figure 3.

Procedure of the proposed methodology.

Steps 1–3

First, a team of K experts is formed. Expert k evaluates the relative priority of criterion i over criterion j by using a linguistic term that can be mapped to a triangular fuzzy number ${\tilde{a}}_{i j} (k)$ , which is usually interpreted as “criterion i is ${\tilde{a}}_{i j} (k)$ times more important than criterion j according to expert k,” or k = 1 ∼ K, i, j = 1 ∼ n, and i ≠ j. To derive the fuzzy priorities of the criteria, the fuzzy sum of the squared error is minimised as follows:

Min \tilde{S S E} (k) = \sum_{i = 1}^{n} \sum_{j \neq i} {(\frac{{\tilde{w}}_{i} (k)}{{\tilde{w}}_{j} (k)} (-) {\tilde{a}}_{i j} (k))}^{2}

(1)

where

{\tilde{w}}_{i} (k)

is the priority of criterion i according to expert k, and (−) denotes fuzzy subtraction. The value of the objective function can be minimised by applying the following fuzzy eigenanalysis method⁴⁴:

d e t (\tilde{A} (k) (-) \tilde{λ} (k) I) = 0

(2)

(\tilde{A} (k) (-) \tilde{λ} (k) I) (\times) \tilde{x} (k) = 0

(3)

where

\tilde{w} (k) = N (\tilde{x} (k))

(4)

where

\tilde{λ} (k)

and

\tilde{x} (k)

are the fuzzy eigenvalue and eigenvector of

\tilde{A} (k)

, respectively; det() is the determinant function; N() is the normalisation function; and ( × ) denotes fuzzy multiplication. Alpha-cut operations (ACOs) can be employed to derive the exact values of the fuzzy priorities as follows⁵⁰:

d e t (A (k) (α) - λ (k) (α) I) = 0; α \in [0, 1]

(5)

(A (k) (α) - λ (k) (α) I) x (k) (α) = 0; α \in [0, 1]

(6)

w (k) (α) = N (x (k)) (α); α \in [0, 1]

(7)

where

A (k) (α)

λ (k) (α)

, and

x (k) (α)

denote the left or right α cuts of the fuzzy variables. By contrast, other existing methods—such as the FGM, FEA, and fuzzy inverse of the column sum method—are subject to inaccuracy and imprecision.^2,22,23,51 In addition, fuzzy relation methods optimise different objective functions.⁵²

The priorities of the criteria are usually presented in the form of a bar chart to facilitate a comparison of the criteria's relative values, as illustrated in Figure 4. By contrast, the fuzzy priorities of the criteria are usually compared using a line graph, as shown in Figure 5. However, presenting the fuzzy priority and the performance of a criterion simultaneously is difficult. Therefore, in the proposed methodology, two XAI techniques, namely the colour management technique and the common expression technique,⁴⁹ are applied, and a gradient bar chart is plotted to illustrate the fuzzy priorities of the criteria (Figure 6), where darker colours represent higher membership levels. The fuzzy priority and (normalised) performance of a criterion are presented simultaneously using a grouped bar chart, as illustrated in Figure 7. Figure 5 can be mapped to Figure 7, as illustrated in Figure 8.

Figure 4.

Bar chart to compare the priorities of the criteria.

Figure 5.

Line graph to illustrate the fuzzy priorities of the criteria.

Figure 6.

Gradient bar chart to illustrate the fuzzy priorities of the criteria.

Figure 7.

Grouped bar chart to compare the fuzzy priorities and (normalised) performance levels of the criteria simultaneously.

Figure 8.

Figure 5 mapped to Figure 7.

Next, $\tilde{A} (k)$ is considered consistent if its fuzzy consistency ratio $\tilde{C R} (k)$ is lower than 0.1:

\tilde{C R} (k) = \frac{\tilde{λ} (k) - n / n - 1}{R I}

(8)

where RI is the random consistency index⁴⁴; otherwise,

\tilde{A} (k)

must be modified.⁵³

Step 4

The fuzzy judgment matrixes of all the experts are aggregated using the FGM as follows⁵⁴:

\begin{aligned} {\tilde{a}}_{i j} = & \tilde{F G M} ({{\tilde{a}}_{i j} (k)}) \\ = \sqrt[K]{\prod_{k = 1}^{K} {\tilde{a}}_{i j} (k)} ≅ {\begin{matrix} (\sqrt[K]{\prod_{k = 1}^{K} a_{i j 1} (k)}, \sqrt[K]{\prod_{k = 1}^{K} a_{i j 2} (k)}, \sqrt[K]{\prod_{k = 1}^{K} a_{i j 3} (k)}) i f & \sqrt[K]{\prod_{k = 1}^{K} a_{i j 2} (k)} \geq \sqrt[K]{\prod_{k = 1}^{K} a_{j i 2} (k)} \\ 1 / (\sqrt[K]{\prod_{k = 1}^{K} a_{j i 1} (k)}, \sqrt[K]{\prod_{k = 1}^{K} a_{j i 2} (k)}, \sqrt[K]{\prod_{k = 1}^{K} a_{j i 3} (k)}) & o t h e r w i s e \end{matrix} \end{aligned}

(9)

By contrast, an aggregation result obtained through other methods, such as the fuzzy weighted average method, may lead to an unreasonable value or membership level.^21,54 The traceable aggregation technique⁴⁹ can be used to explain the aggregation process and result, as illustrated in Figure 9.

Figure 9.

Explaining the aggregation process and result by applying the traceable aggregation technique.

