Enhancing User Experience through Optimization Design Method for Elderly Medication Reminder Mobile Applications: A QFD-Based Research Approach

Abstract

Poor medication adherence among older adults is a widespread problem worldwide. As the population ages, the design of smartphone medication management apps is critical to improving medication adherence among older adults. Taking the design of an elderly medication reminder APP as an example, this study proposes a sustainable design research method that integrates the KANO model, Analytic Hierarchy Process (AHP), Quality Function Deployment (QFD), and PUGH decision matrix. The method collects user demands through in-depth interviews, and applies the KANO model to classify these demands. The hierarchical structure of user needs is established by using AHP, and the priority is sorted according to the weight and importance determined by the judgment matrix. QFD is used to translate user needs into design requirements, and the house of quality matrix identifies key design requirements. Finally, design alternatives are evaluated using Pugh’s concept selection method. The results of this study demonstrate that the integration of KANO-AHP-QFD-PUGH can be effective as a sustainable optimal design approach for the user experience of a medication reminder application for the elderly. This integrated method not only facilitates innovative optimization and sustainability of application design methods but also provides valuable theoretical and practical insights for future drug-assisted design for elderly users.

1. Introduction

With the increasingly apparent trend of global population aging, the proportion of elderly people is continuously rising [1]. As elderly chronic disease patients need to take multiple medications for a long period, taking medication on time and regularly is crucial to improving their health status. However, as the elderly age, their various physical and cognitive abilities decline to varying degrees. Common problems include missing doses, taking incorrect doses, and having a biased understanding of their condition, resulting in medication cessation or reduction [2]. These problems can lead to higher readmission and mortality rates. To improve medication adherence in the elderly and help them address the challenges they may face, such as taking medication at the correct time, understanding the efficacy and side effects of medication, or organizing their medication list, digital interventions have become increasingly important. Mobile-based digital tools related to health can activate reminders for medication, maintain medication history, and furnish medication information indicating the capacity to facilitate self-care for chronic illnesses and enhance medication adherence [3].

Currently, numerous scholars have developed and tested medication management applications, with research on medication reminder applications for the elderly primarily focusing on two aspects. On one hand, there is the development and functional research of medication reminder applications that explore the application functions, interfaces, and preferences of elderly users. On the other hand, there is a study of the usability of existing medication reminder apps for the elderly. Multiple test results indicate that although these reminder applications can significantly improve patients’ medication adherence, they are not user-friendly for elderly patients. Many applications have issues such as complex interface interactions and difficult operations, and the design process does not fully consider the abilities, needs, and limitations of the elderly population [3], resulting in poor usability and low satisfaction among elderly users. Although extensive research has been conducted by many scholars on the needs of elderly users, the usability test results of these applications are not satisfactory. There exists a gap between user needs and design outcomes. Therefore, it is necessary to find a suitable design method to assist designers in better-translating user needs into design requirements during the app design process.

Therefore, this paper proposes an integrated design optimization method, combining KANO-AHP-QFD-PUGH, aiming to help reduce the gap between users and products or services from a design perspective. Using the design of a medication reminder app for the elderly as a research case, and focusing on Chinese elderly users as the target research population, this study aims to validate the feasibility of the proposed design method. It explores the genuine needs of elderly users and accurately translates those needs into design requirements. By selecting the optimal design solution, the study aims to enhance the satisfaction and efficiency of elderly users with such applications, ultimately addressing the issue of poor medication adherence among the elderly.

The subsequent organization of this paper is as follows: Section 2 provides an overview of the theoretical framework, identifies deficiencies in previous studies, and highlights areas that require improvement. Section 3 outlines the research methodologies and procedures employed in this study. Section 4 presents a case study on medication management applications for elderly users. Section 5 analyzes and discusses the findings of the case study and provides relevant recommendations for enhancing the user experience. Finally, Section 6 summarizes the conclusions of the study, discusses its significance, acknowledges limitations, and suggests directions for future research.

2. Literature Review

2.1. Theoretical Background

The QFD model, proposed by Akao in 1966, serves as a planning tool that effectively translates user demands into functional attributes of products through qualitative and quantitative analysis [4]. The fundamental principle of this approach is to utilize the house of quality matrix, which employs data analysis and processing to maximize the incorporation of users’ primary design criteria, as shown in Figure 1. Currently, QFD is commonly used as a research framework in conjunction with other design methods to optimize design schemes, forming a complementary and comprehensive theory:

The Kano model is an effective approach for categorizing and ranking user demands, which is grounded on evaluating the influence of user demands on user contentment [5]. It highlights the non-linear correlation between product performance and user satisfaction [6]. It divides user demands into must-have (M), one-dimensional (O), attractive (A), indifference (I), and reverse (Q). Understanding the attributes of user demands can help designers fully grasp the users’ true demands, making the Kano model a preferred choice for many scholars [7]. The integration of qualitative and quantitative research, utilizing the QFD model, has evolved over time into a sophisticated integrated model [8]. For example, Yongfeng Li and others [9] combined QFD with the Kano model to develop a smartphone APP suitable for the elderly; Kim and others [10] used Kano-QFD to develop the shape design of medical service robots. Shwetank Avikal et al. [11] combined QFD with fuzzy Kano models for the classification of SUV car appearance aesthetic attributes; Xiaoli Wu et al. [4] proposed an approach that integrates the Kano Model, QFD, and Functional Analysis System Technology (FAST) for better stroller design; Akshat Rampal et al. [12] introduced Kano into quality function deployment (QFD) to improve user satisfaction with self-driving electric vehicles. Jianhua Lyu et al. [13] used this model to develop an open-office wooden chair. The Kano-QFD model finds applications in numerous product designs, services, satisfaction improvements, health care, and communication technology equipment improvement. However, when they are used alone, they cannot determine the priority of user satisfaction ranking. Therefore, it is necessary to combine them with other standard decision-making methods to determine which design features have a greater impact on user satisfaction [14].

The Analytic Hierarchy Process (AHP) divides complex decision-making problems into several simple sub-problems at one level by using a tree-like hierarchical structure, and each sub-problem can be analyzed independently [15]. Upon constructing the hierarchy, each component’s relative importance is assigned a weight value, and a pairwise comparison matrix is established to determine the eigenvector and eigenvalue. The eigenvector indicates the priority of each component at each level and can offer decision-makers adequate information for making informed decisions. Additionally, this can assist in organizing the selection criteria, analyzing the weights, and mitigating the potential risks associated with decision-making errors [16].

The application of AHP is very broad, covering almost all areas of decision-making problems. The usability of integrating AHP with QFD, the Kano model, and other research methods has been confirmed in numerous studies. This integration effectively addresses the limitations of using a single method. For example, Rafat Mahmud Hridoy et al. [17] used the joint method of AHP, Kano model, and QFD to try to improve the design of tractor seats in Bangladesh; Angie Wu et al. [18] integrated AHP, QFD, and TRIZ models to innovate the design of motor vehicles in scenic spots; Byanca Porto de Lima et al. [19] integrate the practical application of AHP, QFD, and PROMETHEE methods to the problem of packaging design choice in hesitant fuzzy environments.

