Summary of findings
This review is an update to a previous Cochrane interprofessional education (IPE) review wherein four of the six included studies reported a range of positive outcomes (Reeves 2008). While that review was an improvement from the original Cochrane IPE review that identified no studies for inclusion (Zwarenstein 2000), it marked only a small step forward in establishing the evidence base for IPE due to the small number of studies, methodological limitations, and the heterogeneity of IPE interventions. This updated review is timely not only due to the passage of time but also given the continued interest and investment in IPE by policymakers, educators, healthcare professionals and researchers worldwide.
IPE occurs when members of more than one health or social care profession (or both) learn interactively together, for the explicit purpose of improving interprofessional collaboration or the health/well being of patients/clients (or both). The widespread advocacy and implementation of IPE reflects the premise that IPE will contribute to developing healthcare providers with the skills and knowledge needed to work in a collaborative manner (CIHC 2010; Interprofessional Educ Collab Expert Panel 2011; WHO 2010). Interprofessional collaboration, in turn, is identified as critical to the provision of effective and efficient health care, given the complexity of patients' healthcare needs and the range of health-care providers and organisations. Interprofessional collaboration has been linked to a range of outcomes, including improvements in patient safety and case management, the optimal use of the skills of each healthcare team member and the provision of better health services (Berridge 2010; Reeves 2010; Suter 2012; Zwarenstein 2000).
Professional and academic leaders from diverse countries have developed a shared vision and strategy for postsecondary education in medicine, nursing and public health. This commission called for, among other recommendations, IPE that breaks down professional silos while promoting collaborative relationships (Frenk 2010). Similarly, the World Health Organization (WHO) published a report that outlined the role of IPE in preparing healthcare providers to enter the workplace as a member of the collaborative practice team (WHO 2010). National organisations have created core competencies for interprofessional collaborative practice, positioning IPE as fundamental to practice improvement (CIHC 2010; Interprofessional Educ Collab Expert Panel 2011).
Ideally, IPE should begin in the early training period and extend throughout a person's professional career (Barr 2005). Many examples of IPE at different stages of professional development continue to be published. From this work, it is possible to see that IPE can have an impact on learners' attitudes, knowledge and skills of collaboration (e.g. Charles Campion-Smith 2011; Makowsky 2009; Sargeant 2011). These are important educational outcomes, but not the focus of the current review.
Given the ongoing emphasis on the importance of IPE to collaborative practice and ultimately to healthcare processes and outcomes, ongoing attention is needed to advancing the research evidence related to IPE. It is timely to undertake this updating review to identify whether there are additional studies with research designs that meet the criteria of this Cochrane review, which can further inform the evidence of IPE.
The definition of an IPE intervention used in this review is the following:
- An IPE intervention occurs when members of more than one health or social care (or both) profession learn interactively together, for the explicit purpose of improving interprofessional collaboration or the health/well being (or both) of patients/clients. Interactive learning requires active learner participation, and active exchange between learners from different professions.
The two objectives of this review are:
- to assess the effectiveness of IPE interventions compared to separate, profession-specific education interventions in which the same professions were learning separately from one another;
- to assess the effectiveness of IPE interventions compared with control groups which received no education intervention.
In the first objective we are seeking to understand the effects of IPE better in relation to the current dominant uniprofessional education model, where ideally the control group should receive the same education in a uniprofessional manner. We included the second objective as there was a lack of studies addressing the first objective. Our rationale for doing so was that while studies that do not meet the first objective are not as rigorous as those that do, such studies do nevertheless have value in providing some indication of the effects of IPE.
Criteria for considering studies for this review
Types of studies
Randomised controlled trials (RCTs), controlled before and after (CBA) studies and interrupted time series (ITS) studies.
Types of participants
Health and social care professionals (e.g. chiropodists/podiatrists, complementary therapists, dentists, dieticians, doctors/physicians, hygienists, psychologists, psychotherapists, midwives, nurses, pharmacists, physiotherapists, occupational therapists, radiographers, speech therapists and social workers).
Types of interventions
All types of educational, training, learning or teaching initiatives, involving more than one profession in joint, interactive learning, as described in the IPE definition above.
Types of outcome measures
1. Objectively measured or self reported (validated instrument) patient/client outcomes in the following areas: health status measures; disease incidence, duration or cure rates; mortality; complication rates; readmission rates; adherence rates; satisfaction; continuity of care; use of resources (e.g. cost-benefit analyses).
2. Objectively measured or self reported (validated instrument) healthcare process measures (e.g. skills development, changes in practice style, interprofessional collaboration, teamwork).
Search methods for identification of studies
See: Cochrane Effective Practice and Organisation of Care Group methods used in reviews.
Effective Practice and Organisation of Care Group (EPOC) specialized register (epoc.cochrane.org/epoc-register-studies), July 2006 to 2 August 2011.
The search strategy from the previous IPE Cochrane review was adapted for each of the following databases searched:
- MEDLINE August week 4 2006 to July week 3 2011;
- CINAHL, July 2006 to 2 August 2011.
No language restrictions were placed on the search strategy.
We also handsearched the Journal of Interprofessional Care (2006 to 2011), proceedings from key interprofessional conferences - 'All Together Better Health' (Sydney, April 2010) and 'Collaborating Across Borders' (Minneapolis, October 2007 and Halifax, May 2008) and the grey literature contained on the websites of the UK Centre for the Advancement of Interprofessional Education (date accessed: 15 September 2011) and the Canadian Interprofessional Health Collaborative (date accessed 16 September 2011). In addition, we drew on our international networks to ensure that all relevant published and unpublished work in the field would be identified. These searches generated 76 abstracts. See Figure 1.
|Figure 1. Flow diagram.|
(*Total refers to sum of 1999 review and updates in 2008 and 2012).
