This study consists of 2 steps. In step 1, the SUS was translated and adapted into Swedish. In step 2, the psychometric properties of the Swedish version were tested: data quality, scaling assumptions, acceptability, convergent validity, and internal consistency. The translation, adaptation process, and psychometric evaluation adhered to the COSMIN (Consensus-Based Standards for the Selection of Health Status Measurement Instruments) checklist [31].

The SUS instrument consists of 10 items (statements). The items are divided equally into positively and negatively worded statements. Requests for a response are graded on a scale of agreement ranging from strongly disagree (1) to strongly agree (5). The score is calculated as follows: for positively worded items (1, 3, 5, 7, and 9), the score is the position on the scale (1-5) minus 1, and for negatively worded items (2, 4, 6, 8, and 10), the score is the position on the scale minus 5. Individual item scores, therefore, can range from 0‐4. The sum of all 10 scores is then multiplied by 2.5, resulting in a total score ranging from 0‐100, with higher numbers representing greater usability [7,8].

To date, factor analysis has not been able to show conclusively whether the SUS consists of one factor (usability) or two (usability and learnability) [11]. Efforts to replicate these findings have led to the conclusion that addressing the instrument as 2 dimensional has no practical or theoretical interest. This study, therefore, treats the SUS as a unidimensional instrument of perceived usability [32].

According to Brooke’s [7] original formulation of the SUS, no formal permission is needed for translation, and it can be used free of charge. Further, Brooke [7] allows for the possibility to, in any version, exchange the wording of the scale to a word or expression suitable for the situation. The translation process adopted in this study was inspired by Beaton et al [33]. First, the original instrument was translated from English to Swedish by one of the researchers (CC), proceeding from the versions presented by Bangor et al [34] and Lewis [10]. During the process, perceived difficulties and uncertainties were noted. This first version was then reviewed and discussed by the group, consisting of researchers familiar with eHealth, until consensus was reached. Second, an authorized translator, naive to the research field, carried out a back translation on the instrument and the notes taken during the research process. Finally, the research group reviewed all versions of the translation, with all notes attached, and finalized a second version of a Swedish SUS. This version was then compared to other SUS translations within the Scandinavian countries for content validity and the original version for expressions and conceptions that could have been culturally influenced. Following this last step, the group then decided upon a final version by consensus. The response options were structured in the same way as the original questionnaire, although the phrases “strongly disagree” and “strongly agree” were exchanged for the Swedish equivalents of “totally disagree” and “totally agree.”

This study was conducted in accordance with the Declaration of Helsinki [42] and approved by the Swedish Ethical Review Authority (2021‐05077). The invited parents were recruited through gatekeepers at the pediatric department. Information regarding the study was initially provided orally and then followed up by written information, before written informed consent was obtained by a study nurse. Data were handled confidentially, and participating parents were able to quit without any explanation or impact on the care their children received. Participation in the study was voluntary, and no financial or other form of compensation was provided to the participants.

Both translations (forward and backward) were similar and did not differ substantially. Overall, the conformity of phrasing was high, with some discrepancy for phrasing in statements 2 (I found the system unnecessarily complex) and 8 (I found the system very cumbersome to use). In statement 2, the word “complicated” was suggested in the back translation from the Swedish word “komplex.” In statement 8, the word “awkward” was suggested instead of the original word “cumbersome.”

All such discrepancies were discussed within the research group and a consensus was reached for the final version. The final step, comparing the Swedish version to other Scandinavian versions, generated no further changes. Table 1 shows the original English version and the final Swedish version of the SUS.

A total of 62 individuals were included in this study. Of these, 20 answered the questionnaire, including the SUS, twice. This resulted in a total of 81 measurements included in the analysis. Demographic data for the participants are shown in Table 2.

This study presents a new Swedish version of the SUS that is psychometrically tested. This study seeks to establish the new Swedish version of the SUS as a reliable and valid instrument for assessing system usability. Overall, the psychometric testing showed high data quality, good scaling assumptions, high internal consistency, and fair convergent validity. Together, these analyses support the validity and reliability of the new Swedish version of the SUS.

The translation process of the new Swedish version of the SUS was executed incrementally, both within and outside the research team, involving an authorized translator who conducted a back translation. This method facilitated thorough scrutiny and comparison of the translation from multiple perspectives. The approach also helped reduce the risk of bias and improved the scale’s validity by incorporating multiple viewpoints [43].

The psychometric testing regarding internal consistency showed that the Cronbach α values were satisfactory and that all items contributed to the instrument’s total score. This indicates that the Swedish translation of the SUS is a stable instrument to use in a Swedish context.