The PUGH concept selection method is a qualitative decision analysis tool that was proposed by Stuart Pugh in 1980 [20]. Based on the principle of product concept selection, it converts the process into a matrix method for evaluating concept screening, particularly in the concept design stage. This method allows for scientific analysis of the data and minimizes the personal influence of subjective judgments on the results. Using a standard score, this method assists designers in determining the most preferred solution among the potential options [21]. In 1992, American scholars introduced the PUGH research method into QFD, which sparked further in-depth research [22]. Jing Yang et al. [23] proposed an integrated AHP-PUGH evaluation model and established a comprehensive product evaluation system for conceptual design. Pandit et al. [24] combined PUGH concept selection with QFD to select the most suitable trolley design; Tianlu Zhu et al. [22] conducted research on the design of surgical aids based on AHP, QFD, and PUGH decision matrices; Yongchuan Li et al. [5] utilized the KANO-QFD-PUGH model to optimize the sustainable user experience design of smart home products for the elderly, using smart refrigerators as an example. Despite the combination and application of the Kano, AHP, QFD, and PUGH methods in various fields, a perfect and unified research paradigm has not yet been established. Thus, there remains a certain research space and value for further exploration.

2.2. Current Status of Research on the Application of Medication Reminder in the Elderly

In the past few years, improving the medication compliance of the elderly has become an important topic in the world. Judging from the current research status of medication reminder applications, relevant discussions mainly focus on two aspects: one is the application of medication reminder APPs. Development and functional research; the second is to conduct usability testing and evaluation of the existing medication reminder APP.

Anto’nio Teixeira et al. [25] used a short-term iterative method to develop the “Medication Assistant” application centered on the elderly and interaction design, which includes scenarios and target definition, requirements engineering, design, prototyping, and evaluation of target users by continuously allowing users to test the prototype, discovering advantages and disadvantages and suggestions for improvement, to provide a reliable basis for further improvement and development.

Due to the ongoing global aging trend, the elderly population faces unique challenges in medicine usage [26]. As a result, many scholars have started focusing on the elderly as target users to study the functions, interfaces, and preferences of applications. Susan L Lakey et al. [27] assessed the current use, knowledge, and preferences of medication management tools and supports among 152 older adults; Leah M Haverhals et al. [28] conducted a study on personal health applications (PHA) and findings showed that reliable information links, clear and concise interface, and user-friendly navigation can improve the efficiency of medication management in the elderly; Andrea M. Russell et al. [29] conducted an experimental study to investigate the functional preferences of mobile applications for medication self-management in the elderly. The findings indicated that essential functions such as drug lists, mutual warnings between drugs, medication reminders, and record-keeping were considered crucial by the participants.

This paper studies the usability testing of medication reminder applications mainly focusing on the elderly user group. Several studies have shown that current medication management applications are difficult for elderly users to use [2]. Based on the usability evaluation conducted by Kelly Anne Grindrod et al. [30], it was discovered that the elderly prefer mobile drug management applications with simple interfaces. Rachel Stuck et al. [2] comprehensive review of a popular medication reminder app indicated that its design did not take into account the needs, abilities, and limitations of older adults. The results revealed three main problems: navigation difficulties, poor visibility, and lack of transparency. Jaqueline Donin Noleto et al. [31] compared three medication reminder apps and conducted a senior usability test of the interface to assess its user-friendliness for the elderly. The results showed that these applications lacked design considerations for older adults, and there is a need for improvement in general interface issues. Yongjing Ping et al. [3] conducted a study on the prevalence and correlation of the usage intention of the elderly medication reminder application and found that the main reason for the low usage intention of the elderly is that the interface of the application is complex, user-unfriendly, and dense with instructions.

This may make it difficult for the elderly with less education to register and navigate the app. Furthermore, certain applications exhibit subpar interfaces and lack sufficient instructions in recording medication history and setting medication reminders [32]. Users also expressed concerns about the security of their personal information.

3. Methods

Our research method framework is shown in Figure 2, and we will now proceed to describe the specific research process for each step separately.

3.1. STEP1: User Demand Survey

This research first needs to screen the target users and formulate an interview outline, and then conduct in-depth interviews with the target users to sort out their demands.

3.2. STEP2: Kano Model

Develop a structured questionnaire using the user demands derived from the in-depth interviews of the first step. Aiming at a single functional demand (provide this function/not provide this function), raise positive and negative questions [33], as shown in

Table 1. Kano questionnaire.

Function	Like	Deserved	Neutral	Endurable	Dislike
Provide	√
Does not provide				√

:

The demands are categorized into five items: “like”, “deserved”, “neutral”, “endurable”, and “dislike” (as shown in Table 1). The research results are then classified and counted based on the Kano evaluation list (as shown in

Table 2. Kano evaluation list.

User Demands			Negative Questions
		Like	Deserved	Neutral	Endurable	Dislike
Positive Questions	Like	Q	A	A	A	O
	Deserved	R	I	I	I	M
	Neutral	R	I	I	I	M
	Endurable	R	I	I	I	M
	Dislike	R	R	R	R	Q

) to determine the demand attributes [34]. the demand categories are determined and filtered into must-have, one-dimensional, and attractive user demands, based on the highest frequency among their corresponding categories A, O, M, I, and R. Indifferent and reverse attributes are removed from consideration [9].

3.3. STEP3: Analytic Hierarchy Process (AHP)

3.3.1. Hierarchical Analysis Structure Construction

Analyze the relationships between user demands, establish a systematic hierarchical structure, and create a hierarchy chart. Design a questionnaire based on a judgment scale of 1 to 9, where 1 indicates equal importance, and 9 indicates extremely important (as shown in

Table 3. The 1~9 scale method.

Scale	Implication
1	By comparison, the two elements are equally important
3	By comparison, one element is slightly more important than the other
5	By comparison, one element is more important than the other
7	By comparison, one element is strongly more important than the other
9	By comparison, one element is extremely more important than the other
2, 4, 6, 8	The median value of the above importance

). Based on this judgment scale, pairwise comparisons of first and second-level user demands are conducted to obtain weight values and to process the numerical values [34].