Steps 5 and 6

The COVID-19 pandemic presented an opportunity to assess the sustainability of smart technology applications in healthcare. Some smart technology applications that were effective during the pandemic are now expected to be sustainable, whereas some are not, and for some others, their potential sustainability is unknown. In addition, the occurrence of specific major global events has caused people to re-examine whether they need healthcare-related smart technology applications, and this re-examination has further affected the sustainability of such applications.

Users are generally interested in the outcome rather than the process of evaluating the sustainability of smart technology applications in healthcare. In this study, the sustainability of smart technology application l evaluated subjectively by the experts is ${\tilde{S}}_{l}$ , whereas that evaluated objectively is ${\tilde{O}}_{l}$ ; l = 1 ∼ L. The value of ${\tilde{S}}_{l}$ exists only at certain values of l. In theory, ${\tilde{O}}_{l}$ should be as close to ${\tilde{S}}_{l}$ as possible, as illustrated in Figure 10:

{\tilde{O}}_{l} ≅ {\tilde{S}}_{l}

(10)

Figure 10.

Comparing objective and subjective evaluations of sustainability.

Consequently, the following FNLP problem needs to be solved to derive the fuzzy priorities of the criteria:

Min \tilde{S S E} = \sum_{i = 1}^{n} \sum_{j \neq i} {(\frac{{\tilde{w}}_{i}}{{\tilde{w}}_{j}} (-) {\tilde{a}}_{i j})}^{2}

(11)

subject to

{\tilde{O}}_{l} ≅ {\tilde{S}}_{l}; l = 1 \sim L

(12)

{\tilde{O}}_{l} = \sum_{i = 1}^{n} ({\tilde{w}}_{i} (\times) {\tilde{p}}_{l i}); l = 1 \sim L

(13)

\sum_{i = 1}^{n} {\tilde{w}}_{i} = 1

(14)

{\tilde{w}}_{i} \in R^{+}; i = 1 \sim n

(15)

where

{\tilde{p}}_{l i}

denotes the performance of smart technology application l for optimising criterion i, or l = 1 ∼ L and i = 1 ∼ n. The FNLP problem then needs to be converted into an equivalent crisp problem if it is to be easily solved. First, an ACO is applied to approximate the objective function with

Min S S E ≅ \sum_{α = 0}^{1} (\frac{\sum_{i = 1}^{n} \sum_{j \neq i} ((w_{i}^{L} (α) / w_{j}^{R} (α)) - a_{i j}^{L} (α))^{2} + \sum_{i = 1}^{n} \sum_{j \neq i} ((w_{i}^{R} (α) / w_{j}^{L} (α)) - a_{i j}^{R} (α))^{2}}{2})

(16)

In this case, replacing

w_{i}^{L} (α) / w_{j}^{R} (α)

and

w_{i}^{R} (α) / w_{j}^{L} (α)

with

r_{i j}^{L} (α)

and

r_{i j}^{R} (α)

, respectively, changes Equation (16) into

\begin{aligned} Min S S E ≅ \frac{1}{2} \sum_{α = 0}^{1} \\ (\sum_{i = 1}^{n} \sum_{j \neq i} (r_{i j}^{L} (α) - a_{i j}^{L} (α))^{2} + \sum_{i = 1}^{n} \sum_{j \neq i} (r_{i j}^{R} (α) - a_{i j}^{R} (α))^{2}) \end{aligned}

(17)

where

w_{i}^{L} (α) = r_{i j}^{L} (α) w_{j}^{R} (α)

(18)

w_{i}^{R} (α) = r_{i j}^{R} (α) w_{j}^{L} (α)

(19)

In addition, Equation (13) is equivalent to

O_{l}^{L} (α) = \sum_{i = 1}^{n} (w_{i}^{L} (α) p_{l i}^{L} (α))

(20)

O_{l}^{R} (α) = \sum_{i = 1}^{n} (w_{i}^{R} (α) p_{l i}^{R} (α))

(21)

To satisfy constraint (12),

\sum_{α = 0}^{1} (| O_{l}^{L} (α) - S_{l}^{L} (α) | + | O_{l}^{R} (α) - S_{l}^{R} (α) |) \leq ξ

(22)

where ξ is a threshold.

Furthermore, because ${\tilde{S}}_{l}$ is a TFN,

S_{l}^{L} (α) = (1 - α) S_{l 1} + α S_{l 2}

(23)

S_{l}^{R} (α) = (1 - α) S_{l 3} + α S_{l 2}

(24)

In addition,

w_{i}^{L} (α) \leq w_{i}^{L} (α + Δ α)

(25)

w_{i}^{R} (α) \geq w_{i}^{R} (α + Δ α)

(26)

Finally, the following quadratic programming problem is solved instead:

\begin{aligned} Min S S E ≅ \frac{1}{2} \sum_{α = 0}^{1} \\ (\sum_{i = 1}^{n} \sum_{j \neq i} (r_{i j}^{L} (α) - a_{i j}^{L} (α))^{2} + \sum_{i = 1}^{n} \sum_{j \neq i} (r_{i j}^{R} (α) - a_{i j}^{R} (α))^{2}) \end{aligned}

(27)

subject to

w_{i}^{L} (α) = r_{i j}^{L} (α) w_{j}^{R} (α); i, j = 1 \sim n; i \neq j; α = [0, 1]

(28)

w_{i}^{R} (α) = r_{i j}^{R} (α) w_{j}^{L} (α); i, j = 1 \sim n; i \neq j; α = [0, 1]

(29)

