STRENGTHS AND LIMITATIONS OF THIS STUDY
-
The study applied a service design perspective to engage healthcare professionals as end-users in exploring collaborative care processes.
-
A qualitative explorative approach was used to elicit detailed, context-specific insights from both primary and secondary care professionals.
-
The use of the Consolidated Framework for Implementation Research (CFIR) framework provided a structured lens for analysing implementation-related factors.
-
Only two of the five CFIR domains were included in the analysis, limiting the comprehensiveness of the framework’s application.
-
Findings may not capture perspectives beyond those represented by the selected professional groups and settings.
Introduction
Healthcare worldwide is currently facing major challenges at both the population and organisational levels. At the population level, issues such as ageing, an increase in multimorbidity and increasing health inequity are all significant challenges.1 2 At the healthcare organisational level, challenges include fragmented care offerings, poor communication between primary and secondary care and difficulties maintaining continuity of care across the care continuum. Together, these factors contribute to the growing demand for healthcare access and make it difficult to maintain a high quality of care.3–5 Researchers have stressed the importance of improving processes and services and reducing the fragmentation of care by fostering collaboration and transfer of information between primary and secondary care.6–10 Consequently, numerous interventions have been implemented to improve integration and optimise collaboration, including the development of integrated delivery networks, changes in payment models, as well as the fostering of better healthcare data management.3 11–13 Nevertheless, integration and collaboration remain a major challenge for policymakers, researchers, clinicians and patients worldwide.14–17
A lack of successful integration between primary and secondary care can be attributed to shortcomings in a range of factors such as interdisciplinary teamwork, communication and information exchange, joint guidelines, education, and accessibility and funding.8 14 18 19 A recent study indicated that healthcare professionals must actively change their behaviours and competencies if collaboration between primary and secondary care is to be enhanced.20 21 However, to successfully redesign the primary–secondary care continuum and foster improved collaboration, it is essential to first gain an understanding of users’ (primary and secondary healthcare professionals) needs and assumptions, as well as social and cultural norms related to their (working) environment. Without this insight, new implementations are likely to falter.22–25 This calls for a fundamental transformation in the way healthcare is adapted, moving beyond selective adjustments in professional behaviour or implementing isolated interventions targeting specific issues. Instead, to achieve real progress the system itself must be redesigned, facilitating and sustaining behavioural change among primary and secondary healthcare professionals.24 The significance of this alternative perspective highlights the need for a user-driven redesign and co-creation approach, commonly referred to in the literature as ‘viewing the patient as a user’.26–28 Conversely, primary and secondary healthcare professionals can also be considered ‘users’. Therefore, this article focuses on the consultation and communication service options between primary and secondary healthcare professionals (figure 1).
Primary–secondary care interface visualised as a chain consisting of processes, formed by operations (modified picture inspired by Vissers et al). Dotted lines visualise the moment the decision is made by the healthcare professional to transfer the patient from one process to another process. Yellow lines are the operations related to the interface between general practitioners (GPs) and specialists. Services available services mentioned by healthcare professionals. * In table 1, these services are specified.
Schematic overview of the differences and possible constraints on several levels, concerning the services for collaboration between primary–secondary care
This study is part of the broader The Hague Consortium Of Network CarE Redesign and Transition (CONCERT) research project, which aims to evaluate the impact of implemented healthcare redesign initiatives in the Hague region. The overarching objective is to assess how these initiatives affect referral and consultation practices between primary and secondary care, healthcare outcomes, costs and the experiences of both professionals and patients. Within this context, the current study focuses specifically on identifying the barriers, facilitators and proposed changes encountered by healthcare professionals when collaborating at the care continuum between primary and secondary care, with the ultimate goal of enhancing clinical interaction and collaboration.
Methods
Study design
We conducted a qualitative exploratory study using semi-structured interview data from general practitioners (GPs) and medical specialists (specialists). The study was conducted in an urbanised area in the Netherlands in which care is provided to a diverse population. This article was written in accordance with the standards for reporting qualitative studies (Consolidated Criteria for Reporting Qualitative Research checklist).
