Interprofessional Collaboration among Nurses and Physicians: Making a Difference in Patient Outcome

Deficiencies in collaboration and communication between healthcare professionals have a negative impact on the provision of healthcare and on patient outcomes. Policymakers and healthcare managers, as well as clinicians and practitioners, are aware of this and have a growing interest in improving these relationships. To establish new models of care delivery it is necessary to determine the interventions that are most effective in furthering interprofessional collaboration. This article provides an overview of the evidence base for interprofessional collaboration involving doctors and nurses and new models of care in relation to patient outcomes. Two authors conducted independent literature searches in PubMed, CINAHL, and Cochrane Library and selected fourteen randomised-controlled trials (RCT) for review. All of the RCTs originated from western countries, and the majority tested collaborative care management models against usual care within the elderly population. The major components of the interventions involved individual evidence-based treatment plans, care coordination, health status monitoring, coaching in self-management and promotion of community-based services. They varied between a few days' and three years' duration. Outcome measures incorporated mortality, clinical, functional and social outcomes, and utilisation of medical services. Some studies also used patient-reported outcomes. While the results of the 14 RCTs included were mixed, all but one study reported at least one statistically significant improvement in outcome following interventions based on interprofessional collaboration. More rigorous research in this field and expansion of areas of interprofessional collaboration are needed. Nevertheless, up to now the evidence base of interprofessional collaboration shows promising results in relation to patient outcomes.


Introduction
An increasing volume of literature reports that deficiencies in collaboration and communication between healthcare professionals have a negative impact on the provision of healthcare and on patient outcomes [1][2][3].The consequences reach far beyond stress and frustration levels experienced by professionals; they can result in adverse events such as medication errors and failure to rescue [4][5][6][7].The Joint Commission on Accreditation of Healthcare Organizations (JCAHO) is a voluntary organisation that monitors critical incidents and sentinel events in healthcare settings in the USA, which have been defined as "unexpected occurrences involving death or serious physical or psychological injury, or the risk of thereof" [8].In 2003, JCAHO reported that communication failures among team members are a contributory factor in 60% of sentinel events [9].The factor most influential in reducing these events and their potentially negative effects on clinical outcomes is improvement of relationships among clinicians [3].Policymakers, managers and clinicians therefore have a growing interest in intervening in these relationships through two major approaches: 1) quality and safety improvements by systematically analysing care processes, and 2) interprofessional education and interventions to foster collaboration [10].In particular, new models of care emphasising the need for close interprofessional collaboration and a focus on the use of technology to provide continuous and coordinative care to the aging and chronically ill population in Western countries are often requested in this context [11,12].These approaches are fundamental to delivery of safe quality care.To establish new models of care delivery, policymakers, managers, and clinicians need to know how effective interventions on interprofessional collaboration are.This article sets out to provide an overview of the evidence base for interprofessional collaboration and new models of care in relation to patient outcomes.It is not designed as a formal systematic literature review as recommended by PRISMA Group or Cochrane Library [13].

Methods
To examine the evidence base for the relationship between interprofessional collaboration and patient outcomes, two of the authors searched the literature independently using the following definition to retrieve salient articles: "Interprofessional collaboration exists when two or more members of different healthcare professions work together jointly to solve problems or provide services [10].Furthermore, effective collaboration requires shared power and authority, based on knowledge and expertise and an interaction between subjects with trust, mutual respect and joint contributions to a common goal [14]."Several terms are regularly used to describe working together; these include "teamwork" and "collaboration", and highlight the inconsistencies in defining a complex phenomenon [14].Moreover, terms such as "multidisciplinary", "interdisciplinary", "interprofessional" and "multiprofessional" are often used interchangeably in the literature and their definition depends on people's attitudes and points of view [15].Accordingly, we used most of those terms in different combinations to describe collaborations between healthcare professionals.A set of exclusion and inclusion criteria was used to identify relevant articles for the review process.Papers with a clinical focus on care management between two or more healthcare professions, but comprising doctors and nurses, and on patient outcomes were included.In addition, it was important for us to include only articles describing interventions with inherent collaboration of healthcare professionals.Exclusion criteria were publications not relevant to the topic, those dated prior to 1998, and non-research articles.

