Background

In an era of a world-wide general practitioner (GP) shortage and increased demand for health care services because of chronic illness and ageing, evidence shows that 25–70% of physician tasks could be delegated to non-medical health professionals in advanced roles, especially in primary care [1]. Introducing additional and varied professions into primary care has been deemed an appropriate solution to counteract this shortage while addressing the increased need for primary care services [2,3,4].

Evidence shows that nurses are capable of independently conducting 85% of GP same day appointments [5], providing as high a quality of care and achieving equivalent health outcomes as GPs [6], and contribute to reducing hospitalisations and mortality rates [7]. Particularly nurse practitioners (NP) invoke high levels of patient satisfaction [8, 9]. The titles, training, and experience of NPs vary greatly internationally, leading to them working in various fields and with varying scopes of practice [10]. According to the International Council of Nurses (ICN) “a Nurse Practitioner/Advanced Practice Nurse is a registered nurse who has acquired the expert knowledge base, complex decision-making skills and clinical competencies for expanded practice, the characteristics of which are shaped by the context and/or country in which s/he is credentialed to practice. A master’s degree is recommended for entry level.” [11]

Introducing interprofessional teams results in the need for task re-allocation. This can be done using the concept of skill mix in which professionals with different qualifications collaborate, emphasizing the utilization of professional’s knowledge, experience, and skills to their fullest potential [1]. Alternatively, allocating tasks according to the principle of subsidiarity can lead to an imbalance in workload and dissatisfaction among health professionals, thus perpetuating staffing issues [12]. Evidence to date suggests that the concept of complexity features when allocating tasks in primary healthcare teams and bears a noteworthy impact on interprofessional collaboration [13, 14]. When looking for definitions of complexity within medicine, an evolution of the term has been discovered. Surrogate terms such as comorbidity, multimorbidity or polypathology were often used to describe what today, may be referred to as complex. These terms all refer to a multitude of conditions and or diseases [15]. According to a concept clarification “complexity, as opposing to the previous surrogate terms [sic!], promotes a wider perspective of health by expanding the focus on biology to include the environment and social relations.” [15](p.18) Complexity can also be regarded as a system within which people act. As shown in the Cyenfin Framwork [16] which is based on people operating in one of four systems: simple, complicated, complex or chaotic. In an adaptation which divides various medical fields in to the four systems it is said that holistic medicine represents a complex system. In this model holistic medicine is characterized by informal and interdependent care in which experienced practitioners rely on narratives and metaphors to recognize patterns and make sense of complexity in order to act [17]. Furthermore, complexity can be regarded as a concept according to which professional tasks are allocated, as shown in Kernick’s continuum [18]: the higher the complexity the more educated the health professional. Health professionals range from A to E. A being a GP managing and planning the treatment of patients based on the interpretation and integration of complex clinical, psychological, social, cultural and cost factors in combination with experience and knowledge. Addtionally organizing and coordinating multidisciplinary teams. B being a NP clinically diagnosing and treating less complex cases, active in some areas of chronic care while interacting with other members of the team. C being an Extended Role Practice Nurse providing specific, well-defined, protocol-directed clinical care, for example asthma or contraception management. D being a Practice Nurse providing traditional nurse care, for example the management of minor injuries or immunization. Finally, E being an auxiliary Practice Nurse with limited training performing simple, well-defined tasks such as urine analysis or simple wound dressings [18]. Allocating tasks according to this continuum based on complexity suggests a shift from very separate, different nurse and doctor roles towards a partnership which is inherently flexible. Additionally, when looking at task distribution on a continuum, it is possible to make the most of each professionals’ skills and time thus ensuring health gain in an effective and economic way according to Kernick. The premise of this continuum is, that the less training, the less responsibility and complexity and also the less remuneration.