Context
All interviews were conducted in The Hague, the Netherlands. In the Dutch healthcare system, GPs act as a gatekeeper and are responsible for determining treatment plans or referring patients to specialist care. Such referrals mark the transition from primary to secondary care (figure 1). One approach to supporting GPs in their decision-making—particularly regarding whether to manage a patient within primary care or to refer them to specialist care—is through pre-referral consultations with specialists (table 1). Specialists, in turn, assess whether patients treated in secondary care can return to primary care or require a consultation with the GP, such as during treatment or follow-up. This interprofessional interface is illustrated in figure 1. To facilitate this process, multiple collaboration services are used, each with distinct workflows. Table 1 provides a schematic overview of the current communication and collaboration methods, including their workflows, capabilities and limitations. This overview supports the interpretation of the barriers and facilitators identified during the interviews by highlighting practical differences and potential constraints at various levels.
Perspective
Service (re)design perspective
This study adopts a service design perspective, recognising that existing hierarchical and siloed structures in healthcare struggle to effectively address complex challenges.29 A service (re)design approach is both collaborative and cross-disciplinary and thus has the potential to break down silos.22 23 Service design ultimately leads to co-creation centred on user needs and a shift from a fixed system to continuous development.24 30 Service redesign is not a one-size-fits-all approach; it requires a shift in mindset towards human-centred design and addressing user needs. As such, we considered which stakeholders and collaborative services (telephone, teleconsultation, etc) to include, as well as considering the appropriate perspective. GPs and specialists are the primary ‘users’ of various services designed to support collaboration, facilitate clinical decision-making and promote the integration of primary and secondary care (figure 1 and table 2). A clear understanding of users’ values and needs is vital to the successful implementation of new services and redesign initiatives. This should be complemented by a clear understanding of redesign goals, ensuring that both users and the system and organisation are effectively supported.
Clinician characteristics
Framework
To gain insight into experienced barriers and facilitators of the implementation of current collaboration services (eg, telephone, correspondence) between primary and secondary care, the ‘Consolidated Framework for Implementation Research’ (CFIR) was used to create an interview guide (online supplemental file 1). This framework consists of five domains: (1) intervention characteristics, (2) inner setting, (3) outer setting, (4) individual characteristics and (5) the process of implementation. CFIR is a determinant framework designed to identify barriers and facilitators, guiding the adaptation of implementation strategies and necessary adaptations. In our case, it was applied to services promoting collaboration between primary and secondary care. As the interventions had already been implemented by some practitioners in the study region, the interviews also included implementation outcomes.31 32 These measures help distinguish whether adverse clinical outcomes arise from the intervention itself or from implementation-related factors, drawing on Proctor and colleagues’31 framework to assess barriers and outcomes, such as acceptability, adoption and appropriateness. To obtain a full picture of the overall setting, questions concerning the working environment (care continuum and context) were also included. Due to the high response volume, two separate analyses were conducted. This article focuses on two of the five CFIR domains—Individual Characteristics and Inner Setting—Tension for Change—with particular attention to users’ expectations, as well as gaps and needs regarding collaboration between primary and secondary healthcare professionals. A separate paper will address all five CFIR domains related to (new) collaboration services.
The Individual Characteristics domain explores the barriers and facilitators experienced by individuals working with an innovation. While CFIR encourages the inclusion of a broad range of stakeholders (eg, policymakers, opinion leaders, implementation supporters), this study concentrates on the direct stakeholders, namely primary and secondary healthcare professionals.
The Inner Setting—Tension for Change domain focuses on whether a perceived need for change exists within the organisation. A strong sense of urgency is often required to initiate the redesign of existing innovations. This domain also includes organisational culture, communication patterns and work infrastructure, all of which influence the implementation of change.
Participants
Prior to participant recruiting, heterogeneity among GPs was defined using factors such as the socioeconomic status of the practice area, age, sex, years of experience, practice size and degree of involvement in healthcare redesign. For specialists, heterogeneity was determined based on hospital affiliation (two out of two hospitals in The Hague), surgical vs non-surgical specialities and involvement in healthcare redesign initiatives. Stratified purposive and snowball sampling methods were used to recruit interview participants among both GPs and specialists active in The Hague region.33 Representatives from almost all departments of medical specialists at the two hospitals in the area were contacted via email and invited to participate in interviews. Recruitment was facilitated through personal networks or by randomly consulting specialists who were either personally willing to participate or who referred colleagues. GPs were primarily contacted through personal networks, with additional outreach conducted by locating email addresses through their professional websites to ensure sample diversity. Participants were contacted by email, ‘WhatsApp’ or telephone, and provided with an information letter and consent form. Face-to-face interviews were preferred, with alternative options being telephone or video calls. Face-to-face interviews took place in hospitals or GP practices, while telephone or video interviews also occurred in private settings. Researcher MC, a female GP in training, introduced herself and explained the aim of the study while conducting the interviews.
