Univ Access Inf Soc DOI 10.1007/s10209-016-0503-y

LONG PAPER

Development of a web-based insulin decision aid for the elderly: usability barriers and guidelines Angeline Su Lyn Lum1 • Thiam Kian Chiew1 • Chirk Jenn Ng2 • Yew Kong Lee2 Ping Yein Lee3 • Chin Hai Teo2

•

 Springer-Verlag Berlin Heidelberg 2016

Abstract In recent years, researchers have attempted to shift patient decision aids (PDAs) from paper-based to web-based to increase its accessibility. Insulin decision aids help diabetes patients, most of whom are elderly to make an informed decision to start insulin. However, the lack of usability guidelines applicable for such target group causes developers to struggle to answer the challenging question ‘How can such web service be made usable, and, ultimately, acceptable and accessible for elderly patients?’. Hence, the purpose of this study is to identify the common usability requirements that may facilitate good practices to empower elderly diabetes patients in utilizing a web-based insulin decision aid for their benefit. We set out an

& Angeline Su Lyn Lum [email protected] & Thiam Kian Chiew [email protected] Chirk Jenn Ng [email protected] Yew Kong Lee [email protected] Ping Yein Lee [email protected] Chin Hai Teo [email protected] 1

Department of Software Engineering, Faculty of Computer Science and Information Technology, University of Malaya, 50603 Lembah Pantai, Kuala Lumpur, Malaysia

2

Department of Primary Care Medicine, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia

3

Department of Family Medicine, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, Serdang, Malaysia

approach to use prototyping and retrospective think-aloud techniques to explore web usability barriers that elderly patients may encounter when using an insulin decision aid web site and use the feedback for improving the prototype. Usability requirements were captured iteratively through scoping, brainstorming, prototype, testing and evaluating. The study suggests that the insights from experts and users are equally important to assure the validity of the identified usability guidelines; they reflect the accessibility needs of the aging community while complementing the key requirements of an insulin decision aid. The study contributes to recommend web usability guidelines backed by a series of expert and user evaluations which could be a proactive resource to improve usability, acceptability and accessibility of online insulin decision aids for elderly with diabetes. Keywords Elderly
Web usability
User-centered design
Human–computer interaction
Diabetes mellitus
Insulin decision aid

1 Introduction The growing numbers of elderly in worldwide population have led to the call for a serious attempt to discover how new technologies can be effectively designed to include the usability, accessibility and acceptability requirements of older adults. According to United Nations [1] and the World Health Organization (WHO) [2], the aging of population worldwide will continue to increase for the decades to come. It is estimated that there will be approximately 1.2 billion people aged 60 years and above by 2025, increasing to approximately 2 billion by 2050. In fact, it is reported that older adults are the fastest growing consumer segment

123

Univ Access Inf Soc

of Internet users at present as a result of the growing numbers of elderly who use computers to avoid social exclusion and live more independently [1, 3]. Guidelines such as ‘All possible actions should be visible always’ are far too abstract and may have different interpretations depending on the specific context of use [4, 5]. The definition of guidelines applicable to meet the requirements of a specific group of users and a specific type of application is very much needed. Furthermore, there is ample evidence that developers of various domains are constantly seeking for concrete examples of existing designs as their guide [6, 7]. Hence, the availability of an existing usability guideline proven effective for users with similar backgrounds and needs would be an effective stepping stone when developing a user interface. Developers might be supported more effectively if given a more specific guideline of the captured development context [5]. The traditional software engineering process has involved the development of solutions for problems de novo. However, it is impracticable to start from scratch once critical knowledge on reusing common solutions to common problems is obtained. Thus, the approach of domain-specific software engineering (DSSE), which focuses on aspects of a certain domain and on domainspecific solutions, techniques and patterns, has emerged and found to improve quality, productivity and promote reuse [8]. The present study focuses on the usability guidelines for a web-based insulin decision aid that facilitate insulin initiation decision making. Having usability guidelines for the domain of web-based insulin decision aids for elderly may further support developers to conduct DSSE for future applications that reside in the same domain. The objective of patient decision aids (PDAs) is to help patients to make informed and value-based decision. To date, patients are facing difficulties in making treatment decision because they are often not well informed about the options they may have along with the subsequent pros and cons. To counter this problem, PDAs are designed to improve the amount and thoroughness of informing and decision sharing [9].Throughout the past 30 years, health innovations have been developed to improve patient decision quality [10], and in recent years, researchers have attempted to shift PDAs from paper-based to webbased to increase its accessibility. Despite the advancements in the development of PDAs, the question remains on how to develop an online decision aid tool that is usable for elderly. It is important to consider the patient’s age when designing an online PDA in order to allow users to gain full advantage from it. However, few PDAs take into consideration this factor and are limited by the format of the PDA [10].

123

The insulin decision aid is a form of patient decision aid (PDA) that supports patients with type 2 diabetes to make a decision on insulin initiation. Diabetes mellitus is a common chronic disease worldwide. The global prevalence of diabetes is expected to increase along with aging of populations with the over 60-year age-group expected to show the largest increases [11]. In diabetes care, the use of insulin treatment is an effective way to achieve good glycemic goals and PDAs play an important role in facilitating insulin initiation decision making as the decision to start insulin therapy is often difficult [12, 13]. Given that a broad potential user cohort of web insulin decision aid tools is the elderly population, their usability must give considerable attention to aging factors. A recent study found that there is a conflict between healthcare professionals’ (HCPs) and patients’ expectations of a PDA. On insulin initiation, HCPs tended to focus on information related to the benefit of insulin initiation while patients wanted to know the impact of insulin on their disease and lifestyle. Also, patients’ preference of content presentation often opposes to those of HCPs and this may negatively impact the acceptability of the web site among patients [14]. The key is that patients need to acquire and understand health information provided in an unbiased manner. To ensure this, patients’ beliefs and knowledge about diabetes and insulin need to be accounted when designing an online insulin decision aid. Furthermore, designing for elderly unskilled users is a far greater challenge than designing for experienced frequent users. Therefore, the purpose of this study is to explore what are the web usability barriers elderly patients may encounter in using an online insulin decision aid and propose the identified guidelines to ensure its usability and accessibility for elderly. Section 2 describes how aging factors affect web usability for the elderly. Section 3 describes the web usability requirements in the context of insulin initiation decision making. Section 4 draws out the study method on how data are gathered, while Sect. 5 summarizes the findings from the study including the usability barriers and respective usability guidelines. Section 6 identifies present study limitations and discusses future directions. Lastly, Sect. 7 conclude the paper.