A total of 3069 abstracts were found: 1248 from CINAHL, 285 from EPOC, 1460 from MEDLINE, 76 from handsearching and conference abstracts. After duplicates were removed, 2733 abstracts remained. While the abstract search was sensitive to identifying a high proportion of relevant IPE intervention studies, it exhibited low specificity in relation to differentiating between IPE interventions and other interprofessional teamwork interventions without IPE components, such as continuous quality improvement and total quality improvement initiatives. See Figure 1 for further information.
Data collection and analysis
Three review authors (SR, LP and JG) independently reviewed the 2733 abstracts retrieved by the searches to identify all those that suggested that:
- there was an intervention where interprofessional exchange occurred;
- education took place;
- professional practice, patient care processes or health or patient satisfaction outcomes were reported;
- the intervention was evaluated using an RCT, CBA or ITS design.
Twenty-eight studies were identified from this abstract search as potentially meeting these criteria. The full text of these articles was obtained. These three review authors independently assessed each full-text article to examine whether it met all of the criteria further. Any disagreements and uncertainties were resolved by discussion, and the input of a fourth review author (MZ), who reviewed all of the final papers as a further quality check for inclusion in the review. Nine studies met the outlined criteria; these nine studies were added to the six studies from the previous review for a total of 15 studies.
Assessment of the risk of bias in included studies
Two review authors (SR and LP) assessed the risk of bias for each study using a form with the standard criteria described in EPOC (2002). The 'Risk of bias' assessments are displayed in Figure 2 and Figure 3. The 'Risk of bias' summary is in Figure 4.
|Figure 4. Risk of bias summary: review authors' judgements about each risk of bias item for each included study.|
We did not exclude studies on the grounds of risk of bias, but sources of bias are reported when presenting the results of studies.
Three review authors (SR, LP and JG) extracted the following information from included studies:
- type of study (RCT, CBA, ITS);
- study setting (country, healthcare setting);
- types of study participants;
- description of education programme;
- description of any other interventions in addition to the education;
- main outcome measures;
- results for the main outcome measures;
- any additional information that potentially affected the results.
Ideally, a meta-analysis of study outcomes would have been conducted for this review. However, this was not possible due to heterogeneity of study designs, interventions and outcome measures among the small number of included studies (n = 15). Consequently, the results are presented in a narrative format.
Description of studies
All 15 studies addressed objective number two – to assess the effectiveness of IPE interventions compared with control groups that received no education intervention. Given the major differences between the included studies, a description of each is provided below. A formal calculation of the evidence, including the creation of a 'Summary of findings' table, was not feasible given the lack of overlap among the outcomes reported. The included studies are presented in three sections according to the type of research design they employed.
Randomised controlled trials
Barcelo 2010 described an RCT that aimed to improve the quality of diabetes care in primary healthcare centres using the chronic care model. Forty-three primary care teams based in 10 public health centres participated in this study. Teams were made up mainly of physicians and nurses with other professionals, such as nutritionists and psychologists also participating in some teams. All 10 health centres implemented a clinical information system and provided the opportunity for patients to participate in peer support groups. Beyond this, five health centres were randomly assigned to receive the intervention, and five received no intervention. The intervention consisted of a multifaceted quality improvement initiative during which teams and patients participated in three interprofessional learning sessions within a period of 18 months. These included a structured patient diabetes education programme, training in foot care and in-service training. In each of the three learning sessions, the teams selected specific objectives for 'plan-do-study-act' (PDSA) improvement cycles. The objectives were based on problems identified in the practice of each health centre (e.g. organisation of care, decision support, information sharing). Other aspects of the multifaceted quality improvement programme included support from hospital specialists and a case management advisor. Reported outcome measures included clinical observations (e.g. metabolic control and cholesterol) and adherence to clinical protocols (e.g. conducting periodic foot and eye examinations). The authors reported that multilevel logistic regression models were adjusted for the clustering of participants within health centres.
Brown 1999 undertook an RCT that aimed to examine whether an interprofessional communication skills training programme for physicians, physician assistants, nurse practitioners and optometrists increased participants' ratings of clinicians' communication skills. The healthcare professionals worked for a 'not for profit' group-model health maintenance organisation (HMO) in the US. The IPE intervention, led by two physicians, consisted of two four-hour workshops delivered one month apart with two hours of homework and a telephone call from an instructor inbetween. The intervention involved didactic components, role playing and interactive dialogue. Of the 69 participants (75% of whom were physicians), 37 were randomly assigned to receive the intervention and 32 were assigned to the control group (which received the IPE intervention after the study). Pre- and post-intervention patient satisfaction scores were drawn from routine data collection, which yielded clinician-specific patient satisfaction ratings every six months. The HMO contracted out the routine data collection. The contractor randomly sampled clinical consultations and mailed a questionnaire to the relevant participants within 10 days of each consultation in the sample.
