Improving early referrals for inflammatory arthritis through primary care engagement: a quality improvement initiative

Abstract

Objectives

Early rheumatology referral is crucial for optimal outcomes in inflammatory arthritis, but delays remain common. This quality improvement (QI) project aimed to identify barriers and implement strategies to improve referral adequacy and timeliness within a newly established early referral program. Despite the implementation of a direct referral pathway for primary care physicians, many patients did not meet eligibility criteria, limiting access for those with early inflammatory arthritis.

Methods

Since January 2022, a 1 day/week early arthritis clinic has offered direct and rapid access to rheumatology appointments to 180 primary care physicians (PCPs) across 11 centres. Standardized referral criteria were co-developed with a PCPs committee. A survey assessed PCPs’ confidence, perceived barriers and educational needs using Likert scales and open-ended responses. In response to survey findings, electronic consultations (e-consults), concise referral guidelines and targeted training materials were introduced to support appropriate referrals.

Results

Within the first year, the median referral time decreased from 51 to 15 days, although only 21% of referrals met the criteria. Most PCPs recognized the importance of early diagnosis (67%) but reported limited confidence in managing suspected cases (96% partially or not confident). Reported barriers included non-specific symptoms (48%) and long rheumatology wait times (39%). PCPs expressed strong interest in e-consults (87%) and brief educational tools (78%).

Conclusion

This QI initiative reduced referral delays and identified key barriers to effective early arthritis referrals, including PCPs lack confidence in diagnosis and long rheumatologist wait times. Enhancing communication through e-consults and targeted educational resources may optimize early referral.

Key messages.

• The early referral program reduced median rheumatology referral time from 51 to 15 days.

• PCPs reported low diagnostic confidence but strong interest in targeted rheumatology education and collaboration.

• E-consults improved referral accuracy and reduced delays in access, facilitating PCPs’ and rheumatologists’ communication.

Introduction

RA is a chronic, inflammatory disease with an incidence of 8.3 per 100 000 inhabitants—three times higher for early arthritis—and a prevalence of 0.82% [1]. Prevalence in Spain has been increasing, likely due to an aging population and increased smoking rates, significantly impacting society by reducing patients’ physical function and quality of life [2] and with a substantial economic burden related to work-related disability as a consequence of the disease [3].

Advancements in RA management, including new treatment options and early diagnosis and treat-to-target strategies, have led to better patient outcomes. The early phase of the disease is critical, as inflammation is at its peak, with the highest rates of erosion development and joint inflammation. While inflammation is both treatable and reversible, functional recovery depends on the duration and severity of inflammation. This underscores the importance of the ‘window of opportunity’, during which early intervention is crucial to prevent long-term damage and disability. Consequently, early diagnosis and treatment ideally within the first 3 months [4] have become fundamental principles in rheumatology and extensive evidence shows how significantly they improve RA prognosis [5, 6],

Despite strong support for early treatment and proposed guidelines for early referral [7], delays in diagnosis and treatment persist in Europe [8]. Primary care physicians (PCPs) play a crucial role in early referral to rheumatology, as most patients with inflammatory arthritis initially seek care from a PCP [9]. Research suggests that patients with RA visit their PCP an average of four times before being referred to a rheumatologist [10]. Several factors may contribute to these delays, including lack of awareness about RA’s consequences, insidious symptom onset, atypical and challenging presentations and limited access for PCPs to refer patients, with sometime limited feedback from rheumatologists. Multiple strategies have shown efficacy addressing referral delay as PCP and health professional education programs, triage in combination with referral forms, development of an early arthritis clinic (EAC) and rapid access to rheumatology [11]. However, few of these initiatives have been developed with active input and collaboration with PCPs as important collaborators in early identification and referral.

The objective of the present study was to assess the impact of an early referral program (REUCARE) that combined fast and direct access to an EAC with the implementation of a referral form based on a validated prediction score for risk of RA [12–14]. This program was developed by incorporating PCPs’ perspectives and needs, with the overarching goal of reducing delays in the evaluation of patients with inflammatory arthritis in our rheumatology department compared with rheumatology routine care.