3.3.2. Construction of Judgment Matrix A

Construct the judgment matrix A. In the process of determining the importance of evaluation indicators, it is often difficult to determine the importance of one indicator compared to another. Therefore, the relationship matrix elements, a_i, a_j (i, j = 1, 2, …, n) represent the elements, and a_ij represents the importance of i compared to j. If the comparison is the opposite, the reciprocal 1/a_ij is taken [20]. Based on this, the judgment matrix A can be constructed:

$$A={\left|\begin{array}{cccc}{a}_{11}& a_{12}& \cdots & a_{1n}\\ a_{21}& a_{22}& \cdots & a_{2n}\\ \cdots & \cdots & \cdots & \cdots \\ a_{n1}& a_{n2}& \cdots & a_{nn}\end{array}\right|}_{n\times n}$$

3.3.3. Weight Calculations

Using the judgment matrix, calculate the geometric mean to obtain the average value V_i. Normalize the results to obtain the average value w_i of each weight and the weight vector W [35]. The calculation formula is as follows:

$$V_i=\sqrt[n]{{\prod }_i^n{a}_{ij}}$$

(1)

$$w_i=\frac{\sqrt[n]{\prod _i^n{a}_{ij}}}{\sum _{i=1}^n\sqrt[n]{\prod _i^n{a}_{ij}}}$$

(2)

$$W={\left(w_1,w_2,\cdots ,w_n\right)}^T$$

(3)

3.3.4. Consistency Test of Judgment Matrix

To evaluate the rationality of the weight allocation and avoid errors caused by subjective judgments of decision-makers, it is necessary to conduct a consistency test on the judgment matrix. λ_max is the maximum eigenvalue, CI is the consistency index of the judgment matrix, RI is the average random consistency index of the judgment matrix, and CR is the random consistency ratio of the judgment matrix. If the CR of the judgment matrix A is less than 0.1 or λ_max = n and CI = 0, it is considered to have passed the consistency test [16]. Otherwise, the matrix needs to be modified until its consistency meets the requirements. The calculation method is as follows:

Firstly, calculate the consistency index CI, where n is the number of indicators, using the formula [35]:

CI = λ_max − n/n − 1

(4)

Secondly, look up the random RI values from a reference table (as shown in

Table 4. Average random consistency.

n	3	4	5	6	7	8	9	10	11	12	13	14	15
RI	0.52	0.89	1.12	1.26	1.36	1.41	1.46	1.49	1.52	1.54	1.56	1.58	1.59

) based on the matrix size (n).

Thirdly, calculate the CR. CR is the random consistency ratio of the judgment matrix [36], and is calculated using the formula:

CR = CI/RI

(5)

3.3.5. Determine the Weight Order

Finally, by calculating the elements at each level, the weight values for each requirement are obtained, and the importance ranking of design factors is determined, thus clarifying the key user demands.

3.4. STEP4: Quality Function Development (QFD)

3.4.1. Define Design Requirements and Objectives

By analyzing user demands using the Kano model and AHP, user demands attributes and their importance were identified. The next step is to map the various levels of requirements elements to design requirements, further quantifying the degree of correlation between user requirements and design requirements. This mapping relationship not only ensures the rationality of requirement transformation but also maximizes the implementation of functionality. In the house of quality matrix, different values are used to indicate the strength of correlation. A value of 5 represents a strong correlation, a value of 3 represents a moderate correlation, a value of 1 represents a weak correlation, and an empty space indicates no correlation [37].

3.4.2. Build the House of Quality and Calculate the Importance of Design Requirements

The establishment of the matrix relationship in the house of quality is the core part, which calculates the weight value of various design elements. Assuming that there are n design requirements, with W_i representing the importance value of the i-th demand for the elderly [38]. The degree of association between the i-th elderly demand and the j-th design requirement is denoted as R_ij [39]. Consequently, the importance of the j-th design requirement can be computed using the following formula:

$$H_j={\sum }_{i=1}^n{W}_i{R}_{ij}(j=0,1,2,3,\cdots ,m)$$

(6)

Sort the importance values H_j of design requirements in descending order to determine the critical design requirements. Based on this, plan the design scheme to maximize user satisfaction.

3.5. STEP5: PUGH Decision Matrix

3.5.1. Design Solutions Generation

Based on the house of quality model, the importance of design requirements is determined. Higher scores indicate greater importance in improving the satisfaction of elderly users with medication reminder applications, and vice versa. The obtained design requirements serve as a reference for evaluating design proposals. Medication reminder applications typically consist of a visual interface, interaction mode, and functionality. The gathered user demands include the style and features commonly found in most applications. Among these requirements, the top four design requirements are considered primary elements due to their significance, while the remaining features serve as auxiliary elements [22].

3.5.2. Program Evaluation

To construct a PUGH decision matrix, design evaluation criteria are listed on the left side, while comprehensive weights of design schemes and evaluation criteria are placed at the top. One design scheme is selected as the reference, and a 1–5 point rating system (1 point worse than the reference by a significant margin, 2 points worse than the reference, 3 points equal to the reference, 4 points better than the reference, 5 points significantly better than the reference) is used for evaluation [22]. The final evaluation score is calculated using a formula that takes into account the significance of design requirements and the magnitude of the rating score, as shown below:

$$V_i={\sum }_{i=1}^n{H}_j{\mathcal{V}}_{ij}$$

(7)

The formula determines the final evaluation score of a design concept using the PUGH decision matrix. V_i represents the comprehensive evaluation score of concept i, V_ij denotes the score of concept j on the i-th evaluation criterion, H_i represents the overall weight of the i-th evaluation criterion, and n indicates the total number of evaluation criteria [21]. The design schemes are ranked based on the overall evaluation scores to determine the final design scheme.

4. Case Study

4.1. User Demand Survey

This study employs the qualitative research method of in-depth user interviews to gather original descriptions of target user needs. To ensure the accuracy of the information and the representativeness of the sample, we conducted research on 38 elderly users who met the following criteria: (1) aged between 60–75 years old, taking two or more medications daily; (2) need of long-term or lifelong medication; (3) living independently and possessing self-care capabilities; (4) having experience in using or requiring medication reminder products and services. The sample consists of 17 males and 21 females (user information is provided in

Table 5. User Information Form.

Category	Item	Frequency	%
Gender	Male	17	44.73
	Female	21	55.27
Age	60–65	23	60.53
	65–70	11	28.95
	70–75	4	10.52
Types of daily medications	2 or 3	20	52.63
	More than 3	18	47.37
Use Experience	Use	15	39.47
	Have not used	23	60.53

). The interviews were conducted in a one-on-one format. Before the interviews, three interface designers, two Ph.Ds. in interaction design, and a professor discussed the structure and content of the interviews (as shown in

Table 6. Interview outline.

