The study settings presented in Table 1 were used as a guide for identifying the different stakeholders and key informants. Implementation of PAEHRs may consist of several parts; both PAEHR policy and the practical (often both technical and organizational) implementation, which will likely need to take place on both national and local (healthcare provider) levels (Figure 1).

Depending on the context, activities may vary between the national and local levels, both with respect to policy and practical implementation. The difference between Sweden and the Netherlands will be highlighted below.

We purposefully chose key informants in both contexts that could provide insights into the research questions. Personal and professional networks were used in Sweden, and in the Netherlands, the participants were recruited through the VIPP organization. An element of snowball sampling was also applied, where the initially identified key informants recommended others. In total, 16 key informants chose to participate in the study (see Table 2 for a description of the respondents' roles).

In Sweden, the original developers of the PAEHR Journalen from Region Uppsala, the current responsible organization Inera AB, and other regions and healthcare providers that have implemented the PAEHR Journalen were approached. In the Netherlands, we interviewed the decision-makers of the VIPP program and experienced experts of implementing PAEHRs according to VIPP. In this study, we focused on the project managers' and implementers' perspective, not those of patients and HCPs which have been covered more extensively in the literature already.

Semi-structured interviews were conducted between March and May 2018 via Skype or phone, where possible. All interviews with Dutch respondents were conducted in Dutch. All interviews were performed by the first author, who speaks Dutch but not Swedish. Some Swedish respondents were reluctant to conduct the interviews in English and were offered to (iteratively) answer questions via email in Swedish. Their answers were translated into English with the help of a native Swedish speaker. The remainder of the Swedish respondents participated in English. This applied to interview numbers 4–6 (marked in Table 2) with five respondents.

Semi-structured interview guides were established for each respondent separately based on their role and the context they practice in. The interview guide revolved around the following topics:

The interview guides were based on the Consolidated Framework for Implementation Research (CFIR), which identifies five dimensions that are essential to implement an intervention (6). An overview of the CFIR dimensions and sub-constructs is given in Figure 2.

The framework further provides sub-constructs and related questions for each of the dimensions (6). The structure of the interview guides was determined by the sub-constructs that were deemed relevant for each type of respondent, i.e., respondents on the national or local level for both Sweden and the Netherlands.

To inquire about the factors affecting PAEHR implementation, the following selection of (sub)constructs was used: the strength and quality of evidence for the intervention, the external policies and incentives, the implementation climate, and all stages of the process (planning, engaging, executing, reflecting, and evaluating). The constructs that patient needs and resources and tension for change were used to assess the involvement of patients.

In addition, several other sub-constructs were selected for specific respondent types, such as peer pressure among Swedish regions. The questions proposed by the CFIR framework were adapted to each individual respondent's specific context. The interview guides were further improved iteratively after each interview to ensure that all relevant aspects were covered.

All participants were informed about the goals and risks of the study. If applicable, audio was recorded, and notes were kept. The interviews were transcribed as soon as possible using intelligent verbatim transcription. While listening to the recordings, transcripts were shortened and edited for the sake of clarity before analysis.

The interviews were analyzed by means of content analysis according to Taylor-Powell and Renner (25). The selected CFIR constructs and sub-constructs formed the basis of the codes with which relevant passages from the edited transcripts were categorized. All passages were provided with a condensed meaning unit in English. Related passages and condensed meaning units were collected, compared, grouped, and provided with a description.

We have identified the main barriers and facilitators to developing and implementing national PAEHR policy as described by the respondents from both countries. Barriers exist both on the national level (Table 3) and on the local implementation level (Table 4). One of the main barriers has been resistance from HCPs on both levels in Sweden and on the local implementation level in the Netherlands. This resistance is very much in line with concerns that have previously been described in the literature; concern that patients will misunderstand, take offense or worry unnecessarily, concern that the workload will increase with patients asking questions, and concern that the records' value will be reduced as it can no longer be used as a tool for professional communication.

Regarding policy, both countries had some challenges on the national level. In Sweden, it was difficult to agree on a NRF, and the first version had several “electable rules,” i.e., rules where each region or local healthcare provider had to choose how to implement (20). One could, e.g., choose which types of information to release to patients (notes, lab results, referrals, etc.), also whether to give immediate access or have a 2-week waiting period, and whether to release only signed/validated information or not. This, of course, led to a fragmentation of the otherwise national PAEHR, where it was difficult for patients to understand why they could see lab results from Region A but not from Region B. In the Netherlands, all practical implementation took place at the local level, but the VIPP program provided an important incentive. Here it was, however, difficult to both define goals and measure the hospitals' progress toward these goals.

