Transparency, including the possibility for patients to gain insight into one’s own medical information, is increasingly called for in health care, especially, when it comes to patients’ access to their electronic health records (EHRs) [1]. Many countries (eg, Finland, France, Norway, Australia, Denmark, Canada, United Kingdom, and Sweden) have, in the past, implemented or are now currently implementing patient accessible electronic health records (PAEHRs) [2]. In some countries, these are local implementations at a specific hospital or region, whereas others have national solutions. Differences in strategies and approaches have affected the uptake and impact, and the implementation progress has, in several countries, been slow due to legal constraints [3,4] and concerns about security and privacy among health care professionals [5-7]. A PAEHR often includes the clinical notes written by different health care professionals, as well as other parts of the EHR (eg, lab results, referrals, and medications).

In the United States, the OpenNotes initiative focuses on providing patients access to their notes, specifically. OpenNotes began as a pilot evaluation project that included 105 volunteer primary care physicians and their 19,000 patients [8,9]. The initiative started in 2010 and has since spread throughout the United States [10]. On April 5, 2021, a new federal rule required US health care providers to allow patients access to all the health information in their EHR [11,12]. This new rule mandates rapid, full access to test results, medication lists, referral information, and clinical notes in electronic formats, by request.

In Sweden, the e-service Journalen is a national PAEHR, accessible online via the national patient portal called 1177.se [13]. The PAEHR service accesses the EHR information from most of the various EHR systems used throughout Swedish health care organizations, via a national health information exchange platform [14,15]. Hence, patients have one access point for all their health record information regardless of (1) how many health care providers they have visited and (2) which EHR system their health care providers use [13]. Since the first Swedish region began providing their inhabitants online access to their health records in 2012, all the other regions have connected to the national infrastructure and the PAEHR Journalen. This was not the case from the beginning, though, and the last of the 21 regions connected only in April 2018. In addition, different regions made different choices about how much of their information would be made available to patients; for example, patients receiving care in one region could gain access to both their lab results and notes, whereas patients receiving care in a different region might only be able to access the notes [13,16].

A growing literature on patients’ experiences of accessing their records online reports positive outcomes [10,16]. Patients who read their notes have reported a better understanding of their care plans [10], a sense of greater control over their care [10,16], an improved adherence to medication [17], improved communication with and trust in their clinicians [16,17], and a sense that their care is safer [18].

Despite these benefits, adoption and use can be low [19], and several studies have explored factors that influence adoption [20,21]. User characteristics and perceived benefits of using a PAEHR might influence behavioral intention for use and adoption, but poor usability that increases the effort expectancy can also have an impact. It is, therefore, of interest to further explore how users of the PAEHR Journalen perceive its usability and usefulness; the latter is especially important, considering the differences in the levels of transparency regarding patients’ health information in different regions.

In this study, we analyzed data on usability issues from a national survey conducted among patients who use the PAEHR Journalen. A first analysis of the main results from the survey was published in 2018 [16] and contains an overview of the full survey. Here, we focused only on the usability-related questions as well as some demographic data of the participants. At the time of the study (June to October 2016), not all regions allowed patients access to their records through Journalen, and, among those who did, the level of transparency of this information varied [16]. Table 1 represents an overview of what types of clinical content the health care providers had chosen to allow access to at the time of data collection for this study.

The aim of this study was to explore how the users of the Swedish PAEHR experience the usability of the system and to identify differences in these experiences based on the level of transparency regarding patients’ health information for the region.

During the survey period, 423,141 users logged into Journalen, of which 2587 patients completed the survey (of unique users that logged in, response rate 0.61%). Of all respondents, 62.97% (1629/2587) identified as women and 30.85% (798/2587) as men; 0.39% (10/2587) of respondents chose “other,” and 5.80% (150/2587) did not answer this question. According to use statistics provided by Inera AB (the company providing Journalen and the national patient portal [25]), this reflects the gender distribution of the users in general (in 2016, 60% women and 40% men). Of all respondents, 39.81% (1030/2587) stated that they were working or had been working within health care, and 54.54% (1411/2587) stated that they had no professional relation to health care; 5.64% (146/2587) of respondents did not answer this question. Participants had a higher education level than the general population [16]. Among our participants, 60.57% (1487/2455) had higher education, whereas only 42% of the general Swedish population does [26]. We cannot tell whether this is because users of Journalen are well educated or that people with a higher education represent a subgroup of users who are more inclined to answer a survey. Unfortunately, no data on the general education levels of Journalen users exist.

To sum up, the survey results regarding user characteristics on a national level indicate that most participants were women and that the majority had studied at least 3 years of higher education. In addition, results indicate that many users of Journalen were both patients and medical professionals, at various points in their lives.

In Moll and colleagues’ [16] overview of the survey results, details of the participants’ views of the usefulness and benefits of accessing their health records online are presented in more detail. Overall, patients who answered the survey were positive toward Journalen (Table 4). Participants were asked to rate on a 5-point Likert scale to what extent they agreed to the more general statements, “I think that access to one’s medical records online is generally a good reform,” and “I think that access to Journalen is good for me.” Of all participants, >96% (2454/2541, 96.58% and 2455/2528, 97.11%, for the respective questions) had a positive attitude toward Journalen, answering with either “completely agree” or “partly agree.”

However, a positive attitude toward accessing one’s health records does not say much about the usability of the system, and, therefore, we also present the results of the SUS analysis. Results of the analysis of the SUS questions are first described on a national level. Table 5 presents the results of all participants for the SUS items in the survey, including neutral responses replacing missing answers for participants who responded to at least 1 SUS item.

For all participants, the total mean score for the SUS scale was 79.81 (SD 14.25), which would, according to Bangor and colleagues [24], qualify as a successful system. The median score was 82.5, and the distribution of individual answers is plotted in Figure 1. The scores covered the entire range, from 0 (1 person) to 100 (158 people), but the majority of individuals scored above 60.

Each item can have a score contribution between 0 and 4. Most of the items scored above 3; however, 3 questions stood out with score contributions below 3 (Table 6). All 3 questions that score below 3 related to the complexity of the system and whether functions are well integrated in the system.

To further explore whether the level of transparency in a region would affect the user’s experience of the usability of the system, we made additional SUS analyses based on the 2 regions with the most participants: Region Uppsala (the first to launch, with a high level of transparency) and Region Skåne (an early implementer, with a low level of transparency at the time of the survey). Overall, 692 participants stated that they had received care in Region Skåne, of which 520 responded to at least 1 SUS item and were included in this analysis. However, 520 participants stated that they had received care in Region Uppsala, of which 331 had responded to at least 1 SUS item. Analysis of Region Skåne responses yielded a final score of 79.37, whereas Uppsala’s result was 80.71. The results in mean SUS score were close enough to indicate that no major difference can be seen on the SUS scale based on the level of transparency in the implementation of the PAEHR. A two-tailed t test with equal variance yielded a P value of .16, indicating that the difference was not statistically significant at the standard 5% significance level.

In addition to the SUS, we also asked a question specifically focused on navigation, more precisely, on the participants’ experiences of locating Journalen in the national patient portal. This issue had been brought up as a concern previously; since the national patient portal contains many eHealth services, there were worries that patients might not find Journalen. The majority of the participants (1974/2451, 80.53%) did not have trouble locating the link in the patient portal, whereas 233/2451 (9.51%) expressed difficulty finding Journalen.

To summarize, the results indicate that the users of the Swedish PAEHR Journalen rate the service high (79.81) on the System Usability Scale, yet questions relating to consistency and complexity scored lower. Before discussing the results in more detail, we address some methodological limitations of this study.