2 Web usability and the elderly The insulin decision aid web site provides elderly diabetes patients of 60 years old and above the opportunity to access online information to help them make informed decision about insulin initiation [15]. This has been helpful since traditional healthcare settings have limitations such as lack of time and human resources to provide adequate information for each patient [16, 17]. However, such PDA

Univ Access Inf Soc

does not guarantee that elderly users will succeed in benefiting from the information they seek. Web sites that do not meet the online usability needs of targeted users may prevent them from achieving their goals [15]. While old age is not a barrier to Internet use, agingrelated factors including vision, cognitive and physical abilities may limit the elderly from making full use of the latest technology [18]. Table 1 shows a summary of these aging factors and their impact on web usability for the elderly. Acknowledging this, the National Institute of Aging (NIA) and the National Library of Medicine (NLM) have introduced several guidelines for making web sites senior friendly as shown in Table 2.

3 Web usability within the specific domain of insulin initiation decision making Usability is defined in ISO 9241-11 [22] as the extent to which a product can be used by specified users to accomplish specified goals with effectiveness, efficiency and satisfaction in a specified context of use. As such, in the context of a web-based insulin decision aid, usability requires to take into consideration the areas that affect effectiveness, efficiency and satisfaction of the insulin decision aid in helping elderly patients to decide whether to start insulin intake. Firstly, it is important to understand the functions of an insulin decision aid. Also, it is necessary to understand the factors influencing patients’ decision to start insulin intake from both the patients’ and HCPs’ perspectives. The International Patient Decision Aid Standards (IPDAS) collaboration has identified a set of standards to guide the assessment of PDAs. Table 3 summarizes the important features of an insulin decision aid based on the IPDAS collaboration framework [23, 24]. It is crucial for patients to have the locus of control when making a decision whether to start taking insulin. This can avoid non-adherence after initiation. Often,

patients interpreted insulin PDA content as being persuasive toward insulin over other options and HCPs were supportive of information that favored insulin initiation as clinically, insulin is the best option for blood sugar control [14]. Therefore, care should be taken to ensure that insulin decision aids do not exert psychological influence on patients to start insulin. This influence is exerted as a biased representation of information as depicted in Fig. 1.

4 Study method The study presented in this paper is an exploratory study to learn more about how elderly diabetes patients use the web-based insulin decision aid, to determine their unique functional requirements if any, and to identify specific usability requirements. 4.1 The research team and patient selection process The study team was the synthesis of an active group of researchers, primary care clinicians, a psychologist, elderly patients and designers. Patients play a key role in testing the prototypes created through iterative development. This study involved 13 participants who are patients with type 2 diabetes and are in the midst of making a decision on starting insulin as advised by their doctors. Other inclusion criteria were as follows: – –

Have good English proficiency; Have knowledge on how to use a computer.

On the other hand, the exclusion criteria were as follows: – –

Have cognitive, visual or hearing impairment; Have mental health problems such as depression or anxiety.

English was widely used before Malaysian independence 60 years ago, and the older generation (aged 50 and

Table 1 Aging factors and web usability for the elderly Aging factor

Impact on web usability

Vision

The aging eye causes a decline in visual acuity affecting the ability to see clearly. Vision changes including loss of contrast sensitivity and loss of ability to detect fine details affect the use of web site in terms of reading comprehension and speed, navigation and searches [15, 18–20]

Cognition

Working memory decreases with age. Elderly people often lose the ability to simultaneously remember and process new information, to perform complex cognitive tasks and to comprehend text. Also, they process information more slowly. Therefore, complex navigation and cluttered web pages are not suitable for elderly [17–19]

Physical Abilities

Elderly people experience decrease in motor coordination. Hence, it is difficult for them to scroll down a web page or move and click a mouse. Their movements tend to be less smooth and less coordinated therefore take longer. Additionally, elderly users often have difficulty positioning the cursor of a mouse, in particular, when interacting with smaller objects [11, 15, 18, 19, 21]

123

Univ Access Inf Soc Table 2 Making web sites age-friendly usability checklist Source [18] Web usability

Guideline

Improve readability of online text

Font: • Use a sans serif typeface that is not condensed such as Arial • Use 12 or 14 point type size for body text • Use white lettering on a dark-colored background

Able to present information to mediate elderly increase cognitive skills effectively and decrease memory recall activities.

Text: • Present information in a clear and consistent way to reduce the number of inferences • Use positive statements phrased in active voice • Present text in simple language and break lengthy documents into brief sections Media: • Use text-relevant images • Use short segments to improve performance on older computers • Provide text alternatives for other media such as animation Navigational features: • Use explicit step-by-step navigation procedures to ensure understanding of what follows next • Use consistent layout comprising same symbols and icons throughout • Incorporate text with the icon if possible • Incorporate buttons such as previous page and next page to allow the reader to review or move forward • Use single mouse click over double clicks for information access

Support elderly physical abilities

• Use mouseover technology sparingly • Use pull-down menus sparingly • Incorporate specific scrolling icons on each page • Use large buttons that do not require precise mouse movements for activation

Table 3 Important features of insulin decision aid in relation to IPDAS criteria IPDAS criteria

Insulin decision aid feature

Providing information about options Presenting probabilities

The decision aid describes information including the pros and cons of all the treatment options The decision aid describes the outcome probabilities associated with communicating patients’ blood sugar level risk including specifying the event rates and using the same denominator

Clarifying and expressing values

The decision aid requires patients to identify the concerns that matter most to them and their priority

Guiding/coaching in deliberation and communication

The decision aid provides step-by-step approach to make decision whether to start insulin. It also provides tools (e.g., worksheets) to facilitate patient–doctor discussion/communication

Disclosure

The decision aid includes funding details used for development

Balancing the presentation of options

The decision aid provides the pros and cons of each treatment options

Using plain language (readability)

The decision aid uses simple and clear English language (plain language)