O_{l}^{L} (α) = \sum_{i = 1}^{n} (w_{i}^{L} (α) p_{l i}^{L} (α)); l = 1 \sim L; α = [0, 1]

(30)

O_{l}^{R} (α) = \sum_{i = 1}^{n} (w_{i}^{R} (α) p_{l i}^{R} (α)); l = 1 \sim L; α = [0, 1]

(31)

\sum_{α = 0}^{1} (| O_{l}^{L} (α) - S_{l}^{L} (α) | + | O_{l}^{R} (α) - S_{l}^{R} (α) |) \leq ξ; l = 1 \sim L

(32)

w_{i}^{L} (α) \leq w_{i}^{L} (α + Δ α); i = 1 \sim n; α = [0, 1)

(33)

w_{i}^{R} (α) \geq w_{i}^{R} (α + Δ α); i = 1 \sim n; α = [0, 1)

(34)

\sum_{i = 1}^{n} w_{i}^{L} (1) = 1; i = 1 \sim n

(35)

\sum_{i = 1}^{n} w_{i}^{R} (1) = 1; i = 1 \sim n

(36)

w_{i}^{L} (α) \leq w_{i}^{R} (α); i = 1 \sim n; α = [0, 1]

(37)

w_{i}^{L} (α), w_{i}^{R} (α) \in R^{+}; i = 1 \sim n; α = [0, 1]

(38)

The sustainability evaluation results of several smart technology applications in healthcare can then be compared using a grouped gradient bar chart, which is presented in Figure 11.

Figure 11.

Comparing the sustainability evaluation results for several smart technology applications in healthcare.

Case study

Background

To illustrate the applicability of the proposed methodology, it was employed to evaluate the sustainability levels of eight smart technology applications in healthcare and reported in the literature^{5,9,28,39,55–57} (Table 1). These smart technology applications either ranked first before the COVID-19 pandemic or were widely used during the pandemic.

Table 1.

Eight evaluated smart technology applications in healthcare.

l	Smart technology application
1	Healthcare apps (e.g. smartphone apps, vaccine passports, and contact-tracing apps)
2	Healthcare robots (e.g. sterilisation, monitoring patients’ physiological condition, supporting surgery, dispensing medication, and assisting patients with cognitive challenges and disabilities)
3	Remote temperature scanners (e.g. smart surveillance cameras and wireless medical sensor networks)
4	Smart bracelets (e.g. monitoring body temperature and blood oxygen level)
5	Smart clothing (e.g. smart vests)
6	Smart glasses, spectacles, and contact lenses
7	Smartwatches (e.g. monitoring body temperature, blood oxygen level, heart rate, sleep duration, gestures, motions, step count, and movement)
8	Social-distancing monitors

Application of the proposed methodology

In Step 1, a team of three experts from related fields (an industrial engineering professor, a smart device design engineer, and a medical engineer) was formed to prevent personal bias.²¹

In Chen,² five criteria were used to evaluate the suitability of a smart technology application in healthcare before the COVID-19 pandemic: unobtrusiveness, supporting online social networking, the relaxation of related medical laws, the size of the healthcare market, and correct identification of the needs and situation of a user. Subsequently, in Chen and Wang,⁹ six criteria were considered for the same purpose but focusing on the application during the pandemic: the provision of value-added services, cost-effectiveness, the promotion of healthy mobility, necessity or irreplaceability, ease of implementation and maintenance, and the ability to combine with other smart technologies. After the COVID-19 pandemic, the requirements changed again: after discussion, five criteria were considered critical to the sustainability of a smart technology application in healthcare after the COVID-19 pandemic: correct identification of the needs and situation of a user, the provision of value-added services, cost-effectiveness, necessity or irreplaceability, and ease of implementation and maintenance (Figure 12).

Figure 12.

Critical factors affecting smart technology applications in healthcare.

In Step 2, the experts evaluated the relative priorities of criteria in pairs by using linguistic terms. In addition, the effects of the following major global events were considered:

The shortage of semiconductor chips made the necessity or irreplaceability of a smart technology application especially crucial.

The Ukraine–Russia war had a similar effect.

The fear of inflation has driven smart technology applications in healthcare to become more cost-effective.

The pairwise comparison results are summarised in Table 2.

Table 2.

Pairwise comparison results.

Criterion I	Relative priority	Criterion II
Providing value-added services	Slightly or considerably more important than	Correct identification of the user's needs and situation
Being cost-effective	Slightly or considerably more important than	Correct identification of the user's needs and situation
Being necessary or irreplaceable	Slightly more important than	Correct identification of the user's needs and situation
Being easy to implement and maintain	Considerably more important than	Correct identification of the user's needs and situation
Providing value-added services	Slightly more important than	Being cost-effective
Providing value-added services	Slightly more important than	Being necessary or irreplaceable
Providing value-added services	Slightly more important than	Being easy to implement and maintain
Being cost-effective	Slightly more important than	Being necessary or irreplaceable
Being easy to implement and maintain	Slightly more important than	Being cost-effective
Being easy to implement and maintain	Slightly more important than	Being necessary or irreplaceable
(Expert #2)
Criterion I	Relative priority	Criterion II
Providing value-added services	Considerably more important than	Correct identification of the user's needs and situation
Correct identification of the user's needs and situation	Slightly more important than	Being cost-effective
Being necessary or irreplaceable	Slightly or considerably more important than	Correct identification of the user's needs and situation
Being easy to implement and maintain	Slightly or considerably more important than	Correct identification of the user's needs and situation
Providing value-added services	Slightly or considerably more important than	Being cost-effective
Being necessary or irreplaceable	Equally important to	Providing value-added services
Providing value-added services	Slightly or considerably more important than	Being easy to implement and maintain
Being necessary or irreplaceable	Slightly more important than	Being cost-effective
Being easy to implement and maintain	Slightly more important than	Being cost-effective
Being necessary or irreplaceable	Equally important to or slightly more important than	Being easy to implement and maintain
(Expert #3)
Criterion I	Relative priority	Criterion II
Providing value-added services	Slightly or considerably more important than	Correct identification of the user's needs and situation
Being cost-effective	Slightly more important than	Correct identification of the user's needs and situation
Correct identification of the user's needs and situation	Slightly more important than	Being necessary or irreplaceable
Correct identification of the user's needs and situation	Slightly more important than	Being easy to implement and maintain
Providing value-added services	Slightly more important than	Being cost-effective
Providing value-added services	Slightly more important than	Being necessary or irreplaceable
Providing value-added services	Considerably more important than	Being easy to implement and maintain
Being cost-effective	Slightly more important than	Being necessary or irreplaceable
Being cost-effective	Slightly or considerably more important than	Being easy to implement and maintain
Being necessary or irreplaceable	Slightly more important than	Being easy to implement and maintain