Category	Questions
Basic Information	What is your age, gender, and occupation? Do you have any history of chronic illnesses, as well as allergies, or adverse reactions to any medications? How experienced are you in using smartphones? Have you used any medication reminder applications before?
Function demands	What do you consider to be the biggest challenge in medication management in your daily life? What specific functionalities would you like a medication reminder app to have in order to help you manage and track your medication usage? For example, medication reminder alerts, medication history tracking, the ability to share medication information with your doctor, etc. Do you have specific medication management demands, such as complex medication combinations or timed medication schedules? When using a medication management app, what kind of information and feedback would you expect to receive? For example, medication dosage reminders, warnings for medication interactions, etc. Do you have any other specific demands or expectations that you would like the app to fulfill? What are your concerns and worries regarding the data privacy and security issues when using a medication reminder app?
Operating Habits	When using a smartphone or mobile device, do you prefer to use single-tap or swipe gestures? Which gesture do you find easier to grasp and use? How do you typically handle accidental touches to other functions or screens while using your phone or applications? Are you willing to try and learn new ways of operating and using features? Do you have a specific workflow or sequence of actions when using mobile applications? Are you accustomed to using accessibility features provided by phones or applications, such as magnifiers or voice assistants? Do you have any other thoughts or demands to share?
Interface Preference	Do you have any specific preferences or demands regarding the design of interface icons and buttons in mobile applications? For example, do you prefer larger and clearer icons and buttons? Do you have any specific preferences or demands regarding the font size and color settings in mobile applications? Is it easy to read and distinguish? When it comes to interface layout and design, do you prefer a clean and straightforward interface or a more feature-rich interface? Which type of interface is easier to use and understand?

). Each session lasted approximately 35 min. The entire interview process was recorded using a tape recorder, and detailed notes were taken to capture the perspectives of the interviewer. These annotations are utilized to explain and provide context to the language used by the respondents, ensuring comprehensive and understandable data.

4.2. Kano Model

After obtaining the above information, it is necessary to organize the research results. The results reveal that the surveyed users share many similar demands, with only a few minor differences. Therefore, all user demands are carefully sorted and unnecessary items are removed, while identical items are merged. As a result, a total of 20 original user demands were obtained, which are summarized in

Table 7. Collection of original demands of elderly users.

Serial Number	Original Demands	Serial Number	Original Demands
U1	Larger text display	U11	Easy login process with fewer steps
U2	Easy-to-understand patterns	U12	Medication time reminders and voice alerts
U3	Color uniform	U13	Dosage reminders and voice alerts
U4	Simple interface	U14	Medication precautions reminders and voice alerts
U5	Fewer levels, less memory burden	U15	Remote assistance for family members
U6	Strong fault tolerance	U16	Recording previous medications
U7	Good operational feedback	U17	Ability to add medication instructions and doctor’s orders
U8	Clear and understandable language	U18	Reminder to check medication balance
U9	Voice prompts capability	U19	Online chat with doctors
U10	Operation teaching to aid learning	U20	Addressing personal privacy concerns

.

The 20 original user demands were renamed and defined, and Kano questionnaires were designed based on these demands. Positive and negative questions were set for each demand item (provide this function/do not provide this function). Paper and electronic questionnaires were used in combination to survey elderly target users aged 60–75 years old, with a focus on elderly communities with dense populations. A total of 105 questionnaires were distributed, and the respondents spent a minimum of 200 s responding to the questions. A total of 105 questionnaires were distributed, and the respondents spent a minimum of 200 s responding to the questions. Out of these, 98 questionnaires were considered valid, resulting in a recovery rate of 93.3%.

According to the Kano evaluation table, the results of the demand survey were statistically analyzed, and the elderly people’s requirement attribute categories were obtained, as shown in

Table 8. Kano classification of elderly medication reminder mobile application.

Serial Number	User Demand Elements	Quality Dimension					Total	Quality Type
		M	O	A	I	Q
U1	Larger text display	42	8	5	0	0	55	M
U2	Easy-to-understand patterns	18	5	32	0	0	55	O
U3	Color uniform	8	29	18	0	0	55	A
U4	Simple interface	30	3	22	0	0	55	M
U5	Fewer levels, less memory burden	35	7	13	0	0	55	M
U6	Strong fault tolerance	19	27	9	0	0	55	A
U7	Good operational feedback	20	3	32	0	0	55	O
U8	Clear and understandable language	17	16	22	0	0	55	A
U9	Voice prompts capability	8	40	7	0	0	55	A
U10	Operation teaching to aid learning	32	4	2	0	17	55	M
U11	Easy login process with fewer steps	36	9	10	0	0	55	M
U12	Medication time reminders and voice alerts	41	11	3	0	0	55	M
U13	Dosage reminders and voice alerts	13	15	27	0	0	55	O
U14	Medication precautions reminders and voice alerts	7	12	36	0	0	55	O
U15	Remote assistance for family members	9	3	43	0	0	55	O
U16	Recording previous medications	10	26	19	0	0	55	A
U17	Ability to add medication instructions and doctor’s orders	8	35	9	0	3	55	A
U18	Reminder to check medication balance	3	19	11	0	22	55	I
U19	Online chat with doctors	6	25	24	0	0	55	A
U20	Addressing personal privacy concerns	10	35	7	0	3	55	A

. The must-have attributes (M) were U₁, U₄, U₅, U₁₀, U₁₁, and U₁₂. The attractive attributes (A) were U₃, U₆, U₈, U₉, U₁₆, U₁₇, U₁₉, U₂₀. The one-dimensional attributes (O) were U₂, U₇, U₁₃, U₁₄, U₁₅. The indifferent attributes (I) were U₁₈.

The classification results reveal that most of the elements categorized as must-have attributes (M) share similarities with commonly used applications in the daily lives of the elderly. These similarities are based on the elderly people’s user experience, such as the preference for large fonts, simple interfaces, and one-click login. Therefore, it is crucial to fully consider the operating habits and cognitive abilities of the elderly when designing the application. In comparison to the necessary attributes, the attractive attributes and expected attributes encompass a wider range of functions and usage demands. These include features such as operation feedback, hierarchical conversion, and voice reminders. Meeting these demands is essential as they address the core needs of medication reminder applications and play a significant role in enhancing the satisfaction of elderly users with such applications. The indifferent attributes (I) do not have a significant impact on user satisfaction, but they have the potential to be transformed into attractive attributes in the future. Hence, this study also takes them into consideration without compromising the overall structure of the application.

4.3. Demand Hierarchy Analysis of Mobile Application of Medication Reminder for the Elderly

4.3.1. Hierarchical Construction of Elderly Users Demands

According to the Kano model, the attributes of the elderly people’s demands were classified, and an analytic hierarchy process (AHP) model was established. The goal level is the design of a mobile application for medication reminders for the elderly, the criterion level is the first-level demand elements, which are visual demand (U1), usage demand (U2), and functional demand (U3). The indicator level consists of the 20 s-level user demands belonging to the U1, U2, and U3 sub-levels (as shown in

Table 9. Construction and classification of demand hierarchy.