Basing information on up-to-date scientific evidence

The content of the decision aid is based on clinical evidence

Establishing effectiveness

The decision aid shows improvement in matching features that matter the most to patients and their chosen treatment option

above) has been English-educated [25, 26]. Hence, the participants in this study represent the Malaysian elderly population. The primary care clinicians were patient

123

decision support experts. In this study, they presented the prototypes to patients that they were not attending to (to avoid biased results), conducted the observations,

Univ Access Inf Soc

Missing informaon about other treatment opons

Focus only on benefits of insulin

Start insulin Fig. 1 Psychological factors in favor of insulin initiation

collected, evaluated and shared the users’ feedback with the rest of the team. The researchers developed the prototypes, made necessary modifications and drew out the agreed usability criteria. The designers then provided support on web design issues. Throughout the study, the team were constantly engaged with face-to-face meetings to come to a final expert consensus in identifying the insulin decision aid web site usability barriers encountered by elderly together with the proposed guidelines. 4.2 The usability criteria development process The study process involved a literature review and collaborative development of usability criteria, which comprised an iteration of phases. For this study process, we obtained an ethics approval from the University of Malaya Medical Centre Medical Ethics Committee (MEC Ref No 1038.4). Figure 2 depicts the process of developing the usability criteria, consisting of the following phases: scoping, brainstorming, prototype, testing, evaluating and iterating again from brainstorming to evaluating.

Fig. 2 Process of usability criteria development

First, we began by gathering information regarding web accessibility for elderly and the functionality of insulin decision aid from published literature. Meanwhile, the clinicians delivered the scope of the web-based insulin decision aid. The content of the web site prototype was developed according to the International Patient Decision Aids Standards (IPDAS) and was adapted from the paperbased PDA, ‘Making choices: Should I start Insulin?’ series of patient decision aid books developed by the researchers. The paper-based PDA has been listed in the A to Z inventory of the Ottawa Hospital Research Institute (OHRI) database, a worldwide library for PDAs which have met the IPDASi criteria. Initially, the paper-based insulin decision aid was developed based on literature reviews and input from all stakeholders involved in diabetes care. These included patients, doctors, nurses, pharmacists and policymakers [27]. Then, a study [28] was conducted to explore patients’ and health-care professionals’ views on the content and format of the paper-based PDA. The results of the study [28] reported that patient decision aids with more detailed information generally resulted in slightly higher knowledge and lower ‘feeling uninformed’ scores than those with simpler information, though the differences are small and can be reversed under some circumstances [29]. Therefore, a web site might be better as it addresses more individualized needs as patient can choose the amount of information they want [28] and improve the accessibility of a PDA. Furthermore, a paperbased PDA requires printing and mailing which may hinder wide dissemination and printing and reprinting of the PDA. By comparison, developing a web-based PDA often only needs a single payment for computer programming and web design in the initial phase and its maintenance and updating requires less cost than reprinting of the paperbased PDA [30]. However, there are concerns that webbased PDAs may marginalize older people who may not be IT-savvy [27]. Hence, the web site prototype is designed to investigate and improve the usability of web-based PDA for elderly patients. The web site consists of six main sections namely ‘Diabetes and insulin,’ ‘Concerns about insulin,’ ‘Blood sugar and risk,’ ‘Treatment options,’ ‘Knowing your priority’ and ‘Support and decision’. Additionally, the web site includes a ‘Summary’ section. The group of researchers, clinicians and designers were then engaged in a brainstorming session to further discuss and set the main scope of the web site. The defined scope was to adapt the already available paper-based PDA to be accessible through the Internet with several extensions which include risk communication of patients’ personal blood sugar level and the generation of summary for facilitating patient–doctor discussion. The group was also involved in brainstorm sessions to discuss how to address all the issues that each cycle of usability tests raised. This

123

Univ Access Inf Soc

approach allowed each member to share their expertise and experiences in their respective disciplines creating informed iterations of prototypes, hence, usability guidelines. All constructive insights and resulting consensus were documented for changes in subsequent iterative prototypes. The first prototype was based on the IPDAS framework and the defined scope, while successive prototypes of the web site were developed according to brainstorm outputs. The process was then followed by usability testing of the prototypes. The clinicians presented the prototypes to actual patients during their clinic visits, and observations about the prototypes were screencast as they used them. The insulin decision aid is based on desktop computers as mobile technologies such as smart phones and tablets are small and have small buttons and screens that are hard to see and operate [31]. The limited screen size of smart phones and tablets may likely require more pages or screens to accommodate all the information of a PDA, in other words, may burden the elderly’s motor skills when more scrolling is necessary to navigate between information. Additionally, the bigger screen size of the desktop computers as compared to smart phones and tablets is better in accommodating the vision of older adults. In this study, a screen resolution of 1366 9 768 pixels is used. Besides that, mobile devices often use deep menu structures that place heavy demands on memory [31]. Thus, desktop computers have been used in the present study. Afterward, patients were engaged in a retrospective thinkaloud session whereby they informed the ‘what’, ‘why’ and ‘how’ of their reactions to the web site as their recordings were played back. The clinicians took note of all the feedbacks justifying patients’ actions. The researchers also probed the patients to clarify certain issues which were not explained by patients. The conversations were audiorecorded with consent from the patients. The data gathered from the usability testing were then evaluated by the clinicians who went through the web site by sections, shared their observations and identified clear and apparent patterns. The focus strategy of patterns to look for included usability issues related to aging and factors affecting individual insulin initiation decisionmaking process. Based on their expertise in the field of patient decision making, the clinicians evaluated the extent to which the usability questions (if any) raised were applicable for the next iteration. As said, these insights were brought up in successive brainstorm iterations to allow experts’ analysis. The clinicians described usability elements essential for informed patient decision making. The researchers shared their methodologies in web development while the designers shared their thoughts on web design for older people. This assured a thorough process of identifying the usability criteria. These iterations involved

123

three to six patients for each cycle and continued until the team agreed that the web site was successfully facilitating elderly patients to obtain adequate information to make an informed decision on whether to start insulin therapy. The flowchart shown in Fig. 3 illustrates the usability criteria development process. Table 4 describes the people and activities involved for the three iterations of the design–evaluate–redesign process.