On the basis of the pairwise comparison results, the corresponding fuzzy judgment matrixes were constructed:

\tilde{A} (1) = [\begin{matrix} 1 & 1 / (2, 4, 6) & 1 / (2, 4, 6) & 1 / (1, 3, 5) & 1 / (3, 5, 7) \\ (2, 4, 6) & 1 & (1, 3, 5) & (1, 3, 5) & (1, 3, 5) \\ (2, 4, 6) & 1 / (1, 3, 5) & 1 & (1, 3, 5) & 1 / (1, 3, 5) \\ (1, 3, 5) & 1 / (1, 3, 5) & 1 / (1, 3, 5) & 1 & 1 / (1, 3, 5) \\ (3, 5, 7) & 1 / (1, 3, 5) & (1, 3, 5) & (1, 3, 5) & 1 \end{matrix}]

\tilde{A} (2) = [\begin{matrix} 1 & 1 / (3, 5, 7) & (1, 3, 5) & 1 / (2, 4, 6) & 1 / (2, 4, 6) \\ (3, 5, 7) & 1 & (2, 4, 6) & 1 / (1, 1, 3) & (2, 4, 6) \\ 1 / (1, 3, 5) & 1 / (2, 4, 6) & 1 & 1 / (1, 3, 5) & 1 / (1, 3, 5) \\ (2, 4, 6) & (1, 1, 3) & (1, 3, 5) & 1 & (1, 2, 4) \\ (2, 4, 6) & 1 / (2, 4, 6) & (1, 3, 5) & 1 / (1, 2, 4) & 1 \end{matrix}]

\tilde{A} (3) = [\begin{matrix} 1 & 1 / (2, 4, 6) & 1 / (1, 3, 5) & (1, 3, 5) & (1, 3, 5) \\ (2, 4, 6) & 1 & (1, 3, 5) & (1, 3, 5) & (3, 5, 7) \\ (1, 3, 5) & 1 / (1, 3, 5) & 1 & (1, 3, 5) & (2, 4, 6) \\ 1 / (1, 3, 5) & 1 / (1, 3, 5) & 1 / (1, 3, 5) & 1 & (1, 3, 5) \\ 1 / (1, 3, 5) & 1 / (3, 5, 7) & 1 / (2, 4, 6) & 1 / (1, 3, 5) & 1 \end{matrix}]

In Step 3, the consistency ratios of these fuzzy judgment matrixes were evaluated using Equation (8). To this end, MATLAB R2021 was employed on a PC with an AMD Ryzen 5 3400G CPU and 16 GB of RAM. The execution time was approximately 25 s for each fuzzy judgment matrix. The evaluation results are displayed in Figure 13. Evidently, all the fuzzy judgment matrixes were more or less consistent.

Figure 13.

Fuzzy consistency of the fuzzy judgment matrixes.

In Step 4, the FGM was used to aggregate the fuzzy judgment matrixes. The aggregation process using XAI is illustrated in Figure 14. The fuzzy judgments can be observed to have varied.

Figure 14.

Aggregation process with the FGM.

The aggregation result is expressed as follows:

\tilde{A} = [\begin{matrix} 1 & (0.16, 0.23, 0.44) & (0.32, 0.63, 1.36) & (0.32, 0.63, 1.36) & (0.29, 0.53, 0.94) \\ (2.29, 4.31, 6.32) & 1 & (1.26, 3.30, 5.31) & (0.69, 2.08, 2.92) & (1.82, 3.91, 5.94) \\ (0.74, 1.59, 3.11) & (0.19, 0.30, 0.79) & 1 & (0.58, 1.44, 2.92) & (0.43, 0.76, 1.82) \\ (0.74, 1.59, 3.11) & (0.34, 0.48, 1.44) & (0.34, 0.69, 1.71) & 1 & (0.58, 1.26, 2.71) \\ (1.06, 1.88, 3.48) & (0.17, 0.26, 0.55) & (0.55, 1.31, 2.32) & (0.37, 0.79, 1.71) & 1 \end{matrix}]

Subsequently, the performance levels of smart technology applications in terms of optimising the various criteria were evaluated; the evaluation results are summarised in Table 3.

Table 3.