Target Layer	First-Level Demands	Second-Level Demands	Quality Type
U: Elderly Medication Reminder Mobile Application Project	U1: Visual demands	U11 Larger text display	M
		U12 Easy-to-understand patterns	O
		U13 Color unity	A
		U14 Simple interface	M
	U2: Usage demands	U21 Fewer levels, less memory burden	M
		U22 Strong fault tolerance	A
		U23 Good operational feedback	O
		U24 Clear and understandable language	A
		U25 Voice prompts capability	A
	U3: Function demands	U31 Operation teaching to aid learning	M
		U32 Easy login process with fewer steps	M
		U33 Medication time reminders and voice alerts	M
		U34 Dosage reminders and voice alerts	O
		U35 Medication precautions reminders and voice alerts	O
		U36 Remote assistance for family members	O
		U37 Recording previous medications	A
		U38 Ability to add medication instructions and doctor’s orders	A
		U39 Reminder to check medication balance	I
		U310 Online chat with doctors	A
		U311 Addressing personal privacy concerns	A

).

4.3.2. Construct Elderly Users Demands Judgment Matrix

To ensure the accuracy and diversity of the answers, a total of 25 relevant personnel involved in the design and research of medication reminder apps for the elderly were tested. A combination of questionnaires and open-ended interviews were used to survey the participants, including 6 app developers from internet companies, 8 UI designers, 6 professors researching the direction of aging-oriented transformation in the field of interaction design, and 5 doctoral students.

For the judgment process, the criteria layer judgment matrix U was constructed, and the indicator layer matrices U1, U2, and U3 were constructed in the same way. Professional personnel were invited to compare each pair of hierarchical requirements based on their relative importance and score them according to the scale. Firstly, construct the judgment matrix of the target level U of the medication reminder application, and the score calculation results are as follows:

$$U=\left|\begin{array}{ccc}1.000& 2.000& 0.500\\ 0.500& 1.000& 0.333\\ 2.000& 3.000& 1.000\end{array}\right|\hspace{1em}{w}_1=\left|\begin{array}{c}0.297\\ 0.164\\ 0.539\end{array}\right|$$

Secondly, construct the judgment matrix of U1 visual demands, and the score calculation results are as follows:

$$U1=\left|\begin{array}{cccc}1.000& 3.000& 4.000& 0.500\\ 0.333& 1.000& 2.000& 0.333\\ 0.250& 0.500& 1.000& 0.250\\ 2.000& 3.000& 4.000& 1.000\end{array}\right|\hspace{1em}{w}_1=\left|\begin{array}{c}0.334\\ 0.139\\ 0.080\\ 0.447\end{array}\right|$$

Thirdly, construct the judgment matrix of U2 usage demands, and the score calculation results are as follows:

$$U2=\left|\begin{array}{ccccc}1.000& 5.000& 4.000& 0.333& 2.000\\ 0.200& 1.000& 0.333& 0.200& 0.333\\ 0.250& 3.000& 1.000& 0.250& 0.500\\ 3.000& 5.000& 4.000& 1.000& 4.000\\ 0.500& 3.000& 2.000& 0.250& 1.000\end{array}\right|\hspace{1em}{w}_2=\left|\begin{array}{c}0.254\\ 0.054\\ 0.100\\ 0.451\\ 0.141\end{array}\right|$$

Finally, construct the judgment matrix of U3 function demands, and the score calculation results are as follows:

$$U3=\left|\begin{array}{ccccccccccc}1.00& 2.00& 0.11& 0.125& 0.167& 0.143& 0.20& 0.33& 0.50& 0.25& 5.00\\ 0.50& 1.00& 0.125& 0.11& 0.143& 0.125& 0.167& 0.25& 0.33& 0.20& 1.00\\ 9.00& 8.00& 1.00& 1.00& 3.00& 2.00& 4.00& 7.00& 8.00& 5.00& 9.00\\ 8.00& 9.00& 1.00& 1.00& 3.00& 2.00& 5.00& 5.00& 6.00& 4.00& 8.00\\ 6.00& 7.00& 0.333& 0.33& 1.00& 0.50& 2.00& 4.00& 5.00& 3.00& 6.00\\ 7.00& 8.00& 0.50& 0.50& 2.00& 1.00& 3.00& 6.00& 6.00& 4.00& 6.00\\ 5.00& 6.00& 0.25& 0.20& 0.50& 0.333& 1.00& 3.00& 3.00& 0.50& 5.00\\ 3.00& 4.00& 0.143& 0.20& 0.25& 0.167& 0.33& 1.00& 2.00& 1.00& 3.00\\ 2.00& 3.00& 0.125& 0.167& 0.20& 0.167& 0.33& 0.50& 1.00& 0.33& 2.00\\ 4.00& 5.00& 0.20& 0.25& 0.33& 0.250& 2.00& 1.00& 3.00& 1.00& 4.00\\ 0.20& 1.00& 0.11& 0.125& 0.167& 0.167& 0.20& 0.33& 0.50& 0.25& 1.00\end{array}\right|\hspace{1em}{w}_1=\left|\begin{array}{c}0.028\\ 0.016\\ 0.232\\ 0.218\\ 0.115\\ 0.160\\ 0.073\\ 0.045\\ 0.031\\ 0.066\\ 0.017\end{array}\right|$$

4.3.3. Consistency Test and Importance Ranking

The constructed judgment matrices were subjected to consistency testing to identify any logical errors. If CR < 0.1, the matrix was considered consistent and passed the test. However, if CR > 0.1, the matrix required revision. Based on the CR values in the table, all consistency indicators of the constructed matrices exceeded 0.1, indicating that they passed the consistency test.

To determine the importance ranking, we first calculated the weights of each primary indicator. Then, we separately calculated the within-group weights and total weights of each required element in the secondary indicators. Finally, the total weight score was used to determine the overall ranking and importance level of each requirement (as shown in

Table 10. Weight arising from the AHP Analysis.

Criteria	Consistency Test	Weight	Sub- Criteria	Quality Type	Weight within Group	Rank within Group	Overall Weight	Overall Rank	Consistency Test
U1	$\mathsf{\lambda }$max = 3.009 CR = 0.009 CI = 0.005	0.297	U11	M	0.321	2	0.107	3	$\mathsf{\lambda }$max = 4.081 CR = 0.030 CI = 0.027
			U12	O	0.142	3	0.047	8
			U13	A	0.087	4	0.029	12
			U14	M	0.450	1	0.150	2
U2		0.164	U21	M	0.254	2	0.085	4	$\mathsf{\lambda }$max = 5.244 CR = 0.055 CI = 0.06
			U22	A	0.054	5	0.018	15
			U23	O	0.100	4	0.033	11
			U24	A	0.451	1	0.150	1
			U25	A	0.141	3	0.047	9
U3		0.539	U31	M	0.028	8	0.009	18	$\mathsf{\lambda }$max = 11.812 CR = 0.053 CI = 0.081
			U32	M	0.016	9	0.005	20
			U33	M	0.232	1	0.077	5
			U34	O	0.218	2	0.073	6
			U35	O	0.115	4	0.038	10
			U36	O	0.160	3	0.053	7
			U37	A	0.073	5	0.024	13
			U38	A	0.045	6	0.015	16
			U39	I	0.031	7	0.010	17
			U310	A	0.066	6	0.022	14
			U311	A	0.017	10	0.006	19

). Referring to the table, U₁₂ (0.0473), U₂₅ (0.0470), U₁₄ (0.1501), and U₂₄ (0.01504) have the same values as shown. To determine the final weight value, we compare the values up to the fourth digit after the decimal point.

