Causes and consequences of rejected or returned referrals from general practice

Effective healthcare comprises collaboration between primary and secondary healthcare, where general practitioners (GPs) handle most patient contacts and act as gatekeepers to secondary healthcare [1]. An increasing proportion of referrals from GPs to hospitals has been returned. This may pose a threat to the continuity of patient trajectories across healthcare systems [2-4].

The desire to provide more care in general practice has brought about reimbursement schemes such as capped hospital budgets, which are likely incentivising more restrictive referral management [5-8].

Since 2001, Danish hospitals have employed GPs as part-time liaison officers to facilitate cross-sectoral collaboration. Liaison officers assist their GP colleagues in communicating with hospital departments to ensure seamless patient pathways, and they regularly meet with chief physicians [9].

This study aimed to uncover the causes and consequences of referral returns from the GP liaisons’ perspective.

Methods

Design

This was a qualitative study conducted in January and February 2024. Individual in-depth phone interviews explored the GP liaisons’ experiences and attitudes towards returned referrals in the Region of Southern Denmark, which has a population of 1.2 million citizens. The design was guided by the Consolidated Criteria for Reporting Qualitative Research (COREQ) checklist.

Setting

In Denmark, most healthcare services are tax-funded. Nearly all citizens are registered with a general practice [1]. Outpatient hospital care requires a referral, usually issued by the GP [1]. A hospital physician triages the referrals to determine the appropriate specialist, the necessary initial investigations and the acceptable delay, or whether the referral should be returned to the GP. An extended description of the setting and the liaisons is shown in the Supplementary Material.

Participants

We used purposive convenience sampling to ensure variation among liaisons from all five hospitals in the region, equal gender distribution, differing levels of experience as both GP and liaison, and inclusion of both single-handed and group practices.

Participants were invited via email, and we concluded recruitment after ensuring predefined variation criteria, resulting in 20 out of 36 eligible liaisons participating. Demographics are shown in the Supplementary Material.

Data collection

The interviews followed a semi-structured interview guide that was developed and pilot-tested by the authors. The liaisons were provided with the guide in advance. After the initial two interviews, adjustments to the interview guide were considered, but no modifications were deemed necessary. The first author conducted the interviews. These were transcribed, and the transcripts were sent to each informant for correction and approval before analysis. The guide and detailed analysis are included in the Supplementary Material.

Qualitative analysis

We conducted a systematic text analysis as described by Malterud [10]. Initially, an overview of the data was established by reading the compiled text. Two authors (ME, NK) then individually generated preliminary themes and then agreed on final themes. Meaningful units in the text were identified and sorted into code groups. Each code group was reviewed and divided into subgroups (Figure 1). The subgroups were merged into subtheme headings around newly emerged core themes. The themes and subthemes were further developed through reflective discussions among all authors, ME, NK, JL and CBM, all of whom are experienced clinicians and researchers representing both general practice and hospitals. This approach ensured that the overall findings remained valid and preserved the integrity of their original context. The first 16 interviews were analysed before the final four to test for data saturation.

Trial registration: not relevant.

Results

The thematic analyses consolidated three themes and six sub-themes (Figure 1).

Patients

Effect on patients

The liaisons reported that returned referrals affected patients differently. In some cases, it had no adverse consequences as the health issues were resolved within general practice or by another healthcare provider. In other cases, the return of the referral caused delays in diagnosis and treatment, led to prolonged and complex healthcare trajectories, and occasionally had considerable health and social repercussions for the patient.

The liaisons argued that hospitals primarily focus on diagnosis, whereas the individual context of the patients’ health problems receives less attention.

For some liaisons, the impact of rejected referrals on social inequality was a considerable concern. Socially deprived patients faced uncertainty and prolonged waiting when referrals were rejected, whereas more privileged patients often accessed private care, improving their access but potentially increasing the risk of unnecessary procedures.

“... some well-off patients turn to the private sector, while the less resourceful are left in a limbo of uncertainty.”

Non-standardisable problems

The liaisons commented that referral criteria were designed for standardisable patients with clear and well-defined health issues. This challenged referrals of patients with atypical symptoms, multimorbidity or patients who were difficult to diagnose. Liaisons reported that hospital departments increasingly retreated from the initial diagnostic processes, occasionally forcing GPs to admit patients inappropriately to the emergency department for diagnostic assistance.

The liaisons called for hospital functions with broader competencies and closer collaboration on diagnosing patients with complex or unclear conditions. They suggested implementing a “red button” in referrals, which the GP could activate in case of high-risk patients, ensuring that a limited number of patients could not be denied access to hospital care.

”It is increasingly difficult to refer patients with non-specific symptoms; those who fit a standardised pathway benefit, while the complex patients risk repeated rejections”.

The GP-patient relationship

Returned referrals, which are accessible online, may challenge the GP-patient relationship. The patient might experience mistrust or question the GP’s work if the wording of the return letter suggests that the referral was poorly or inadequately written.

However, some liaisons noted that the GP-patient relationship may also occasionally benefit from returned referrals, as it confirms to overly demanding patients the irrelevance of specialist involvement.

Organisation

Division of tasks

Return of referrals transfers tasks from hospitals to general practice. Therefore, some GPs felt overwhelmed, while others accepted more tasks provided that financial resources followed. They agreed that, given sufficient resources, some new medical tasks could be managed within general practice. However, other tasks were considered impractical to handle in this setting due to their rarity, complexity or the lack of access to necessary tests and equipment.

The liaisons reported that some referrals were returned because specific departments were facing workload challenges. Anticipating rejection, some GPs became less inclined to submit similar referrals.