5 Summary of findings Our findings, while preliminary, may encourage future studies on web usability for elderly specifically in the domain of online insulin intake decision aids. We ceased recruiting participants when thematic saturation was reached. Thematic saturation is defined as the ongoing collection and analysis of data until no new concepts emerge with successive interviews [32, 33]. In our research, thematic saturation was achieved after three iterations of the usability criteria development process. The initial designs were the outcome of several discussions sessions among the HCPs, designers and researchers. Through the pilot testing through a retrospective think-aloud process, the HCPs could understand the rationale of patients’ actions and reactions to the PDA web site. The HCPs also found out what designs have succeeded or failed to achieve the goal of informed patients’ decision making. The successful designs were retained or refined, while the failed designs were revamped in the subsequent discussion session among the HCPs, designers and researchers. Interestingly, through this study, we found that some web elements may influence elderly users to refrain from starting insulin therapy. We also found that involving user and getting their feedback from early prototypes and iterative development improves the quality of iterative phases with regard to a decrease in the number of iterations involved in the design–evaluate–redesign cycle to obtain thematic saturation. This is compliant with the ‘golden rules’ of interface design guidelines [34]. Also, when identifying web usability criteria insulin decision aids for elderly patients, the insights from experts and users are equally important to assure the validity of the identified usability guidelines. The findings were in accordance with the beta testing conducted with 13 patients, whose profiles were as follows: age range 49–75 years, (median age of 65 years), 10 male and 3 female with 9 having a secondary education degree and 4 graduate/postgraduate degree. The heterogeneity of elderly and middle-aged patients (the age of patients ranged from 49 to 75) who participated in the study helps us to

Univ Access Inf Soc Fig. 3 Flowchart of the usability criteria development process

better understand the behavioral differences that might exist between younger and older adults and how these might affect the insulin decision aid web site’s usability. From this study, we observed a variety of notable differences in how older adults interacted with the prototype web site. Furthermore, the diversity of the patients’ age in our study can be explained, as the term elderly people itself encompasses a strongly inhomogeneous group of people

and even small subsets of these groups have a great diversity of needs and functionality [35, 36, 37]. 5.1 Web usability barriers There are several web usability barriers identified from this study. The barriers can be classified into two categories, namely aging factors and psychological factor. Aging

123

Univ Access Inf Soc Table 4 Design–evaluate–redesign process Iteration/ phase

People and activities involved

First iteration Design

Evaluation

• Clinicians defined the scope of the web-based insulin decision aid such as the content of the PDA which conforms to the IPDAS framework • Researchers and designers developed the initial web site prototype based on the given scope • Clinicians presented the initial web site prototype to the participants during their clinic visit • The participants used the insulin decision aid without intervention from the clinicians. They went through each section of the web site prototype (User Assessment) • Clinicians and participants were engaged in a retrospective think-aloud session. Participants informed the ‘what’, ‘why’ and ‘how’ of their reactions to the web site as their video-recorded reactions were played back. The clinicians took note of the feedback • The clinicians evaluated the extent to which the usability questions (if any) raised were applicable for the next iteration

Subsequent iteration (second and third iteration until thematic saturation is achieved) Design

• The clinicians collated and described to the researchers and designers on the findings on usability elements essential for informed patient decision making • The three parties, clinicians, researchers and designers discussed and reached a consensus on the revisions to be made for the next iteration • Researchers and designers redesign the prototype based on the consensus made

Evaluation

• Clinicians presented a new revision of the web site prototype to the participants during their next clinic visit • Participants reassessed the redesigned prototype and engaged in a retrospective think-aloud session with the clinicians • The clinicians noted and evaluated the findings

factors can be further classified in vision, cognitive and physical abilities. 5.1.1 Aging factors Aging factors, vision, cognitive and physical abilities are the more apparent group of factors that cause web site usability problems with elderly patients. Usability issues that fall into this category disable users from taking full advantage of the web site as they are not able to make use of all the features available. Vision The optimal usability of the web site is affected by the elderly user’s ability to see clearly especially for diabetes patients who are prone to comorbidities such as cataract. We found the elderly patients struggled to read the information from the web site when the font size is \20 pixels. In terms of colors and backgrounds, it was found that as opposed to the ‘Senior-Friendly Usability Checklist’ [18], elderly patients highlighted that lighter color backgrounds with white lettering help them to see well compared to darker backgrounds. Additionally, elderly patients had a slow reading speed. They had trouble digesting the information from the slideshow playing at a speed of 3 s per slide in the initial prototype. Other barriers identified included having difficulties in locating radio buttons when not outlined and difficulties finding text visible when not filled with solid color like the numberings shown in Fig. 4.

123

Another significant barrier identified under this sub-category was the negative effect of that gradient-filled color on elderly users’ vision. Cognition Cognitive skills deteriorate in elderly patients. The study shows that it is difficult for them to identify possible actions on a web page. For example, Fig. 4 shows the interface of the first web site prototype landing page. From our observations of the usability test performed, we found that patients encountered difficulties in navigating from the landing page, not knowing that the ‘Diabetes & Insulin’ image button is clickable. With refinement made to resolve this usability issue, the interface of the landing page of the second prototype was altered as shown in Fig. 5. The new interface includes a blinking text (‘Start From Here’) to inform users of where to click. This time the results show improvement with fewer patients reporting the issue of clicking on the button. However, the blinking text raised an issue of readability. In the third prototype as shown in Fig. 6, a static text (‘Click here to start’) replaces the blinking text. The results of the usability test show that patients had no difficulty clicking on the button. Additionally, in compliance with the ‘Senior-Friendly Usability Checklist,’ our findings show that elderly patients had difficulty to review or move forward from a page when buttons are not incorporated with text. Only one out of six patients in the first pilot testing iteration realized the

Univ Access Inf Soc

Fig. 4 Prototype 1—landing page of insulin decision aid web site

Fig. 5 Prototype 2—landing page of insulin decision aid web site

presence of the buttons when text was not incorporated in them, as shown in Fig. 7. Besides that, patients find it difficult to understand what follows when the buttons’ designs are not consistent as shown in Fig. 8. Patients misinterpreted the buttons when the designs were not consistent; thus, they encountered difficulty in grasping the navigation flow.