4.3.4. Result Analysis

According to the weight results of the indicators for the elderly medication reminder application, the indicator weight rankings of the demand layer are as follows: U3 functional demands > U1 visual demands > U2 usage demands. Therefore, when designing the elderly medication reminder application, it is important to fully consider the functional needs of the elderly.

Functional demands are the most basic and core part of the entire application. According to the table, the design elements that need to be emphasized in this part include U₃₃ Medication time reminders and voice alerts, U₃₄ Dosage reminders and voice alerts, U₃₆ Remote assistance for family members, and U₃₅ Medication precautions reminders and voice alerts. Visual demands are key elements in improving the ease of use for the elderly. Through previous research experience, it has been found that the same application lacks care for the elderly in terms of interaction interface design, such as difficulties in navigation and low identification, etc. Therefore, in the process of interface design, attention should be paid to whether the interface of the entire application is concise, and whether the font size and icons are easy for the elderly to read and understand. Finally, usage requirements, combined with the elderly’s operating habits, can greatly improve their efficiency and reduce the difficulty of learning to use. Hence, it is important to fully consider the cognitive ability of the elderly in the design process. Using easy-to-understand language descriptions, fewer interface levels, and voice prompts can better enhance the satisfaction of elderly users.

4.4. Quality Function Deployment

4.4.1. Define Design Requirements

Based on the weight of each demand obtained from the Kano model analysis and analytic hierarchy process in the previous stage, the user demand is then transformed into the design requirement through the mapping relationship. In the refining process, 4 interface designers, 5 UI designers, and 5 doctoral students majoring in user experience jointly participated in the comprehensive summary of the characteristics of the mobile APP for elderly medication reminders and finally obtained 3 first-level design requirements and 14 s-level design requirements. The following tables describe in detail the refining process of visual demands (

Table 11. Relational mapping between visual demands and design requirements.

First-Level User Demands	Second-Level User Demands	First-Level Design Requirements	Second-Level Design Requirements
U1: Visual Demands	U11 Larger text display	D1: Interface Design	D11 Interface font design
	U12 Easy-to-understand patterns		D12 Interface icon design
	U13 Color unity		D13 Interface color design
	U14 Simple interface		D14 Interface layout design

and

Table 12. The specific content of interface design requirements.

First-Level Design Requirements	Second-Level Design Requirements	Specific Design Contents
D1: Interface Design	D11 Interface font design	Font selection, font style, font size, leading and kerning, text color
	D12 Interface icon design	Icon style, graphic design, size and proportion, color and matching,
	D13 Interface color design	Primary color, secondary color, lightness, purity, contrast, and color balance of color
	D14 Interface layout design	Content organization, navigation and menus, content typography, grid systems, responsiveness and adaptability, images and media, white space and negative space, visual hierarchy

), usage demands (

Table 13. Relational mapping between usage demands and design requirements.

First-Level User Demands	Second-Level User Demands	First-Level Design Requirements	Second-Level Design Requirements
U2: Usage Demands	U21 Fewer levels, less memory burden	D2: Operation Design	D21 Information architecture design
	U22 Strong fault tolerance		D22 Confirm & Return operation design
	U23 Good operational feedback		D23 Feedback design
	U24 Clear and understandable language		D24 Intelligibility & Legibility design
	U25 Voice prompts capability		D25 Voice broadcast design

and

Table 14. The specific content of operation design requirements.

First-Level Design Requirements	Second-Level Design Requirements	Specific Design Contents
D2: Operation Design	D21 Information architecture design	Page hierarchy design (home page, first-level page, second-level page, third-level page), page journey and user path, accessibility design
	D22 Confirm & Return operation design	Prompts for confirming operations, clear labels and buttons, prompts for canceling and undoing operations, prompts for error handling, recovery, and backup
	D23 Feedback design	Visual feedback: success, error, or guidance hints Sound feedback: button sound effect, operation prompt sound effect Haptic feedback: Vibration prompts and feedback
	D24 Intelligibility & Legibility design	Clear and clear language, concise text, correct wording, correct use of punctuation
	D25 Voice broadcast design	High-quality speech synthesis, smooth broadcast sound, text-to-speech conversion, broadcast speech rate and intonation, customizable voice settings, broadcast content selection

), and function demands (

Table 15. Relational mapping between function demands and design requirements.

First-Level User Demands	Second-Level User Demands	First-Level Design Requirements	Second-Level Design Requirements
U3: Function Demands	U31 Operation teaching to aid learning	D3: Function Design	D31 Operation teaching
	U32 Easy login process with fewer steps U311 Addressing personal privacy concerns		D32 User management and security
	U33 Medication time reminders and voice alerts		D33 Reminder settings
	U34 Dosage reminders and voice alerts
	U35 Medication precautions reminders and voice alerts
	U36 Remote assistance for family members		D34 Social function
	U310 Online chat with doctors
	U37 Recording previous medications		D35 Medicine management
	U38 Ability to add medication instructions and doctor’s orders U39 Reminder to check medication balance

and

Table 16. The specific content of function design requirements.

First-Level Design Requirements	Second-Level Design Requirements	Specific Design Contents
D3: Function Design	D31 Operation teaching	Operation instructional videos and demonstrations
	D32 User management and security	One-click login, personal account information security, data backup, and synchronization
	D33 Reminder settings	Visual and voice reminders of medication time, dosage, and medication precautions, medication refill reminder
	D34 Social function	Contact relatives and friends, emergency contacts and help function, and communicate with doctors online, health advice and information, Connect with other wearable devices (medication reminder or health monitoring)
	D35 Medicine management	Medication records and reports, medication inventory and management, medication information and instructions, personalized medication management plans

) into design requirements and the specific design contents of 14 s-level design requirements.

In the process of transforming the mapping relationship between functional demands and design requirements, because each user demands belongs to a different category of design requirements, the second-level user demands are placed according to the classification of the second-level design requirements.

4.4.2. Matrices Solution to Determine the Relationship between User Demands and Design Requirements

Based on the elderly user demands (Table 9) and design requirements (Table 12, Table 14, and Table 16), a house of quality matrix model is constructed to illustrate the relationship between the two. In the house of quality model, the “left wall” represents the elderly user demands, while the “ceiling” represents the design requirements. The correlation between user demands and design requirements is depicted using scores. A score of 5 indicates a strong correlation, 3 indicates a moderate correlation, 1 indicates a weak correlation, and a blank space indicates no correlation. The scores were assigned by 4 elderly target users and experts in interaction design and interface design. The results of this assessment are presented in

Table 17. User demands and design requirements house of quality.