Campbell 2001 described an RCT that evaluated an interprofessional training programme for emergency department (ED) physicians, nurses, social workers, hospital administrators and representatives from local domestic violence service organisations. The intervention aimed to increase the identification of acutely abused women in EDs, and improve staff and institutional responses. The two-day programme, developed and implemented by violence prevention organisations, involved didactic instruction, role play, team planning and team work to develop a written action plan. Participants from each ED were asked to meet before and after the training. The programme addressed systems change and coalition building as well as provider attitudes and skill building. The attendees were expected to collaborate in order to implement system changes in their respective EDs, including implementing training for ED staff. The instructors were available for telephone assistance during the implementation phase. Six EDs were randomly assigned to receive either the IPE intervention (three hospitals) or to be in a control group that received no intervention (three hospitals). Follow-up data were collected at nine to 12 months and 18 to 24 months.
Helitzer 2011 reported an RCT that evaluated the effects of an intervention aimed at improving patient-centred communication skills and proficiency in discussing patients' health risks. Twenty-six primary care professionals (physicians, physician assistants and nurse practitioners) based in a single academic setting participated in the intervention. A total of 12 professionals were allocated to the intervention group and 14 to the control group. The intervention consisted of training focused on patient-centred communication about behavioural risk factors and included a full day of IPE, individualised feedback on video-taped interactions with simulated patients, and optional workshops to reinforce strategies for engaging the patient. Data were gathered from patients on professionals' patient-centred communication behaviour during two clinic visits that were held at six and 18 months following the intervention.
Nielsen 2007 described a cluster RCT study to evaluate the effectiveness of a teamwork training intervention in reducing adverse outcomes and improving the process of care in hospital labour and delivery units. Fifteen hospitals took part in this study, seven as intervention sites and eight as control sites. Participants included labour and delivery room personnel from obstetrics, anaesthesiology and nursing (n = 1307). The intervention consisted of a three-day instructor training session comprising four hours of didactic lessons, video scenarios and interactive training covering team structure and processes, planning and problem solving, communication, workload management and team skills. The intervention also included assistance with creation and structure of interprofessional teams at each intervention site, which entailed facilitators conducting onsite training sessions to structure each unit into core interprofessional teams. In addition, a contingency team, a group of physicians and nurses drawn from practitioners that were on call during a 24-hour period, were trained to respond in a co-ordinated way to obstetric emergencies. Data were gathered on adverse maternal or neonatal outcomes as well as clinical process data from 28,536 deliveries.
Strasser 2008 described a cluster RCT aimed at evaluating the effects of an IPE intervention on team functioning in stroke rehabilitation units. A total of 227 staff on 14 intervention teams and 237 clinical staff on 15 control teams participated in this study. All teams had representatives from medicine, nursing, occupational therapy, speech-language pathology, physiotherapy and social work. The team training intervention consisted of a multi-phase IPE programme delivered over six months, including: an interactive workshop emphasising team dynamics, problem solving, the use of performance feedback data and the creation of action plans for process improvement. The intervention also included follow-up telephone and video-conference consultations. Patient outcomes data (functional improvement, community discharge, length of stay) were gathered from 579 stroke patients treated by these teams before and after the intervention.
Thompson 2000a described a group RCT aimed at evaluating the effectiveness of IPE and a clinical practice guideline aimed at improving the recognition and improvement of depression in primary care practices. A primary care physician, practice nurse and community mental health nurse delivered the four-hour IPE seminars to general practitioners and practice nurses in groups of two or three practices when convenient. Teaching was supplemented by video-tape recordings, small-group discussion of cases and role play. The educators were available for nine months after the seminars to facilitate guideline implementation and promote use of teamwork. Fifty-nine primary care practices were assigned to the intervention group (29 practices) or control group (30 practices). Practices in the control group received the IPE intervention after the study had been completed. Data were collected six weeks and six months after patient visits.
Thompson 2000b undertook a cluster RCT to examine the effectiveness of a one-year intervention linked to improving identification of domestic violence and the collaborative management of primary care clinics. The intervention for teams of physicians, nurse practitioners, physician assistants, registered nurses, practical nurses and medical assistants, consisted of two half-day IPE sessions, a bimonthly newsletter, clinic educational rounds, system support (posters, cue cards, questionnaires) and feedback of results. Five primary care clinics were randomly assigned to receive the intervention (two clinics) or to the control group (three clinics). Data were collected at baseline, nine to 10 months, and 21 to 23 months.
Controlled before and after studies
Janson 2009 reported a CBA study aimed at improving the care and outcomes of people with type 2 diabetes by improving the care delivered by interprofessional teams. Participants consisted of interprofessional teams of 120 learners (56 second/third-year medicine residents, 29 second-year nurse practitioner students and 35 fourth-year pharmacy students) who delivered team-based diabetes care to 221 people. The control group consisted of 28 traditional-track internal medicine residents who provided usual care to 163 people. The study was undertaken in two general medicine clinics. The intervention involved weekly didactic presentations, clinical discussions and clinic visits with participants. A quality improvement approach was offered by planning and implementing projects using the plan-do-study-act model. The intervention group also received quarterly patient panel reports on process of care benchmarks and clinical status markers.
Morey 2002 presented a CBA study to evaluate the effectiveness of a programme aimed at improving collaborative behaviour of hospital ED staff (physicians, nurses, technicians and clerks). The intervention consisted of an emergency team co-ordination education course as well as implementation of formal teamwork structures and processes. A physician-nurse pair from each ED was involved in developing and implementing the curriculum. The course consisted of eight hours of instruction in one day. The format was lectures, discussion of behaviours, practical exercises and discussion of video-segments. Teamwork implementation involved forming teams by shift and delivering care in a team structure. Each staff member completed a four-hour practicum in which teamwork behaviours were practised and critiqued by an instructor. Staff supported the adoption of collaborative behaviour during normal shifts. This teamwork implementation phase lasted six months. Nine hospital EDs self selected to receive either the IPE intervention (six EDs, 684 clinicians) or act as a control (three EDs, 374 clinicians). Control group departments received the intervention at a later date. Data were collected at two four-month intervals following the training.