Methods

Study design

The rheumatology department at our academic institution provides 20–30 appointments per day for new patients during routine outpatient clinics. Referrals are typically generated by PCPs or by other specialties, including orthopaedics, internal medicine, rehabilitation, endocrinology and the emergency departments. Patients may be scheduled through the standard referral pathway or via requests for priority evaluation.

According to data published by the Madrid Regional Health System in January 2022 (https://servicioselectronicos.sanidadmadrid.org/LEQ/ConsultaEspecialidades.aspx), the median waiting time for a new patient to access our rheumatology department at Hospital General Universitario Gregorio Marañón was 51.43 days. Given that such delays may adversely affect outcomes in patients with suspected inflammatory arthritis, we develop a new referral pathway (REUCARE) in January 2022 that enabled direct access from PCPs to a 1-day/week EAC. In total, 180 PCPs work in 11 primary care centres affiliated with our academic centre serving a population of 31 000 people. Patient distribution varies, ranging from 19 023 at Centro de Salud Valdebernardo (11 PCPs) to 40 174 at Centro de Salud Pavones (23 PCPs). All PCPs were informed of this initiative by the coordinator of each primary care centre and the Director of Care Continuity at our institution.

The program began in January 2022, integrating the EAC with rapid and direct access for PCPs in the specialized multidisciplinary centre, Centro de Enfermedades Inflamatorias Mediadas por la Inmunidad, for patients with immunomediated inflammatory diseases. This holistic model fosters interaction with other specialists and PCPs and enhances comprehensive patient evaluation and management through improved interdisciplinary communication and supports educational activities for both PCPs and patients.

PCPs could directly refer patients with inflammatory arthritis to the EAC using simple clinical referral criteria developed in consensus with a PCP committee and approved by the Southeast Local Research Committee in October 2021. The EAC early referral criteria are at least two swollen joints, symptom duration of <2 years and no prior rheumatology diagnosis in relation to the current symptoms. The EAC allocated six dedicated appointments within a half-day clinic every Tuesday for the assessment of patients with a high suspicion of inflammatory arthritis as identified by PCPs according to the predefined referral criteria. Notably, implementation of the EAC did not entail additional rheumatology personnel or costs, rather it was achieved through the reallocation of existing medical resources.

Patients are evaluated in an initial screening visit and are excluded if the screening suspected gouty arthritis, viral arthritis, osteoarthritis or fibromyalgia. They may be referred to the regular outpatient rheumatology clinic if necessary and if a rheumatology evaluation is not required, they are provided with recommendations and are discharged. Patients fulfilling inclusion criteria were invited to participate in the study and after signing the informed consent form they completed two questionnaires on a tablet, a referral form based on patient-reported outcomes (PROs) and a multidimensional health assessment questionnaire (MDHAQ).

Patients’ questionnaires

The MDHAQ is a PRO-based questionnaire that is useful in different rheumatologic diseases [15] and adapted from the standard HAQ [16]. It collects data on physical function, psychological variables, sleep quality, depression, fatigue, symptoms and medical history. It includes 10 activity-related questions, scored from 0 to 3, with higher scores indicating worse physical function. Additionally, the MDHAQ includes three visual analogue scales (VASs) [pain, patient global assessment (PtGA) and fatigue], a 60-symptom checklist and a self-reported joint count based on the RA Disease Activity Index (RADAI), covering 16 joint groups with pain scores from 0 to 3 (total 0–48). Demographic data such as sex, birth date, ethnicity and education level are also included. The Routine Assessment of Patient Index Data 3 (RAPID3) [17], derived from the MDHAQ, includes three measures—physical function, pain and PtGA—each scored on a scale of 0–10 (total 0–30).

Referral form

The referral form is based on a prediction rule for undifferentiated arthritis developed in an early arthritis cohort [13] and includes nine clinical variables: sex, age, symptoms localization, morning stiffness (in minutes), tender joint count (out of 68), swollen joint count (out of 66) and three laboratory markers: CRP level, RF positivity and anti-CCP antibodies. This referral form was previously validated in an early arthritis cohort and includes clinical variables to facilitate completion by patients, showing face validity in the content and feasibility for implementation in the PCP setting.

The original prediction score ranges from 0 to 13, with 94% of patients scoring <6 not developing RA after 1 year [18]. A cut-off score of ≥8 has shown a positive predictive value of 93–100% across three European cohorts [13].