User Demands		Importance	Design Requirements
			D1				D2					D3
			D11	D12	D13	D14	D21	D22	D23	D24	D25	D31	D32	D33	D34	D35
U1	U11	0.111	5	1	3	1			5	5	3	1		1	1	3
	U12	0.046	1	5	5	3		3	5	5	5	3	3	3	1	1
	U13	0.027	3	5	5	3	3		1			3	3
	U14	0.149	1	3	3	5	3		3	1	3	5	5	3	1	1
U2	U21	0.085				3	5		3			1
	U22	0.018				1	1		5	1	5		3
	U23	0.033				3		1	3		5		5
	U24	0.150	1					3	1	5	1	3	3
	U25	0.047		1				5		3	3			5	3
U3	U31	0.008	1			1		3		3	5	5
	U32	0.005				1					1		5
	U33	0.079	1	3	1	3		5	5	5	5			5		1
	U34	0.074	1	1	1	3		5	5	5	5			5	1	1
	U35	0.039	1	1				5	3	5	3			5	1	3
	U36	0.055	1	1				5	1	5	5		1	3	1	1
	U37	0.023				1		3	3	3	5			1	5	5
	U38	0.014	3						1	3					1	5
	U39	0.010						1			3				1	5
	U310	0.021	1			1		3			3		1	1	5	1
	U311	0.005						1					5	1	3	1
Importance			1.301	1.376	1.299	1.964	0.968	2.263	2.878	3.217	2.969	1.649	1.764	2.107	0.876	1.117
Total Importance Ranking			10	9	11	6	13	4	3	1	2	8	7	5	14	12
Rank in group			3	2	4	1	5	4	3	1	2	3	2	1	5	4

.

4.5. PUGH Decision Matrix

In this project, the PUGH decision matrix technique was employed to evaluate the generated plans for the elderly medication reminder app. The scores obtained were utilized to identify the optimal plan, aiming to ensure a satisfactory user experience and enhance the efficiency of usage for the elderly.

4.5.1. Creating Design Solutions

Based on the house of quality matrix’s prioritized ranking of design requirements, the highest-ranked design requirements serve as the main reference for the design of medication reminder applications for the elderly, which include “D₂₄ Intelligibility & Legibility design”, “D₂₅ Voice broadcast design”, “D₂₃ Feedback design”, “D₂₂ Confirm & Return operation design”, “D₃₃ Reminder settings”, and “D₁₄ Interface layout design”. To avoid the limitations and one-sidedness of the design, we have integrated the design standards and requirements of interface functionality layouts with high usage and recognition rates in current reminder applications, and have designed the following 4 schemes (as shown in Figure 3, Figure 4, Figure 5 and Figure 6).

4.5.2. Program Evaluation

Based on the design requirements, the PUGH decision matrix is created to evaluate the interaction interface and functional usage of four high-fidelity design schemes. Scheme C is selected as the reference scheme. Formula (7) is employed to calculate the scores, ensuring rationality and objectivity in the evaluation process. The final results are obtained by ranking the schemes based on their total scores and importance. The evaluation results show that Scheme B achieved the highest score of 84.997 points (as shown in

Table 18. Correlation between user needs and evaluation elements.

Design Requirements		Importance	Ranking	Alternative
				Project A	Project B	Project C	Project D
D1	D11	1.301	8	2	3	3	2
	D12	1.376	7	1	5	2	3
	D13	1.299	9	2	5	5	2
	D14	1.964	6	3	4	2	2
D2	D21	0.968	11	2	5	2	4
	D22	2.263	5	2	4	3	3
	D23	2.878	3	2	3	5	4
	D24	3.217	1	4	5	2	3
	D25	2.969	2	3	4	2	3
D3	D31	0.026	16	2	4	5	3
	D32	2.272	4	2	5	4	3
	D33	1.244	10	3	4	2	4
	D34	0.426	14	2	3	2	2
	D35	0.895	12	5	4	3	2
	D36	0.630	13	4	3	5	3
	D37	0.170	15	2	5	2	3
Total Score				62.976	99.659	71.297	70.898

), followed by Scheme C with 82.097 points, Scheme D with 76.956 points, and Scheme A with the lowest score of 69.699 points. By using this scoring method, we can select the best design solution and incorporate it into the application development process.

To validate the effectiveness of the design solution, a comparative analysis was conducted between this design solution and existing medication reminder apps available in the market. In this experiment, three different apps, namely Meddify, iCare · Take medicine reminder, and Scheme B, were selected for satisfaction comparison. Five participants aged 60 and above, with prior experience using medication reminder apps, were recruited to take part in the experiment. Each participant was asked to spend 10 min experiencing each app and, upon completion, rate each app based on the design requirements. Ratings were given on a scale of 1 to 5, where 1 represented “very dissatisfied”, 2 represented “dissatisfied”, 3 represented “neutral”, 4 represented “satisfied”, and 5 represented “very satisfied.” The final scoring results show that Scheme B has the highest score, which is more in line with the needs and satisfaction of the elderly (as shown in

Table 19. Comparative analysis of Meddify, iCare Take medicine reminder, and Scheme B.

Design Requirements		Meddify	iCare · Take Medicine Reminder	Scheme B
D1	D11	2	3	4
	D12	3	1	3
	D13	4	2	4
	D14	3	2	3
D2	D21	4	3	3
	D22	2	3	4
	D23	4	2	4
	D24	2	5	5
	D25	3	2	4
D3	D31	4	3	3
	D32	5	3	2
	D33	4	2	5
	D34	3	4	3
	D35	4	3	4
Total Score		47	38	51

).

5. Discussion

5.1. Optimization of Design Method by Integrating the KANO-AHP-QFD-PUGH Model

This paper proposes an integration of QFD theory with the Kano model, AHP, and PUGH decision matrices to enhance the objectivity and accuracy of design decisions and outcomes. The Kano model captures user demand information through surveys and qualitative feedback but lacks a quantitative measurement of the importance of different user needs. By incorporating AHP, subjective evaluations of various needs can be converted into weights using a quantitative judgment matrix and calculation process, thereby addressing the limitations of the Kano model [15]. The QFD model facilitates the transformation of user demands into specific design requirements, determining the prioritization of design demands through the construction of a house of quality [36]. These output design requirements serve as evaluation criteria in the Pugh decision matrix, enabling a structured comparison and assessment of different designs. The combination of QFD and PUGH methods enables decision-makers to comprehensively understand the strengths and weaknesses of each design option, mitigating subjective biases and yielding more objective evaluation results [22]. Combining the 4 methods of the Kano model, AHP, QFD, and Pugh decision matrix can provide comprehensive requirements identification, prioritization, decision support, risk reduction, and quality assurance, thereby helping the design team to better meet user demands and improve design quality and user satisfaction.

