Mobile health (mHealth) plays a vital role in delivering health education and disease management advice and in monitoring patients’ health in many ways [1,2]. mHealth development is beneficial not only for populations with diseases but also for active individuals, athletes, and older adults [3-6]. The 2 main factors that influence the effective use of mHealth apps for self-care are perceived usefulness and ease of use [7]. Well-designed mHealth apps have proved their cost-effectiveness by educating and empowering patients as well as by improving the medication adherence of patients [8-10]. However, reviews have shown that there is insufficient evidence for the quality of mHealth apps.

The target users of mHealth apps can be categorized into either patients or health care providers. mHealth apps for patients are specifically designed for those who intend to manage their health, for example, health behaviors, while mHealth apps for health care providers are developed for delivering health care services, for example, medication prescription, laboratory orders, consultation, and patient education [11]. Some terms need to be customizable in the usability questionnaire to represent specific target users of mHealth devices. In addition to categorization according to target users, mHealth apps can be categorized according to interactive functions, that is, interactive mHealth app or standalone mHealth app [11]. Interactive mHealth apps enable users to send and receive information from their health care providers. Patient-provider communication can be synchronous or asynchronous. Standalone mHealth apps contain reminders or progress charts that collect health or activity information; however, the data are not shared with the health care providers immediately. Therefore, there are limited interactions in standalone mHealth apps [11].

Any mobile app should be designed with good usability or the app should be easy to use without errors and be able to achieve its objectives effectively. The characteristics of good usability are (1) efficiency: the comprehensiveness of a product enables users to achieve their goals, (2) satisfaction: users have positive feedback about the product, (3) learnability: the product is easy to learn, (4) memorability: the usage mechanism of the product or the system is easy to remember even after the users have not been using it for a while, and (5) low error rate: low errors can prevent disasters [12,13]. However, the above usability characteristics are not specific to mHealth apps because of several phone issues, for example, small screen or screen overload [14].

A new usability model has been proposed to enhance the framework within the context of mobile apps and it is called PACMAD (People At the Centre of Mobile App Development) [15]. This model identifies 3 factors that should be considered when evaluating usability: (1) user, (2) task, and (3) context of use that can affect the overall usability of a mobile app [15]. The added element of “cognitive load” in PACMAD assesses the user ability to perform additional tasks, for example, exercise while using the mobile device [15]. Therefore, usability should be based on effectiveness, efficiency, satisfaction, learnability, memorability, and cognitive load with very minimal errors [15].

Several usability scales have been evaluated based on the importance of usability in developing mobile apps. The mobile app rating scale and the user version of the mobile app rating scale include some usability components to assess the quality of mobile apps [16,17]. However, they does not assess the usability by the end user and they have a broader scope [16,17]. Taking this into consideration, mobile apps that are intended for health purposes (mHealth apps) should be assessed with a more appropriate usability questionnaire. Therefore, a new usability scale called the mHealth App Usability Questionnaire (MAUQ) was developed [18]. This MAUQ can assess the ease of use, interface, satisfaction, and the usefulness of mHealth apps to the end users (either patients or health care providers) and the type of interaction between the patients and health care providers (standalone or interactive).

To the best of our knowledge, only 1 study has reported the usability of a Malay-translated mobile app [19]. However, the questionnaire used was the system usability scale questionnaire, which caters to general software systems. As more health interventions using mobile apps are being developed, this study aims to translate and validate a Malay version of MAUQ for future mHealth app research and usage.

The content validity index for the relevancy (Table 1) and clarity (Table 2) of M-MAUQ was calculated to be 0.983 and 0.944, respectively. Meanwhile, the face validity index (Table 3) for understandability was calculated to be 0.961. The content validity index and face validity index scores above 0.79 indicate that the items in the questionnaire are relevant for domain, clarity, and understandability for the target user [26]. The modified kappa agreement for every item in M-MAUQ demonstrated excellent agreement (κ=0.76-1.09).

Table 2

Content validity index of item clarity and modified kappa agreement by 10 experts.

Of the 59 participants, 51 (the calculated sample size was 48) signed the web-based consent form and completed the questionnaire for the reliability study. The age of the participants ranged from 22 years to 25 years, and all of them were students (females, 47/51, 92%). The mean usability score for MyFitnessPal using M-MAUQ was 0.802 (SD 0.104), which indicated good usability. The Cronbach α values for total, subscale ease of use (Q1-Q5), subscale interface and satisfaction (Q6-Q12), and subscale usefulness (Q13-Q18) were .946, .893, .942, and .742, respectively. The value of Cronbach α if an item was deleted (refer to Table 4) remained highly consistent without significant difference, indicating good internal reliability of the developed questionnaire. The final M-MAUQ can be viewed in Multimedia Appendix 1.