Despite the level of importance assumed by the concept of complexity in existing literature and the theoretical constructs, to date there is no overview, which provides practical guidance for practitioners on the precise use of complexity as a factor for task allocation. By looking at recently published studies in which NPs have been introduced into primary care and are collaborating with GPs the use of complexity can be examined and insights into possible methods for task allocation gained. Additionally, information regarding tasks performed by NPs working collaboratively in different primary care settings and countries may contribute to understanding the role of NPs further. This may be particularly helpful for practices seeking to implement or enhance skill mix. Therefore, the objective of this narrative review was to investigate the reported use of complexity as a factor for task allocation among GPs and NPs working collaboratively in primary care by collecting and analysing existing evidence based on quotes referring to complexity.

Methods

A protocol was written to guide the methodological process following the PRISMA statement [19]. Evidence pertaining studies set in primary care and describing the collaboration between NPs and GPs were searched for in scientific databases. This was deemed an appropriate method of reaching the goal of creating an overview of how complexity is used in task allocation in models of primary care which offer some model of shared care. Literature was examined in a broad manner with the goal of linguistically detecting the term complexity or related terms and analysing the context.

Information source and search strategy

Database searches were carried out in PubMed and CINAHL in November 2019 using search terms built upon three concepts: nurse, role, and GP. For the search in PubMed the terms comprised Medical Subject Headings (MeSH) and free text words combined using Boolean operators and truncations as seen in Table 1. Here, the time and language restrictions were included in the search terms. For the search in CINAHL, MeSH terms were replaced with Exact Subject Headings (MH). Additional filters were put in place as follows: scholarly journals, published dates: July 2006 – November 2019, languages: English and German. Forward cited literature and bibliographies of the resulting literature were searched manually to complete the selection.

Table 1 Search strategy

Eligibility criteria

There are two sets of eligibility criteria, which can be seen in Table 2. Stage 1 criteria were applied to assess titles and abstracts and stage 2 criteria to assess full texts. There were no restrictions regarding the study type because the concept of complexity is not bound to a specific study design. Furthermore, it was unclear how much literature would suit the inclusion criteria and therefore, imposing minimal restriction led to a comprehensive search of all up-to-date literature.

Table 2 Eligibility Criteria

Stage 1 criteria stated that abstracts in a language other than English or German, published outside the range of July 1st 2006 to November 30th 2019, of non-scientific articles and opinion papers, featuring no nurse, a professional in training or specialised multidisciplinary physician, in a setting other than primary care must be excluded. Languages had to be restricted to those that the authors could understand without the need of a translator because funding was limited. Furthermore, NPs are predominantly established in English-speaking countries and in countries that mainly publish in English, e.g., the Netherlands. The time frame was chosen to include up-to-date concepts applied in current health care systems.

Stage 2 criteria stated that the full text must explicitly mention NPs. The rationale for focusing on NPs was because it is a term widely used to describe advanced nursing roles who may have the potential not only to practice collaboratively but also independently within a team. Hence, the possibility for NPs tasks to differ from GPs in complexity warrants further investigation. Additionally, at least one of the terms: complex, difficult, minor or easy must explicitly be mentioned in the context of task allocation. These terms were validated by conducting a search with multiple synonyms and antonyms for complexity (complex, complicated, intricate, difficult, simple, easy, uniform, and minor). Two random samples of 50 studies each were searched for all synonyms to evaluate which ones would yield the studies relevant to the research question. Studies containing the terms complicated, intricate, simple and uniform were excluded, as they did not provide relevant results. Lastly, the relevant text passages had to be part of the studies’ own findings and not part of a reference to another study. Only if the defined terms were present in the correct context was the full text read and considered for inclusion.

Study selection

The study selection was carried out by two independent reviewers as follows: Upon completing the database searches the resultant studies were transferred into the reference manager EndNote© and de-duplicated according to the guidelines by Bramer et al. [20]. Then the application of the eligibility criteria took place. The search tool in Adobe Acrobat Reader DC© was used when applying Stage 2 eligibility criteria before the eligible full texts were read. The same process was applied to the resultant forward cited literature. Once the process was completed the reviewers compared their results. If a study was excluded or included differently, the study was discussed with a third reviewer until a consensus regarding its allocation was reached.