Rask 2007 presented a CBA study that aimed to evaluate an interprofessional fall management quality improvement project in nursing homes. Participants consisted of 19 interprofessional falls teams (made up of a nurse, physiotherapist or occupational therapist, certified nursing assistants, a member of maintenance staff). The control group comprised 23 falls teams. The intervention consisted of a full-day interprofessional workshop and a second workshop approximately one month later to address arising challenges. Organisational interventions were also provided in the form of seeking leadership buy-in and support, providing a designated facility-based falls co-ordinator, and ongoing consultation and oversight by advanced practice nurses with expertise in falls management. Data were gathered on process of care documentation, trends in fall rates and changes in physical restraint use.
Weaver 2010 described a CBA study that evaluated an intervention aimed at improving teamwork for operating room staff based at two community-based hospitals. In total, 55 professionals participated in the intervention: 29 in the intervention group (three surgeons, nine nurses, three surgical technicians, 12 anaesthesiologists, two physician assistants); and 26 in the control group (two surgeons, 18 nurses, three surgical technicians, three anaesthesiologists). The intervention consisted of one four-hour session that included didactic presentations and interactive role-playing activities between participants aimed at improving their knowledge and skills of teamwork and collaboration. Data were gathered by observed changes in collaborative behaviour (frequency of team briefings in which information was shared among team members and patient care was planned).
Young 2005 presented a CBA study that evaluated effects of a consumer-led innovation aimed at improving the competence of mental health practitioners working in community mental health provider organisations. The practitioner intervention for psychiatrists, nurses, therapists, case managers, residential staff, mental health workers, and administrative support involved six educational components held over a one-year period that included presentations, discussions, small groups and role-playing techniques, as well as three or four full-day follow-up visits to sites. An additional 16 hours was also spent with staff at the sites. The intervention was developed and delivered by two people who were consumers of mental health services. The innovation also involved a consumer-focused intervention. The study was conducted at five organisations in two states; one organisation in each state received the intervention (total of 269 mental health practitioners, 151 in intervention groups and 118 in control groups). Data were collected at baseline and one year.
Time interrupted series studies
Hanbury 2009 described an ITS study that aimed to test the effectiveness of a theory of planned behaviour intervention to increase community mental health professionals' adherence to a national suicide prevention guideline. The intervention was delivered to 49 participants. The intervention comprised three components designed to target normative beliefs. First, a presentation that contained factual statements, statistics and graphs taken from key government publications highlighting and supporting the guideline evidence base. Second, an interprofessional group discussion was facilitated to ensure that positive normative beliefs were emphasised and any negative normative beliefs challenged. Third, interprofessional group work based on two real life vignettes was undertaken by participants. Data in the form of aggregated, monthly audit adherence data were collected for nearly four years (28 months before the intervention and 18 months afterwards) to evaluate patterns of adherence to using the national suicide prevention guideline. Data from a control site was also included to evaluate the level of adherence.
Taylor 2007 presented an ITS study that assessed the effects an intervention designed to improve the delivery of standard diabetes services and patient care. Professionals based in a single primary care clinic participated in the study. An eight-hour intervention was delivered to participants. The intervention consisted of a range of interactive activities (task redistribution, standardised communication methods and decision-support tool development) that aimed to improve interprofessional communication, teamwork, workflow organisation and information exchange in order to enhance the care of 619 people with diabetes. Data were collected from medical records. Using 1805 clinic visits completed during the study period (160 pre-intervention clinic days and 122 post-intervention clinic days), diabetic services and associated patient outcomes were evaluated for adherence to the American Diabetes Association periodicity recommendations and treatment targets: quarterly blood sugar; quarterly blood pressure; annual low-density lipoprotein; annual urine microalbumin; and annual lower extremity amputation prevention check.
Risk of bias in included studies
For the eight studies that were RCTs, four met five of the nine EPOC 'Risk of bias' criteria (Brown 1999; Nielsen 2007; Thompson 2000a; Thompson 2000b). Three of the five CBA studies met five of the nine EPOC 'Risk of bias' criteria (Janson 2009; Morey 2002; Young 2005). The EPOC 'Risk of bias' criteria have seven elements for ITS studies and one of the two studies met four of the seven EPOC 'Risk of bias' criteria (Hanbury 2009).
Randomised controlled trials
Four of the eight RCTs reported adequately protecting against contamination (Campbell 2001; Strasser 2008; Thompson 2000a; Thompson 2000b). All of the RCTS demonstrated adequate similar baseline outcome measurements. Only one study was inadequate with regards to baseline characteristics being similar (Campbell 2001). Inadequate allocation concealment was an issue in four of the RCTs, with studies either failing to conceal allocation or not making this clear (Barcelo 2010; Campbell 2001; Helitzer 2011; Thompson 2000b). The same four RCTs were unclear or failed in their reporting of adequate sequence generation (Barcelo 2010; Campbell 2001; Helitzer 2011; Thompson 2000b). Four RCTs were unclear or inadequate with regards to the adequacy of blinding in the assessment of outcomes (Barcelo 2010; Helitzer 2011; Nielsen 2007; Strasser 2008). Three RCTs were unclear or had evidence of selective outcome reporting (Barcelo 2010; Brown 1999; Thompson 2000a). All RCTs had evidence of other bias.
