JMIR Publications

ABSTRACT

Background:

In the current digital landscape, ensuring optimal usability is one of the most crucial factors determining the success of any mobile application. Questionnaire-based usability evaluations represent a highly prevalent methodology for this purpose. To date, questionnaires have been developed to assess the general system usability, however, there are hardly any questionnaires specifically designed to assess the usability of mHealth apps. The most widespread, mHealth App Usability Questionnaire (MAUQ), has been developed in four versions, according to the type of app (interactive or standalone) and the target user (patient or provider).

Objective:

The objective of this study was to translate and validate the English version of the MAUQ (standalone, for patients) into a Spanish version (S-MAUQ)

Methods:

The methodology employed here follows that proposed by Sousa and Rojjanasrirat, which comprises four stages. The initial stage of the process entails a translation, harmonisation and adaptation procedure, resulting in the production of a pre-final S-MAUQ. The second and third steps entailed content validation (by 10 experts) and face validation (by 12 target users), which were conducted to evaluate the relevance and clarity of the questionnaire items, respectively. The I-CVI, S-CVI, I-FVI and S-FVI, as well as the modified kappa statistic, κ, were employed to evaluate interrater agreement among the raters, taking into account the probability of agreement by chance, Pc. It is also important to correctly interpret the formula used to calculate the probability of agreement (Pc), as this has been incorrectly applied in numerous similar studies, leading to erroneous results. In the final stage of the process, the reliability of the S-MAUQ was evaluated by a sample of 61 participants. These individuals were required to install the Yazio application, utilise it to complete the specified task sheet, and respond to the S-MAUQ. Subsequently, the Cronbach α value for the entire questionnaire and its three subscales was calculated. A value of 0.70 or above is deemed acceptable, indicating good internal consistency.

Results:

The S-CVI was calculated to be 0.778. However, two items demonstrated κ values below 0.4, indicating poor agreement, while two others exhibited κ values between 0.4 and 0.6, suggesting fair agreement. This suggests the necessity for further refinement of the questionnaire. However, following the second round of analysis, two items (#14 and #15) were identified as exhibiting a markedly low level of relevance and clarity, accompanied by a similarly low κ (0.4≤κ≤0.6, interpreted as "fair"). The results and subsequent analysis with the panel of experts highlighted the necessity to eliminate both items from the pre-final Spanish questionnaire (S-MAUQ’). Ultimately, the content validity index of the general scale (S-CVI) was determined to be 0.881. The face validity index for the final S-MAUQ was 0.927, indicating that the items are clear and understandable for the non-expert target user. Furthermore, the modified kappa statistic for each item in S-MAUQ demonstrated excellent agreement (κ>0.74), supporting the construct validity of the instrument. The Cronbach α value for the entire questionnaire and for its three subscales was 0.87 and 0.74, 0.74 and 0.64, respectively. These results indicate that the S-MAUQ exhibited good internal reliability.

Conclusions:

Two items (#14 and #15) exhibited a markedly low level of relevance and clarity and must therefore be eliminated from the Spanish questionnaire. The final S-MAUQ meets the validation criteria, demonstrating reliability and validity comparable to that of the original version. Consequently, the S-MAUQ is suitable for assessing the usability of mHealth applications in Spanish in future studies.