Performance levels of smart technology applications.

l	${\tilde{p}}_{l 1}$	${\tilde{p}}_{l 2}$	${\tilde{p}}_{l 3}$	${\tilde{p}}_{l 4}$	${\tilde{p}}_{l 5}$
1	(4, 5, 5)	(3, 4, 5)	(1.5, 2.5, 3.5)	(3, 4, 5)	(3, 4, 5)
2	(1.5, 2.5, 3.5)	(1.5, 2.5, 3.5)	(0, 0, 1)	(1.5, 2.5, 3.5)	(0, 1, 2)
3	(1.5, 2.5, 3.5)	(3, 4, 5)	(1.5, 2.5, 3.5)	(1.5, 2.5, 3.5)	(1.5, 2.5, 3.5)
4	(0, 1, 2)	(1.5, 2.5, 3.5)	(1.5, 2.5, 3.5)	(3, 4, 5)	(1.5, 2.5, 3.5)
5	(0, 1, 2)	(0, 1, 2)	(0, 1, 2)	(0, 1, 2)	(3, 4, 5)
6	(1.5, 2.5, 3.5)	(4, 5, 5)	(0, 0, 1)	(1.5, 2.5, 3.5)	(0, 0, 1)
7	(1.5, 2.5, 3.5)	(3, 4, 5)	(1.5, 2.5, 3.5)	(3, 4, 5)	(1.5, 2.5, 3.5)
8	(4, 5, 5)	(3, 4, 5)	(1.5, 2.5, 3.5)	(3, 4, 5)	(3, 4, 5)

In Step 5, the sustainability levels of several smart technology applications in healthcare were subjectively evaluated in accordance with their effectiveness (or lack thereof) during the COVID-19 pandemic; the corresponding results are shown in Table 4.

Table 4.

Subjective results of the evaluation of several smart technology applications.

l	Smart technology application	${\tilde{S}}_{l}$
1	Healthcare apps (e.g. smartphone apps, vaccine passports, and contact-tracing apps)	(4, 5, 5)
2	Healthcare robots (e.g. sterilisation, monitoring patients’ physiological condition, supporting surgery, dispensing medication, and assisting patients with cognitive challenges and disabilities)	(3, 4, 5)
3	Remote temperature scanners (e.g. smart surveillance cameras, and wireless medical sensor networks)	(4, 5, 5)
4	Smart bracelets (e.g. monitoring body temperature and blood oxygen level)	Unknown
5	Smart clothing (e.g. smart vests)	(0, 1, 2)
6	Smart glasses, spectacles, and contact lenses	Unknown
7	Smartwatches (e.g. monitoring body temperature, blood oxygen level, heart rate, sleep duration, gestures, motions, step count, and movement)	Unknown
8	Social-distancing monitors	(0, 0, 1)

In Step 6, an FNLP model for deriving the fuzzy priorities of the criteria was constructed and converted into a nonlinear programming problem to be solved using a branch-and-bound algorithm with Lingo on the same PC as that mentioned previously. For this purpose, ξ was set to 8 × 20 = 160. The execution time was less than 1 min. The derivation results are summarised in Figure 15, which presents the result of $S S E^{*} = 439$ . Based on the derived fuzzy priorities of the criteria, the sustainability of a smart technology application in healthcare could be evaluated using the following formula:

\begin{aligned} {\tilde{O}}_{l} = (0.232, 0.260, 0.372) (\times) {\tilde{p}}_{l 1} (+) \\ (0.346, 0.353, 0.364) (\times) {\tilde{p}}_{l 2} \\ (+) (0.117, 0.162, 0.238) (\times) {\tilde{p}}_{l 3} (+) \\ (0.056, 0.138, 0.182) (\times) {\tilde{p}}_{l 4} \\ (+) (0.053, 0.087, 0.335) (\times) {\tilde{p}}_{l 5} \end{aligned}

(37)

Figure 15.

Derived fuzzy priorities of criteria.

In Step 7, the sustainability of each smart technology application in healthcare was evaluated using Equation (13). The results are shown in Figure 16. The subjective evaluations are shown in the same figure for comparison if available.

Figure 16.

Sustainability levels of eight smart technology applications in healthcare.

Discussion

Based on the experimental results, the following items are discussed:

The sustainability levels of smart technology applications in healthcare were defuzzified using the centre-of-gravity method⁵⁸ to facilitate comparison. The results are summarised in Table 5. The most sustainable smart technology application in healthcare was discovered to be healthcare apps, followed by smartwatches and remote temperature scanners.

Smart clothing was found to be the least sustainable smart technology application in healthcare.

Owing to the impact of major global events, the threshold for users to accept a smart technology application in healthcare was discovered to be 3.0. Consequently, only the three most highly rated smart technology applications in healthcare will likely be sustainable.

This research is timely because the COVID-19 pandemic has provided a considerable amount of evidence regarding the usefulness of several smart technology applications in healthcare. Additionally, the end of the COVID-19 pandemic has given rise to renewed interest in novel smart technology applications in healthcare. The present research provides a theoretical basis for related development. Furthermore, some major global events have severely affected the supply and demand of smart technologies, and this reality requires must be urgently considered.

Compared with existing methods, the proposed methodology is novel in that it can assess the sustainability levels of smart technology applications both subjectively and objectively. In addition, this study considered the impact of major global events on the sustainability levels of smart technology applications, an impact that has not been investigated in previous studies.

To further elaborate on the effectiveness of the proposed methodology, the experimental results were compared with the conclusions of several previous studies. First, the criteria used by other studies to assess the suitability or sustainability of smart technology applications in healthcare are different from those used in the present study. In addition, the priorities of the criteria in other studies have been unequal, as shown in Table 6, which illustrates how people's needs have changed over time.

