Interprofessional collaboration in primary care: what effect on patient health? A systematic literature review

The development of primary care, defined as a model of care that supports first-contact, accessible, continuous, comprehensive and coordinated person-focused care, is a global priority [1]. Studies have already shown that most patients are treated in a primary care setting [2, 3]. Most patients suffering from common diseases such as hypertension, diabetes, chronic obstructive pulmonary disease and asthma only consult a primary care provider [4]. The ageing of populations, the growing importance of chronic pathologies, the international shortage in the health care workforce [5], and the growing complexity of care pathways call for development of new modalities of practice in primary care.

Functioning in a primary care team, based on Interprofessional Collaboration (IPC), is widely supported by the health authorities in France [6], similar to existing models in several countries [7]. IPC is defined as several health workers from different professional backgrounds providing comprehensive services by working with patients, their families, and other caregivers [8]. This definition can be supplemented by the need for contact, negotiation and interaction among health care professionals. As this concept is recent and vast, several terms have been used, but the term IPC is the most currently used [9]. Part of these IPC teams regroup professionals in the same practice. Many practices work as a team of GP's, nurses (including Advanced Nurse Practitioners), paramedics (including Advanced Clinical Practitioners), clinical pharmacists, physiotherapists, physician associates and others. In this model, an integrated and collaborative approach to patient care has been developed. The literature on this subject shows a recent emulation with numerous articles in many journals [1, 7‐11]. Several authors have reported that working as a team is a source of satisfaction among professionals [7, 10]. A 2018 literature review reported how primary care teams were formed [11], but it did not report any information on the effect of these organizations on patient-centred outcomes.

Some authors have focused on the team-based approach for specific pathologies for which collaborations largely mobilize specialists and hospital professionals, but these investigations concern secondary rather than primary care services [12‐14]. While a recent literature review [15] investigated the effect of IPC in a primary care setting on adults with diabetes and/or hypertension, we wanted to further investigate which areas of treatment and primary care team organization had an effect on patient-centred outcomes.

A better understanding and better characterization the composition of IPC teams and the health fields in which this collaboration would be relevant and effective for patients is important for decision-makers and professionals who wish to engage in the evolution of their practice. Which professionals are involved in IPC? For which treatments and illnesses have results been obtained? For which treatments and diseases have we not obtained convincing results?

The objective of this study was to assess the effects of IPC grounded in primary care setting on patients-centred outcomes.

This systematic review was conducted according to the PRISMA guidelines [16]. We searched for studies published between January 1st, 1995 and March 1st, 2021. To ensure that we found all research articles published by the various health professionals, we chose to increase the number of search engines normally used. The following databases were searched: PubMed, Embase, PsycINFO and CINAHL. Additional articles that were found by hand searching the references were also reviewed. The following research algorithms were used 1) PubMed: ("Intersectoral Collaboration"[Mesh] OR "Cooperative Behavior"[Mesh] OR "Patient Care Team"[Mesh:NoExp]) AND ("Primary Health Care"[Mesh]) AND ("Outcome and Process Assessment, Health Care"[Mesh]); and 2) Embase, PsycINFO and CINAHL: « intersectoral collaboration», « cooperative behaviour», « patient care team» AND « primary health care» AND « outcome and process assessment, health care». First, the titles were reviewed, and then the abstracts and full texts of the selected articles were reviewed independently by two reviewers with the Abstrackr tool [17]. Any disagreements were resolved by consensus; MJ, MA, and JFH resolved any remaining disagreements.

After reading the full texts, we included the following studies in this analysis:

The exclusion criteria were as follows:

One reviewer (MA) independently extracted data using a prepiloted form and was supervised by a second reviewer (CB), and the following data were collected: study country, pathology(ies) studied, intervention, the number of patients included, design, study duration, main outcome measures, and patient outcomes (selected on the basis of frequency of reporting and clinical relevance). For consistency and clarify of presentation, the results centred on patient outcomes are grouped within 3 categories in the remainder of the manuscript: 1- patients at cardiovascular risk, 2-polypathological and elderly patients, and 3- patients with mental health problems, chronic pain and unexplained complaints.

The quality of the studies was then assessed using the revised Downs and Black Checklist [18]. The checklist includes 27 items on reporting (10 items), external validity (3 items), internal validity (13 items), and power (1 item). Similar to others studies, the power item was modified regarding whether a power analysis was described (0 = not reported, 1 = reported). The maximum possible score is 28 for randomized studies and 25 for nonrandomized studies. Quality was categorized by using the following Downs and Black score ranges: strong (21-28), moderate (14-20), limited (7-13), and poor (≤ 7) [19].