On the other hand, Fig. 9 shows the refined buttons with both text and consistency. With these buttons, the patients were able to navigate and make full use of the buttons without any usability problem. Other usability barriers associated with the decrease in cognitive skills seen in elderly patients were also identified. First, patients could not capture and process important

123

Univ Access Inf Soc

Fig. 6 Prototype 3—landing page of insulin decision aid web site

Fig. 7 Button without text incorporated

information when the webpage was too cluttered. Apart from that, it was observed that elderly patients were often confused if instructions were omitted from actions required of them. For instance, patients provided feedback in the retrospective think-aloud session indicating that they were unaware of the action to select the checkbox ‘Yes, I’m

123

concerned’ as shown in Fig. 9, although it is indeed one of their concerns. This usability barrier occurs when the checkbox is not incorporated with instructive text. We also found that apart from using text-relevant images and short segments to improve performance as stated in the literature, elderly patients are not familiar with media (i.e., auto-

Univ Access Inf Soc

Fig. 8 Button with no consistency

Fig. 9 Button with text and consistency incorporated

play slideshow). All patients continued to review forward the webpage after the first slide. They tended to miss out the rest of the important information in successive slides of the slideshow. Moreover, elderly users could not comprehend actionable icons such as the datepicker and

dynamically generated input fields. As a result, they failed to obtain risk communication results which were delivered based on their expressed value. Physical abilities The results of this study show that it is difficult for elderly patients to scroll down a web page or

123

Univ Access Inf Soc

move and click a mouse. They need a longer amount of time to coordinate their movements of the mouse and tend to be less smooth. Hence, the more movements required the more difficulties they encountered. Additionally, the elderly have a difficulty positioning the cursor of a mouse during interactions with smaller objects such as buttons, radio buttons, checkboxes and text fields. Both drop-down and mouseover menus are identified as barriers for elderly patients to obtain the information they seeking for. 5.1.2 Psychological factors Through the usability study, several psychological barriers were also identified that made elderly patients refrain from using the insulin decision aid web site. One of the significant barriers is presentation of information in a persuasive manner. For example, the section title ‘How do you start insulin?’ had a negative psychological impact on patients. They refrained from going through the section because the persuasive communication style led them to misinterpret that using the web site entailed they agreed to start insulin therapy instead of supporting them to make a decision. Besides that, the use of the word ‘already’ also acts as a barrier raised in this usability study. Patients tend to misinterpret ‘already’ as ‘all ready.’ Thus, they are more reluctant to continue to use the web site. It is important to take note of all of these web usability barriers to ensure elderly patients have sufficient support to make their decision. Table 5 summarizes the web usability barriers identified in this study. 5.2 Proposed usability guidelines Acknowledging the web usability barriers faced by elderly patients, certain usability guidelines were identified that can be useful resources for developers when developing insulin decision aids. We considered the usability of the decision aid domain by using the IPDAS framework as well, and all of the patients responded to it positively. Table 6 shows the proposed usability guidelines to improve the usability of the insulin decision aid web site. The effectiveness of our proposed usability guidelines for elderly patients using this insulin decision aid web site has been shown to increase its acceptability to reach thematic saturation after three iterations of the design–evaluate–redesign process.

6 Discussions The study has shown that collaborative work is needed in the process of eliminating web usability barriers in a specific context, in this case, insulin decision aid web sites

123

for the elderly. Involving the experts as well as the end users in creating the usability guidelines is crucial, given that such collaboration helps resolve any difficulties or differences between the two parties. This study helps to generate a set of usability guidelines that can lead to faster development times and greater user satisfaction. Furthermore, this set of usability guidelines is significant in a way that it differs from other existing guidelines in terms of its scope. Unlike the guidelines available at present, the proposed guidelines focus on addressing aging as well as psychological factors. We have achieved identifying usability attributes to realize the core values of the domain investigated. The resulting usability guidelines in this study complement the key requirements of an insulin decision aid which include avoiding bias toward the benefit of insulin and having a non-persuasive communication style. Indeed, the findings from this study will contribute to overcoming usability problems that elderly users may encounter in seeking online information and providing support to make an informed decision on insulin initiation. Our study is motivated and driven by the concept of domain-specific software engineering that enables developers to exploit common usability requirements for common applications that share common functionalities for common users. The set of guidelines is deemed to provide domain knowledge and promote reuse for the development of similar applications within the same domain. As such, developers can easily reuse the usability guidelines, taking advantage of the commonalities among applications for the insulin decision aid for elderly patients. Although there are abundant studies reporting on the elderly’s difficulties with web pages and providing elderlyspecific design guidelines, design guidelines targeted specifically for the context of insulin decision aids are yet to be available. Hence, these guidelines are deemed to be too vague or general, offering little in the way of specific guidance to designer when developing insulin decision aids for elderly people. The presentation of information for a web-based insulin decision aid should not only take into account the esthetics aspect but should also consider effective ways to design the web site. This shall ensure it aids the elderly to engage in a more informed decisionmaking process. This is acknowledged in the proposed guidelines under the criteria of psychological factor. Additionally, the proposed guidelines set out a few usability guidelines which are not included in the existing age-specific guidelines. There are some guidelines in the ‘aging factors’ section of Table 5 which are different from existing web design guidelines for the elderly. For example, existing guidelines [15, 18, 38] often address font size and font type to improve the readability. However, our proposed guidelines consider the font style as well as color as it is suggested that non-filled fonts should not be used to

Univ Access Inf Soc Table 5 Insulin Decision Aid Web site usability barrier for elderly patient Category

Web usability barrier

Aging factor Vision

• Font size of \20 pixels is less visible • Blinking text is distracting and difficult to read • Gradient colour reduces visibility • Duration of slideshow (3 s) is too short for reading • Non-filled text is not readable • White lettering on dark background reduces readability

Cognitive

• Inconsistency of content format with real-world practice challenge elderly patients’ working memory (e.g., patients tend to input date in dd/mm/yy format, and any difference to it will burden their working memory) • Important information is missed out when a page is too cluttered with lengthy text • Icons like ‘PDF’ and datepicker are not comprehensible • Inconsistency of interface challenges elderly patients to process new information and actions (e.g., navigation confusion occurs when buttons are not consistent) • Top navigation menu causes elderly patients to skip sections • Buttons without text incorporated are missed out