Controlled before and after studies
Allocation concealment was an issue for all CBA studies. Four of the CBA studies did not address incomplete outcome data (Janson 2009; Rask 2007; Weaver 2010; Young 2005). Two of the studies did not demonstrate adequate sequence generation (Janson 2009; Weaver 2010); or selective outcome reporting and adequate blinding (Rask 2007; Weaver 2010). All CBA studies ensured baseline outcome measurements were similar with the exception of one (Weaver 2010). Two studies did not report similar baseline characteristics (Weaver 2010; Young 2005); or that the study was adequately protected against contamination (Rask 2007; Weaver 2010). Only two studies were free of other bias (Janson 2009; Morey 2002) (see Figure 2).
Interrupted time series studies
Both ITS studies were adequate for pre-specifying the shape of the intervention effect and for the intervention to be unlikely to affect data collection. Taylor 2007 was unclear in their reporting of whether the intervention was independent of other changes, and were inadequate with regards to selective outcome reporting. Hanbury 2009 did not address all incomplete outcome data. Both ITS studies were not free of other bias (see Figure 3).
Effects of interventions
Effects of IPE interventions reported in each of the studies are presented by the research design each employed.
Randomised controlled trials
The results of the study by Barcelo 2010 indicated that the proportion of people with good glycaemic control (glycosylated haemoglobin (HbA1c) < 7% (53 mmol/mol)) among those in the intervention group increased from 28% to 39% after the intervention (p value < 0.05). The proportion of people achieving three or more quality improvement goals increased from 16.6% to 69.7% (p value < 0.001) among the intervention group while the control group experienced a non-significant decrease from 12.4% to 5.9% (p value = 0.118).
In the study by Brown 1999, the communication skills training programme did not improve patient satisfaction scores. Based on an average of 81 responses for each of the 69 participating clinicians, there was no significant difference in the mean satisfaction scores for the intervention and control groups: each group showed a very small increase in mean scores on 9-point scales (intervention group 0.03 points and control group 0.05).
The results in Campbell 2001 study indicated that the EDs that received the intervention to improve responses to acutely abused women recorded significantly higher levels on all components of the "culture of the emergency department" system-change indicator (e.g. appropriate protocols; materials such as posters, brochures, medical record intervention checklists and referral information available to staff; and staff training) (F = 5.72, p value = 0.04) and higher levels of patient satisfaction (F = 15.43, p value < 0.001) than the EDs in the control group.
Helitzer 2011 reported that the intervention generated significant and persistent changes in patient-centred communication in the intervention group. After six months, a significant difference was found in scores for patient-centredness, which favoured the intervention group (F(1, 20.59) = 8.43, p value < 0.01). After 18 months, the intervention group's significantly higher patient-centredness scores were sustained (F(1, 17.16) = 5.48, p value = 0..032).
Nielsen 2007 found overall no statistically significant differences between the intervention and control groups. Data on adverse outcome prevalence were similar in the control and intervention groups, both at baseline and after implementation of teamwork training (9.4% versus 9.0% and 7.2% versus 8.3%, respectively). However, the time from the decision to the incision for an immediate caesarean delivery was significantly shorter in the intervention group (p value = 0.03). In addition, one process measure, the time from the decision to perform an immediate caesarean delivery to the incision, differed significantly after team training (33.3 minutes versus 21.2 minutes, p value = 0.03).
Strasser 2008 reported a significant difference in improvement in motor score between the intervention group and the control group (13.6% of people in the intervention gained more than 23 points, p value = 0.038). There was no significant difference for the other two outcome measures (p value = 0.1) for both. The proportion of people who had had a stroke making greater than the median functional gain increased by 4.4% in the intervention group, whereas it decreased by 9.2% in the control group, lending further support to the effect of the intervention. At the same time, the intervention had no measurable effect on participants' length of stay.
Thompson 2000a reported no differences between the intervention and control groups in relation to the recognition of depressive symptoms in their evaluation of the effectiveness of an IPE and clinical practice guideline intervention. The outcome for people diagnosed with depression did not significantly improve at six weeks or six months after the intervention.
Thompson 2000b reported that following the intervention, documentation of domestic violence incidents increased by 14.3%. It is also stated that there was a 3.9-fold relative increase of documentation at nine months in intervention clinics compared to the control sites. Overall case finding increased by 30%, but this was not statistically significant. Recorded quality of domestic violence patient assistance did not change.
Controlled before and after studies
Janson 2009 reported that, at study completion, intervention group participants more frequently received assessments of HbA1c (79% versus 67%; p value = 0.01), low-density-lipoprotein cholesterol (69% versus 55%; p value = 0..009), blood pressure (86% versus 79%; p value = 0.08), microalbuminuria (40% versus 30%; p value = 0.05), smoking status (43% versus 31%; p value = 0.02), and foot examinations (38% versus 20%; p value = 0.0005). It was also reported that intervention group participants had more planned general medicine visits (7.9 = 6.2 versus 6.2 = 5.7; p value = 0.006) than did control group participants.
The results of Morey 2002 evaluation of the effectiveness of an interprofessional teamwork training programme on collaborative behaviour in EDs, showed a statistically significant improvement in quality of observed team behaviours between the intervention and control groups following training (p value = 0.012). The clinical error rate significantly decreased from 30.9% to 4.4% in the intervention group (p value = 0.039).