Table 5.

Defuzzified sustainability levels.

l	$C O G ({\tilde{O}}_{l})$
1	4.204
2	2.088
3	3.193
4	2.452
5	1.453
6	2.810
7	3.392
8	1.973

Table 6.

Priorities of criteria in several studies.

Priority rank	Chen²	Chen and Wang⁹	This study
Easiness	Not considered	5	1
Being necessary or irreplaceable	Not considered	4	2
Being cost-effective	Not considered	3	3
Being value-added	Not considered	2	4
Correct identification	3	Not considered	5
Healthy mobility	4	1	Not considered
Compatibility	Not considered	6	Not considered
Unobtrusiveness	2	Not considered	Not considered
Legally compliant	1	Not considered	Not considered
Market size	5	Not considered	Not considered

Furthermore, the suitability or sustainability evaluation results related to smart technology applications in healthcare have differed among several studies, as detailed in Table 7, which indicates that some smart technology applications are merely tentative or have now proven impractical.

The one similar conclusion reached by both the present study and previous studies is that smartphone applications and smartwatches are consistently recognised as useful smart technology applications, whereas remote temperature scanners, although widely used, have been neglected.

Table 7.

Suitability or sustainability evaluation results related to smart technology applications in healthcare in several studies.

Rank	Chen²	Chen and Wang⁹	This study
Healthcare apps (e.g. smartphone apps, vaccine passports, and contact-tracing apps)	1/2	1/3	1
Healthcare robots (e.g. sterilisation, monitoring patients’ physiological condition, supporting surgery, dispensing medication, and assisting patients with cognitive challenges and disabilities)	Not compared	7/9	6
Remote temperature scanners (e.g. smart surveillance cameras and wireless medical sensor networks)	Not compared	8	3
Smart bracelets (e.g. monitoring body temperature and blood oxygen level)	7	4	5
Smart clothing (e.g. smart vests)	13	Not compared	8
Smart glasses, spectacles, and contact lenses	11	Not compared	4
Smartwatches (e.g. monitoring body temperature, blood oxygen level, heart rate, sleep duration, gestures, motions, step count, and movement)	3	2	2
Social-distancing monitors	6	6	7
Wireless medical sensor networks	Not compared	5	Not compared
Smart connected vehicles	4	Not compared	Not compared
Smart smoke alarms	5	Not compared	Not compared
Smart wheelchairs	8	Not compared	Not compared
Smart defence technologies	9	Not compared	Not compared
Smart wigs	10	Not compared	Not compared
Smart toilets	12	Not compared	Not compared

Conclusion

During the COVID-19 pandemic, some smart technology applications were more effective than had been expected, whereas others were less effective than had been expected and thus may not be widely used in the future. Therefore, the sustainability of smart technology applications in healthcare needs to be evaluated. To this end, a hybridising subjective and objective fuzzy group decision-making approach was proposed in this study. After applying the proposed methodology to evaluate the sustainability levels of eight smart technology applications in healthcare, the following conclusions were drawn:

The application of healthcare apps, smartwatches, and remote temperature scanners in healthcare is expected to be highly sustainable. By contrast, the sustainability of smart clothing and other applications in this field was concluded to be questionable.

After optimisation, the subjective evaluations of sustainability were fairly close to the objective results.

Major global events have raised the threshold for users to accept smart technology applications in healthcare. Consequently, only three smart technology applications in healthcare will likely be sustainable.

The contributions of this research are described as follows:

Because the performance of smart technology applications during the COVID-19 pandemic was considered, the sustainability of these applications could be subjectively evaluated and corrected based on their objective performance. By contrast, most related studies have conducted assessments only from the subjective perspective.^2,9,55

This study considered the impact of several major global events during the COVID-19 pandemic on the sustainability of smart technology applications. This topic had not been considered in previous related studies.

In addition, this research adds the following scientific value:

This study found that the priorities of the criteria for assessing the sustainability of smart technology applications have shifted considerably because of the COVID-19 pandemic.

After considering the impact of several major global events, this study found that users are placing a relatively strong emphasis on substitutability so that they may choose relatively cheap and widely available healthcare products and services.

Finally, this study has the following limitations:

The process for solving the formulated FNLP problem was somewhat complicated.

The application of XAI techniques and tools was limited to only a few steps.

Although this study assessed the sustainability of smart technology applications in healthcare, how these applications can be improved is another key issue. In addition, subjective and objective evaluation results can be integrated in ways other than those described in this paper.^59–62 Furthermore, a consensus reached among experts is sometimes insufficient and thus needs to be enhanced.^63,64 These issues constitute suggestions for future research.

Footnotes

Contributorship

Both authors contributed equally to the writing of this paper.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethical approval

Not required.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The publication fee of this paper was support by Ministry of Science and Technology, Taiwan.

Guarantor

Not required.

ORCID iDs

Tin-Chih Toly Chen

Min-Chi Chiu

References

Brandt

Samuelsson

Töytäri

, et al. Activity and participation, quality of life and user satisfaction outcomes of environmental control systems and smart home technology: a systematic review. Disabil Rehabil: Assist Technol 2011; 6: 189–206.

Chen

. Assessing factors critical to smart technology applications to mobile health care − the fgm-fahp approach. Health Policy Technol 2020; 9: 194–203.

Waheed

Shafi

. Successful role of smart technology to combat COVID-19. In: Fourth international conference on I-SMAC, 2020, pp.772–777.

Jahangiri

Chobar

Ghasemi

, et al. Simulation-based optimization: analysis of the emergency department resources under COVID-19 conditions. Int J Ind Syst Eng 2021; 1: 1–10.