Data collection
The CFIR guide, including domains and constructs, was adapted by researcher MC to align with the research focus. Multiple questions were formulated for each CFIR domain, as outlined in the interview guide (online supplemental file 1). However, only a selection of the most relevant CFIR questions was included. Questions related to specific interventions were deliberately excluded as they will be addressed in a separate publication. The guide was pilot tested with three pioneering health professionals involved in interprofessional consultation services. Following the pilot interviews, no changes were made to the interview questions. However, it became apparent that the questions elicited extensive responses from participants, requiring the interviewer to actively guide them in providing well-structured answers. The interviews were conducted in Dutch between August 2021 and October 2022. They were audio-recorded after obtaining oral consent, with written consent collected at the end of the interviews. When interviews were conducted by phone or video, specialists had limited time and signed the informed consent digitally (online supplemental file 2). Pilot interview data were included in the overall analysis. Field notes were not taken, there were no dropouts and no repeat interviews were required. Data collection concluded upon reaching saturation, whereby no new topics emerged in two consecutive interviews. Additionally, the research group confirmed adequate representation, ensuring complete heterogeneity as previously described. Transcripts were not returned to the participants.
Data analysis
Semi-structured interviews were recorded, transcribed, pseudonymised and analysed iteratively. An external company transcribed the interviews. Transcribed data were coded in Atlast.ti 22, applying a mixed deductive-inductive approach. Initially, a deductive coding strategy was employed based on the CFIR, whereby data segments were categorised under predefined CFIR constructs. However, when relevant content did not align with existing CFIR categories, an inductive approach was adopted. New subthemes were developed to capture these insights while still aligning thematically with the overarching CFIR domains. For example, within the ‘intervention characteristics’ domain, distinct subthemes were created to reflect different interventions. These refinements reflect the inductive nature of adapting the coding scheme based on empirical findings. This part of the analysis is elaborated in a separate paper focusing on network care. To ensure reliability, two researchers (MC and SvS) independently coded all transcripts, comparing codes until consensus on code names was reached. Differences in inductive coding were discussed and resolved. If disagreements persisted, a third independent researcher (RCV) was consulted. This was followed by the identification of the main barriers and facilitators in communication and collaboration within the interprofessional interface. No feedback of participants regarding end results was provided.
Patient and public involvement
None.
Results
A total of 37 primary and secondary healthcare professionals, who were considered key users of collaboration services (eg, referrals, telephone communication) were interviewed. Of the 39 GPs originally approached, 14 were interviewed, while 23 of the 29 specialists we approached were eventually interviewed (see table 2). The analysis identified essential requirements for change based on the expectations and needs of healthcare professionals concerning collaborative services. Table 1 presents these various communication methods (collaboration services) in a structured format, highlighting differences across multiple dimensions, including the direction of communication, response time and the mode of support data transfer. The responses were first categorised according to the CFIR domains (Individual Characteristics and Inner Setting—Tension for Change). To enhance clarity and enable analysis throughout the full primary–secondary care collaboration process, responses were further subdivided according to specific operational stages within the collaboration process, such as consultation or follow-up (figure 1). This was followed by identifying overarching themes related to the answers provided and grouping them accordingly. This approach resulted in four key domains with subthemes, subdivided per operation and CFIR domain (online supplemental file 3). The fourth key area addresses main barriers that may result in patients not receiving adequate care during transitions between primary and secondary care or vice versa, which might be a result of needed improvements in the first three domains.
The following key domains were identified (online supplemental file 3):
-
Software and record integration.
-
Seamless personal interaction.
-
Eliminating a sense of ‘us vs them’.
-
Gaps in continuity of care.
Software and record integration
GPs and specialists require accurate and easily accessible information. Several software-related issues were mentioned in CFIR domains Inner Setting—Tension for Change and Individual Characteristics—collaboration and expectations. This is particularly relevant to the operational steps referral and discharge, in relation to the digital collaboration service used for referrals (figure 1).