Figure 1
Flow diagram on phases of review process (according to PRISMA statement [13]).
The initial search included two bibliographic databases (PubMed, CINAHL) with the following key words and Boolean operators (OR, AND) being used as appropriate: "multi-/interdisciplinary collaboration", "medicine and nursing", "physicians and nurses" and "new models of care medicine/nursing".A preliminary reading of the abstracts showed that the articles used a broad range of research designs, with few randomised controlled trials.For that reason a second search was conducted in PubMed and Cochrane Library.In this search, the terms "interdisciplinary collaboration", "interdisciplinary health team", "physicians and nurses", "interprofessional collaboration" and "cooperative behaviour" were used together with the following limits: published within the last ten years, clinical trial, meta-analysis, randomised controlled trial (RCT), review, and restricted to English and German articles.Out of a total of 451 articles, 22 suitable publications meeting the above-mentioned criteria were selected on the basis of the abstract.Randomised controlled trial was added as a selection criterion, since all Cochrane reviews regarding collaboration complained of the lack of RCTs.Two reviewers independently read the complete articles and selected fourteen studies for the final sample, again applying the inclusion criteria specified above.Disagreements on eligibility were resolved by consulting a third reviewer; a majority needed to agree on the inclusion of an article (fig.1).A data-driven thematic analysis was used to identify prominent themes, and findings of the selected studies were summarised and structured accordingly under thematic headings [16].

Designs, settings and patient populations
All of the fourteen RCTs were conducted in Western countries (table 1); nine from the USA [17][18][19][20][21][22][23][24][25][26], three from Europe [27][28][29], and one each from Australia and Canada [30,31].The study samples included between 50 and 500 patients per group.Randomisation referred either to patients or to interventions of primary care physicians.Thus, separation of intervention and control groups sought to prevent contamination of treatment between groups.One study from England included an additional nested qualitative component within the trial, namely patient and practitioner interviews [28].Eight studies applied a repeated measurement to measure the impact of the intervention over a longer time frame.
Nine RCTs tested collaborative care management models against usual care within the elderly population; seven studies focused on chronic diseases such as heart failure, multimorbidity and Alzheimer's disease.A single study investigated a collaborative model for paediatric asthma [22].All but two studies implemented an intervention in outpatient care.Nikolaus and colleagues [27] studied the effect of a comprehensive geriatric assessment and post-discharge home intervention for patients with acute illnesses in a geriatric hospital in Germany.Naglie and colleagues [31] were interested in the outcomes of postoperative interprofessional care management for patients with hip fracture at a university-affiliated acute care hospital in Canada.

Intervention
The major components of the interventions, based on interprofessional collaboration, included new models of care with bio-psycho-social assessment, development of an individual evidence-based treatment plan, coordination of care, monitoring of health status, coaching patients in chronic disease self-management and promotion of community-based services.Depending on the scope of the intervention, the duration of the programmes included in this review varied between a few days (e.g., a two-visit educational intervention at an outpatient clinic for children with asthma [22]) and two to three years of implementation (e.g., introduction of a system for collaborative primary care for elderly people aged 65 or over [18,25,26] or for patients with bipolar disorder [23,24]).

