In order to support the evidence-based treatment of patients with multimorbidity, recommendations of the 2017 DEGAM CPG “multimorbidity” (9) were implemented in gp-multitool.de. Main functionalities comprised standardised assessments, which should aide GPs’ information management. Additionally, medication reviews and talks between GP and patient about assessment results were supported by the tool. Recommendations of the DEGAM CPG “multimorbidity” and their implementation in gp-multitool.de are shown in Table 1.

The implemented assessments were standardised, ie, each of them consisted of a predefined set of questions mostly with predefined answer categories. Assessments comprised two validated instruments for the assessment of social contacts (14) and psychiatric symptoms (15). Additionally, based on the literature (16–21), seven instruments were developed by the study team. They comprised prioritisation of treatment goals, assessments of control preferences, involvement of other healthcare professionals, activities and participation, treatment burden, problems with medication, pain, and other health complaints. The final versions of these instruments are shown in Boxes S1 through S7 in Supplementary Material.

We designed gp-multitool.de as web application for GPs and their practice teams. It facilitates sending hyperlinks to the standardised assessments by email. The assessments then can be completed by patients on any mobile or stationary digital device with browser and internet access. Screenshots of the questionnaires on mobile devices are shown in Figure 2. It is also possible to conduct and document assessments during consultations. Moreover, assessment forms can be printed by the GP, filled out by patients using pencil and paper and afterwards entered manually in the digital tool by the practice staff. GPs need to approve finished assessments, talk with patients about the results and they can document their mutual decisions in gp-multitool.de. A screenshot of these functions is shown in Figure 3. Additionally, the GP can access a history with older assessments and decisions.

The concept for gp-multitool.de also comprised other functions including easy access to structured literature search and a database of clinical practice guidelines relevant for primary care, which can be updated within gp-multitool.de by the study team. Moreover, systems for managing practices using gp-multitool.de, users within the practices and patients under treatment were implemented. A site map of gp-multitool.de including all available functions can be found in Figure 1.

Measures of data protection and data security were defined in a data protection concept including data protection impact assessment. We based this concept on the Standard Data Protection Model of the Conference of the Independent Data Protection Authorities of the Federal and State Governments in Germany (22) and the Technical Guideline for Security Requirements for eHealth Applications of the German Federal Office for Information Security (23). Additionally, the company NSF Prosystem GmbH, Hamburg, was consulted regarding requirements resulting from national and European law regulating medical devices. The application gp-multitool.de was programmed by Trinidat Software-Entwicklungs-GmbH, Düsseldorf, who also provide technical support.

In order to facilitate patient and public involvement, a project advisory board was recruited, which consists of representatives of the patient organisation “Bundesarbeitsgemeinschaft Selbsthilfe e.V.”, the caregiver organisation “Wir pflegen e.V.”, the physician organisation “Deutscher Hausärzteverband Landesverband Niedersachsen e.V.”, the professional society DEGAM, the Schleswig-Holstein Association of Statutory Health Insurance Physicians, the health insurance company “Techniker Krankenkasse” and the private company “GAIA AG”. The board advised the research group on how to establish interests and concerns of patients and care providers in the design, implementation and evaluation of gp-multitool.de. For example, it commented on the recruitment process for the focus groups, assessed data protection and discussed functionality and usability of gp-multitool.de before the focus groups were conducted.

The focus groups were conducted and analysed by three researchers. JN is a female health educator working as junior researcher at university. VJ is a female physician working in a regional hospital. IS is a male sociologist, doctor of philosophy and private lecturer for epidemiology and health care research working as senior researcher at university. JN and VJ were trained in qualitative research and supervised by IS. No researcher had any relationship with any study participant and, before participation, study participants had no relevant knowledge about the researchers.

Study participants were recruited in two steps. In the first step, 663 GPs in the region of Hamburg, Germany, and rural surroundings were identified using the websites of the regional Associations of Statutory Health Insurance Physicians and invited by mail to participate in the study. GPs were eligible if they worked as statutory health insurance physician. In case of interest, GPs declared their consent in a contract with the study centre.

The second step consisted in selecting patients by applying convenience sampling. Participating GPs were asked to identify up to six eligible patients from their practice and to invite them during their regular consultations to participate in the study. Patients were included, if they were 65 years or older and suffered from at least three chronic diseases. Patients were excluded if they had no capacity to consent (e.g., in case of dementia), if they were not able to participate because of medical reasons (e.g., severe psychiatric disease), if they were not able to hear, if they had insufficient German language skills or if they lived in a nursing home. Patients received study information by their GP and had to sign an informed consent sheet before participation.