Physician evaluation, laboratory tests and imaging assessments

A comprehensive physical examination is conducted, including joint counts, laboratory tests and a baseline musculoskeletal ultrasound, as per protocol. The medical evaluation also incorporates the RheuMetric checklist, which includes a physician global assessment (PGA) using a 10-cm VAS scale (0 = excellent, 10 = very poor) and three subscales assessing inflammation, irreversible damage and distress (e.g. fibromyalgia, depression) [19, 20].

Referral adequacy

The time from referral to baseline visit at the EAC was calculated and the adequacy of the referral was evaluated. The screening visit was performed by the same rheumatologist to confirm whether the three predefined criteria were all met. Referrals fulfilling all criteria were classified as adequate, whereas those not meeting any of the criteria were classified as inadequate. Percentages of adequate vs non-adequate referrals were calculated and reasons for non-adequacy were identified with a descriptive analysis of the referral pattern. Additional information related to the primary care centre and patient’s environment, which may impact referral, was collected. This included the distance in kilometres to the primary care centre, the number of PCPs per primary care centre and an estimate of patient income based on zip code, used as a surrogate for healthcare access. Higher-income areas may be associated with better health outcomes due to improved access to healthcare, healthier environments, higher levels of education and greater health literacy. These factors can contribute to lower smoking rates, greater adherence to medical advice and greater seeking of medical care. Income levels based on zip codes were categorized as follows: 0 (≤€27 525), 1 (≤€31 294), 2 (≤€46 313) and 3 (≥€54 318).

PCP survey

A low proportion of adequate referrals represented the main implementation challenge encountered during our program’s rollout, prompting us to investigate the needs and expectations of PCPs through a survey. The survey was developed in collaboration with PCP investigators to explore potential areas of improvement (Supplementary Data S1). The survey was design to explore contextual factors contributing to the program’s efficiency and areas of improvement. The survey was distributed by e-mail with a letter of invitation and a link to the digital survey through the primary care centres. It included five domains covering PCP demographics, previous knowledge about RA, level of confidence, factors and potential barriers influencing referral and interest in rheumatological training. Data were captured using 4-point Likert scales, yes/no questions or free text and were analysed descriptively.

Ethical approval

This study was performed in accordance with the ethical standards of the responsible committee on human experimentation and the Helsinki Declaration of 1975, as revised in 1983. Research ethics committee approval for the protocol was obtained prior to commencing the study (REU-CAR O3/2021.21/2022).

Results

From 328 patients referred to our EAC from January 2022 to November 2023, 70 (21%) fulfilled the inclusion criteria: 23 (33%) with undifferentiated arthritis, 24 (34%) with RA, 17 (24%) with SpA including psoriatic arthritis and 6 (9%) with other diagnoses such as scleroderma, SS and lupus. Patients in the RA group were older, more frequently female, with positive ACPA and RF, higher CRP levels and a higher mean score for the referral form (16.2 vs 13.6 for undifferentiated arthritis, 10.0 for SpA and 11.1 for other diagnoses). A mean score >8 was used as a cut-off with a positive predictive value of 93–100% for RA diagnosis (Table 1).

Table 1.

Demographics and clinical characteristics at baseline according to diagnosis.

Characteristics	Undifferentiated arthritis	RA	SpA	Other
Patients, n (%)	23 (32.8)	24 (34.3)	17 (24.3)	6 (8.6)
Age, years, mean (s.d.)	44.7 (10.7)	57.3 (17.2)	50.1 (7.8)	53.5 (9.1)
Female, n (%)	18 (78)	20 (83)	7 (41)	4 (67)
RF positive, n (%)	12 (53)	15 (64)	2 (14)	2 (33)
ACPA positive, n (%)	3 (13)	18 (73)	0	0
ESR, mean (s.d.)	15.3 (13.3)	27.8 (16.7)	29.4 (24.2)	18.3 (14.7)
CRP, mean (s.d.)	0.5 (0.3)	2.1 (1.1)	1.7 (1.1)	0.4 (0.2)
PtGA, mean (s.d.)	4.4 (2.7)	4.9 (2.5)	6.4 (2.7)	5.2 (1.9)
PGA, mean (s.d.)	1.9 (1.5)	2.5 (1.9)	2.7 (1.8)	1 (0.7)
Disease activity indices
DAS28-CRP, mean (s.d.)	1.9 (0.6)	2.7 (0.8)	2.0 (0.7)	1.7 (0.4)
CDAI, mean (s.d.)	7.5 (6.2)	12.6 (9.7)	6.9 (6.3)	6.6 (4.3)
RAPID3, mean (s.d.)	9.9 (6.5)	11.5 (7.6)	15.7 (5.2)	10.0 (4.4)
Referral score, mean (s.d.)	13.6 (10.8)	16.2 (10.1)	10.0 (7.5)	11.1 (9.6)