Deficiencies of referral software
GPs reported shortcomings in their referral software, including the failure to include all available lab results from the patient record and the cumbersome manual process involved when attaching correspondence or background information (online supplemental file 3, table 1, Q=quote: Q2-GP3, Q6-GP11). This issue is illustrated by the following comment from a GP:
Q2-GP3: “The main handicap […] when you refer from a selected episode with [referral software] is that only the lab results from that episode are included. And that, of course, sometimes makes no sense […].” (Previously collected laboratory values may still be relevant, including values associated with a different episode of care)
Conversely, specialists mentioned a perceived lack of information in referrals, such as the absence of correct laboratory results and additional clinical information (online supplemental file 3, table 1, Q7-S25, Q8-S19, Q20-S19, Q21-S20). This concern is exemplified by the following specialist’s observation:
Q8-S19: “Then it (the referral) says ‘anaemia’. So I look for an HB level, but instead I find cholesterol and blood sugar mixed in with a bunch of other lab results. But the HB – the one thing that actually matters – is often missing. And then I think, well you know, it’s just not a (complete) set of information you’re providing […].
This issue acts as a barrier and hampers understanding among professionals, undermining collaboration (online supplemental file 3, table 1, Q1-GP2, Q16-S31, Q17-GP1, Q3-S36, Q9-S21, Q10-S32).
Real-time up-to-date information
Inaccurate data synchronisation and software integration were often-mentioned barriers to collaboration. Improvements in information and communication technology are vital for GPs and specialists because current manual processes of data transfer risk information loss, such as delayed discharge letters with insufficient and outdated information for primary care follow-up. This concern was voiced by a GP as follows:
Q15-GP6: “Then you miss out on a lot of information because these systems don’t talk to each other. That’s obviously a huge shortcoming […].”
A specialist similarly highlighted the issue from their perspective.
Q14-S36: “The back end of the collaboration is clearly incredibly poor.”
Another deficiency concerning real-time, up-to-date information exchange was infrequent letter-based communication. Patients often see GPs in parallel with specialists. Immediate data sharing would enable professionals to promptly adapt to each other’s current treatment strategies. At present, there is no official service for brief updates within the care continuum, meaning that GPs must rely on sending brief updates via teleconsultation or via an (un)secure email. Specialists have two options for updates: either to send an additional letter or make a call, although neither of these options is commonly used. To resolve this issue, a shared electronic health record has been proposed, but at the very least compatible systems are needed that allow professionals to access a patient’s current treatment. The importance of such access is underscored by the following GP’s reflection:
Q18-GP13: “Sometimes you see (as a GP) the patient much more often, for example in between consultations (in secondary care) for other health problems. But you are often unaware of what is going on in secondary care. […] That would be valuable. […] And, yes, to promote some sort of team sentiment.”
Seamless personal interaction
Our findings highlight the need for seamless personal interaction between healthcare professionals. This was mentioned in the CFIR domain Inner Setting—Tension for Change, across all operation steps related to interprofessional collaboration (referral, consultation service (phone call), treatment/discharge, follow-up and overall process). Similar needs were also observed within the individual characteristics domain, particularly in relation to the use of the consultation services phone and chat applications.
Missing interactive bidirectional consultation
Both GPs and specialists emphasised the critical importance of an interactive bidirectional consultation service for professionals. Bidirectional consultation is considered essential for exchanging a patient’s medical history, treatment information, follow-up details and timely updates on specific cases. The professionals declared that it would foster shared responsibility and better collaboration, particularly concerning patients with chronic conditions. The existing shortage of continuity in healthcare professionals (due to part-time work, temporary replacements and staff shortages) complicates clear, direct communication, as well as the embedding of umbrella agreements (online supplemental file 3, table 2, Q26-GP13, Q24-HA19, Q27-S11). One specialist reflects on the impact this has on informal collaboration:
Q23-S24: “[…] I really miss that, the quick switch between the GP and the specialist.” (i.e. the ease of initiating peer-to-peer discussions outside of formal referral pathways.)
Currently, three channels of bidirectional consultation with rapid reply are available: (1) live telephone consultation, (2) ‘Siilo’, a secured chat feature on private telephones, and (3) ‘Zorgdomein Chat’ (see table 1). However, none of these options fully meets the professionals’ specific needs. Although telephone consultation most closely resembles the preferred model of collaborative service, GPs often find them unsatisfactory. Difficulties include challenges in directly reaching specialists due to mismatched schedules, problems getting in contact with the responsible caregiver in secondary care and the frequent lack of a direct contact number (online supplemental file 3, table 2, Q34-GP3). The following GP comment illustrates the frustration with current communication channels:
Q31-GP9: “… (consultation by telephone) that can also be annoying, because sometimes it takes a long time to get hold of someone. And sometimes you would prefer not to bother a specialist.”