Rask 2007 reported that several key areas of documentation regarding assessment and management of fall risk factors improved. All except two were statistically significant for the intervention teams. Fall rates were not significantly different for the intervention nursing homes (p value = 0.92) and were significantly positive (p value = 0.008) for the control sites. Restraint use decreased substantially during the project period, from 7.9% to 4.4% in the intervention nursing homes (a relative reduction of 44%) and from 7.0% to 4.9% at the control sites (a relative reduction of 30%).
Weaver 2010 reported that intervention participants engaged in significantly more team pre-case briefings after attending training (F [1, 147] = 35.01, p value < 0.001). There was also a significant increase in the proportion of information sharing (e.g. intervention team members were more willing to speak up and participate during briefings) (F [1,128] = 11.47, p value < 0.001). This pattern was also present in the frequency of care plan discussions (F [1,145] = 5.00, p value < 0.05).
Young 2005 reported that in comparison to mental health practitioners in the control group, practitioners in the intervention group reported significantly higher scores in relation to the following competencies: teamwork (r = 0.28, p value = 0.003); holistic approaches (r = 0.17, p value = 0.06); education about care (r = 0.22, p value = 0.03); rehabilitation methods (r = 0.25, p value = 0.007) and overall competency (r = 0.21, p value = 0.02).
Time interrupted series studies
Hanbury 2009 reported that the intervention did not significantly increase adherence to the national guideline. Multiple regression was used to calculate the proportion of variance in intention accounted for by the predictors, and identify the most significant predictor. The intervention was found to account for 58% of the variance (adjusted R
Taylor 2007 found that the intervention achieved improvements in microalbumin testing (+7.40%, p value = 0.001) and HbA1c testing (+3.80%, p value = 0.029). A significant increase in microalbumin levels that were at target (+3.87%, p value = 0.018), and a significant decrease in HbA1c levels that were also at target (–3.81%, p value = 0.011). It is unclear in the reporting if the intervention is independent of other changes. In addition, outcomes were not assessed blindly.
In total, this review included 15 studies, locating nine new studies, which were added to the six studies from the previous update (Reeves 2008). This small growth of eligible studies marks continued development of the IPE field, as the first IPE review found no eligible studies (Zwarenstein 2000).
Seven of the studies reported positive outcomes in the following areas: improvements in diabetes clinical outcomes and healthcare quality improvement goals (Barcelo 2010); improvements in patient-centred communication (Helitzer 2011); improved clinical outcomes for people with diabetes (Janson 2009); collaborative team behaviour and reduction of clinical error rates for ED teams (Morey 2002); increased rates of diabetes testing and improved patient outcomes (Taylor 2007); improved mental health practitioner competencies related to the delivery of patient care (Young 2005); and improved team behaviours and information sharing for operating room teams (Weaver 2010). Three of the studies also reported that the gains attributed to IPE were sustained over time: eight months (Morey 2002) and 18 months (Barcelo 2010; Helitzer 2011).
In addition, four studies (Campbell 2001; Rask 2007; Strasser 2008; Thompson 2000b) reported a mixed set of outcomes. As well as reporting positive outcomes in relation to changes in professional practice and patient satisfaction, Campbell 2001 found no differences in the identification rates of victims of domestic violence between their intervention and control groups. While Rask 2007 reported improvements in care documentation and decreases in the use of restraint for people in nursing homes, they found no change in fall rates. Despite reporting functional gains for patients, Strasser 2008 also reported no significant difference in length of stay or rates of community discharge for stroke rehabilitation patients. Thompson 2000b found that documented asking about domestic violence significantly increased, yet the increase in case finding was not significant.
Four studies reported that the IPE interventions had no impact on either healthcare processes or patient health care or outcomes: Brown 1999 found no significant difference in the improvement of routinely collected patient satisfaction scores between intervention and control groups; Hanbury 2009 reported that the intervention did not significantly increase adherence among participants; Nielsen 2007 reported no statistically significant differences between the intervention and control groups; and Thompson 2000a reported that there were no differences between the intervention and control groups in relation to the recognition or treatment of patients with depression.
Although overall the results indicate some positive outcomes related to IPE, its effectiveness remains unclear at this time due to the heterogeneity among the 15 studies as well as their methodological limitations, as outlined above. The studies were heterogeneous in relation to the objectives and format of the educational intervention, the existence of other interventions in addition to the education, and the clinical areas and settings. The IPE component in these studies ranged from a few hours, to a few days, to longitudinal programmes that were delivered over one year or more. The professional mix of participants also varied from surgeons, nurses, surgical technicians, anaesthesiologists and physician assistants (Weaver 2010), to nurses, physiotherapists, occupational therapists, nursing assistants and maintenance staff (Rask 2007). The aims of the interventions also varied. For example in studies by Brown 1999 and Helitzer 2011, the emphasis was on communication between clinicians and participants, whereas other studies explicitly focused on interprofessional team work in the context of particular settings (ED, operation room) (e.g. Morey 2002; Weaver 2010).