The studies that assessed the effect of IPC on patient outcomes could be grouped into 3 categories, depending on whether the patients 1) were at cardiovascular risk, 2) were elderly and/or had polypathology, or 3) had mental or physical disorders. One study also aimed to improve prevention care. Our review of the literature did not find any studies that evaluated the effect of IPC on patient outcomes in the following fields: orthopaedics and the musculoskeletal system, cancer care, paediatrics, current infectious diseases, and patient health monitoring. Only one study analysed cancer screening in addition to monitoring comorbidities. The positive effect of IPC on patient outcomes has been widely described in patients at vascular risk. Of 28 studies, only 5 reported no significant effect on their primary outcome measures. For the 2 other categories (elderly and/or polypathological patients, patients with physical or mental disorders), the reported effects varied from one study to another. Among 18 studies that assessed the effect of IPC on the outcomes of elderly or polypathological patients, only 10 studies reported positive effects on the primary outcome measure, and 11 out of 12 reported positive effects in the category of patients with mental disorders. In this last category, 10 studies focused on patients with depressive syndrome. The majority of studies reported clinical improvement in patients. The majority of the proposed interventions relied on 3 main health professionals: general practitioners, pharmacists and nurses. Other professions were less frequently included in the studies.

In general, studies evaluating the effect of IPC are difficult to compare since the interventions are often very different, and the designs and the evaluation criteria vary, making it impossible to conduct a meta-analysis of the data. The number of studies is often very limited: many fields of care have not been the subject of any study. Some pathologies are only approached in a very isolated way, e.g., chronic pain and anxiety disorders. Finally, for the two topics best covered (cardiovascular risk and depressive syndrome), the differences among the studies remain significant and limit the opportunity to aggregate the results. It is notable that the judgement outcomes are also very different. Compliance [44], patient satisfaction [42], or improvement in blood pressure or HbA1c levels [25, 29] cannot be compared. Finally, even when the studies analysed a bioclinical measure such as HbA1C levels, the authors chose different judgement criteria: the rate of prescription of an examination [51], the examination completion rate [30, 40, 49], the variation of the result over time or the rate of patients reaching their objective for this measure [22]. This complicates the comparison of studies. Smith did not show any effectiveness of his intervention on the HbA1c level in diabetic patients (the primary outcome measure) but showed that the proportion of patients who carried out the recommended monitoring examinations had increased [45]. For the 15 studies evaluating the effect of ICP on arterial hypertension, the diversity of judgement outcomes still remains significant, depending on whether the authors chose to analyse the mean systolic blood pressure value in the intervention arm [29], the change in systolic blood pressure over one year [34, 43], the change in diastolic blood pressure over one year [36], the proportion of patients with controlled systolic blood pressure (less than 140 mmHg) [43], the proportion of patients for whom the systolic blood pressure was measured (care procedure indicator) [26], the proportion of patients for whom an anti-hypertensive treatment dose adjustment was carried out [44], the evolution of the mean 24-h SBP [24], or the evolution of the SBP at home [48].

The clinical impact of the effects observed can also be questioned. For example, the study by Heisler [34], a large randomized study (4100 patients), showed a reduction of 2.4 mmHg in blood pressure thanks to the help of the pharmacist, which is clinically limited. With regard to depression, one can wonder about the benefit of a 3.6-point reduction in the Hamilton score for patients, even if the result is statistically significant [81].

As in other studies, we found that GPs, nurses and pharmacists were the most represented professionals in the IPC teams in the primary care setting [11, 85].

Several authors have shown that IPC would increases patient use and costs of care [86]. These indicators of the use and costs of the health care system are important from the perspective of the decision-maker, but they must be interpreted with caution. On the one hand, less use of to the health care system can testify to better individual health with lower morbidity (a reduction in the number of visits to the emergency room for example, or a reduction in the number of hospitalizations) [38, 66, 87]. On the other hand, these indicators can also attest to a positive effect with regard to issues of compliance and improvement of patient health: in many patients with chronic diseases monitored in primary care, the recommendations emphasize the prevention of serious events through the implementation of reinforced medical monitoring, resulting in an increase in the use of health care system by patients. Thus, the increase in the number of consultations with a health care professional may be appropriate during a period of antihypertensive treatment dosage adjustment to allow the achievement of the objectives [29].

Many studies have shown that IPC can improve the management of patients at cardiovascular risk. Other studies have investigated the effect of IPC in polypathological elderly patients and in patients with mental or physical disorders. For these pathologies, the number of studies remains limited, and the results are heterogeneous. Researchers should be encouraged to perform studies based on comparative designs: it would increase evidence on the positive effect and benefits of IPC on patient variables.