Abbas

Michael

. COVID-19 contact trace app deployments: learnings from Australia and Singapore. IEEE Consum Electron Mag 2020; 9: 65–70.

Ennafiri

Mazri

. Internet of things for smart healthcare: a review on a potential IOT based system and technologies to control COVID-19 pandemic. In: The proceedings of the third international conference on smart city applications, 2020, pp.1256–1269.

Stella

AjČeviĆ

Furlanis

, et al. Smart technology for physical activity and health assessment during COVID-19 lockdown. J Sports Med Phys Fitness 2021; 61: 452–460.

Khan

Kushwah

Singh

, et al. Smart technologies driven approaches to tackle COVID-19 pandemic: a review. 3 Biotech 2021; 11: 1–22.

Chen

Wang

. Recommending suitable smart technology applications to support mobile healthcare after the COVID-19 pandemic using a fuzzy approach. Healthcare 2021; 9: 1461.

10.

Ali

Singh

Javaid

, et al. A review of the role of smart wireless medical sensor network in COVID-19. J Ind Integr Manag 2020; 5: 413–425.

11.

Hassankhani

Alidadi

Sharifi

, et al. Smart city and crisis management: lessons for the COVID-19 pandemic. Int J Environ Res Public Health 2021; 18: 7736.

12.

Bown

. How the United States marched the semiconductor industry into its trade war with China. East Asian Econ Rev 2020; 24: 349–388.

13.

Gui

Wang

. Risk assessment of port congestion risk during the COVID-19 pandemic. J Mar Sci Eng 2022; 10: 150.

14.

Zhang

. An analysis on the crisis of “chips shortage” in automobile industry–based on the double influence of COVID-19 and trade friction. J Phys Conf Ser 2021; 1971: 012100.

15.

Seiler

. Weighting bias and inflation in the time of COVID-19: evidence from Swiss transaction data. Swiss J Econ Stat 2020; 156: 1–11.

16.

Bluszcz

Valente

. The economic costs of hybrid wars: the case of Ukraine. Def Peace Econ 2022; 33: 1–25.

17.

Shirazi

Kia

Ghasemi

. A stochastic bi-objective simulation–optimization model for plasma supply chain in case of COVID-19 outbreak. Appl Soft Comput 2021; 112: 107725.

18.

Kamran

Kia

Goodarzian

, et al. A new vaccine supply chain network under COVID-19 conditions considering system dynamic: artificial intelligence algorithms. Socioecon Plann Sci 2022: 101378. doi: 10.1016/j.seps.2022.101378.

19.

Vaníček

Vrana

Aly

. Fuzzy aggregation and averaging for group decision making: a generalization and survey. Knowl Based Syst 2009; 22: 79–84.

20.

Zhang

Zhong

, et al. Extended TODIM for multi-criteria group decision making based on unbalanced hesitant fuzzy linguistic term sets. Comput Ind Eng 2017; 114: 316–328.

21.

Chen

TCT

. Advances in fuzzy group decision making. Cham, Switzerland: Springer, 2021.

22.

Buckley

. Fuzzy hierarchical analysis. Fuzzy Sets Syst 1985; 17: 233–247.

23.

Chang

. Applications of the extent analysis method on fuzzy AHP. Eur J Oper Res 1996; 95: 649–655.

24.

Gunning

Stefik

Choi

, et al. XAI – Explainable artificial intelligence. Sci Robot 2019; 4: eaay7120.

25.

Tjoa

Khok

Chouhan

, et al. Improving deep neural network classification confidence using heatmap-based eXplainable AI, https://doi.org/10.48550/arXiv.2201.00009 (2022).

26.

Kakderi

Oikonomaki

Papadaki

. Smart and resilient urban futures for sustainability in the post COVID-19 era: a review of policy responses on urban mobility. Sustainability 2021; 13: 6486.

27.

Lin

Hsieh

Chen

, et al. Apply IOT technology to practice a pandemic prevention body temperature measurement system: a case study of response measures for COVID-19. Int J Distrib Sens Netw 2021; 17: 15501477211018126.

28.

Hall

Studdert

. “Vaccine passport” certification – policy and ethical considerations. N Engl J Med 2021; 385: e32.

29.

Heijmans

. Singapore PM pushes for living with COVID, without the fear, https://www.bloomberg.com/news/articles/2021-10-09/singapore-premier-pushes-for-living-with-covid-without-the-fear (2021).

30.

Research and markets. Mobile health (mHealth) market

–

Growth, trends, COVID-19 impact, and forecasts (2021–2026), https://www.researchandmarkets.com/reports/4520220/mobile-health-mhealth-market-growth-trends (2021).

31.

Wego Travel. Travel without vaccine in 2022: can you still travel If you’re not vaccinated?, https://blog.wego.com/travel-without-vaccine/ (2022).

32.

Wang

Chen

TCT

. Assessing and comparing COVID-19 intervention strategies using a varying partial consensus fuzzy collaborative intelligence approach. Mathematics 2020; 8: 1725.

33.

Kulkarni

Vishwanath

Shah

. Implementing a real-time, AI-based, face mask detector application for COVID-19, https://developer.nvidia.com/blog/implementing-a-real-time-ai-based-face-mask-detector-application-for-covid-19/ (2021).

34.

Goodarzian

Taleizadeh

Ghasemi

, et al. An integrated sustainable medical supply chain network during COVID-19. Eng Appl Artif Intell 2021; 100: 104188.

35.