Despite three studies sharing a focus on improving diabetes care (Barcelo 2010; Janson 2009; Taylor 2007), each employed a different research design: an RCT (Barcelo 2010), a CBA (Janson 2009) and an ITS (Taylor 2007). The interventions were different: from a single eight-hour IPE session (Barcelo 2010), to three workshops (Taylor 2007), to weekly seminars (Janson 2009). The participants also varied, from physicians, nurses, nutritionists and psychologists based at 10 public health centres (Barcelo 2010), to 120 students (medical residents, senior nurse practitioner and pharmacy students) (Janson 2009), to an existing team of professionals (who were not identified) based in a single clinic (Taylor 2007). These few examples are some indication of the degrees of heterogeneity and why it is difficult to summarise and identify key elements of successful IPE.
Eight of the studies (Barcelo 2010; Campbell 2001; Janson 2009; Morey 2002; Nielsen 2007; Rask 2007; Thompson 2000b; Young 2005) contained multi-faceted interventions, of which the IPE was only one component. The other interventions included team restructuring, tools such as posters and questionnaires, measurement and feedback, and consumer-directed interventions. In these studies, the authors commented on the importance of system change and the time and resources required to facilitate it (Campbell 2001), the need for leaders who support teamwork within organisations (Morey 2002; Rask 2007) and the use of quality improvement projects (Barcelo 2010; Janson 2009).
Methodologically, the studies shared a common key limitation. All comparative studies (RCTs and CBAs, n = 13) compared the effects of the IPE interventions with control groups that received no educational intervention. As a result, it is difficult to assess the effects of the IPE. Furthermore, most of the included studies involved small samples (defined as fewer than 100 individually randomised practitioners or fewer than 20 randomised clusters), which limited their ability to provide a convincing level of generalisable evidence for the effects of the IPE interventions.
It is also worth noting that there was little evidence of preliminary studies to optimise the IPE interventions and evaluation strategies. IPE interventions are complex, multifaceted interventions in which the components may act both independently and interdependently. Guidance on the development and testing of complex interventions stresses the importance of stepwise work to understand the context for the intervention fully, and optimise the design and implementation of the intervention and evaluation before proceeding to a trial (Craig 2008).
When planning future trials of IPE, thought should be given to the following dimensions: better randomisation procedures, allocation concealment, larger sample sizes and more appropriate control groups. Importantly, studies should include at least one common outcome for measurement of teamwork to enable a formal weighing up of the evidence; in addition, the remainder of the outcomes should include a clear patient health outcome rather than only process measures. Given that IPE is delivered by two or more providers, future trials should have cluster randomised designs, and researchers are advised to be thoughtful about their unit of analysis. In addition, given a lack of evidence on the impact of IPE on resources (e.g. costs and benefits), attention is needed in this area.
While uniprofessional education remains the dominant model for delivering education for health and social care professionals, IPE is increasingly becoming common. Advocacy and implementation of IPE reflects the premise that IPE will contribute to developing healthcare providers with the skills and knowledge needed to work in a collaborative manner (Barr 2005; CIHC 2010; Interprofessional Educ Collab Expert Panel 2011; WHO 2010). Interprofessional collaboration, in turn, is identified as critical to the provision of effective and efficient health care, given the complexity of patients' healthcare needs and the range of healthcare providers and organisations. In relation to implementing IPE at differing stages of the professional development continuum, it is worth remembering that pre-qualification IPE can be regarded as an investment in the future and, in general, studies with short periods of follow-up would not be expected to detect effects on patient outcomes or healthcare processes, which would be difficult to pinpoint, due to a wide variety of potentially confounding variables. Measuring patient outcomes or care process outcomes arising from IPE after qualification (e.g. during continuing professional development and quality improvement initiatives) is more feasible. But it still presents methodological challenges, particularly identifying the influence of IPE within multifaceted interventions and, further, identifying key attributes of effective IPE.
Although this review located nine new IPE studies (which were added to the six studies from the last update) their heterogeneity limits the conclusions we can draw from this work. Nevertheless, a continued increase in eligible studies represents a further positive step forward in establishing a robust evidence base for the effects of IPE on professional practice and healthcare outcomes.
Implications for practice
Our first IPE review published in 1999 found no eligible studies, our 2008 update located six studies, and this update located a further nine studies. At 15 eligible studies, this demonstrates that the IPE field is growing steadily in terms of publishing rigorous IPE research (those employing RCTs, CBA or ITS designs). Although these studies reported a range of positive outcomes, the heterogeneity of IPE interventions means it is not possible to draw generalisable inferences for the effects of IPE. Despite marking a step forward in beginning to establish an evidence base for IPE, more rigorous IPE research (those employing RCTs, CBA or ITS designs) is needed to demonstrate evidence of the impact of this type of intervention on professional practice or healthcare outcomes, or both.
Implications for research
Despite a growth of IPE studies in the past few years, most of this research does not employ rigorous designs. Future RCTs explicitly focused on IPE with rigorous randomisation procedures and allocation concealment, larger sample sizes and more appropriate control groups would improve the evidence base of IPE. A focus on understanding the use of IPE in relation to resources is also needed. These studies should also include data collection strategies that provide insight into how IPE affects changes in healthcare processes and patient outcomes as research to date has not sufficiently addressed this critical issue.
To improve the quality of evidence relating to IPE and patient outcomes or healthcare process outcomes, the following three gaps will need to be filled: studies that assess the effectiveness of IPE interventions compared to separate, profession-specific interventions; RCT, CBA or ITS studies with qualitative strands examining processes relating to the IPE and practice changes; and cost-benefit analyses.
We completed a substantive update of review from 2008 to 2011. Nine new studies were found and added to the six studies located from the previous update.
Amber Fitzsimmons, doctoral student at University of California, San Francisco, for her help with preliminary abstraction notes of some of the included studies.
