To which non-physician health professionals do French general practitioners refer their patients to and what factors are associated with these referrals? Secondary analysis of the French national cross-sectional ECOGEN study

The increasing prevalence of chronic diseases, multimorbidity [1, 2], and widening health needs [3, 4] is creating challenges in primary care, causing it to continually expand and change. In the context of this changing role, well-organised primary care helps reduce costs, improves user satisfaction, population health outcomes and equity, and strengthens health system performance [5]. Considering this, numerous countries have initiated health system reforms built around restructuring primary care provision, integrating patient care pathways, and enhancing interprofessional collaborations [6‐9].

General practitioners (GP) play a key role in most health care systems since they are often responsible for coordinating patient care pathways, including referrals to non-physician health professionals (NPHPs) [10] and both the primary and secondary care sector [11]. In some countries they act as gatekeepers with GP referral validating health insurance reimbursement for consultations with other health professionals, including some NPHPs.

GP referral to NPHPs is a key indicator for multiprofessional practice [12]. Examining GP referral frequency to different NPHPs, and factors associated with referrals, can provide information about how primary care is organised and works.

Despite growing interest in developing multiprofessional practice in primary care, literature about NPHP referrals in general practice is still very scarce, apart from studies on nurse referrals or referrals for specific chronic diseases [13‐17]. This knowledge would be useful for policymakers and all those involved in planning for future healthcare workforce requirements and developing multiprofessional practice in primary care teams [12].

NPHP referrals in general practice probably differ depending on the health care system. Therefore, any data on this topic should be carefully interpreted according to the national context, before being compared with data from other countries when available. In France, as in many other health care systems, GPs, nurses, and pharmacists form the core of primary care, along with various other health professionals, such as dentists, physiotherapists, midwifes, or podiatrists [18, 19]. A 2009 paper described the French primary care system as a “professional non-hierarchical model”, like Germany or Canada, characterised by low level territorial organisation, coexistence of different practice types (solo practices, mono- and multiprofessional groups, including GPs, specialists, and NPHPs) and mainly private practices. This description was compared with “professional hierarchical models”, where GPs have a strong, long-standing, and more formal gatekeeping role (such as in the UK, Netherlands, Australia, or New-Zealand) or “normative hierarchical models”, where primary care is legally defined and based on multiprofessional territorial organisations (such as in Catalonia in Spain, Finland, or Sweden) [20]. Since then, French health policies and professional leaders have supported implementing multiprofessional practice and organisations in primary care, resulting in approximately 15% of GPs practicing within interprofessional health care teams, and a majority being engaged in various forms of multiprofessional practice. Furthermore, in France, GP referrals validate health insurance reimbursement for consultations with some NPHPs such as physiotherapists, nurses, speech therapists or podiatrists, while no GP referral is required to see pharmacists, dentists or midwifes. Except for specific situations, care from dieticians, psychologists or osteopaths is not reimbursed and patients must cover the full cost.

Building on original, nationwide, practice-based observational data, this study aims to describe French GP referral frequency to various NPHPs and identify factors associated with referral rates.

The ECOGEN database contains 20,613 consultations, 45,582 consultation results and 98,847 healthcare procedures.

Our findings show that French GPs refer 6.8% of patients to NPHPs, with physiotherapists, podiatrists, and nurses accounting for 85.3% of referrals. We found higher referral rates are associated with older, retired patients, with multiple health problems, and longer consultation durations. Specific associations and trends were observed for referrals to nurses, physiotherapists, and podiatrists.

To our knowledge, no data are available from other countries that may be directly compared to ours in terms of referral rates. Publications concerning NPHP referrals mostly rely on declarative data, which are subject to different biases (such as memory or social desirability), rather than observational practice-based data, and do not report consultation-scale data that would allow GP referral rates to be estimated [24, 25].

In France, the health system requires a GP to prescribe physiotherapist, podiatrist, or nurse treatments which the national health insurance then reimburses. This would explain why GPs refer to these three NPHPs most often, at least explicitly, as opposed to a pharmacist, a midwife, or a dentist for whom formal referral is not needed. GPs in France prescribe medication to patients in 80.7% of consultations, making pharmacists possibly the most referred to profession [26], but this referral is not explicitly discussed with the patient. Importantly, the GP referral process does not cover visits to nutritionists, psychologists, or osteopaths, which the French national health insurance will not reimburse. Patients most often decide to consult these non-reimbursed professionals either upon GP advice or through a self-referral process explaining the lower formal referral rates to these NPHPs in our study.

In addition to the issue of reimbursement, there may be other barriers to patients consulting these non-reimbursed NPHPs. These barriers include NPHPs, such as psychologists and dieticians, often being less accessible in comparison to other health professionals since they have shorter opening hours, rarely perform home visits and there are fewer of them meaning distribution is reduced [27]. Furthermore, acceptability may be an issue for some patients since there is still stigma surrounding eating or weight disorders and mental health problems and patients can find these disorders difficult to accept [28‐31].

