The use of electronic health records (EHRs) is increasing worldwide [1,2]. It has been associated with improvements in health care quality and patient safety [3]. In international literature, different terms are used interchangeably to refer to electronic clinical documentation, such as electronic medical records, electronic patient records, or EHRs [4]. In this paper, we use the term EHRs to refer to different types of electronic documentation of patient health data. EHRs are digitized medical records used in clinical health care within an organization [5]. EHRs are linked to organizations (eg, an EHR that is used by the staff in an intensive care unit [ICU] of a hospital), as opposed to personal health records. Personal health records are characterized by the fact that patients can manage them themselves and provide access to others [6]. EHRs can electronically gather and record both administrative and health-related information as well as store, transmit, and display information from various sources [7]. Traditionally, health-related information in EHRs includes a medical history and medication orders, vital signs, or laboratory results [3,8]. Administrative information may include age, sex, or International Classification of Diseases codes [9]. Depending on the context, EHRs include different submodules, such as medication display, the display of vital signs, or diagnostic information [10]. For example, different content is more critical for work in an ICU than for work in a palliative care unit. Depending on the context, there are EHRs specific to each area of medical care to ensure optimal documentational support [11]. It is useful to ensure that information can be transferred within the units of a hospital and between health care institutions. However, this diversity of records still poses challenges for interoperability [12].

In recent years, technological progress has led to extreme improvements in the field of EHRs in terms of design and functions [13]. EHRs can be used to minimize costs and workload with the help of shared, location independent, and clear documentation [14] and to improve collaboration and coordination between different professions and individuals [15]. In addition to the digitization of previously paper-based documentation, electronic decision support systems and the use of predefined clinical guidelines and standards can support quality improvement based on the latest health care knowledge [3,16].

However, the solitary implementation and use of EHRs in isolation will not guarantee that the quality of care improves. The literature suggests that EHRs with poor usability or functionality may have unintended consequences for their users and patients [17]. For example, the lack of adaption to workflows [18,19] and user needs [20], poor usability [21], and unstructured data sets in EHRs lead to high cognitive demands on users [22]. These aspects are associated with work-related stress, fatigue, and burnout for the main user groups of EHRs: nurses [23,24] and physicians [25-27]. Furthermore, poor usability has been associated with patient harm [28]. For example, it can make it difficult for health care providers to access necessary medical data for the treatment of patients or lead to misinterpretation of available data. This can lead to misdiagnosis, incorrect treatment, or unsuitable medication for a patient’s condition. This can put the patient’s health and well-being at risk [28]. Therefore, user acceptance is essential for successful implementation, actual use, and user satisfaction of EHRs [29]. Expected usefulness, technical concerns, technical problems, and expected workflow challenges can facilitate or hinder technology acceptance [30].

To promote international joint development projects, globally valid standards have been drawn up. For example, there is International Organization for Standardization (ISO) 9241-210, which focuses on the ergonomics of human-system interaction. It is an important standard for classifying and demanding usability engineering measures in product development processes such as the development of EHRs. Unfortunately, these standards are often only partially complied with, leading to the various abovementioned problems.

In addition, the involvement of users is necessary to adapt EHRs to the needs of health care professionals and to ensure their acceptance [31]. This is increasingly being addressed, resulting in an expansion of projects involving future users in EHR development. Existing reviews have focused on the involvement of users in technology development. In recent years, several reviews have been published to address the involvement of users in the development of different health-related technologies [32,33]. The focus of these reviews has been, on the one hand, on the involvement of different user groups, such as older people [34-39], people living with dementia [40], or patients with chronic diseases [41]. Specific to these groups is the fact that their cognitive and physiological characteristics must be addressed in the development of digital technologies. On the other hand, reviews cover different use cases of technologies such as mobile health [42], serious digital games for health promotion [43], or for the treatment of depression [41]. In addition, other reviews cover more general aspects of user involvement for the development of health-related technologies [32,33]. Despite the empirical evidence supporting the need for user involvement in the development and implementation of EHRs, this topic is largely excluded from the reviews. For example, in 3 reviews covering generic aspects of user involvement, no study focused on EHRs [32,33,44]. Different approaches such as participatory design or co-design are common practices to involve users in the development of new technologies. For example, on the one hand, participatory design actively and creatively involves both users and designers and thus includes different individual qualifications [31,45]. This approach can be defined as “...a process of investigating, understanding, reflecting upon, establishing, developing, and supporting mutual learning between multiple participants in collective ‘reflection-in-action’. The participants typically undertake the two principle roles of users and designers where the designers strive to learn the realities of the users’ situation while the users strive to articulate their desired aims and learn appropriate technological means to obtain them” [45]. On the other hand, an approach such as co-design is defined as an “active collaboration between stakeholders in the design of solutions to a pre-specified problem” [46]. Although these 2 approaches are often used interchangeably and synonymously, they differ in how much choice is given to users in the development of a technology. It can be assumed that the participatory design approach gives users more influence than the co-design approach.