Conversely, specialists also encounter challenges when trying to reach GPs by telephone. Delays due to an extensive menu of options and queuing result in specialists refraining from calling, hanging up or using emergency numbers. Furthermore, when specialists manage to contact an assistant, in some instances the GP is either unavailable or fails to return the call (online supplemental file 3, table 2, Q31-GP9). Several specialists describe the practical barriers that discourage further attempts to consult the GP, for instance:
Q32-S26: “Look … the queue (with all the dialling options) at the GPs … that’s a huge hurdle for specialists. That’s actually the barrier that often leads us to avoid consulting the GP.”
The second option (secured chat app ‘Siilo’) meets several needs of the professionals. However, it lacks official integration into the electronic health records, requires manual data input, operates on private phones (resulting in 24/7 after-work hours availability) and currently there is no financial compensation for using this service (online supplemental file 3, table 2, Q38-S17, Q39-S33). The third option (‘Zorgdomein’ Chat) is barely used due to difficulties in reaching it, a lack of integration with patient records and a lack of widespread adoption. A specialist reflects on the unintended workload that can arise from alternative communication platforms:
Q38-S17: “The only disadvantage of Siilo for us is, it sounds a bit lame, but if people find you through Siilo, you will get a lot of questions. And of course, there is no payment or anything in return. So, then you’re doing a lot of work (for free). You actually do teleconsultations via Siilo.”
Eliminating a sense of ‘us versus them’
The urge to eliminate the sense of ‘us vs them’ was mentioned in the CFIR domain Inner Setting—Tension for Change and Individual Characteristics (collaboration, expectations). It is notable that this need appears at almost every step (operation) of the care continuum between primary and secondary care (processes).
Shared responsibility
Both GPs and specialists highlighted the need to move away from isolated approaches and prioritise shared decision-making and shared responsibility in treatment planning. To guarantee continuity of care, this is especially important when developing new treatment plans for patients with chronic conditions. Both professional groups express a willingness to adopt interventions recommended by the other party. However, challenges include a lack of understanding of each other’s needs, a lack of trust and differences in treatment guidelines that may result in interruption of the continuity of care (online supplemental file 3, table 3, Q46-GP1, Q51-GP13, Q52-GP9, Q56-S29). This need for mutual agreement is emphasised by a GP:
Q46-GP1: “What I think is, that we should agree much more with each other (i.e. primary and secondary healthcare professionals about shared patients).”
Another GP illustrates the potential for more proactive collaboration:
Q52-GP9: “What should happen more often is that the specialist calls and says […] what are you doing, GP? Or should we approach it this way?”
To strengthen shared responsibility, both parties propose granting patients access to one single specialist consultation as part of their primary care treatment. For chronic conditions, they suggest incorporating an annual teleconsultation—consisting of one digital question and one response—into the integrated chronic patient primary care programme. Gynaecologists and paediatricians, in particular, express strong enthusiasm for multidisciplinary team meetings with primary care; however, these meetings do not currently include GPs (table 1).
Awareness of preferences
Specialists stressed the facilitation of mutual awareness of each other’s preferences via referral documentation and discharge letters. They preferred clear referral questions with a good summary and expected more involvement of GPs in palliative care. Conversely, GPs preferred complete specialists’ consultations, including diagnostic tests, which would enable them to report back to a patient and explain the tests conducted. However, turning these ideas into reality is challenging. Some GPs felt that specialists have little interest in primary care and that the opportunities to discuss treatment for regular patients are insufficient. Equally, specialists mention that some GPs send unstructured referrals without clear summaries (online supplemental file 3, table 3, Q54-GP1, Q55-S15, Q57-GP11, Q59-GP3, Q61-S30). One specialist illustrates the frustration caused by vague or overly lengthy GP records:
Q44-S25: “[…] then you just have so little information to go on, and then you get this whole SOEP (GP record), this three-page summary of SOEP. And it just says, ‘chronic abdominal pain’, and then I have to make sense out of all these things … Yeah, I think it’s terrible.”