Each focus group with GPs had between 4 and 6 participants and 5.3 ± 0.8 participants on average. In total, 32 GPs participated. Of those, three (9.4%) were 40 years or younger, four (12.5%) were between 41 and 50 years old, 22 (68.7%) were between 51 and 60 years old and three (9.4%) were older than 60 years. Most GPs (25; 78.1%) were women and 7 (21.9%) were men. Mostly experienced GPs were included into the sample. Only five GPs (15.6%) had been working for ten years or less in primary care, while six (18.8%) had been working between eleven and 20 years, 17 (53.1%) between 21 and 30 years, and four (12.5%) since more than 30 years.

Each focus group with patients had between 3 and 8 participants and 5.5 ± 2.1 participants on average. Of the 33 patients who participated, ten (30.3%) were between 65 and 70 years old, 19 (57.6%) between 71 and 80 years old, and four (12.1%) were 81 years or older. The fraction of women (19; 57.6%) in the sample was a little higher than the fraction of men (14; 42.4%). Educational level of twelve patients (36.4%) was primary, while 19 (57.6%) attained secondary education and two (6.1%) tertiary.

Four GP groups and four patient groups were used to discuss benefit, wording and risk of bias of the standardised assessments, before they were implemented in gp-multitool.de. The interviews were based on pre-defined semi-structured interview guides (24), which reflected our research aims and the CPG recommendations. Different interview guides were used for focus groups with GPs and patients (25, 26). Discussed stimuli in GP focus groups included: •

Which information is beneficial for making treatment decisions with older patients with multimorbidity and which information is unnecessary or even harmful? Why?

After programming of the prototype was finished, gp-multitool.de was tested and discussed in the remaining two GP groups and two patient groups. In the GP groups, all relevant functions with focus on patient management, management of assessments and medication reviews, data collection, presentation of results, and documentation of treatment decisions were tested and discussed afterwards. Stimuli in GP focus groups included: •

Which problems do you expect when using gp-multitool.de? Which improvements would you like to see? Why?

Study participants were contacted by email or telephone in order to coordinate dates for focus groups. Of the GP groups, one was conducted in person at the research institute and five were conducted online via WebEx. Of the patient focus groups, three were conducted in person at the research institute, two in person at a community centre in a rural area and one was conducted online via WebEx. Each focus group lasted approximately two hours. Most focus groups were moderated by two researchers (IS, JN) and in the other a third researcher (VJ) was involved. During the focus groups, a student assistant took field notes. All focus group discussions were audio recorded and transcribed verbatim.

Transcripts were analysed using structuring content analysis (27) based on deductive and inductive coding. In a first step, categories were extracted from the interview guide. During coding, the category system was refined using the analysed data and relevant literature (28). Two researchers (JN, VJ) independently coded the data and developed the first category system. Codes and examples were discussed between JN, VJ, and IS until consensus was reached. Based on this consensus, the final category system was developed. MAXQDA Analytics Pro 2022 (version 22.3.0) was used to support data analysis.

For both GPs and patients, the relevance of assessments and functions was an intensively discussed theme. GPs did not want to raise false hopes by exploring topics they could not offer to help with. They also wanted to avoid unnecessary and time-consuming functions. Validity and suitability of the assessments were frequently discussed, but mostly by GPs. Sometimes, they suggested simplifying scales and response categories or including residual categories. Patients supported such changes.

Understandability was relevant for both, GPs and patients, and also a frequently discussed theme. In particular, they addressed possible misunderstandings due to the wording of items. Adequacy and acceptability were a thematic focus in less than half of the assessments. GPs and patients discussed if some items might be too intimate or even discriminatory or if they might overtax patients intellectually. Correctness of the items was intensively discussed, but mostly regarding the clustering of health complaints in one questionnaire.

GPs and patients assessed and in most cases confirmed usability of all functions, but patients suggested changes in default settings and GPs recommended to facilitate data exchange with GPs’ practice management systems. Reliability was an important aspect in the discussion of the management of assessments, patients and users, where study participants pointed to a few minor bugs that needed to be fixed. There was little discussion regarding data security. While some GPs considered it an important topic, most patients were unconcerned with this issue and open to share their data.

The application “gp-multitool.de” was conceptualised as digital support for the implementation of the DEGAM CPG “multimorbidity”. As such, the concept was evidence-based. Patients and public have been involved via the project advisory board. And needs and wants of the target groups have been integrated by considering the feedback given in the focus groups described here.

In the focus groups, both, the healthcare providers’ and consumers’ perspectives have been taken into account which facilitates exploring complementary or even contradictory views (29). Also, patients and GPs of different sexes, age groups, experience and educational levels are represented in our sample. Moreover, intersubjective comprehensibility and credibility of the results have been enhanced by two researchers independently coding the material and consensus regarding the category system between three researchers.