In existing studies, the design of the methodology varies depending on resources, time period, and technology. When using participatory design or co-design methods, methodological choices must be made [46]. For the planning of similar research projects and a sensible use of diverse methods, it is crucial to provide an overview. Within the framework of a scoping review, we therefore investigated which forms of user involvement have been used to date, under what circumstances, and with what results. The result can also be used to facilitate guidelines for the involvement of health professionals in the development of EHRs. The overview in the metadata table in Multimedia Appendix 1 [47-115] is intended to be particularly helpful in this regard. In addition to the range of possibilities, specific classifications can also be made as to which method is helpful and for which objective.

The review was guided by the question: “How are health care professionals involved in the development of EHRs?”

The Methods section is reported as recommended by the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 statement [33]. The presentation of this scoping review is based on the methodological specifications by Peters et al [116].

This methodology was developed using an a priori scoping review protocol [117]. The decision to use a scoping review methodology was based on the identification of a gap in current knowledge [117] and the need for an overview of different methods without any assessment. The result should be a narrative account with a focus on the different ways and methods of involving end users in the development of EHRs.

The methodology of scoping reviews is gaining popularity, particularly in the field of health care [118]. Whereas systematic reviews aim to synthesize collate empirical evidence on a focused research question and present the evidence from the reviewed studies [119], scoping reviews map the existing literature on a topic area [120]. In addition, scoping reviews provide a descriptive overview [121] and are therefore an appropriate method for addressing the research question.

This review aimed to provide an overview of the existing ways and methods of user involvement in the development of EHRs in the literature. The four specific objectives of this review were (1) to conduct a systematic search of the published literature for studies focusing on user involvement in the development of EHRs, (2) to present the characteristics and range of methods used in the identified manuscripts, (3) to explore the reported challenges and limitations of the methods, and (4) to make recommendations for the further development of the approach to the development of EHRs and to improve the consistency with which these types of studies are conducted and reported.

Planning for the review began in January 2021. The review was conducted and evaluated from June 2021 to April 2022. Four people were involved in carrying out the review (JL, CJ, SK, and TSB). JL and CJ had experience in conducting (scoping) reviews. CJ, SK, and TSB worked with prospective users to develop an outpatient EHR, an inpatient EHR, and a cross-sectoral EHR for pediatric palliative care with future users.

Studies were eligible for inclusion if they described the involvement of health care professionals in the development of EHRs. This explicitly included studies that examined a specific EHR. However, excluded studies focused on the general workload resulting from the use of different EHRs in different institutions or other parameters related to different EHRs, regardless of their design. Articles published in languages other than English were excluded. Manuscripts that described a process without performing it were excluded from the scoping review. Gray literature was not included because of the focus on research projects, although this was included in the scoping review methodology [122].

SK and JL formulated the inclusion and exclusion criteria and discussed them with TSB and CJ. The inclusion and exclusion criteria are listed in Textboxes 1 and 2.

The search was carried out in the PubMed, CINAHL, and Scopus databases. A supplementary search was carried out in Google Scholar. The final search was performed by JL and SK on March 17, 2021, using the search strings from Multimedia Appendix 2. The forward-and-backward citation tracking [123] was then carried out by JL using Scopus and Google Scholar.

First, an initial limited search was conducted in a selection of relevant databases to analyze possible terms in the title and abstract to identify keywords describing the articles. This was followed by a search of all databases using all identified keywords. SK and JL formulated the basic idea of the review and conducted the initial searches. Afterward, SK, JL, and TSB developed the search terms. The search terms used were based on two main categories: (1) search terms around the term EHRs and corresponding synonyms as well as Medical Subject Headings terms (PubMed) and subject headings (CINAHL) were used, and (2) search terms around the term participatory design with corresponding synonyms and Medical Subject Headings terms or subject headings were used. The search strings for PubMed, CINAHL, Google Scholar, and Scopus are shown in Multimedia Appendix 2.