Outcomes measured and effects reported
The most common outcome measures included mortality, clinical, functional and social outcomes, and utilisation of medical services.A majority of studies (n = 9) also involved patient-reported outcomes such as quality of life, activities of daily living and satisfaction with care.
Five studies reported no difference in mortality between the intervention and usual care groups [17,21,25,27,31] whereas Schraeder and colleagues [18] reported a 49% reduction in all-cause mortality in the treatment group during the second year (odds ratio: 0.51, 95% CI: 0.29-0.91,p = 0.02).Moreover, Inglis and colleagues [30] observed median survival in patients with chronic heart failure almost twice that of the control group (40 versus 22 months respectively, p <0.001) and fewer deaths overall (adjusted relative risk: 0.74, 95% CI: 0.53-0.80,p <0.001) after up to ten years' follow-up.In five studies an improvement in physical, emotional or social functioning was demonstrated in the intervention group [19,24,[27][28][29], though four studies showed no differences between groups [17,21,25,31].Similarly mixed results were reported regarding utilisation of medical services.Again, five studies showed a significant reduction in medical service use in the intervention group compared to the control group [17,20,22,27,30], and two studies reported mixed results.Counsell and colleagues [25] observed that following the intervention there was a significant reduction in emergency department visits without hospitalisation in their low-income elderly population, but no group differences for hospital admission rates.Bauer and colleagues [24] reported a trend towards lower hospitalisation rates for any reason among psychiatric patients in the intervention group during the second year of the study, and a significant reduction in the third year (34 % versus 48%, respectively, p = 0.02).However, two American studies reported no differences regarding hospitalisation rates or length of hospital stay between groups [18,21].
With regard to patient-reported outcomes, the geriatric intervention group in a German study [27] showed a significantly higher score of self-perceived health and life satisfaction than the control group (p = 0.04).In addition, chronically ill seniors in another American study [17] reported an increase in social activities compared with a decrease in the control group (p = 0.04).Two studies reported no group differences in activities of daily living [21,25], whereas Melis and colleagues [29] showed a significant improvement in functional abilities during the first six months; after this time, however, the effects were no longer significant.Furthermore, four studies showed that participants who experienced collaborative care management models were significantly more satisfied with their care than usual-care recipients [20,[24][25][26].