Limitations include low response rate of GPs and self-selection of participants, which could result in higher probability for participation of individuals with certain characteristics (30, 31). For example, patients with higher symptom burden and GPs with higher affinity for digitalisation could have been more motivated to participate. Still, some participating GPs were sceptical or even rejected the idea of digital exchange of information between GPs and patients. Another group that may be underrepresented in our study are patients who do not have the necessary resources and skills to handle digital assessments. Therefore, sometimes the discussion of suitability did not include the concerned persons.

In order to receive advice of the target groups regarding how to improve gp-multitool.de, we discussed the quality of the tool in semi-structured focus group discussions. With this approach, we neglected standardised methods for quality assessment of technology like UTAUT (“unified theory of acceptance and use of technology”) (32) and possibly missed specific details on the acceptance of gp-multitool.de.

In our study, GPs and patients discussed separately, i.e., patients could not reply to comments of GPs and vice versa. Therefore, misunderstanding of patients about the perspective of GPs and misunderstandings of GPs about the perspective of patients could not be identified in our study. Also, it is possible that views on data protection and other aspects of the tool are varying in other regions of the world due to different cultural identities and healthcare systems.

Focus group discussions are often used for developing new scales (33–35), or to assess face and content validity of existing scales (36–38). Also, qualitative studies can support the development of digital tools. For example, in a study from Norway, interviews, focus groups and a workshop were used to accompany content and system development of a digital tool supporting shared decision making in the health care for patients with chronic conditions (39). In our study, a similar approach was used by refining content, functionality, usability and reliability of gp-multitool.de via focus group discussions. In line with other studies (40, 41), our results showed that comprehensibility and acceptability needed to be improved.

In addition to the wording and content of the questionnaires, the participants also discussed the relevance of the surveyed topics. Similar to a Swedish focus group study on attitudes towards digital questionnaires to prepare medical appointments for patients with diabetes (42), both, physicians and patients welcomed that gp-multitool.de supports a more active participation of patients in medical decision making. In contrast to the cited paper, the specific wording of items was of high relevance to the physicians in our study. For example, they wanted to avoid misunderstandings or to raise false expectations in their patients with multimorbidity, which punctually led to extensive rewording and revising of the assessments.

One factor that was particularly discussed by the participating GPs was the risk of revealing problems that, on the one hand, cannot be ignored, but, on the other hand, could not be addressed, e.g., due to time constraints. In another interview study (43) using a digital tool to assess patient-reported outcomes prior to healthcare visits, many issues were uncovered by healthcare professionals, but they had little time to resolve these issues, which was experienced as challenging by study participants. In gp-multitool.de, we revised content and functions, which lead to undesirable side effects from the perspective of the target groups.

In our study, there was a large difference between GPs and patients regarding the subjective importance of data security. While GPs generally stressed the relevance of security aspects, many patients did not see data security as a priority. Other studies on this subject show conflicting results. For example, in a German mixed methods study on privacy and data protection in eHealth, older patients with worse health condition had less concerns about privacy and data protection than younger and healthier patients (44). Another mixed methods study about privacy concerns of older adults using voice assistant systems came to similar conclusions (45). In contrast, several other studies stressed a higher awareness and more cautious behaviour among older people (46–48).

In many countries including Germany, policy makers advocate for speeding up processes of digitalisation, particularly in the medical care system (49–52). This requires the development of manifold digital solutions. Digital tools can help to reduce the high level of complexity, which often applies to the care of older patients, in particular, in case of multimorbidity (53–55). Our study indicated that focus groups can be used to customize digital tools according to the needs and wants of the target groups and thus, improve content, functionality, usability, and reliability of such tools. However, this is only a first step in participatory development. The next steps are to pilot and evaluate the digital tools in everyday care to assess if they are feasible for implementing in healthcare, and effective in improving patient-relevant outcomes (60, 61).

Central aim of gp-multitool.de was evidence-based digital support for mutual prioritisation of treatments between GPs and patients. Many studies indicated deficits in the knowledge needed for this process. For example, a study from the Netherlands (56) showed low agreement in prioritisation of treatment goals of patients and the perception of both their GPs and medical specialists. Similar results were found for other topics assessed in gp-multitool.de. For example, GPs had little knowledge on patients’ social relations and feelings of loneliness (57). And information management regarding medication adherence needed to be improved, both regarding methods for assessing non-adherence (58), and GPs’ and patients’ knowledge of adherence problems (59). In our focus group study, both, patients and GPs confirmed that gp-multitool.de can be one relevant approach for overcoming these deficits in multiple dimensions.