Garson

Furlong

. Disrupters and defenders: What the Ukraine war has taught us about the power of global tech companies, https://institute.global/policy/disrupters-and-defenders-what-ukraine-war-has-taught-us-about-power-global-tech-companies (2022).

36.

Whitelaw

Mamas

Topol

, et al. Applications of digital technology in COVID-19 pandemic planning and response. Lancet Digital Health 2020; 2: e435–e440.

37.

Davis

. Integrating digital technologies and data-driven telemedicine into smart healthcare during the COVID-19 pandemic. Am J Med Res 2020; 7: 22–29.

38.

Jaiswal

Agarwal

Negi

. Smart solution for reducing the COVID–19 risk using smart city technology. IET Smart Cities 2020; 2: 82–88.

39.

Inn

. Smart city technologies take on COVID-19. World Health 2020; 841: 1–10.

40.

Djuric

. Apple watch – Original repair, https://zh.ifixit.com/Device/Apple_Watch (2022).

41.

Mardani

Hooker

Ozkul

, et al. Application of decision making and fuzzy sets theory to evaluate the healthcare and medical problems: a review of three decades of research with recent developments. Expert Syst Appl 2019; 137: 202–231.

42.

Adebanjo

Laosirihongthong

Samaranayake

. Prioritizing lean supply chain management initiatives in healthcare service operations: a fuzzy AHP approach. Prod Plan Control 2016; 27: 953–966.

43.

Peng

Krishankumar

Ravichandran

. A novel interval-valued fuzzy soft decision-making method based on CoCoSo and CRITIC for intelligent healthcare management evaluation. Soft Comput 2021; 25: 4213–4241.

44.

Saaty

. Axiomatic foundation of the analytic hierarchy process. Manage Sci 1986; 32: 841–855.

45.

Kahraman

Ruan

Doǧan

. Fuzzy group decision-making for facility location selection. Inf Sci (NY) 2003; 157: 135–153.

46.

Wang

. Evaluating new product development performance by fuzzy linguistic computing. Expert Syst Appl 2009; 36: 9759–9766.

47.

Lin

Hung

Lin

, et al. Green suppliers performance evaluation in belt and road using fuzzy weighted average with social media information. Sustainability 2017; 10: 5.

48.

Chen

Chiu

. A fuzzy collaborative intelligence approach to group decision-making: a case study of post-COVID-19 restaurant transformation. Cogn Comput 2022; 14: 531–546.

49.

Lin

Chen

TCT

. An intelligent system for assisting personalized COVID-19 vaccination location selection: Taiwan as an example. Digital Health 2022; 8: 20552076221109062.

50.

Chen

Lin

Chiu

. Approximating alpha-cut operations approach for effective and efficient fuzzy analytic hierarchy process analysis. Appl Soft Comput 2019; 85: 105855.

51.

Ahmed

Kilic

. Fuzzy analytic hierarchy process: a performance analysis of various algorithms. Fuzzy Sets Syst 2019; 362: 110–128.

52.

Olabanji

Mpofu

. Appraisal of conceptual designs: coalescing fuzzy analytic hierarchy process (F-AHP) and fuzzy grey relational analysis (F-GRA). Results Eng 2021; 9: 100194.

53.

Zhang

Kou

, et al. On priority weights and consistency for incomplete hesitant fuzzy preference relations. Knowl Based Syst 2018; 143: 115–126.

54.

Wang

Chen

Yeh

. Advanced 3D printing technologies for the aircraft industry: a fuzzy systematic approach for assessing the critical factors. Int J Adv Manuf Technol 2019; 105: 4059–4069.

55.

Chen

Chiu

. Smart technologies for assisting the life quality of persons in a mobile environment: a review. J Ambient Intell Humaniz Comput 2018; 9: 319–327.

56.

Kaiser

Al Mamun

Mahmud

, et al. Healthcare robots to combat COVID-19. COVID-19: Prediction, Decision-making, and Its Impacts, 2021, pp.83–97.

57.

Nichols

. Disinfecting robots to fight coronavirus run into travel bans, https://www.zdnet.com/article/disinfecting-robots-to-fight-coronavirus-run-into-travel-bans/ (2021).

58.

Van Broekhoven

De Baets

. A comparison of three methods for computing the center of gravity defuzzification. IEEE Int Conf Fuzzy Syst 2004; 3: 1537–1542.

59.

Chen

Lin

. A fuzzy-neural system incorporating unequally important expert opinions for semiconductor yield forecasting. Int J Uncertain Fuzziness Knowl-Based Syst 2008; 16: 35–58.

60.

Chen

Lin

. Smart and automation technologies for ensuring the long-term operation of a factory amid the COVID-19 pandemic: an evolving fuzzy assessment approach. Int J Adv Manuf Technol 2020; 111: 3545–3558.

61.

Chen

Wang

Chiu

. Assessing the robustness of a factory amid the COVID-19 pandemic: a fuzzy collaborative intelligence approach. Healthcare 2020; 8: 481.

62.

Chen

Wang

. Analyzing the impact of vaccine availability on alternative supplier selection amid the COVID-19 pandemic: a cFGM-FTOPSIS-FWI approach. Healthcare 2021; 9: 71.

63.

Zhang

. Personalized individual semantics-based consistency control and consensus reaching in linguistic group decision making. IEEE Trans Syst Man Cybern: Syst 2021; 52: 5623–5635.

64.

Zhang

. Consensus reaching with consistency control in group decision making with incomplete hesitant fuzzy linguistic preference relations. Comput Ind Eng 2022; 170: 108311.