Another specialist underscores the importance of well-written referrals for patient care:
Q55-S15: “And I also think that you are failing your own patient as a GP if you do not write a good referral.”
Q59-GP3: “The receiving side (specialist) is often not aware of what the GP sees. So, you would actually wish that the specialist could see how things are handled on the GP’s side. They really don’t know.”
For GPs, a major barrier to collaboration with outpatient clinics is organisational inefficiency. They expressed frustration over referred patients being advised to ‘call their GP for an emergency referral’, which puts GPs in a difficult position regarding their doctor–patient relationships. Additionally, GPs highlighted inconsistencies in follow-up arrangements between outpatient clinic assistants and specialists, often resulting in patients turning to their GP for further guidance (online supplemental file 3, table 3, Q43-GP10). One GP illustrates this frustration with the following remark:
Q43-GP10: “And especially the outpatient assistants who say: ‘oh yes, if it is urgent, you can call the GP, they can arrange it urgently.’ I find that very annoying.
Transfer of responsibilities
A significant challenge in primary–secondary care collaboration is the transfer of responsibilities from secondary to primary care. These responsibilities are often communicated by correspondence without consulting the primary healthcare professional. According to GPs, the practice of secondary healthcare professionals directing orders to primary care professionals should be abandoned (online supplemental file 3, table 3, Q49-GP6, Q50-MS17, Q53-GP9, Q60-GP7). One GP expresses this concern by highlighting the lack of coordination in follow-up planning:
Q48-GP1: “[…] if you have changed something, why don’t you [the specialist] give that form to the patient? Schedule it – renal function check in one week’s time – and put it in your own day planner […].”
Another GP reflects on the discomfort caused by unilateral instructions in discharge letters:
Q49-GP6: “[…] at the bottom of the letter is often ‘action by the GP’ which is supposed to be some sort of instruction, like ‘this is for you’. And you think, ‘Yes, but maybe I don’t have any options for that. […]. I find that difficult, like I’m now apparently to do this, but there has been no interaction about it.”
From a different perspective, specialists feel that including action points in discharge letters helps guide GPs and reduces their burden. One specialist acknowledges the good intentions on both sides, while also pointing to the need for better communication:
Q50-S17: “[…] I think both parties are actually always very well-intentioned, but sometimes, because of the communication we have now, some friction can arise […]. So, I would suggest an accessible way of communicating with the primary care and basic care.”
Organisational obstacles
Professionals also often highlighted organisational difficulties that limit the collaboration. For instance, the fact that decisions on service redesign and implementation of new agreements and interventions are made by small working groups and committees. Even when these decisions reach the process of secondary care (specialists) through the hospital governance structure, integrating them into the process of primary care (GPs) poses challenges, especially across all the individual practices in The Hague. Despite the existence of a central GP organisation, collaboration is hindered by a lack of mutual agreement on issues such as lack of feedback on the content of the referral and discharge letters, clearly defined roles in who is responsible for what and policy on triage and appointments in secondary care (online supplemental file 3, table 3, Q45-G12, Q47-S11, Q58-S22). One GP illustrates the consequences of delayed communication from specialists:
Q45-G12: “[…] the specialists often wait before sending a letter, I understand that […] but sometimes a patient is in a huge process and then it is difficult, because then they come to me and then they actually expect me to be aware of what they are into and I don’t know that at all.”
A specialist highlights the importance of direct contact over relying solely on organisational structures:
Q58-S22: “[…] Because GPs sometimes also feel that we don’t want something or that we are pushing it off or vice versa […]. We just want to work well together. […] that goes much easier if you have contact with each other more often, rather than relying on certain organizations.”
Gaps in continuity of care
The previous three domains may lead to situations where continuity of care is no longer guaranteed and creates a lack of responsibility regarding the patient. These issues were highlighted in the domains of Inner Setting—Tension for Change and the Individual Characteristics (Collaboration), primarily affecting the operational steps at the beginning (referral) or at the end (discharge/follow-up) of the process (figure 1).
No secondary care appointment
If a GP refers a patient who then fails to schedule a specialist appointment for whatever reasons, such as anxiety or language barrier, the GP will not be informed. Because hospitals often require patients themselves to arrange appointments, uncertainty arises regarding responsibilities for patient care. One GP, working in a deprived area, illustrates the consequences of this gap in communication:
Q62-GP11 (working in a deprived area in which not everyone always speaks Dutch): “[…] You have already put a lot of effort into making sure that you plan is clear to the patient […]. Then sometimes someone will return after a year … And then you think to yourself, ‘Really? You haven’t seen the specialist?’”