The search in Google Scholar used a substantially shortened search string, as the search engine cannot process longer, complex search strings. This resulted in several results that did not meet the inclusion criteria. Therefore, using Google Scholar’s sort by relevance function, only the first 250 results were checked for eligibility, of which 47 were selected.

All citations were imported into the bibliographic manager EndNote (Clarivate), and duplicate citations were automatically removed, with further duplicates removed if found later in the process. The citations were then imported into the software [124] to subsequently check the relevance of the titles and summaries and to characterize the data of the full articles. Rayyan provided blinded checking and automatically displayed matching inclusions, exclusions, and conflicts after blinding was turned off.

First, the titles and abstracts were checked by SK and JL to ensure compliance with the inclusion criteria. Differences were discussed with TSB. Subsequently, TSB and JL screened titles and abstracts for the forward-and-backward citation tracking results, and the differences were discussed with CJ.

All citations deemed relevant after title and abstract screening were obtained for subsequent review of the full-text article. For articles that could not be obtained through institutional holdings that were available to the authors, attempts were made to contact the authors of the source and request the article. In addition, articles were requested via interlibrary loan.

SK and JL screened the full texts; the differences were discussed with TSB. TSB and JL screened the full texts of the forward-and-backward citation tracking results; the differences were discussed with CJ.

At this stage, studies were excluded if they did not meet the eligibility criteria. After reviewing approximately 25 articles independently, the reviewers met to resolve any conflicts and to ensure consistency among the reviewers and with the research question and purpose [125]. The excluded studies were appropriately labeled with the reason for exclusion to improve traceability.

Categories were formed deductively. This was based on a systematic review by Vandekerckhove et al [33] that focused on electronic health interventions. The aim of the review by Vandekerckhove et al [33] was to report and justify participatory design methods in empirical eHealth studies for further development of the methodology. The decision to follow this review was based on its comprehensive presentation and its fit for the research question pursued here. However, the categories were supplemented by inductive categories that emerged from reviewing the material. The categories can be named as “factual categories” according to Kuckartz [126], designating specific facts in the included studies. All codes were reviewed, coded, and discussed in regular meetings by TSB, CJ, and SK.

Owing to the diversity of study designs and the research questions, a quality assessment was not performed. Following the approach of Vandekerckhove et al [33], an assessment of the sufficiency and design of reporting was conducted. To improve comprehensibility, the inductive categories were supplemented by key questions (based on the definitions of the categories that were created and constantly refined during the analysis process) and served to represent the collected data items. This was partly based on the categories in the study by Vandekerckhove et al [33], whose review dealt with eHealth interventions. For example, category 1 in this review was developed based on the category “eHealth intervention” by Vandekerckhove et al [33] and category 3—study participants—was based on the category “stakeholder types” by Vandekerckhove et al [33]. Category 4—methods and materials—was based on the category “tools of participatory design” by Vandekerckhove et al [33]. The other categories were derived from the material itself, as described earlier in the Data Analysis section. This resulted in the following division, which was used to structure the method representation:

The study selection is described in Figure 1.

The initial search resulted in a total of 23,446 hits (PubMed: n=8281, 35.32%; CINAHL: n=1846, 7.87%; Scopus: n=13,319, 56.81%). In addition, 47 records were extracted from other sources (Google Scholar) after screening the first 25 pages of approximately 7710 results. From a total of 23,446 hits from the initial search and these 47 additional records, 19,002 (81.04%) hits remained after duplicate reduction.

Of these 19,002 articles, 18,830 (99.09%) articles were excluded during title and abstract screening. The remaining 172 texts were subjected to full-text screening, resulting in a total of 74 titles.

The forward-and-backward citation search yielded a total of 2769 hits (755 by forward citation and 1985 by backward citation). Automatic deduplication reduced the number of hits to 2665. Manual duplicate reduction led to a final result of 2625 hits. After title and abstract screening (34 texts remaining), full-text screening was carried out, resulting in 23 articles.

These 23 articles from the forward-and-backward citation search were included in the final assessment along with the previous 74 studies from the initial search. Of these 97 studies, 27 (28%) studies were excluded because of insufficient information and duplicates. The remaining 70 articles were included in the evaluation and can be found in the metadata table in Multimedia Appendix 1.