Physical abilities

• Elderly patients struggle to position the mouse cursor on smaller objects such as radio buttons • Mouse movements are less smooth (i.e., scrolling is a difficult action) • Filling multiple text fields burdens the elderly’s motor skills

Psychological factor

• Presentation of information in persuasive manner causes reluctance to use the web site • The word ‘already’ is perceived as ‘all ready,’ thus, misinterpretation of the purpose of the web site as to prompt insulin initiation occurs

increase visibility for the elderly. Apart from that, the proposed guidelines suggest that the duration for each slide is best to be longer than 7 s to improve readability: The present study has found that older adults have a difficulty reading the content of each slide if the slideshow is played below the mentioned duration. This is in contrast with existing guidelines which only incorporate other media such as images and video. Guidelines on the use of common web elements such as slideshow are not discussed in existing guidelines. With the advancement of web elements, it is important to set usability guidelines on the use of these new web elements to also cater for the needs of elderly users. The findings also show that elderly users are not familiar with the auto-play function of the slideshow. They tend to navigate to other sections of the webpage without realizing they miss out the content of the slides after the first. Thus, the proposed guidelines suggest using a play button instead of the auto-play feature for slideshows and avoiding autoloop of the slideshow that may cause confusion with regard to which the last slide is. Moreover, the elderly are not familiar with the use of the accordion menu. They tend to miss out the web content when the accordion menu is not collapsed. On the other hand, they accept content well when the graphics used are outlined. This design is not

addressed in existing guidelines [15, 18, 38] which only emphasize the graphic download time and text-relevancy. These usability guidelines are contextual as they ensure the important content needed for users to make informed insulin initiation decisions. The proposed usability guidelines are also considered an improvement over the existing guidelines which are not targeted toward the context of insulin decision aids. To improve the elderly’s memory, the proposed usability guidelines also address content format. We found that the ease of use of the insulin decision aid web site improves when the date format is ‘dd/mm/yyyy’ that matches with the real routine of noting down the date and time of the glucose test. Since PDAs are meant to provide sufficient information presented in a designated sequence for patients to make an informed health-care decision, the proposed guidelines address navigation structure to block users from jumping sections to prevent them from missing out any information. Such usability design is not highlighted in existing guidelines which are more general and only highlight the ease of navigation [15, 18, 38]. Addressing the cursor positioning of the mouse is difficult for older adults. The proposed guidelines suggest that auto-jump cursor for text fields improves usability for elderly users. This adds to the existing guidelines, which only specify the use of large

123

Univ Access Inf Soc Table 6 Usability guidelines for Insulin decision aid web site for elderly patient Criteria/justification

Web usability guideline

Aging factor Improve visibility and readability

• Use font size of at least 20 pixels for body text • Avoid blinking text. Use static texts instead • Use homogeneous solid color over gradient • Set duration of slideshow above 7 s for each slide • Avoid using non-filled texts • Use white lettering on light background

Mediate decreasing cognitive skills

• Use consistent format that match real-world practice (e.g., input date in dd/mm/yy format and input date field comes before blood sugar level, HbA1c input field) • Separate sentences or use bullet points to avoid lengthy text • Consider alternatives to replace less meaningful icons (e.g., use text fields instead of datepicker and replace download PDF file icon with text incorporated button) • Uphold design consistency throughout the web site (e.g., use consistent navigation button design to help elderly users recall) • Control navigations to avoid skipping important sections (e.g., step-by-step navigation path support users to understand what follows) • Incorporate buttons with text • Avoid using circle style type for bullet points to avoid confusion with radio buttons • Avoid auto-play slideshow. Use play button instead • Use instructive text to support user comprehension (e.g., use ‘Click here to start’ instead of ‘Start here’) • Avoid jargons and technical terms Provide text explanation if such terms are necessary

Reduce burden on motor skills

• Increase size of input fields for easy cursor positioning • Make clickable items (e.g., buttons) easy to target and hit • Use auto or parallax scrolling via ‘Previous’ and ‘Next’ buttons to reduce scrolling and mouse movements • Use consistent placing of navigation buttons to reduce mouse movements • Use auto-focus on multiple text fields to reduce mouse cursor movements and positioning • Present information in columns instead of in rows to reduce scrolling • Avoid elements requiring mouse movements (e.g., dropdown and mouseover menus) • Avoid dynamically generated input fields that result in more complex cognitive tasks (e.g., display all input options at once)

Psychological factor

• Phrase words in a non-persuasive manner (e.g., replace ‘How do you start insulin?’ with ‘Before making a decision, you may want to know how insulin is used’ and ‘Why do you need insulin?’ with ‘Why is there a need for insulin?’) • Use words that are not easily misinterpreted (e.g., use the word ‘now’ instead of ‘already’ which may be perceived as ‘all ready’) • Avoid use of ambiguous sentences (e.g., treatment option: ‘Make no change’ may be interpreted as ‘continue with current treatment or insulin does not help change for better’)

IPDAS Providing information about options

• Include all treatment options with equal details: Make no change (continue with current treatment) Follow stricter diet, exercise program and medication advice Start insulin injection Add another type of diabetes tablet Start another type of injection (non-insulin) Use alternative medicine (e.g., supplements, herbs and traditional medicine) • Provide information of the risks and benefits of all the treatment options in an unbiased manner

123

Univ Access Inf Soc Table 6 continued Criteria/justification Presenting probabilities

Web usability guideline • Support informed decision by providing risk communication indicating possible chances of having complications associated with patients’ blood sugar level to enable user to forecast what they are likely to experience • Use event rates specified by the population in periods based on patient’s latest HbA1c reading (e.g., number of people get and avoid complications out of 100 people within 10 years) • Use visual objects to illustrate probabilities (e.g., HbA1c value vs time chart and ‘smiley faces’ to depict numbers

Clarifying and expressing values

• Incorporate worksheets for patients to express their concerns and priority in managing their disease • Clarify the treatment options that are ideal based on their current treatment

Guiding/coaching in deliberation and communication

• Provide information step-by-step from information of treatment options to support and decision

Disclosure

• Provide funding information

Balancing the presentation of options

• Provide balanced information by allowing comparison of advantages and disadvantages between treatment options

Using plain language (readability)

• Include summary (e.g., printable.pdf file) to allow user to facilitate discussion of personal values with doctor and family members