GPs and nurses often support patients presenting with mental health or eating disorders, since they commonly have expertise in these areas [32‐35]. Some of these patients would also benefit from psychologist or dietician consultations [36, 37]. In France, patients with chronic diseases such as diabetes can also be supported by health professional teams including GPs and public health nurses, with interesting results for some intermediate outcomes (such as glycaemia and adherence to follow-up tests). However, morbidity and mortality results are still lacking [38, 39]. Regardless of the team configuration, role clarification is essential to prevent conflicts between team members and implement effective interprofessional care [40, 41]. However, the implications of overlapping tasks for patients and health professionals remain largely unknown.

Importantly, not all patients need referring to an NPHP. GPs provide comprehensive and patient-centred care for many patients, alone or with other health professionals. They play an essential role aiming to ensure the limited available health care resources are allocated equitably to those who require further care. In this respect, we observed that increased age and multiple health problems managed were associated with high referral probability. This association has already been observed with GP referrals to dieticians and nutritionists in Australia [25] and is consistent with chronic condition and multimorbidity prevalence increasing with age [42, 43] and the increased need for multiprofessional care for these patients [44, 45]. In addition, we found consultation duration was longer (3 min on average) if a referral ensued. This may suggest that complex, comorbid patients require more referrals, or the extra duration may reflect the time needed to discuss the referral with the patient and write the referral letter.

Furthermore, the higher rate of physiotherapist referral in female patients has already been described in the literature [46]. The literature also reveals that female physicians have longer consultation durations, and are more likely to make follow-up arrangements and referrals and perform female prevention procedures [47]. This could partly be explained by physician-patient gender concordance [47].

GPs not actively creating networks with local NPHPs [48] or GPs and NPHPs not working together in multiprofessional centres [49] have been previously suggested as potential barriers to multiprofessional practice. In our study, co-location seemed to have a limited influence, if any, on referral frequency, since only podiatrist referrals were slightly higher in multiprofessional centres versus solo or mono-professional group centres. Overall, patient characteristics and clinical situations impact referral frequency more than GP-related factors suggesting that patients who need referral most are indeed those who are most often referred.

Several authors describe the referral process as a first step in interprofessional collaboration or teamwork [50, 51]. It is therefore important to acknowledge that our study does not enable in-depth analysis of the degree of collaboration, or the conditions required to implement efficient interprofessional practice in primary care. Furthermore, GP practices may not always be aligned with optimal care and observing practice referrals is not sufficient to determine what their referral frequency should be. However, improved understanding of the frequency and factors associated with GP referral to NPHPs provides useful information for health policy development support including planning for future healthcare workforce requirements and implementing interprofessional teams in primary care.

Although the ECOGEN study is slightly dated, to our knowledge, it is the only study which has reported data on NPHP referrals in France and equivalent data do not exist for other countries. Study sample size is large with 128 GPs and 20,613 patient consultations. Included GPs were representative of French GPs in terms of age, gender, practice location and annual consultation numbers [17, 26]. However, all participating GPs were GP internship supervisors. Compared with other GPs, they have similar continuing professional development participation rates, and their patient characteristics don’t differ, but they more often work in group practice and have shorter weekly working hours. The potential impact of these differences on our results is difficult to appreciate. Furthermore, the proportion of female GPs has increased steadily since the ECOGEN study was performed and was 44.2% in 2019 [data upon request from French health insurance (CNAMTS)] versus 34% in the ECOGEN study, which may slightly influence referral frequencies.

Data collection completeness was excellent since most patients consented to participate (only 0.8% (n = 168) of consultations were not included due to lack of consent). Data entry was reliable as no significant difference was observed between the 4.7% double recorded consultations (mean difference: 0.002; p = 0.69) [17].

However, study design introduces an information bias, as only new and explicit referrals were considered in the data collection. This may underestimate referral rates, especially for nurses, where prescriptions are commonly written for several months. Furthermore, the factors associated with referral rates are based on univariate statistical analysis since this study aimed to describe GP practice from an organisational perspective. Modelling patient referral probability with a more clinical perspective would require multivariate analyses.

Beyond the general observation suggesting that national health system characteristics may be a strong determinant for NPHP referrals, our data are specific to France and applicability to other settings is not possible without caution. Should data from other countries become available, it would be interesting to compare them with the ECOGEN data and determine the similarities or differences according to health care system organisation, including NPHP reimbursement. Future investigations should also explore the implications of receiving care from different health professionals, according to specific situations and needs, as well as implications for providers who are potentially filling in for specialist care due to limited accessibility. Finally, future research should address modalities of interprofessional collaboration and teamwork rather than just referrals.

In France, GPs refer around 1 out of 15 patients (6.8%) to NPHPs. Referral frequency is associated with patient characteristics and clinical situations more than GP-related factors, suggesting that patients needing referral most are most often referred. Physiotherapists, podiatrists, and nurses are the most common referrals because the French national health insurance only reimburses treatment costs for these three NPHPs upon GPs referral. Health system legislation and NPHP reimbursement appear to be strong determinants for NPHP referrals. This means there is room for change if health policies aim to support multiprofessional care development.