Errors in response to teleconsultation
There is no automatic reminder in the GP system to notify the GP of a reply to a teleconsultation. As this is done manually, it can lead to mistakes: (1) no one notices that a response has not been sent, resulting in no follow-up. (2) There is a response, but it is not acted upon due to the volume of incoming correspondence or the number of GPs working on a case.
Letter sent to the wrong GP
Discharge letters from hospital admissions or secondary care outpatient departments are frequently sent to the wrong GP. GPs often receive such letters, along with results and follow-up care information of patients not registered with them. This poses privacy issues and can lead to gaps in patient follow-up.
No follow-up after discharge
Specialists often include recommendations and outline follow-up actions in discharge letters. However, GPs, who are frequently overwhelmed by the volume of correspondence, noted the risk of missing specific instructions due to the need for manual processing of correspondence. While most specialists assume their recommendations are acted upon, the absence of fast and efficient communication with the GPs, combined with a lack of shared responsibility, creates uncertainty about whether the requested follow-up actions are being carried out (online supplemental file 3, table 4, Q64-GP9, Q65-S29, Q67-GP13). One GP illustrates how delayed correspondence can undermine timely follow-up:
Q63-GP10: “Recently I received a letter: Apologies that this letter is late. It was actually a month late. And at the bottom of the letter, it said: (thus a month ago), ‘Please check magnesium levels in two weeks’ time’.”
Another GP reflects on the risk of important instructions being overlooked in the daily workflow:
Q67-GP13: “So yes, you hope that they can process the mail and also extract those instructions and write that down, […] it can sometimes slip through. Or, let me say: that has happened sometimes.”
Lack of continuity of care for chronic patients
GPs often face challenges when patients with chronic conditions transition between primary and secondary care, as this process requires re-registration in integrated care programmes. To improve continuity of care and streamline patient management, there is a clear need to integrate secondary care into these existing programmes. A specialist illustrates the lack of feedback and uncertainty that can arise in this process:
Q65-S29: “But you would actually like to hear back, like, ‘Okay, I’ll pick it up. Now it just goes into a black hole. And then you just hope that things go well.”
Discussion
This study identified barriers and proposed changes across four key domains, from a user perspective by primary and secondary healthcare professionals, that deserve consideration when redesigning the primary–secondary care continuum: (1) software and record integration, (2) seamless personal interaction (3) eliminating the sense of ‘us vs them’ and (4) gaps in continuity of care. Barriers in the first three domains may contribute to challenges in maintaining continuity of care, as identified in the fourth domain.
Understanding healthcare professionals’ needs and preferences, as well as assumptions, is crucial to enhancing collaboration across the entire primary–secondary care system. The service design perspective underscored the importance of involving healthcare professionals as users when redesigning collaboration services. Previous research highlighted the importance of considering the user perspective when seeking to enhance collaboration between primary and secondary care and has emphasised factors like personal and social considerations and the pivotal role of GPs in identifying locally relevant and feasible change.34 35 Various studies have emphasised the role of professionals in enhancing collaboration at this care continuum, indicating that success hinges on frontline professionals.20 21 28 35 36 While behavioural change and the collaboration of frontline healthcare professionals are both crucial, in our opinion solely relying on behavioural change will not guarantee a successful shift towards high-quality collaboration. Conversely, little structural attention is paid to the user experience of professionals on both sides of primary and secondary care of the interprofessional interface on a practical level. Our findings highlight that healthcare professionals in both primary and secondary care frequently encounter similar obstacles, particularly due to system-level shortcomings and inadequate collaborative support tools. These limitations prevent them from delivering the levels of care they aspire to. Participants expressed a strong need for closer and more direct communication, particularly in the management of chronically ill patients, where up-to-date information and accessibility are crucial. Importantly, the preferred model of collaboration appears to be shared across care domains; however, it is not sufficiently supported by existing systems and infrastructures. Frustrations regarding each other’s work methods—often rooted in fragmented or poorly integrated software systems—were evident. In primary care, GPs reported dissatisfaction with patients being returned accompanied by orders, recommendations and redistributed responsibilities, which contribute to increased workload and perceived imbalance in the interprofessional relationship. Investment is needed in a bidirectional communication system that allows for horizontal, peer-level consultation, supported by accurate and timely exchange of patient information to enable seamless personal interaction and a more equal way of collaboration. However, improving collaboration requires more than the development of a new communication tool. Unlike existing research, which often focused on specific aspects, for instance, primary or secondary care, specific diseases, patient categories or services,8 17 37 our study identified underlying structural needs across the process of primary and secondary care and all the operation steps, which underscores the imperative of avoiding the isolated redesign of specific services used for collaboration between primary and secondary care. The healthcare system operates as a complex adaptive ecosystem, where each operation significantly influences the next (figure 1).38 A critical re-evaluation of the broader system and its underlying processes across the many interconnected facets within the care continuum is essential. Focusing on a single operation step in isolation risks perpetuating dysfunction elsewhere. A meta-level approach is needed—one that addresses the system as a whole, rather than its individual components. Interestingly, and in contrast to literature reports, financial constraints were not cited as barriers to collaboration by the professionals in our study,14 35 39 suggesting a gap between policy discourse and frontline perceptions of healthcare collaboration.