• Show all treatment options regardless of the suggested treatment option based on personal value (not bias toward insulin treatment) • Use simple and clear language

Basing information on up-to-date scientific evidence

• Use content based on published evidence (e.g., IPDAS Collaboration Framework)

Establishing effectiveness

• Assessed in evaluation phase to show evidence that the web site supports patients in improving their knowledge about treatment options and making informed decision on insulin initiation

buttons to minimize precise mouse movements needed to activate them [15]. Another aspect that makes the proposed usability guidelines to stand out from existing guidelines for the elderly is that it focuses more on ‘how’ rather than ‘what’ information is to be presented to users, specifically older adults. This is how guidelines in the ‘IPDAS’ section of Table 5 differ from the IPDAS criteria. Although the IPDAS guideline provides a framework for the development of a PDA, it does not provide guidance on the usability guidelines to use; neither does it explicitly link the development of a PDA to its usability evaluation for older adults. For this reason, the set of usability guidelines described in this study is developed to complement the IPDAS guideline in developing PDA to improve its usability for the elderly. As such, the elderly will be able to take full advantage of the insulin decision aid web site with ease of use. This study has certain limitations. One is that the study is confined to elderly persons who are proficient in English. Hence, it is possible that some insightful feedback is compromised when many are ‘linguistically isolated’ because of low English proficiency. In the future, web sites should be fully translated to other languages to avoid isolation of end users. Given the majority of population in Malaysia is Malays (51%) [39] and diabetes is prevalent among Malays (23.8%) [40], the Malay language ought to be prioritized. Also, the study was limited to assessing a

small group of participants as this is only an initial nonstatistical study. A larger sample size could be used in follow-up studies in order to meet a valid level of significance to measure against the expected effects obtained from this pilot study. Future research should also include a comparison study between the proposed guidelines and other existing guidelines to further assess the usefulness of this proposed set of guidelines. In addition, this study can be extended to investigate the effect of the proposed guidelines on the accessibility and usability of the insulin aid web sites among elderly users, from the perspective of comparing web sites designed with and without implementation of the set of guidelines.

7 Conclusion It is important for web developers to understand well and integrate good web usability guidelines when developing an online insulin decision aid for the elderly. This is to assure that the web site meets the elderly needs and is able to help them to take full advantage of the online resources. Past research has addressed web usability to cope with the needs of older people; however, web usability for such a specific context is yet to be explored. This study has accomplished its goal to derive a comprehensive set of usability guidelines for designing and evaluating web sites in the context of insulin decision aid web sites for the

123

Univ Access Inf Soc

elderly. The guidelines were derived based on several iterations involving the following recursive steps: web site prototyping, usability testing, retrospective think-aloud, experts’ analysis. The results of this study draw the necessary web usability guidelines in order to accommodate vision, cognition and motor skills of older adult users. It is suggested that usability guidelines should not only cater to aging factors, but psychological factors as well, to help elderly patients make an informed decision. The usability guidelines improve user satisfaction and effectiveness of the insulin decision aid web site, as information is presented in a non-persuasive communication style, as neutrality is critical in insulin decision-making process. Their validity is backed by already published work and robust evaluations involving both end users and experts comprising of healthcare professionals, developers and designers. With the involvement of different and important parties accounting for both end users’ and experts’ perspectives, the comprehensibility of the proposed guidelines is assured. As such, this study is a significant step in eliminating usability barriers in the context of insulin decision aid web site for elderly diabetes patients. It can serve as a proactive resource for web developers to use in future web development, to reduce cost and web development time. Acknowledgements This research is funded by the University of Malaya Research Grant (UMRG) RP003E-13ICT ‘‘Mobile Health Monitor to Improve Healthcare for Diabetic Patients’’ and University of Malaya Research Grant (UMRG) RP015B-13HTM ‘Development and Pilot-testing Decision Making in Insulin Therapy (DMIT) Patient Decision Aid for the Internet.’

References 1. United Nations: Enhancing the voice of older persons. http://www. un.org/en/development/desa/news/social/older-persons.html (2011). Accessed 20 Dec 2015 2. World Health Organization: 1 October: International Day of Older Persons Older people: a new power for development. http:// www.who.int/ageing/events/idop_rationale/en/index.html (2011). Accessed 20 Dec 2015 3. Zaharias, P., Koutsabasis, P.: Heuristic evaluation of e-learning: comparing two heuristic sets. J. Campus Wide Inf. Syst. 29(1), 45–60 (2011) 4. Barber, J., Bhatta, S., Goel, A., Jacobson, M., Pearce, M., Penberthy, L., Shankar, M., Simpson, R., Stroulia, E.: AskJef: integration of case-based and multimedia technologies for interface design support. In: Artificial Intelligence in Design’92, pp. 457–475. Springer (1992) 5. Henninger, S.: A methodology and tools for applying contextspecific usability guidelines to interface design. Interact. Comput. 12(3), 225–243 (2000) 6. Lewis, C., Olson, G.: Can principles of cognition lower the barriers to programming? In: Empirical Studies of Programmers: Second Workshop, pp. 248–263. Ablex Publishing Corp (1987)