5.2. Optimal Design Strategy for Elderly Medication Reminder

5.2.1. Visual Design Strategy

For the visual design requirements of the medication reminder application for the elderly, priority should be given to the simplicity of the interface design (D₁₄ Interface layout design). Displaying an excessive amount of information on a single page can burden the memory of elderly users and distract them during usage. Therefore, the interface layout should prioritize the core functions and content of the application, followed by the design of display fonts (D₁₁ Interface font design). Using characters that are excessively large or bold can result in limited content being displayed and may not be visually pleasing. Similarly, characters that are too small or thin can pose difficulties for elderly users in reading. Hence, it is crucial to determine the appropriate character style and size range from the outset of the design process. The design of icons within the application (D₁₂ Interface icon design) is also essential for content recognition. Abstract icons can impede the cognitive abilities of the elderly, while a flat design approach proves more effective in overcoming cognitive barriers. Overall, the design should aim for simplicity, clarity, and easy recognition. Additionally, maintaining a consistent interface color scheme (D₁₃ Interface color design) is desired. Excessive color combinations can lead to confusion and hinder the visual experience. Although colors such as red and yellow enhance interface recognition, their strong stimulation may cause eye discomfort during prolonged use. Therefore, when considering recognition, it is important to account for the physiological characteristics of the elderly.

5.2.2. Operation Design Strategy

Regarding usage requirements, elderly users need language descriptions that are easy to understand (D₂₄ Intelligibility & Legibility design). They are concerned about whether they can fully understand the text and operational prompts displayed in the app. Otherwise, they will experience greater psychological anxiety. Therefore, in the design process, we should first consider the cognitive and memory levels of most elderly people regarding text understanding. During use, the elderly will continuously confirm their actions’ correctness or incorrectness through the app’s timely feedback, such as sound, visual, and tactile feedback (D₂₃ Feedback design). These feedback modes will enhance the elderly users’ perception of safety, allowing them to know the task or project they are undertaking. The prompt to confirm and return operations (D₂₂ Confirm & Return operation design) is an effective error prevention measure. Due to the decline in cognitive ability and physiological function, elderly users have a higher likelihood of miss-operation than younger people. Therefore, more confirmation information is needed to prevent errors. Voice broadcasting (D₂₅ Voice broadcast) is particularly favored by elderly users, as they believe that playing voice information can effectively reduce their visual burden. Furthermore, they can receive medication reminders at any time while doing other things. Additionally, there should be fewer page hierarchies (D₂₁ Information architecture design), as multiple hierarchical information structures can create memory difficulties for the elderly. Basic functions and buttons are more suitable for the primary page, saving users’ search time, and improving ease of use and efficiency. This research result is similar to Rahul Malhotra et al.’s findings on low usage rates and willingness to use medication reminder applications for the elderly [3]. A complex interface, dense usage instructions, and language barriers greatly affect elderly users’ willingness to use [40].

5.2.3. Function Design Strategy

Finally, we also found that in terms of functional requirements, older adults are more concerned about the most basic and primitive functions of this application—reminders of medication time and dosage (D₃₃ Reminder settings). They are also very willing to share their medication information with their family members or complete medication reminders and notifications from their family members through electronic devices (D₃₄ Social function). According to the summary of the changes in reports on medication adherence challenges by Mushfique Ahmed et al. [41], reminders may become more feasible, especially reminders from family members. The main reason for this is that many older adults feel a strong sense of loneliness, and receiving care and greetings from their family and friends will make them feel warm. Therefore, in future designs, the connection between older users and their relatives and friends should be emphasized.

In addition to the above requirements, according to the experimental discussion results of the mobile application functional preferences for medication self-management in older adults by Andrea M. Russell et al., medication records and medication instructions (D₃₅ Medication management) are considered to be one of the most important functions [31]. Due to the COVID-19 pandemic, many older users with chronic diseases often have other complications and need to take multiple medications at the same time, so it is necessary to ensure the mutual warning effects between medications, making the design of these two functions particularly important. The privacy issues and convenience of logging in (D₃₂ User management and security) for older users also need to be considered. A previous research report in Singapore stated that privacy issues are a major barrier to the widespread adoption of technology, especially in the healthcare field [41]. Providing offline data or privacy settings is necessary, and the login method is the first step for users to start using the app [3]. Therefore, more convenient login methods can effectively save time and improve usability, such as logging in by connecting with other social accounts, otherwise, it may reduce users’ willingness to continue using the app.

6. Conclusions and Limitations

Due to the impact of the COVID-19 pandemic, older adults face a heightened memory burden when it comes to medication use. In order to help improve the medication compliance of the elderly and the usability, efficiency, and satisfaction of the medication management APP for the elderly, this study attempts to improve the design problems of the medication management APP and proposes a possible integrated Kano-AHP-QFD-PUGH The continuous optimization design method not only solves the limitations of the single design method, but also makes up for the shortcomings of the medication management application in the aging transformation, and improves the user experience of the elderly.

In the case study, we demonstrate how to obtain user needs through qualitative research and classify them into Kano attributes. We then use AHP to construct a hierarchical structure of user needs and determine the weight rankings of each need. QFD method maps user needs to design requirements and uses a quality house to determine the importance of each design requirement, ultimately identifying key needs. Finally, PUGH concept selection is used to verify the feasibility of the design scheme, and the results indicate that Project B is the best design solution.

The entire design process focuses on the elderly user group and uses a combination of qualitative and quantitative methods to integrate their special needs into the app design. The results of this study have both theoretical and practical significance. It not only provides a theoretical basis for designers to develop app design methods in the future but also offers specific targeted design solutions. This study demonstrates that this research framework can effectively be applied to the age-appropriate transformation of medication reminder apps for the elderly, and is not only suitable for the elderly but also for other user groups researching sustainable user experience optimization design for apps.

In future endeavors, it is essential to gather additional data on the requirements of elderly users in various regions. Simultaneously, conducting user testing and usability research based on the current findings is crucial. Furthermore, placing greater emphasis on addressing the emotional needs of the elderly and alleviating any psychological burdens associated with technology products will enhance their overall user experience. Finally, we should also pay attention to the willingness of elderly users to continue using the medication reminder APP, and improve user experience and satisfaction through continuous iterative updates.

This study still has some limitations. Firstly, in terms of data collection, due to constraints such as time and resources, the selection of research subjects was confined to elderly communities in first- and second-tier cities in China. Consequently, the number of research samples was limited, which could introduce bias and impact the representativeness of the research findings. Secondly, during the application of the integrated method of KANO-AHP-QFD-PUGH, certain aspects of the research process required judgments based on the researcher’s expertise and personal subjective experience. Consequently, subjective factors might have influenced the research outcomes. Lastly, regarding generalizability, significant variations exist in living environments, geographical locations, economic levels, and cultural backgrounds across different regions, countries, and cultural contexts. These variations can lead to corresponding differences in user needs, potentially limiting the universal applicability of the research findings.