DAS28-CRP: 28-joint DAS with CRP; CDAI: Clinical Disease Activity Index.

The mean time from PCP referral to the baseline visit at the EAC was 15 days (s.d. 8.7), with a significant reduction vs the median waiting time for a new patient to access our rheumatology department (51.43 days) according to data published by the Madrid Regional Health System in January 2022 (https://servicioselectronicos.sanidadmadrid.org/LEQ/ConsultaEspecialidades.aspx).

The centres with higher rate of referral were Centro de Salud Pacifico (12%), Centro de Salud Pavones (11%), Centro de Salud Torito (10%) and Centro de Salud Valdebernardo (10%), with an adequate referral in 35%, 6%, 7% and 21%, respectively.

Referral pattern

A description of the referral pattern is presented in Fig. 1. Patients who did not meet the referral criteria were mainly patients with osteoarthritis (35%), non-specific arthralgia (15%), gout (9%), fibromyalgia (9%) and polymyalgia rheumatica (6%). Patients in the adequate referral group were younger (50.4 vs 55.6 years, P = 0.03), with no differences in income level and primary care centre characteristics in terms of size and distance to our EAC.

Figure 1.

Referral patterns for patients not fulfilling referral criteria. Percentages are given for five more common reasons for referral

PCP survey

E-mail invitations were sent to 180 PCPs through their primary care centre coordinator, with a response rate of 15%. More respondents were women (76%), 65% were >50 years of age, 69% had been in practice for >20 years, 73% reported having 30–40 patients per day, 25% had >40 patients per day, 50% had 1–5 patients and 9% had 6–10 patients newly diagnosed with RA in the last year. The majority (83%) reported additional rheumatological training besides medical school, including during a fellowship (55%), specific courses (50%), rheumatology journals (45%) and conferences (25%), however, they were not familiar with the incidence and prevalence of the disease. Around 67% were aware of the importance of early diagnosis and treatment, however, they felt not very confident making the initial diagnosis (61% ‘somewhat confident’ and 35% ‘not confident’).

The main reasons for delay from the PCPs’ perspective were non-specific initial symptoms of RA (48%) followed by a long rheumatology waiting list (39%) (Fig. 2A). The main reasons for referral to rheumatologist were clinical presentation suggesting RA (65%), positive RF (46%) and elevation of acute phase reactants (46%) (Fig. 2B). Most PCPs considered the waiting time for rheumatology prolonged (65%) or very prolonged (22%) and reported none (9%), insufficient (30%) or limited feedback information from rheumatologists (30%). Although 77% indicated they used additional resources to improve their knowledge in RA, all participants stated that they would be interested in receiving more information on early diagnosis, with the preferred resources being a specific e-consult (87%), updated mini-guidelines (78%) and face-to-face educational meetings with rheumatologists. In terms of the scope of the educational meetings, they were mainly interested in improving their clinical abilities (91%), interpretation of laboratory and imaging studies (96%), physical examination (83%) skills and initial therapy (73%) in patients with inflammatory arthritis.

Figure 2.

Reasons for (A) delay and (B) motivation for referral to rheumatology from the PCPs’ perspective.

The main areas in which PCPs were willing to collaborate with rheumatologists in the care of patients with RA and other inflammatory arthritis were early diagnosis (83%), adherence to treatment (70%), control of adverse events of treatment (65%), follow-up of patients with adequate control or in remission (65%), comorbidity management (57%) and control of flares (43%).