Data collection process and narrative analysis

An initial, random sample of five included studies was selected for the development of an extraction sheet. Once the data extraction sheet was adapted sufficiently the included studies were reviewed systematically. Firstly, familiarization with the included studies took place and quotes in which complexity featured were located and extracted. Secondly, information to summarise the use and narrow context of complexity was gathered; in the narrow context only information directly from the paragraph in which complexity was used was taken into account. Thirdly, similarities and differences across studies were recognized in the broad context; in the broad context the entire publication was taken into account. The third step was an iterative process based on a narrative analysis following the Cochrane Consumers and Communication Review Group Guidelines [21]. The narrative analysis was chosen because a meta-analysis was not possible as the data stem from a wide range of study designs and capture various interventions as well as non-interventions, which are not conducive to being pooled and analysed. To support the narrative analysis the “Guidance on the Conduct of Narrative Synthesis in Systematic Reviews” [22] was consulted.

Results

Study selection

As shown in Fig. 1 representing the PRISMA flow diagram the database searches delivered 5255 studies upon de-duplication. Titles and abstracts were screened which resulted in 4240 studies being excluded. Whereupon 1015 full texts were screened resulting in a further exclusion of 983 studies, leaving 32 studies for inclusion. During data extraction, a further 63 forward cited studies were obtained. The same process was performed, which resulted in an exclusion of 33 abstracts and 25 full texts, leading to the additional inclusion of a further five studies. Finally, 37 studies were included.

Fig. 1
figure 1

PRISMA 2009 Flow Diagram (overall)

Study characteristics

As can be seen in Table 3 the included studies date from 2007 [53] to 2019 [28,29,30, 52, 60]. Twenty-three are from the US [23, 25, 31,32,33,34,35, 37,38,39,40, 42,43,44,45,46, 48, 50,51,52,53, 55, 58], two from Australia [24, 57], five from the Netherlands [26, 27, 49, 54, 56], one from Norway and Finland [28], five from Canada [29, 30, 36, 41, 59] and one from the UK [60]. Eight studies are qualitative in design [23,24,25,26,27,28,29,30] and twenty-one are quantitative [31,32,33,34,35,36,37,38,39,40,41,42,43,44,45, 47,48,49,50,51,52]. Additionally, there are four studies in which mixed methods are applied [53, 54, 59, 60], one review [55], one country comparison [56], one case study [57], and one perspective [58]. As also shown in Table 3, complexity is used to describe patients (cases, populations, individuals, patient panels) in twenty-three studies, their needs and conditions (problems, complaints) in twenty-two studies, and health professionals’ tasks in five studies.

Table 3 Summary of included studies with regard to complexity

Results of the narrative analysis

The understanding of the use of complexity was based on the broad context consisting of four aspects: patient population, setting, professionals and NP role taking information from the entire study into account as seen in Table 4. Excerpts of the respective text passages can be found in Appendix Table 5.

Table 4 Analysis of included studies with regard to complexity

Patient population

Patient populations are either specific or unspecific. Twelve studies included specific patient populations, two of which were older adults [27, 28], seven diabetics [31, 33, 39, 40, 43, 51, 52], two chronically ill [47, 48], and one home-bound [45], whereas twenty-five studies included unspecific patient populations consisting of general primary care patients.

In geriatric care NPs have been reported to be competent in performing assessments in adults requiring complex care, despite this however, the reality of introducing NPs into general practice may be that GPs focus on more complex geriatric care [27]. Alternatively, they may take on autonomous roles within their scope of practice managing complex geriatric care cases [28]. In diabetes care, complexity can be used to distinguish between medically complex patients, those with comorbidities, receiving GP care and socially complex patients, those effected by poverty, and consequences of dementia and depression, receiving NP care [40, 43]. Alternatively GPs may treat all complex cases while NPs provide supplemental care [39] or disease prevention measures resulting in improved care, for example in terms of adherence to diabetes care guidelines [31]. In certain Veteran Health Associations (VHA) NPs provided entire diabetes care independently with or without delayed physician supervision, which can be considered as the treatment of patients with high decision-making complexity [33, 51, 52].