Implementers from both countries faced technical barriers when implementing PAEHRs. In Sweden, the first pilot projects struggled to find secure authentication methods, but in later implementations, the technical challenges related mainly to connecting the EHR systems to the national HIE platform:

“The first barriers were strictly technical, making sure that we had the right protocols from the supplier of the EHR and making sure that everything worked fine in that integration.” (Sweden, local level)

Connecting to the national HIE platform was not only described as a technical challenge; it could also be quite costly, which kept smaller private healthcare providers from connecting:

“[…] healthcare providers in our region can use whatever EHR system they want. It is expensive to make changes in the systems unless they are big. That's why there are healthcare providers who are still not connected. All hospitals and most health centers are connected.” (Sweden, local level)

Dutch implementers were dependent on their IT suppliers for implementing a successful PAEHR and achieving the VIPP goals:

The Dutch interviews revealed no facilitators on the national level, likely due to the fact that all practical implementations took place at the local hospital level. In Sweden, the National HIE platform as well as the updated, more strict version of the national regulatory framework (24) was described as facilitator (Table 5).

On the local implementation level (Table 6), the national infrastructure was also described as a facilitator in Sweden, with the use of national protocols and contracts making it easier for regions and private healthcare providers to integrate. Social aspects were also important Swedish facilitators, where involvement of different stakeholders, learning from peers who were also implementing, and a gradual implementation process were described as beneficial:

Involvement of HCPs in the implementation process was described as a facilitator, but there were also some respondents who highlighted the importance of daring to proceed despite resistance from HCPs, indicating that this may be a double-edged sword.

Strong decision-makers and involvement of different stakeholders (including patients and HCPs) were highlighted as facilitators in the Dutch context, too (Table 6):

In the Dutch interviews, the EHR vendors' role was more prominent, and their contribution to addressing, e.g., security issues was considered a facilitator:

Some facilitators and barriers were considered critical. In the Swedish context, the central management of the patient portal and HIE was seen as essential. In the Netherlands, involvement of end-users (both patients and HCPs), implementing toward a clear future goal, involvement of vendors and IT departments, and strong leadership were considered critical success factors. In the Dutch context, the dependence on collaboration with system vendors was also seen as a critical weakness.

The different aspects of patient involvement in PAEHR policy development and implementation in Sweden and the Netherlands are displayed in Table 7 (national level) and Table 8 (local implementation level). In Sweden, patients' wishes and preferences regarding digital access to the EHR were analyzed during the early deployment of the PAEHR Journalen. Regions implementing Journalen also attempted to do so but used fewer means to explore the patients' preferences. In addition, as most decisions relating to the national patient portal and the PAEHR were centralized to the national organization, there was a sense of loss of control on the local/regional level, making the incentives for engaging patients in decision-making limited. In the Netherlands, patient involvement mainly took place at the implementation level, even though hospitals face multiple barriers when doing so. Little to no patient involvement was carried out when developing VIPP.

Swedish and Dutch developers and implementers of national PAEHR policy have reported on many different barriers, facilitators, critical success factors, and aspects of patient involvement. These outcomes are compared and linked to existing literature in order to interpret them and give recommendations.

The strength of this study is that policymakers and implementers were chosen as respondents, while previous studies have mainly focused on the viewpoints of patients and/or HCPs.

Limitations include the sampling method and conduct of the interviews. Even though attempts were made to interview as many respondents as possible, the respondents were sampled though convenience sampling, and it was not investigated how well the implementers represented the whole implementation level in both countries, yet a limited number of people have deeper insights into the implementation process and the individuals interviewed in each context can be considered experts. It is important to note also that the respondents from the implementation levels from both countries were implementing PAEHRs in different settings within healthcare. The interview respondents that implement VIPP all represent hospitals, while Swedish county councils that implement Journalen are responsible for all levels of healthcare. This may, of course, influence the experienced barriers and facilitators, similar to the way other contextual differences do. Some Swedish respondents were offered to answer the interview questions via email in Swedish, even though the researcher responsible for data collection was not proficient in this language. The answers were translated together with the last author (who is a native Swedish speaker), yet the email interview format led to less detailed questions and answers than in other interviews. However, the respondents' expertise was considered important, and their answers confirmed results from the more in-depth in-person interviews.