Multimedia Appendix 1 includes the metadata of the included studies.

The review had four objectives: (1) to conduct a systematic search of the published literature for studies focusing on user involvement in the development of EHRs, (2) to present the characteristics and range of methods used in the identified manuscripts, (3) to explore the reported challenges and limitations of the methods, and (4) to make recommendations for further developing the approach and improving the consistency with which they are conducted and reported. Therefore, the main focus of the review was to examine in which settings which participants were involved with which methods and materials and which frameworks were used. Furthermore, the frequency and design of the development and an overview of the competences of the respective researchers involved in the development were examined. To the best of our knowledge, this is the first review to describe the methodological aspects for involving health care professionals in the development of EHRs.

The characteristics of EHRs addressed in the included studies covered a variety of different aspects. On the one hand, a large number of studies addressed a comprehensive EHR, whereas on the other hand, many studies addressed only individual modules of an EHR. The wide range of characteristics in this review was largely because of the broad inclusion criteria, which were designed to provide a comprehensive picture of the methodological aspects of professional involvement in the development of EHRs. This leads to a better description of the complex field of EHRs and their methodological aspects. However, this sometimes makes it difficult to compare the interventions. In terms of setting, it was found that most of the included studies were conducted in general hospitals and ICUs. The authors suggest analyzing studies in a specific setting and how the participants are involved there in the future to cover the methodological aspects, such as in ICU or in palliative care. Future studies could also focus on the individual modules of an EHR, for example, medication modules.

In terms of participants, the studies mainly included physicians and nurses. In terms of multidisciplinary care, it would be desirable for all health care professionals (including physiotherapists, occupational therapists, speech therapists, social workers, nursing assistants, etc) to record their activities and observations in joint documentation and to be able to view them mutually. In addition, sharing EHRs can facilitate communication [148] (eg, by sending messages within a program and assigning tasks). With this in mind, it is surprising that only these 2 professional groups were so intensively involved. It would be desirable for further studies to include all professional groups and to design EHRs to meet their needs. However, most of the studies did not specify which health care professionals were involved in the development. This was partly because of an imprecise naming of the participants and partly because of the lack of naming. Future studies should specifically describe the demographic characteristics of the participants, which may lead to a better assessment of the results [33,36]. This is important because demographic variables have a strong influence on the acceptance of EHRs and the level of competence in using EHRs and digital health technologies in general [149,150]. Therefore, it is recommended that study investigators collect key demographic variables from participants and present them in tabular form to improve the interpretation of the results.

The methodology of the studies was balanced between qualitative, quantitative, and mixed methods. However, in 63% (44/70) of the studies, no terminology was used to describe the design of user involvement in more detail. This adds to the imprecision of the presentation in terms of the level of involvement and a qualitative assessment of the methodology. Although 23% (16/70) of the included studies that mentioned user-centered design as an approach will be examined to see if this was really implemented, most of the studies remain vague about user involvement. This again supports the broad search strategy of the review but also points to qualitative ambiguities of implementation.

The frequency of user involvement varied widely, and in most of the studies (40/70, 57%), users were involved only at one point. This shows that a true participatory design or co-design, as it is called for, is rather rare and fuels the suspicion of sham participation, where user requirements are collected but no iteration is performed to test the fit. Another problem with user involvement in development is that it often occurs at only one point in the development of new technologies. This problem is often referred to as “project-based temporality” in the involvement of users [151]. Therefore, it is recommended that users are involved in the development of new technologies at all stages of development over a longer period [151].

Most of the methods vary widely. Think-aloud approaches were often used to obtain user feedback on an EHR. The continuation of this method, near-live testing, which also increases the likelihood of a good fit, was used in only one study. In the case of questionnaires, individual, nonvalidated questionnaires were frequently used, which reduced the quality of the results. The most commonly used questionnaires were mainly oriented toward usability (System Usability Scale) in 16% (11/70) of the studies and the cognitive load (Nasa Task Load Index) in 7% (5/70) of the studies. Cognitive load refers to the amount of mental effort required by a person to perform a specific task and the associated required capacity of the human working memory [152]. It is imperative to consider cognitive load in the development of EHRs and in the implementation of appropriate solutions in clinical practice, as cognitive load has been linked to the development of burnout and distress [27]. It is recommended that user involvement studies use a mix of methods from the fields of telling, making, and enacting [153]. This emphasizes the impact of user involvement through the exchange of current and future practices and the sharing of needs (telling). For successful user involvement, it is crucial that future users develop something (making), contributing to the existence and design of a new technology. In addition, by involving users, it should be possible to transfer ideas into reality by creating a simulation to test them [153]. Accordingly, different questions require different methods from each field, but a mixed methods approach ensures a diversity of perspectives.