When observing prescription patterns of GPs caring for chronically ill patients, it seems they care for more complex cases, because they prescribe more and newer medications compared to NPs. This assumption is derived from the fact that co-morbid patients require more medication [47]. This is in keeping with the concept of previous surrogate terms for complexity being comorbidity or multimorbidity [15]. On the other hand, the complexity of chronically ill patients with multiple chronic diseases may be a possible indicator for shared care involving both GPs and NPs equally, especially after recent hospitalization or new diagnosis [48]. In home-based care in which the patients’ medical as well as psychosocial needs must be met, the need for team-based models of care, in which NPs may care for the most complex patients, are promoted [45]. The use of complexity within unspecific patient populations was also broad, however, no clear trend was discernible.

Setting

The settings are either specific or unspecific. Eleven studies include specific settings, two of which are Patient Centred Medial Homes (PCMH) [25, 44], three Community Health Centers (CHC) [36, 41, 42], five Veteran Health Associations (VHA) [33, 37, 51,52,53], and one home-based care setting [45], whereas twenty-six studies describe unspecific settings such as general primary care practices.

Both PCMHs, which are considered enhanced models of primary care aiming to improve quality, invoke better experiences and reduce costs [44], and CHCs, which are community-led, non-profit organizations delivering health as well as social and community services [41], are conceptualized with interprofessional teamwork in mind. In these settings, complexity may be used to allocate medically complex patients to GP care [42, 44] while NPs refer patients to GPs when conditions exceed their scope of practice or range of competence [36] and care for more socially complex patients to minimise consultations with GPs [41]. If however, NPs take on a lead clinician role they may care for all types of complex patients [25].

In the VHA, the largest integrated healthcare system in the US [37, 53], GPs initially casted doubt on the appropriateness of NPs substituting GPs and expressed the need for GP supervision, especially in complex cases [53]. And according to patient encounters, GPs did treat slightly more complex cases [37]. However, NPs increasingly fill similar roles as GPs, working independently and treating similarly complex patients [51, 52], albeit with some delayed GP supervision [33]. As mentioned above in a team-based model of care in a homebound setting, which predicates complexity based on medical and psychosocial needs, NPs may care for patients independently [45].

Similar to unspecific patient populations, the use of complexity is broad within unspecific settings and no clear trend is discernible.

Professionals

Studies with more than two types of health professionals and only two types in collaboration are distinguished. Twelve studies include more than two health professionals [25,26,27,28, 30, 36, 42, 45, 51, 56, 58, 59], while only two health professionals were described in twenty-five studies [23, 24, 29, 31,32,33,34,35, 37,38,39,40,41, 43, 44, 47,48,49,50, 52,53,54,55, 57, 60].

In teams consisting of more than two types of health professionals, including other nurses, NPs are among the highest qualified, thus substituting GPs as lead clinicians providing complex care [25, 27, 28, 51] or managing complex patients within a shared care model [45]. In teams consisting of more than two health professionals and NPs are the only type of nurse or perform similar tasks to a nurse, they may treat less complex patients [26, 30, 56, 58] and improve overall access for them [59]. This depends on their legal scope of practice [42] and practice demands [36].

In teams consisting of only two health professionals NPs may treat less complex patients [23, 24, 29, 39, 43, 47, 49, 53,54,55] or more socially complex patients [40, 41]. This may lead to increased patient access to primary care [54], increased time for GPs to treat (medically) complex cases [34, 35, 44, 53,54,55], reduced referrals to specialists [35], and increased patient outcomes [31]. On the other hand, NPs may also treat complex patients themselves [32, 33, 37, 48, 50, 52, 57, 60].

The question of which professional treats complex patients may not be answered the same way among professionals themselves. Both professionals self-reportedly treat complex patients. However, not many GPs report that NPs treat complex cases [38].

NP roles

NP roles are either described within a distinct model of care or they are unspecified. In seven studies a distinct model of care [26, 32, 39,40,41,42, 49] is illustrated, whereas in thirty it is not.