Convergent validity showed fair correlation between the SUS and domain 5 of the eHLQ, which measured the motivation to engage with digital services (I find that digital technology support me in taking care of my health). This supported our hypothesis that a generally positive attitude toward digital solutions would correlate with a high usability score. This hypothesis is also in line with previous research, which shows that positive expectations directed toward a product generate positive subjective usability ratings [14].

There was an indication of a ceiling effect for 3 items on the Swedish version of the SUS (items 4, 8, and 10). They presented above 70% in item frequency distribution and a high mean total score, which is in line with studies indicating a generally high score for the SUS [10]. Previous studies have also shown ceiling effects of the SUS but only for the total score [15,44], thus the responses on the different items can not be justly compared between studies. This study, however, explores the SUS in much more detail and in line with recommended psychometric evaluations, as it explores the psychometric properties for each item separately [12,37]. Further, the distribution was excessively skewed. This could be an expected result of the nonnormal distribution, since negatively skewed variables are assumed to have a ceiling effect [37,45]. Also, the robustness of skewed distribution as the sole indicator of ceiling effect has been questioned [45].

However, based on the skewness result of the items, the question arises as to whether the range of the response options is wide enough (5-point scale) and if the wording in the response options (“strongly disagree” to “strongly agree”) is a sufficient description. It could be interesting to expand the number of response options but, arguably, there could be obstacles with revising such a widely used and widespread instrument.

Regarding the high mean scores for both the total sum and the items of the SUS, it should be noted that the population had a mean age of 33 years and are, therefore, used to digital solutions. This could have influenced their experience of the app. Therefore, the high item scores of the SUS could also indicate a product that is perceived to have usability for this group. A recommendation for future studies would be to test the instrument in other contexts and with different age groups. This app was designed to support users who were in exactly the same situation as the participants in this study: parents in a specific situation. Going home with a child after hospitalization can be stressful; the parents in this study reported this eHealth solution as supportive [46]. The SUS has previously been tested for its sensitivity in different digital solutions and has shown a high scoring (ceiling effect) within the best-of-class products [44].

The Swedish government has envisioned Sweden as a leading country of eHealth in the near future. It has declared the need for individual users to act as the cocreators of such solutions [30]. Within eHealth, there has already been a call for a participatory design to increase the equality of access to digital solutions [47]. Even though digital technology and eHealth are intended to enhance access to health care, for example, despite geographical distances. It may also result in the opposite, as people have different knowledge in using digital solutions. For example, younger people tend to use and access digital solutions differently than older people, which is why solutions should be adapted to the intended end users [48]. In this study, the population was younger, as previously discussed, which could have influenced the result. To have a project design of cocreation often means that end users are involved from the start of the project as collaborators. Coproduction, on the other hand, often involves the end users during the implementation phase [49]. Arguably, the end users are invaluable in all steps of the process, and the usability of products needs to be evaluated where instruments such as the SUS can be useful [50]. This study could therefore be regarded as enhancing the possible participation of end users in the future development of digital solutions and eHealth in a Swedish context.

In conclusion, this study indicates that this developed and psychometrically tested Swedish version of the SUS can be recommended for use within the Swedish adult context.

There are some methodological challenges in this study. First, the sample could be regarded as small (N=62), although it was sufficient according to the recommendations [31,33]. A larger sample might enable different methods of analysis, such as Rasch analysis, which might offer deeper psychometric insights.

Second, the 1-year interval for data collection could be considered long in a rapidly changing world. This length of time could influence how the app was perceived. That said, the digital product was not changed to any great extent during this period, so the perception of usability should not have been influenced to any great extent. Additional data collection was carried out after COVID-19 restrictions were essentially withdrawn in Sweden. There is, however, always the possibility of perceptions fluctuating, as the product could be used in various ways according to the circumstances of the parents.

Third, there was an intention to pursue a data quality test for reproducibility as some of the participants answered the SUS twice. Regrettably, the test-retest sample (n=20) was regarded as too small in this study, and such analyses need to be considered in future studies. In addition, the use of imputation can be discussed. There was a low number of missing data, with a noncomplete frequency of only 9 responses in total. The percentage of missing data per item was also low across all items. This resulted in only 6 imputed scores across the items, and based on mean score for the item, this was regarded as not influencing the result noticeably.

This study presents a new Swedish version of the SUS. It is the first study to carry out a psychometric evaluation of a Swedish SUS, to establish the Swedish version of the SUS as a reliable and valid instrument for assessing system usability. Overall, the psychometric testing showed high data quality, good scaling assumptions, high internal consistency, and fair convergent validity, all of which support the validity and reliability of the new Swedish version. The results from this study are promising. They raise the possibility that the Swedish version of the SUS could be used to evaluate digital health solutions. To further strengthen the usability of the scale, we suggest additional analysis on data that evaluate more complex eHealth technology and include a wider participant age group, both younger and older.