The survey revealed contextual factors contributing to the program’s reduced efficiency, such as the rapid saturation of available slots for new early arthritis patients by inadequately referred cases. To address this issue, we introduced electronic consultations (e-consults) and distributed targeted training materials for PCPs, as suggested by the survey findings. The e-consults allowed us to prescreen and prioritize appropriate referrals for in-person baseline visits, leading to a more efficient and sustainable operation of the program within our department.

Discussion

Early diagnosis and treatment of RA are supported by strong, widely accepted evidence and have been endorsed by many organizations, including the Spanish Society of Rheumatology [21]. However, most patients in our area are not seen within the first 4 months after symptom onset. This situation may worsen in the future due to a projected decline in the rheumatology workforce, coinciding with demographic shifts that will lead to an increase in patients with rheumatological conditions [22]. Given the limited resources, it is essential to encourage changes in workforce distribution and improve practice efficiencies, ensuring direct access for patients in whom delays significantly impact optimal management.

To address this issue, we implemented an early arthritis program based on a team-based approach in collaboration with PCPs. This program has significantly reduced referral delays by 29%, facilitating the screening of patients with inflammatory arthritis through rapid access to an EAC and close collaboration with PCPs.

Our program includes a PRO-based referral form, easily completed by patients on a tablet, which showed higher scores in RA patients compared with those with other inflammatory arthritis. Although patients currently complete this form during the screening visit at the hospital, it could be valuable in the primary care setting to help identify those at higher risk for RA or other inflammatory arthritis. Primary care patients often self-report symptoms such as persistent joint pain, swelling and stiffness—common in both inflammatory arthritis and non-musculoskeletal conditions. This overlap creates a challenge in PCP settings, highlighting the need for new approaches to ensure accurate and early diagnosis, facilitating a treat-to-target strategy [23]. Additionally, retrospective studies show that primary care visits increase as an RA diagnosis approaches [24], making this period ideal for implementing early detection strategies.

Incorporating the early referral form into routine care enhances the quality of care and communication between patients and doctors while promoting early detection of RA and other inflammatory arthritis. However, patient participation in care and research remains a challenge for both patients and clinicians. There is still a need to develop strategies that support patient-centred care, offering tailored assistance to both patients and their PCPs. Our collaborative approach between PCPs and rheumatologists fosters patient empowerment and health education, ultimately improving the quality of care.

During the initial months of the EAC, we noted a low adequacy of referrals, with only ≈21% of patients meeting the predefined inclusion criteria. Consequently, the dedicated appointments were rapidly occupied, necessitating additional efforts to maintain an optimal referral time of <2 weeks. This early observation during the pilot phase prompted implementation of the survey among PCPs aimed at identifying their needs and improving the referral process. Although the survey response rate was not optimal, it provided valuable insights to identify contextual factors contributing to the program’s reduced efficiency, such as the rapid saturation of EAC appointment slots due to inadequate referrals. For example, PCPs proposed to introduce an e-consult system for patients with suspected early arthritis, designed to facilitate communication between PCPs and the rheumatologist coordinating the EAC. E-consultations have been associated with improved access to hospital care and a reduction in unnecessary hospital referrals [25], making them a promising digital health tool to enhance early detection of inflammatory arthritis. The e-consult system has been incorporated in our EAC, thereby reducing workload pressures and enhancing the appropriateness of referrals. The e-consults now enable prescreening and prioritization of appropriate referrals for in-person baseline visits, leading to a more efficient and sustainable program within our department.

A combined approach—direct EAC access, rapid e-consult communication and ongoing PCPs education—has contributed to maintain the observed reduction in referral time without increasing workload or personnel demands.