Models of care involve NPs in the role of either a usual provider, a substitute or a supplement. NPs working as usual providers manage their own patient panels independently. Similarly, NPs functioning as substitutes also manage their own patient panels, however have the possibility to access consultations with a GP similar to any other GP working in a group setting. NPs working as supplements have almost no overlapping tasks with GPs, and thus provide supplemental care. An influential factor on the role of a NP is legislative scope of practice, which may range between full, partial, or restricted scope and defines the range of services provided by a NP.

The evidence shows that NPs as usual providers may treat socially complex patients [40]. Similarly as substitutes they may care for more socially complex patients to minimise consultations with GPs on medical complexity [41] or they may be substitutes only for minor ailments [49]. Alternatively, they may treat under-served patients, who do not differ in medical complexity compared to GPs’ patient panels [32]. Despite potentially full scope of practice, NPs may function as supplements and are less involved in complex care [42]. This scenario may result in as good as or better results in patient care than exclusive GP care [39].

The use of complexity is broad where there is a distinct model of care and, where there is a lack thereof. Furthermore, it is noteworthy that NPs substituting GP tasks is not synonymous with them treating an equivalent patient panel.

Discussion

Main findings

The use of complexity as a factor for task allocation is generally inconsistent. However, trends were recognized: Complexity is used to describe patients, their needs, and health professionals’ tasks. The understanding of the use of complexity as a factor for task allocation between NPs and GPs is based on the patient population (specific vs. unspecific), the setting (specific vs. unspecific), the numbers of health professionals involved (two vs. more than two), and the NP role (distinct model of care vs. no model). Despite similarities in these areas, the tasks which NPs take on range from minor to complex. So for example, a NP’s role may be described as that of a GP substitute, yet only substitute non-complex care, or alternatively take on an entire patient panel with the same complexity as a GP. However, a distinction between medical and social complexity is noticeable throughout all included literature, with a tendency towards GPs treating more medical complexity, while NPs treat more social complexity.

Interpretation & comparison with existing literature

Allocating tasks according to complexity and the professionals’ ability to deal with said complexity is reflected in Kernick’s continuum: This is in keeping with some results of the included studies: complexity is used to allocate patients to health professionals according to their educational ability to treat complex cases [25, 56]. However, given that the included studies originate from countries in which, the NP profession is being developed or is well established and mostly includes an education on Master’s level, Kernick’s continuum is not evident in all practice settings. Hence, when practitioners are considering which tasks to allocate to NPs, it may be indicated to have thorough knowledge of their educational ability, which may define to what extent or in which context complex tasks can be performed.

Similarly the way in which complex systems are displayed in the Cynefin Framework is highly relevant for many of the included studies in primary care. An adaptation of the framework states that all four systems are represented in primary care. GPs are said to manage the simple and complicated systems, including performing therapeutic procedures and prescribing medication. NPs are said to manage complex systems, including the management of chronic illness by supporting and empowering patients to change attitudes, beliefs, and behaviors [61]. This is in keeping with findings in this systematic review: assuming GPs manage medical complexity in simple and complicated systems and NPs manage social complexity in the complex system, i.e. the distinction between medical and social complexity is recognisable. Here it may be of value for practitioners to consider in which system they consider themselves to be active and how tasks can be divided accordingly.

An influential factor on the NP role is legislative scope of practice, which varies largely among countries and regions and informs NP training as well as competencies. However, broadening the legal scope of practice and hence the educational curriculum are not the only steps needed for NPs to care for complex patients. As seen in an example from the Netherlands, where NPs are allowed and able to care for patients with complex conditions, might not do so based on the conceptualisation and traditions of the practice setting [26]. In Indicating that changes in extended areas are needed for NPs to fulfil their potential in practical settings. This observation is supported by Weiland who reports that political, social and professional changes need to take place for NPs to meet society’s health care requirements [62]. Therefore, it is not merely an issue of legislative adjustment, but a matter of developing practice dynamics in which practitioners play a vital role. A further concern is the reimbursement system in place, which may encourage or discourage the employment of NPs in primary care [63]. Tasks may be assigned to a professional based on remuneration to the practice rather than according to actual skills according to a review on facilitators and barriers influencing GP and NP teamwork [64].