123

7. Domeshek, E. A., Kolodner, J. L.: A case-based design aid for architecture. In: Gero J. (ed.) Artificial Intelligence in Design 92’’ (1992) 8. Taylor, R.N., Medvidovic, N., Dashofy, E.M.: Software Architecture: Foundations, Theory and Practice. Wiley Publishing, Chichester (2009) 9. Holmes-Rovner, M., Nelson, W.L., Pignone, M., Elwyn, G., Rovner, D.R., O’Connor, A.M., Coulter, A., Correa-de-Araujo, R.: Are patient decision aids the best way to improve clinical decision making? Report of the IPDAS Symposium. Med Decis Mak 27, 599–608 (2007) 10. Ng, C.J., Lee, Y.K., Lee, P.Y., Abdullah, K.L.: Health innovations in patient decision support: bridging the gaps and challenges. Aust. Med. J. 6(2), 95 (2013) 11. Hedden, T., Gabrieli, J.D.: Insights into the ageing mind: a view from cognitive neuroscience. Nat. Rev. Neurosci. 5(2), 87–96 (2004) 12. Westphal, S.A., Palumbo, P.J.: A case for introducing insulin early in the treatment of type 2 diabetes mellitus. Insulin 1(2), 65–69 (2006) 13. Zainuddin, N.A., Zulkarnain, A.K., Tahir, A.: Psychological insulin resistance (PIR) among type 2 diabetes patients at public health clinics in federal territory of Malaysia. Int. Med. J. Malays. 10(2), 7–12 (2011) 14. Lee, P. Y., Khoo, E. M., Low, W. Y., Lee, Y. K., Abdullah, K. L., Azmi, S. A., Ng, C. J.: Mismatch between healthcare professionals’ and patients’ views on a diabetes patient decision aid: a qualitative study. Health Expect (2015) 15. Becker, S.A.: A study of web usability for older adults seeking online health resources. ACM T Comput. Hum. Int. (TOCHI) 11(4), 387–406 (2004) 16. Korytkowski, M.: When oral agents fail: practical barriers to starting insulin. Int. J. Obes. Relat. Metab. Disord. 26, S18–S24 (2002) 17. Lee, Y.K., Lee, P.K., Ng, C.J.: A qualitative study on healthcare professionals’ perceived barriers to insulin initiation in a multiethnic population. BMC Fam. Pract. 13(1), 28 (2012) 18. Hodes, R.J., Lindberg, D.A.: Making Your Web site Senior Friendly. National Institute on Aging and the National Library of Medicine, Washington (2002) 19. Arch, A.: Web accessibility for older users: a literature review (2010) 20. Echt, K.V., Morrell, R.W.: Designing web-based health information for older adults: visual considerations and design directives. Older Adults Health Inf. World Wide Web, 61–87 (2002) 21. Caserta, R.J., Abrams, L.: The relevance of situation awareness in older adults’ cognitive functioning: a review. EUR Rev. Aging Phys. A 4(1), 3–13 (2007) 22. ISO/IEC: 9241–14 Ergonomic requirements for office work with visual display terminals (VDT)s—part 14 Menu dialogues, ISO/ IEC 9241-14: 1998 (E) (1998) 23. Elwyn, G., O’Connor, A.M., Bennett, C., Newcombe, R.G., Politi, M., Durand, M.A., Drake, E., Joseph-Williams, N., Khangura, S., Saarimaki, A., Sivell, S.: Assessing the quality of decision support technologies using the international patient decision aid standards instrument (IPDASi). PLoS One 4(3), e4705 (2009) 24. Ng, C.J., Lee, P.Y., Lee, Y.K., Khoo, E.M., Chen, W.S., Low, W.Y., Abdullah, K.L., Samad, A.A., Azmi, S.A.: Should I Start Insulin? (English), 2nd edn. University of Malaya, Kuala Lumpur (2012) 25. Yahaya, A., Yahaya, N., Ismail, S.: Factors contributing to proficiency in English as a second language among Chinese students in Johor Bahru. Elixir Psychol. 41, 5837–5848 (2011) 26. Hanapiah, M. F.: English language and the language of development: a Malaysian perspective. In: International Conference IPBA, Lembah Pantai, Kuala Lumpur (2002)

Univ Access Inf Soc 27. Ng, C.J., Lee, Y.K., Lee, P.Y., Abdullah, K.L.: Health innovations in patient decision support: bridging the gaps and challenges. Australas. Med. J. 6(2), 95–99 (2013) 28. Lee, P. Y., Khoo, E. M., Low, W. Y., Lee, Y. K., Abdullah, K. L., Azmi, S. A., Ng, C. J.: Mismatch between health-care professionals’ and patients’ views on a diabetes patient decision aid: a qualitative study. Health Expect. (2015) 29. Feldman-Stewart, D., O’Brien, M.A., Clayman, M.L., Davison, B.J., Jimbo, M., Labrecque, M., Martin, R.W., Shepherd, H.: Providing information about options in patient decision aids. BMC Med. Inform. Decis. Mak. 13(Suppl 2), S4 (2013) 30. Ng, C.J., Mathers, N., Bradley, A., Colwell, B.: A ‘combined framework’ approach to developing a patient decision aid: the PANDAs model. BMC Health Serv. Res. 14(1), 1 (2014) 31. Goodman, J., Brewster, S., Gray, P.: Older people, mobile devices and navigation. In: HCI and the Older Population, pp. 13–14 (2004) 32. Strauss, A.L., Corbin, J.M.: Grounded Theory in Practice, 2nd edn. Sage, Thousand Oaks (1998) 33. Silverman, D.: Doing Qualitative Research: A Practical Handbook. Sage, Thousand Oaks (2000) 34. Strong, G. W., Gasen, J. B., Hewett, T., Hix, D., Morris, J., Muller, M. J., Novick, D. G.: New Directions in Human-Computer Interaction Education. Research, and Practice, report sponsored by NSF and ARPA (1994)

35. Aehnelt, M., Peter, C., Mu¨sebeck, P.: A discussion of using mental models in assistive environments. In: Proceedings of the 5th International Conference on Pervasive Technologies Related to Assistive Environments, p. 37. ACM (2012) 36. Gregor, P., Newell, A. F., Zajicek, M.: Designing for dynamic diversity: interfaces for older people. In: Proceedings of the fifth international ACM conference on Assistive technologies, pp. 151–156. ACM (2002) 37. Zajicek, M., Brewster, S.: Design principles to support older adults. Univ. Access Inf. Soc. 3(2), 111–113 (2004) 38. Kurniawan, S., Zaphiris, P.: Research-derived web design guidelines for older people. In: Proceedings of the 7th international ACM SIGACCESS conference on Computers and accessibility, pp. 129–135. ACM (2005) 39. Department of Statistics Malaysia: Department of Statistics 2000, Putrajaya (2000) 40. Letchuman, G.R., Wan Nazaimoon, W.M., Wan Mohamad, W.B., Chandran, L.R., Tee, G.H., Jamaiyah, H., Isa, M.R., Population Distribution and Basic Demographic Characteristics, The 2000 Population and Housing Census of Malaysia, Zanariah, H., Fatanah, I., Ahmad Faudzi, Y.: Prevalence of diabetes in the malaysian national health morbidity survey III 2006. Med. J. Malays. 65(3), 180–186 (2010)

123