Discussion
Interprofessional collaboration is a common strategy to achieve desired quality outcomes in an effective and efficient manner in a complex array of health services [32].Nowadays, improved interprofessional collaboration is essential to facilitate information flow and the coordination and provision of healthcare within an increasing diversity of disciplines where one health professional can no longer meet all patient needs [33][34][35].
This literature review provides an overview of the evidence base of interprofessional collaboration and its relationship to patient outcomes.Our main finding was that the vast majority of the eligible publications indicate at least one improved outcome following interventions based on interprofessional collaboration.
Most of the studies included introduced new models of care with promising approaches to service delivery, yet some have methodological shortcomings such as sample sizes inadequate to reach satisfactory power levels or longer time frames over which outcome is measured, resulting in patient withdrawal and missing data.Nevertheless, the studies' longer time frames were probably beneficial in making it possible to introduce complex interventions properly and judge the sustainability of programme effects.An additional weakness of several studies was that some outcome measures were collected by nurses who were included in the intervention and not blinded to participants' status, a factor which may have had an impact on the results.More importantly, most of the models tested were multifaceted, rendering separation of the impact of collaboration from other parts of the intervention impossible.The research instruments applied in the RCTs included mainly measured patient-related outcomes such as the Barthel Index, the Short Mental Status Questionnaire, and the National Institutes of Health Stroke Scale.Other databases such as death registries and medical record systems were also used.Collaboration was not a study endpoint and hence questionnaires specific to this issue were not used.
Overall the relationships between teamwork and patient outcomes seem to be difficult to investigate with RCTs, since we can only compare a facilitated, more structured collaboration with a routine form of collaboration.In a previous Cochrane review on interventions to promote collaboration between nurses and doctors, the authors concluded that rigorous evaluations are difficult to conduct because the interventions are complex and the intermediate processes are difficult to assess [36].This review only included two out of 31 studies which tested interventions in academic hospital settings with acceptable methodological quality [37,38] 6-and 12-month follow-up: intervention group was about twice as likely to rate their overall care as "high quality" (p = 0.006).Primary care physician were more satisfied with their interactions with patients and their families (p <0.05) and nurses expressed a consistently high job satisfaction.
processes.A few years later a second systematic review on the effects of interventions on interprofessional collaboration and education [10] yielded fourteen studies in the area of collaboration.Of the studies included five reported no difference in outcomes and nine showed improved results related to their interventions, suggesting a positive impact of interprofessional collaboration interventions on healthcare pro-cesses and outcomes [10].
Although RCTs are the gold standard in establishing a sound evidence base in quantitative research, results from other studies may be crucial in adequately interpreting this evidence, formulating research questions and designing future studies.RCTs may establish causal relationships but do not identify causal processes, which are also important in health services research.Hence investigations into the English health services increasingly use qualitative research methods alongside RCTs in order to gain a more comprehensive understanding of the impact of health service delivery and organisation [39].Complex practice settings often require more diverse methodology.
For example, observational studies of the processes and performances of patient care teams show that teamwork is an important factor in the causation and prevention of adverse events [7].A very recent publication on the impact of surgical team behaviours on patient outcomes reported increased odds on major complications or death when the team less frequently shared information during intraoperative phases and handoff phases [40].Moreover, all measures of poor teamwork across all intraoperative phases were significantly associated with deteriorated patient outcome after adjusting for the American Society of Anaesthesiologists (ASA) score.Therefore, restriction to RCTs alone is a limitation of our review.
Barriers to good interprofessional relationships described in the literature include time pressure, lack of explicit descriptions or of understanding of each other's roles and tasks, poor organisational support, absence of clear leadership, different traditions and professional values, different aims and priorities, and vertical management structures with discriminatory power structures [35,[41][42][43].To identify factors that enable nurse and physician collaboration, an American study investigated the literature and additionally interviewed 141 physicians, managers and staff nurses from 44 clinical units in five hospitals [44].Structural enablers included joint nurse/physician practice committees, integrated patient records, joint practice record review, and the use of protocols or critical pathways in the care of specific patient groups.With regard to interpersonal relationships and interactions, they mentioned trust, respect, shared leadership, recognition of unique contribution, collegiality, and open communication as en-abling factors [44].
In summary, although the studies included reported mixed results, all but one study identified at least one positive and statistically significant effect of the collaborative care models tested.Nevertheless, additional research is needed in this field involving different patient populations and including inpatient care.Two interesting studies are currently under way.One is a cluster randomised controlled trial in five tertiary-care hospitals in Toronto to investigate an intervention with the aim of improving collaborative communication between healthcare professionals and patient-centred care [45].The second study is a cluster randomised controlled trial in primary care in Australia which seeks to determine the impact of nurse and general practitioner partnership on the quality of care and patient outcomes for patients with chronic obstructive pulmonary disease [46].
Up to now the evidence base of interprofessional collaboration shows promising results in relation to patient outcomes, thus justifying further implementation of new models of service delivery in clinical practice which supports collaboration between professionals.

Funding / potential conflict of interest
No funding.No conflict of interest.

Review article
Swiss Med Wkly.2010;140:w13062 Swiss Medical Weekly • PDF of the online version • www.smw.ch

Table 1
. The two trials could not demonstrate effects on mortality and only moderate gains in healthcare Abbreviations: ADL= Activities of daily living/ED = emergency department/ y = years of age/ TIA = transient ischaemic attack/CI = confidence interval Swiss Medical Weekly • PDF of the online version • www.smw.ch

model of care, providers Outcomes measures Participants Selected results
Care management model significantly improved the profile of health and prevention (p <0.0001) in the intervention group.Moreover the effect size for each domain was positive with the highest for stroke knowledge (0.98) and the smallest for neuromotor function (0.10).PDF of the online version • www.smw.ch

Table 1
Abbreviations: ADL= Activities of daily living/ED = emergency department/ y = years of age/ TIA = transient ischaemic attack/CI = confidence interval

model of care, providers Outcomes measures Participants Selected results
, 18-and 24-month follow-up: intervention group had significant improvement in quality of care and reduced ED utilisation rate (p = 0.03).No group differences for hospital admission rates, ADL or death.