Implications

To further clarify the allocation of complex care within interprofessional teams, role descriptions for all health professionals in primary care need to be developed given the country’s health care system and legislative framework. Factors which could be integrated and clarified in a role description are job titles, training, and tasks which could be determined with a functional job analysis [65]. This may be a vital step towards redesigning the system and changing the culture of team work which is evidently needed given the introduction of NPs in primary healthcare [66]. Whether a common yet individually adaptable model for multiple countries would be a viable option is unclear from these results.

Further, role understanding can be encouraged to allocate complex tasks appropriately. This, along with a collaborative work environment, can be facilitated through interprofessional education [67] which “occurs when two or more professions learn about, from and with each other to enable effective collaboration and improve health outcomes” [68]. Additionally, clearer regulations with regard to scope of practice, reimbursement and accountability could enhance skill-mix by increasing NP participation in primary care [63]. As shown in the included literature, various mixed models of care in which roles are mutually understood and skills are appropriately distributed according to regulations can lead to increased efficiency of patient care [54, 59].

Future research could include an overview of university curricula and role descriptions in practice in various countries. This may lead to more knowledge regarding the possibility of creating an over-reaching NP role description including concrete references to the allocation of complex medical and social care, which could be applicable across countries. Furthermore, researchers should determine if a clear allocation of complex care is associated with higher job satisfaction. We also anticipate that clarity could improve role identity and self-confidence among NPs, especially, if inexperienced GPs see the potential value of NPs to the team and support their development [69]. Lastly, improving knowledge about complexity might allow policymakers to develop more transparent and fairer remuneration systems for NPs and GPs.

Limitations

First, the data-derived extraction sheet might have introduced a risk of extraction bias. For example, education was not explicitly represented as a criterion upon which the complexity of tasks could be allocated. However, using this method allowed an extraction process, which remained true to the data at hand, hence, it is to be understood that insufficient information regarding educational level was given in the included literature.

Second, heterogeneity in terminology led to included studies predominantly originating from the US. The term, as well as the profession of NPs, were born in the US in the 1960s, thus an abundance of literature and experience are available. Even though the term NP is well known, other countries have used different terminology. Heterogeneity in terminology may have also led to the exclusion of forward cited literature. In some original sources, the concept of complexity was not explicitly featured, and the citing authors interpreted a described situation as being “complex”. Additionally, restricting the literature search to English and German articles may have resulted in missing publications. Furthermore, we may have missed historical concepts of complexity that were not mentioned in any article that was published within our limited time frame of 13 years.

Third, it can be assumed that tasks are shared among GPs and NPs in primary care settings for which the methods are not published in scientific journals but in policy documents, particularly in less developed countries. Hence, these methods are not visible in the presented results.

Lastly, the lack of restriction on study design meant that heterogeneity in the types of included studies occurred. Consequentially the results are not directly comparable. However, this was not considered a major issue given that the way the search was structured, the aim was to find in which context the word “complex” was used. Hence, the methodological soundness of the individual studies has limited bearing on the statement referring to complexity.

Conclusion

This narrative review delivers an overview of the varied use of complexity and can be used as a point of reference when practitioners are seeking methods for task allocation in a collaborative primary care setting. Complexity has a broad and inconsistent use as a factor for task allocation. However, the findings show, that complexity as a concept is prominent in primary care not only because of increasing rates of chronic illness in an ageing population but also because collaborative practice is on the rise. There is a slight trend towards NPs treating socially complex patients and GPs focusing on medically complex cases. Furthermore, complexity is used to describe patients, their conditions and professional’s tasks. Hence, it may make sense to distinguish a “complex patient” or “complex condition” in terms of medical or social complexity to allocate tasks between GPs and NPs. Task allocation based on complexity can be observed based on patient populations, the setting, the involved health professionals and the roles they take. So, not only can the complexity of the patient and their condition be assessed when allocating tasks but also how many and what types of health professionals are available to provide care. This means that a NP may be one of the highest qualified and therefor may take on a complex caseload, similar to that of a GP or may share complex care according to the given professional abilities. To a large extent however, task sharing according to complexity is also influenced by overreaching legal frameworks which in turn influence education, competencies and team-work culture within practices.