It is also important to consider how new implementations and behaviours might integrate into the dynamic healthcare landscape. Awareness of the continuous alignment and mutual influence of processes is therefore key to success. We hope that professionals and policymakers will clearly define requirements and prioritise future solutions to these barriers before implementing new interventions.
When redesigning the primary–secondary interprofessional interface from a service design perspective, we advise taking all user needs into account, including for those of patients, policymakers and insurers.31 Questions that remain unanswered at this point include prioritising of addressed user needs over existing agreements, as well as issues related to a currently rigid system. Exploring the potential benefits of an interactive digital interprofessional consultation system is therefore crucial to the successful enhancement of collaboration between primary and secondary care.
Considering the aim of our study, we focused on detailed, context-specific insights from both primary and secondary care professionals, using implementation frameworks and a service design perspective to engage healthcare professionals as end-users in exploring collaborative care processes. This can be considered a methodological strength. However, when interpreting the study results, several barriers need to be considered. Using the CFIR framework, we initially analysed new collaboration service interventions such as teleconsultation and joint consultation. However, due to the substantial volume of responses regarding network care related to the CFIR domain ‘Inner Setting—Tension for Change’, we conducted a separate analysis to ensure clarity. Consequently, this article does not cover all CFIR domains. Instead, it focuses mainly on the barriers and facilitators affecting collaboration between GPs and specialists. Perspectives beyond those of the selected professional groups and settings are not included, which limits the scope of identified improvements. The focus of this study was on understanding implementation processes rather than measuring outcomes. Besides, as with most qualitative explorative research, the findings are not intended to be generalisable beyond the specific context in which the study was conducted. Also, our recruitment strategy in which GPs were primary contacted through personal networks could have introduced sample bias, as those more engaged or interested in the topic may have been more likely to participate. Contacting GPs proved difficult due to their heavy workload and limited accessibility. Interviews were therefore conducted in a variety of settings convenient to the interviewee, which may have influenced the interview. The study is geographically limited to one city, which might impact generalisability. On the other hand, barriers related to electronic health records and related assumptions are likely experienced more widely and are therefore context independent.
Conclusions
The findings suggest that healthcare professionals in both primary and secondary care frequently encounter similar obstacles in collaboration, particularly due to system-level shortcomings and inadequate collaborative support tools. This leads, among other things, to increased workload, miscommunication and ultimately impacts the quality of care. The study stresses that improving collaboration between GPs and specialists cannot be achieved by simply adjusting or renewing individual services; rather, it requires a comprehensive rethinking of the entire referral and back-referral process. The key domains, ranging from technological improvements to behavioural changes, are deeply interconnected within the entire collaboration. A broader, more integrated approach is essential for improving the overall quality of patient care, streamlining workflows and ultimately enhancing health outcomes.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants. Ethical and research governance approval for this project was obtained by a waiver from the Medical Research Ethics Committee (reference number/ethics ID: N21.093), judging the project not subject to Dutch law on medical research. All participants received information and gave oral informed consent to participate in the study and to be audio-recorded before the interviews. Concerning the interviews by telephone or video, written informed consent was obtained later because it had to be sent digitally, which included agreement to be interviewed and the use of anonymous quotes in publications.