The liaisons called for flexible referral processes, focusing on dialogue rather than causing cycles of returns and re-referrals. They recommended that departments should request additional information from patients or GPs instead of returning referrals. They questioned why some departments returned referrals for potentially critically ill patients, opting instead for an alternative patient pathway, particularly when that pathway was managed within their own department.

The liaisons reported that some departments suggested referring patients to other departments, only for the receiving departments to return the referrals as well, leaving the patients caught in a loop. They called for more coordinated, accessible and mutually agreed-upon referral criteria, acknowledging the cumulative information burden placed on general practice by multiple receiving departments.

Collaboration

Access to specialists

The liaisons generally praised the good collaboration between GPs and most departments. They highlighted the availability of telephone consultations with specialists as a positive development that helped reduce referral returns. They also noted that the return letters had improved in quality and more often described feasible alternatives to hospital care. But a few departments still have a culture unsupportive of collaboration.

From patient transfer to dialogue

The liaisons reported that the quality of GPs’ referrals had improved recently, but noted that the increasing demand for irrelevant details was leading to more returns. They understood the need for thorough referrals but expressed frustration over excessive “checklist-mania” that did not intuitively contribute to qualify triage.

The liaisons had observed a noticeable shift towards management of more patients within general practice, guided by specialist advice, rather than transferring the patient to a hospital. They expected that this approach would transform referrals into professional dialogues about patient care. Some liaisons suggested referrals be labelled as “inquiries” to emphasise the collaborative nature of the process, adding that while this would strengthen the professionalism in general practice, it would probably also keep the responsibility for more patients within general practice, leading to increased workload.

However, an overly restrictive admission culture could lead to an unhealthy dynamic between consultants and GPs, where GPs might feel that they needed to 'plead for treatment' for their patients, disrespecting their role as healthcare gatekeepers.

The liaisons reported that referral return letters can be valuable learning opportunities for GPs. Some liaisons reused plans learned from previous return letters, but also stated that "development through rejection" can be frustrating. It was mentioned that return letters may teach GPs to write referrals that include irrelevant but requested information while omitting relevant information that would expose the complexity of the case and thus increase the risk of a return.

Discussion

Summary of findings

The benefits of returned referrals included the facilitation of cost-effective healthcare and enhanced GPs’ professional expertise, while the harms included delay, reduced quality and complication of some affected patients’ healthcare. Referrals of patients with non-specific symptoms, whom the GPs found difficult to diagnose, were returned more frequently than other referrals. Returns also affected patients from rural areas and patients of low social status, thereby increasing inequity in healthcare. In some cases, returns could cause mistrust between GPs, hospitals and patients and impair the GPs’ functioning as gatekeepers. According to the liaison GPs, the increase in returned referrals was primarily driven by four factors: increased specialisation of care; standardised patient pathways that do not accommodate the growing population with complex health problems; resource constraints in hospital departments; and economic incentives for hospitals to shift healthcare tasks to general practice.

The liaisons suggested that, to reduce returns, some referrals could be substituted by improved cross-sectoral collaboration based on dialogue between the advising specialists and GPs. However, this requires adequate resourcing in general practice and a willingness among hospital departments to acknowledge their shared responsibility for care delivered beyond hospital walls. Patients with non-standardisable problems should be prioritised, and it should be possible to designate certain referrals as non-returnable.

Strengths and limitations

This study was based on qualitative data and therefore cannot be used to quantify problems related to referral returns. We ensured a broad variation of the liaison officers, developed and followed an interview guide and conducted tests to ensure data saturation, which may be seen as strengths. However, it remains unknown which of the identified views are shared by or could be nuanced by the triaging hospital consultants or the patients. The liaisons also work as GPs, which may make it difficult for them to distinguish between insights gained in their role as GPs and those gained as liaisons. There is also variation in the liaisons’ personal experiences and professional contexts. However, we consider this a strength, as this qualitative analysis aimed to explore potential pros and cons related to referral returns.

The study was conducted in a single Danish healthcare region. However, comparison with international experiences suggests that the findings may be cautiously transferable to other settings.

Comparison to other studies

In the UK and Switzerland, inflexible patient pathways also put frail patients with complex health problems at risk of rejection [2, 4, 11, 12]. Half of patients subsequently diagnosed with cancer initially attend general practice without symptoms meeting the criteria for the specific cancer pathways and risk that their referrals are returned [13-15].

The Royal College of General Practice and others support the liaisons' perception that referral returns can aggravate health inequity [4, 16, 17].

The risk of undermining the doctor-patient relationship is also a concern in Norway and the UK when shared decision-making between GPs and patients is overridden by third-party decisions [11, 12, 18].

The Danish GP liaison officers found that specialist advice centred around the patient with the GP as a coordinator was highly beneficial. In the UK, the reported experiences were more nuanced, and limitations have been identified [19]. The agenda behind the advice setup probably reflects the perceived outcome. The liaisons suggested that hospital departments should transfer patients to fellow departments rather than return them to the GP. However, other studies have found this approach questionable and possibly leading to unnecessary specialised care and a lack of coordination [20].

Implications for research and/or practice

The overall quality of patient care can undoubtedly be improved by fostering a more collaborative and respectful approach to referrals from both GPs and hospital consultants, e.g., by expanding timely available specialist advice. A critical review of referral criteria and an enhanced focus on frail and complex patients with non-standardisable problems should be considered. Further research is needed to quantify the underlying reasons and consequences of referral returns.

Ethical approval

All participants received written and verbal information about the study and consented orally to participate. The participants were not paid by the research project. According to Danish legislation, this project needed no approval from the National Scientific Ethical Committee. As the study was based on a small sample of informants, we deleted all wording from statements and quotations which could theoretically identify the informants.