Frameworks, theories, and guidelines were very rarely used. Moreover, the results showed a rather low level of consideration of the theoretical underpinnings to the detriment of the quality of the studies. This is particularly problematic because the use of such frameworks can structure the development and make the replicability of results between different studies comparable. Furthermore, it is problematic that a large number of studies did not refer to theories and models. This would also provide a theoretical basis for the development and make the quality of the results more comprehensible [32]. It is particularly surprising that only 9% (6/70) of the studies referenced to ISO standards. These standardizes the process for the development of new software. It would be useful to refer to these standards in future studies and to highlight the stage of development of the respective technology [154,155].

The researchers’ skills have rarely been documented. A multidisciplinary research and development team should consist of individuals with different skills from different health care disciplines, methodological disciplines, and social disciplines. This allows for optimal design and support of different stakeholders during development and implementation [156]. Similar to the sometimes imprecise description of demographic characteristics of the study participants, the skills of the investigators should be described. Owing to the interdisciplinary nature of technology development research projects, it would be advisable to have multidisciplinary teams and to identify the respective competencies and experience in technology development.

To be able to interpret these results, it is necessary to describe several limitations of this study. First, the search strategy was limited by its focus on empirical, scientifically published work. The involvement of health care professionals in the development of EHRs may not always be published in scientific journals. Therefore, this review is a first step on the topic of involving health care professionals in the development of EHRs. In further studies, it might be interesting to include gray literature and databases with a focus on technology-oriented research and engineering (eg, Institute of Electrical and Electronics Engineers). However, the heterogeneity of the quality of the publications must be taken into account. In addition, EHRs are mostly developed by large digital technology companies. It can be assumed that these companies often involve users in the development but do not produce publications or perform actual research. Therefore, it can be assumed that there was publication bias. It would be necessary in the future to survey such large companies on how they involve users in the development of EHRs.

Second, it should be noted that the screening process was limited by the definitions of user involvement, which accordingly shaped both the search terms and the inclusion and exclusion criteria. Although a wide range of terms were used, it cannot be ruled out that individual manuscripts that were coherent in terms of content, and therefore would have led to different results, were not included because of the lack of used terminology.

Third, it should be taken into account that some studies published their results in several manuscripts and did not briefly review the entire development or implementation process. In our evaluation, we were only able to consider the described frequency of user involvement from the information provided in the included manuscripts. However, it is possible that the included manuscripts each report only a subset of the study project; whereas in the overall study project, users were much more frequently involved.

Finally, one of the findings, namely the sometimes-low transparency of reporting, also directly points to a limitation in terms of analysis and conclusions—drawing conclusions on the basis of reporting in manuscripts has limited validity, as there is no way to ensure that the actual methodological considerations and intentions correspond to what was presented in the manuscript.

Further research can help to improve the methodological framework for involving health care professionals in the development of EHRs. Particular attention should be paid to the rationale for the methodological choices. It is also crucial to combine different methods from the fields of telling, making, and enacting; involve users at several points in time to avoid sham participation; and strive for maximum user orientation. The growing interest in “design through design research” [157] should be encouraged but with conditions to promote high-quality developments. Little knowledge can be gained from publications with low reporting quality in terms of transferability and quality assessments. It would be useful for studies to report the aspects more precisely. Specific reporting guidelines for reporting the results of technology development studies would be helpful, as is the case for many other types of studies [158]. Process evaluations should be used in a standardized manner to improve study quality.

In addition, in future studies, it would be necessary to examine in more detail the outcomes of the participatory design with users. In this context, questions should be answered regarding the specific outcomes that have been improved by the involvement of users. How these outcomes were measured and how, for example, improvements to the software were evaluated in different iterations should also be analyzed.

Studies involving health care professionals in the development of EHRs have used various approaches. This paper provides an overview of the approaches in different fields of development with the inclusion of diverse users. Often, however, there is no specific approach, framework, or theory underlying the procedure and there is missing or inaccurate information in the reporting.