The main limitation of our early arthritis program is that we rely on PCP participation. Despite multiple communication channels, regular feedback and program updates via newsletters, some primary care centres referred only a limited number of patients. One of the objectives of our study was to assess whether characteristics of PCPs might influence referral patterns. However, due to the high turnover of physicians in the primary care centres affiliated with our hospital, it was not feasible to evaluate certain variables collected in the survey, such as age, gender and years of clinical experience. We were only able to analyse structural and contextual factors, including the size of the primary care centre (in terms of number of PCPs), its distance from the hospital and the income level of the neighbourhood served. These variables did not impact the adequacy of referral. Additional variables not identified in our study could explain why the percentage of appropriate referrals was almost five times higher in some centres compared with others, despite having a similar referral rate (≈10–11%). Some potential variables might be the workload or training level of PCPs, but these data were not collected in our study. Nevertheless, this information could be considered in the future to design additional improvement strategies, including knowledge transfer from centres with better referral adequacy to others. Referral adequacy rates have been explored in patient with SpA in Spain and agreement between rheumatologists and PCPs was very poor for variables like inflammatory back pain, which are crucial for diagnosis, highlighting the importance for training programs for PCPs [26].

An additional limitation of our study was the low survey participation. A recent systematic review on survey response rates in patients and healthcare professionals reported that physicians generally have lower response rates compared with patients, with an average of ≈51% when surveys are distributed by e-mail [27]. Reminders could not be sent, as the survey was distributed through an official institutional channel, and there is a high survey burden already placed on PCPs in our area. Although we recognize this as a limitation of our study, the information obtained from the survey was highly informative and directly supported implementation of the e-consult, in line with PCPs’ recommendations.

Additionally, part of our data is derived from patient self-reported questionnaires and structured visits, so some clinical information may not be fully accurate. Although we actively promoted the screening process and offered frequent, accessible visits, some patients may have been missed. Since this screening intervention is non-invasive, the risk of adverse events is minimal. However, there is a possibility of unnecessary specialist care. As rheumatology evaluations were only recommended for high-risk patients and treatment was initiated at the rheumatologist’s discretion, most participants appeared to benefit from the assessment.

Concerning a high turnover among PCPs in our region, continuous dissemination of project materials and updates is essential. To this end, we developed a dedicated website (https://www.reucare.com/inicio) that provides access to project resources, survey results and educational material.

In conclusion, we implemented an early referral program that significantly shortens the delay in referral for patients with inflammatory arthritis from PCPs to rheumatology. This quality improvement study highlights how direct referral pathways in collaboration with PCPs can shorten the time to rheumatology assessment without requiring additional resources. The study also generated tools—including an e-consult pathway, PCP educational materials and an autoreferral tool for patients—that may be adaptable to other healthcare systems. Developing this program is a strategic initiative for our Public Health System, optimizing hospital resources for patients with inflammatory arthritis requiring costly treatments. Given the increasing workload in rheumatology across Europe due to population ageing, increasing musculoskeletal disease burden and insufficient specialist numbers, strategies that can be implemented in different health systems to optimize health resources are crucial. Early identification of inflammatory arthritis is essential to initiate timely therapy, prevent irreversible damage and improve disease control. Collaboration with PCPs is crucial in achieving this goal.

Supplementary material

Supplementary material is available at Rheumatology Advances in Practice online.

Data availability

The datasets generated and/or analysed during the current study are available from the corresponding author upon reasonable request. Due to patient privacy concerns and ethical restrictions, some data may not be publicly shared. Anonymized data may be made available to qualified researchers in accordance with institutional and ethical guidelines.

Authors’ contributions

I.C. is the principal investigator of the project and conceived the idea for this article. I.C., J.M.C., J.M.B., J.R. and J.M.A.G. designed the study, planned analyses and interpreted the results. J.O.G. and G.F.D. participated in the design as advisors from primary care. C.B., M.S.L., T.S. and M.T.R. were involved in data acquisition. IC extracted the data, performed the analyses and wrote the first draft of the manuscript. All authors critically reviewed the manuscript and agreed with the submission. I.C. had full access to all data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Funding

This work was supported by grants PI20/00847, PI23/01226, RD21/0002/0034 and RD24/0007/0029 from Instituto de Salud Carlos III and co-financed by the European Regional Development Fund.

Disclosure statement: I.C. reports consulting and/or speaker fees from Alfasigma, Boehringer Ingelheim, BMS, Lilly, Galapagos, Gebro Pharma, Gilead, GSK, Pfizer and UCB, not related to the submitted work. C.B., M.S.L., J.O.G., M.T.S., T.R., J.M.C., J.M.B., G.F.P., J.R. and J.M.A.G. declare no conflicts of interest.