Data and analyses
This review has no analyses.
Appendix 1. MEDLINE search strategy
MEDLINE search strategy
1 (interprofession$ or inter-profession$).tw.
2 (interdisciplin$ or inter-disciplin$).tw.
3 (interoccupation$ or inter-occupation$).tw.
4 (interinstitut$ or inter-institut$).tw.
5 (interagen$ or inter-agen$).tw.
6 (intersector$ or inter-sector$).tw.
7 (interdepartment$ or inter-department$).tw.
8 (interorgani?ation$ or inter-organi?ation$).tw.
9 interprofessional relations/
11 (multiprofession$ or multi-profession$).tw.
12 (multidisciplin$ or multi-disciplin$).tw.
13 (multiinstitution$ or multi-institution$).tw.
14 (multioccupation$ or multi-occupation$).tw.
15 (multiagenc$ or multi-agenc$).tw.
16 (multisector$ or multi-sector$).tw.
17 (multiorgani?ation$ or multi-organi?ation$).tw.
18 exp professional-patient relations/
19 (transprofession$ or trans-profession$).tw.
20 (transdisciplin$ or trans-disciplin$).tw
22 (education$ or train$ or learn$ or teach$ or course$).tw.
23 exp education, continuing/
24 exp education, graduate/
26 21 and 25
27 program evaluation/
28 "health care outcome?".tw.
29 (education$ adj outcome?).tw.
31 26 and 30
Appendix 2. CINAHL search strategy
CINAHL search strategy
(TX interprofession*) or (TX inter-profession*) or (TX interdisciplin*) or (TX inter-disciplin*) or (TX interoccupation*) or (TX inter-occupation*) or (TX interinstitut*) or (TX inter-institut*) or (TX interagen*) or (TX inter-agen*) or (TX intersector*) or (TX inter-sector*) or (TX interdepartment*) or (TX inter-department*) or (TX interorgani?ation*) or (TX inter-organi?ation*) or (MH interprofessional relations) or (TX team*) or (TX multiprofession*) or (TX multi-profession*) or (TX multidisciplin*) or (TX multi-disciplin*) or (TX multiinstitution*) or (TX multi-institution*) or (TX multioccupation*) or (TX multi-occupation*) or (TX multiagenc*) or (TX multi-agenc*) or (TX multisector*) or (TX multi-sector*) or (TX multiorgani?ation*) or (TX multi-organi?ation*) or (MH "Professional-Patient Relations+") or (TX transprofession*) or (TX trans-profession*) or (TX transdisciplin*) or (TX trans-disciplin*) AND (TX education*) or (TX train*) or (TX learn*) or (TX teach*) or (TX course*) or (MH "education, continuing+") or (MH "education, graduate+") AND (MH "student performance appraisal+") or (MH "course evaluation") or (MH "program evaluation") or (MH "evaluation research+") or (MH "health care outcome*") or (education* N1 outcome*)
Lack of Evidence
Received 20/04/2003 13:47:02
I am assuming this excellent work is a follow up from earlier published material from 1999 (J. Int. Care 13 (4)417-4). What I cannot understand is why, therefore is IPE still 'flavour of the month'? We wouldn't push ideas forward without adequate evidence of effectiveness first! Isn't anyone else out there brave enough to concur with the authors? I certify that I have no affiliations with or involvement in any organisation or entity with a direct financial interest in the subject matter of my criticisms.
Thank you for your positive comment. The article to which you refer is indeed a print version of this Cochrane review, and we will note that in the review. We would like to stress that the 'absence of evidence of effect is not evidence of absence of effect' (Cochrane Reviewers' Handbook 4.1.5, section 9.7). We therefore suggest that interprofessional education (IPE) interventions ought to be implemented widely, but ONLY in the context of rigorous evaluations, ideally randomised controlled trials of their effects. This is not as difficult as it might at first seem, and we would encourage those who are interested enough in IPE to want to subject it to reliable test to contact us or other groups of researchers with randomised controlled trial experience for advice and help.
Merrick Zwarenstein [on behalf of the reviewers.]
The most recent update to this review is published in Issue 1, 2008. The update now has 6 studies. However, it still remains very difficult to draw conclusions about the effectiveness of this intervention and we continue to require further research in the area.
Alain Mayhew [on behalf of the authors and the editorial staff and team]
Jane Warner, Practice Nurse
Last assessed as up-to-date: 3 August 2011.
Protocol first published: Issue 3, 2000
Review first published: Issue 1, 2001
Contributions of authors
SR, LP and JG searched and reviewed the literature and extracted data with input from MZ. SR interpreted the data and wrote the main draft of the review with input from LP, JG, DF and MZ. MZ Is guarantor for the review.
Declarations of interest
Sources of support
- Center for Innovation in Interprofessional Education, University of California, San Francisco, USA.
- Li Ka Shing Knowledge Institute of St Michael's Hospital, Canada.
- Continuing Education and Professional Development, Faculty of Medicine, University of Toronto, Canada.
- Institute of Health Sciences Education, Queen Mary University London, UK.
- Institute for Clinical Effectiveness, Toronto, Canada.
- No sources of support supplied
Medical Subject Headings (MeSH)
*Interprofessional Relations; *Patient Care; *Patient Care Team; *Professional Practice; Attitude of Health Personnel; Health Personnel [*education]; Randomized Controlled Trials as Topic; Treatment Outcome
MeSH check words