This study described the translation, adaptation, and validation of the English version of MAUQ into the Malay version. The results of our study revealed a high level of content validity and face validity for M-MAUQ (standalone for patients). The content validity index was high for all individual items (item-level content validity index >0.900) and for overall score (scale content validity index >0.944) in term of relevancy and clarity, which exceeded the recommended benchmark of 0.79 [21]. Moreover, our kappa statistic indicated excellent agreement between expert raters for content validity and among user raters for face validity (κ>0.77). This study also had high reliability for the total domains of M-MAUQ, with Cronbach α>.90.

A questionnaire is one of the well-known methods for usability testing [27], but developing a new one might require concerted effort by the members of a research team, extra cost, and a lot of time [19]. Thus, adaptation of established, appropriate, and available questionnaires with documented validity in other languages is recommended [19]. Appropriate translation and adaptation of an instrument are essential for ensuring the equivalence between the original and translated versions translated according to the accepted standards and at the same time being culturally and conceptually appropriate [28]. Literal translation might cause misunderstanding of the questions posed since it does not anticipate the cultural sensitivity or the cultural influence of the question. Regardless of the presence of a multi-ethnic community in Malaysia, Bahasa Malaysia or Malay is the national language, which justifies validating a Malay-translated version for MAUQ [29]. In this study, we translated, culturally adapted, and validated an 18-item questionnaire called the M-MAUQ. This study adheres to the comprehensive and detailed guidelines for the translation, adaptation, and validation process documented by Sousa and Rojjanasrirat [20], which was adopted by Marzuki et al [19]. Translating some technical words such as navigation and interphase were challenging since these words were more familiar in English than in Malay. Moreover, a direct translation of the words of “social setting” from the questionnaire might raise questions from the user regarding the meaning. Thus, the involvement of certified translators who are experts in the translated and original language is essential. Harmonization among experts is essential to ensure that the user has understood the sentence correctly. A few examples were added to make the questions more clear and discussions were held with the original authors of MAUQ to ensure the consistency of each item in both languages.

Validation is an essential process for assuring that the measures in the translated version of the questionnaire are equivalent to those of the construct of the original version. Numerous studies have established content validity by using content validity index to quantify the validity of the questionnaire [27]. Content validity index is used primarily by researchers because it is simple to measure, is understandable, provides details for each item, and can be used to modify or delete instrument items [21]. Meanwhile, face validity focused more on the appearance and the understandability of the questionnaire by the target user. The validation process involved both experts and target users to ensure that different opinions from the 2 groups could be collected and discussed, and a consensus could be reached [21]. Modified kappa informed the agreement properties among the raters [30], and high scores in both content validity index and face validity index suggest that M-MAUQ had been translated accurately and adopted appropriately by our local users.

The internal consistency of the questionnaire is one of the reliability components. Internal consistency was determined by calculating the Cronbach α coefficient that represents the extent to which the items are measuring the same things [31,32]. The results of the reliability of M-MAUQ were comparable to those of the English MAUQ, which was tested among 128 participants in the Greater Pittsburgh area and the Cronbach α coefficients for the 3 subscales ranged from .717 to .908 [11].

This study utilized the web-based data collection method owing to the current COVID-19 pandemic, wherein mass gathering is not encouraged. Participants were gathered through virtual meetings using the Google meet platform. Although virtual meetings offer few advantages such as flexible time schedules and flexibility in attending the meeting from each one’s own comfort zone, some limitations were identified. Technical issues such as sudden internet breakdown or slowdown could have led to communication breakdown. However, participants who had problems with internet connection were contacted later in separate virtual meetings. The other limitations were as follows. First, the convenient sampling method was used, wherein participants were recruited from 1 university in the state of Pahang, Malaysia, thereby making it difficult to generalize the findings of this study to a wider population in Malaysia. Second, this research only focused on healthy young adults and the findings may not be generalizable to older adults or anyone with health-related issues. Third, additional reliability steps of test-retest would be beneficial for measuring the stability of respondent attributes.

In conclusion, our study shows that M-MAUQ has excellent validity and reliability. M-MAUQ can be used to assess the usability of mHealth apps among healthy young adults who are native Malay speakers. We suggest conducting a criterion validation study for future research [33,34] as this would be able to predict an outcome for another measure or domain of the M-MAUQ. Moreover, this questionnaire can be added in the future usability rating scale, as suggested in the Good Practice Guidelines on Health Apps and Smart Devices [35].